Roger D. Woodard-The Textualization of The Greek Alphabet-Cambridge University Press (2014)

Th e Te x t u a l i z a t i o n o f t h e G re e k A l p h a b e t
In this book, Roger D. Woodard argues that when the Greeks first began to use the
alphabet, they viewed themselves as participants in a performance phenomenon conceptually modeled on the performances of the oral poets. Since a time older than
Greek antiquity, the oral poets of Indo-European tradition had been called weavers
of words their extemporaneous performance of poetry was word weaving. With
the arrival of the new technology of the alphabet and the onset of Greek literacy, the
very act of producing written symbols was interpreted as a comparable performance
activity, albeit one in which almost everyone could participate, not only the select few.
It was this new conceptualization of and participation in performance activity by the
masses that eventually, or perhaps quickly, resulted in the demise of oral composition
in performance in Greece. In conjunction with this investigation, Woodard analyzes
a set of copper plaques inscribed with repeated alphabetic series and a line of what
he interprets to be text, which attests to this archaic Greek conceptualization of the
performance of symbol crafting.
Roger D. Woodard is Andrew van Vranken Raymond Professor of Classics and
Professor of Linguistics at the University of Buffalo (The State University of New
York). His visiting positions have included appointments at the American Academy in
Rome, Oxford University, the Centro di Antropologia e Mondo Antico dellUniversit
di Siena, the Max-Planck-Institut fr Wissenschaftsgeschichte in Berlin, and the MaxPlanck-Institut fr evolutionre Anthropologie in Leipzig. He is author or editor of
many books, including Myth, Ritual, and the Warrior in Roman and Indo-European
Antiquity; The Cambridge Companion to Greek Mythology; Indo-European Sacred
Space: Vedic and Roman Cult; Indo-European Myth and Religion: A Manual; Ovid:
Fasti (with A. J. Boyle); The Cambridge Encyclopedia of the Worlds Ancient Languages;
Greek Writing from Knossos to Homer: A Linguistic Interpretation of the Origin of the
Greek Alphabet and the Continuity of Ancient Greek Literacy; and On Interpreting
Morphological Change: The Greek Reflexive Pronoun.
The Textualization of the Greek

Alphabet
Roger D. Woodard
University of Buffalo
With a chapter by
David A. Scott
32 Avenue of the Americas, New York, NY 100132473, USA

Cambridge University Press is part of the University of Cambridge.
It furthers the Universitys mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.
www.cambridge.org
Information on this title: www.cambridge.org/9781107028111
Roger D. Woodard2014
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published2014
Printed in the United States of America
A catalog record for this publication is available from the British Library.
Library of Congress Cataloging in Publicationdata
Woodard, RogerD.
The textualization of the Greek alphabet / Roger D. Woodard ; with a
chapter by David A. Scott.
pages cm
Includes bibliographical references and index.
ISBN 978-1-107-02811-1 (hardback)
1. Greek language Alphabet. 2. Greek language History. 3. Greek literature History
and criticism. I. Scott, David A., 1948 II. Title.
pa 273.w 664 2014
481.1dc23 2013015500
ISBN 978-1-107-02811-1 Hardback
Additional resources for this publication at www.cambridge.org/9781107028111
Cambridge University Press has no responsibility for the persistence or accuracy of url s
for external or third-party Internet Web sites referred to in this publication and does not
guarantee that any content on such Web sites is, or will remain, accurate or appropriate.
For Ed Brown, teacher and friend, and to the memory

of his beloved Nickey
Contents
Preface
Acknowledgments
Abbreviations
page xi
xiii
xv
1 bac kground 1
2 the associative structure of the copper plaques . ................... 15
2.0 Introduction
2.1 Alpha
2.2 Beta
2.3 Gamma
2.4 Delta
2.5 Epsilon
2.6 Digamma
2.7 Zeta
2.8 Eta
2.9 Theta
2.10 Iota
2.11 Kappa
2.12 Lambda
2.13 Mu
2.14 Nu
2.15 Xi
2.16 Omicron
2.17 Pi
2.18 San
2.19 Qoppa
15
15
24
27
28
29
31
34
36
46
56
69
72
75
79
83
83
83
89
97
vii
viii
Contents
*
2.20 Rho
2.21 Sigma
2.22 Tau
3 physical
98
101
103
and chemical examination of the
copper plaques . ...................................................................................
107
3.0
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
107
109
110
111
113
113
116
117
117
David A.Scott
Introduction
Electron Probe Microanalysis
Binocular Bench Microscopy
X-radiography
Optical Metallography
The Patina
Scanning Electron Microscopy
X-ray Diffraction Analysis
Conclusions
4 the
4.0
4.1
4.2
4.3
syntagmatic structure of the copper plaques . ................ 119
Introduction
Narrow Orthographic Transcriptions of the Copper Plaques
Broad Orthographic Transcriptions of the Copper Plaques
Alphabetic Variation in the Copper Plaques
5 langue
5.0
5.1
5.2
5.3
et criture ............................................................................. 140
Introduction
Arbitrariness: Part1
Distinctiveness and Ambiguity
Arbitrariness: Part2
5.3.1 Greek Writing
5.3.2 Semitic and Egyptian Writing
5.4 Alphabetic Order
5.5 Language and Non-language
6 of
119
120
127
136
styluses and withes . ....................................................................
6.0 Introduction
6.1 (Ml)
6.1.1 (Smil) and (Ml)
6.1.2 The Common Origin of (Smil) and (Ml)
6.2 (Luzd)
6.2.1 (Lugos) and Plaiting / Weaving
6.2.2 (Lugos) as a Synonym for (Agnos)
6.2.3 On Samos: Part1
6.2.4 (Lugos) and the Oldest Trees
140
140
146
150
151
154
161
171
177
177
178
180
189
192
192
195
198
200
201
Contents
6.2.6 (Sanis) and (Smilis)

6.2.7 (Smilis) and (Smil)
6.2.9 A Lexical Matrix of Cult and Writing as (Cult) Performance
6.2.10 (Lugizd): A Secondary Formation
6.2.11 (Luzd): A Primary Formation
6.3 (Se) and (abgd)
6.4 On Samos: Part 4
An Addendum on MS 22, Lines 10/11
7 the
ix
202
204
210
213
218
220
221
223
224
warp and weft of writing . ...................................................... 227
7.0 Introduction
7.1 Confusion of Language and Script
7.2 Poetic Weaving
7.2.1 Pindar and Bacchylides
7.2.2 Archaic Greece
7.2.3 Common Indo-European Tradition
7.3 Weaving of a Written Text
7.3.1 Weaving of Alphabetic Letters
7.3.2 Latin Alphabetic Interweaving
7.4 Dionysius of Halicarnassus: Literary, Linguistic,
and Alphabetic Weaving
7.5 St. Jerome and Alphabetic Interweaving
7.6 West Semitic Alphabetic Interweaving
7.7 Alphabetic Interweaving and Division
7.8 Greek Alphabetic Interweaving and the Copper Plaques
7.8.1 A Geometric Subset within the Interwoven Alphabet
7.8.2 The Nu-Iota-Mu Subset within the woven Alphabet
7.9 Greek Alphabetic Interweaving beyond the Copper Plaques
7.9.1 Dotted Omicron
7.9.2 Square Theta and Omicron
7.10 At the Juncture of the Alphabetic Substrings
7.11 The Woven Alphabetic Text
7.12 The Performance of the Alphabet
7.13 Zeus of the Sign
An Excurses on (Sma), (Smaleos), and
(Smantr)
7.14 Homers Bane
227
228
228
228
230
232
233
235
235
239
246
247
249
252
252
256
258
258
260
262
263
264
266
267
288
Notes
291
Bibliography
343
Index
357
Preface

1

,
5
.
,
.
Homer, heroes bard, was deathly vexed inIos
1
When by the Muses some boys did a riddle weave;
With Nectar, Nereids of Sea anointed him andlaid
Him dead beneath a rocky ledge on shore,
For Thetis he had glorified, her son as well, and other
5
Heroes fights, and deeds of Laertes son of Ithaka.
Blessed among the islands of the sea Ios is, for it hashid,
The tiny isle, the Muses and the Graces star.
Greek Anthology 7.1 (Alcaeus of Messene)
The story goes that Homer died when fisher boys on Ios posed him a riddle that he could not unknot. Something like this: Those we caught we left
behind; those we did not catch we brought back with us. What is it? It was
no fish, as Homer seemed to imagine, but lice. And they say that this cleverly
woven web its creation inspired by the very Muses proved to be Homers
undoing when he could not tease apart its fibers.
xi
xii
Preface
But I think not. It was no weaving of a riddle that was Homers bane but
the weaving of the alphabet when the Muses began to show their favors far
and wide to practically anyone who could scratch out its symbols not just
to some boys on Ios. It was a woven viral hexameter that did himin.
There are many indications of this, not least of which is the great variability of letter shapes that were employed in early Greek inscriptional writing,
within individual inscriptions, producing variegated graphemic patterns
text, literally. This variability has long been noticed. The prominence of the
back-and-forth twining lines of boustrophedon is another indication. But the
triad of copper plaques with which this study begins and that provides a continuous thread passing through the narrative of this work, I will argue, drags
this metaphoric weaving out and lays it before us, making it unmistakably
recognizable, gathering our attention to the alphabetic fabric that might have
otherwise escaped our attention.
A word about that. While (1) the set of copper plaques etched with abe
cedarium after abecedarium and (2) the concept of the weaving of alphabetic strands are interlinked in this work, each is a distinct phenomenon. The
investigation of each of the two constitutes a separate study. Each forms the
centerpiece of a separate thesis. One thesis does not depend on the other, but
one informs the other. One thesis concerns the interpretation of a particularly unique set of documents; the other thesis addresses the earliest Greek
conceptualizations of alphabetic writing. The thoughtful reader is asked to
evaluate them separately.
This book has affiliated with it a dedicated Cambridge University Press
Web site: www.cambridge.org/9781107028111. There the reader will find the
following images: (1) my hand-annotated X-ray images of each side of each
plaque; (2) scans of the surface of each side of the plaques in the collection
of Martin Schyen, which were produced at the University of Oslo with the
assistance of Professor Jens Braarvig; and (3) the three tables and seventeen
figures that are referenced by Professor David A. Scott in Chapter3.
Acknowledgments
There are many to whom the author needs and wishes to offer his gratitude
for assistance, encouragement, and insight provided in various valuable ways.
To attempt to name all would inevitably result in the regrettable omission of
some; but special thanks must be expressed to Martin Schyen of Oslo and
Irma Wehgartner of the Martin-von-Wagner-Museum in Wrzburg; David
A. Scott (who contributed the third chapter of this work) and his conservation staff at the J. Paul Getty Museum in Los Angeles; Bernard Comrie and
his colleagues at the Max-Planck-Institut fr evolutionre Anthropologie
in Leipzig; Jens Braarvig of the University of Oslo, Department of Cultural
Studies and Oriental Languages; David Porter, Kerry Christensen, Edan
Dekel, Meredith Hoppin, Amanda Wilcox, and other members of the Williams
College Classics Seminar of 2008, especially Paul A. Woodard, for sharing
the podium with me; Temple Wright and Erika Bainbridge of the Center
for Hellenic Studies Library in Washington, D.C.; James Clackson of Jesus
College, Cambridge; President Hermione Lee and the Fellows of Wolfson
College, Oxford; Robert Parker of New College, Oxford; Maggie Sasanow
and Charles Crowther at the Centre for the Study of Ancient Documents,
Oxford; Anna Davies, Philomen Probert, John Penney, and Andreas Willi of
the Oxford Philology Seminar; Eleanor Dickey of the University of Exeter;
Brent Vine and his colleagues in the UCLA Departments of Classics and
Linguistics; and Leonard Chiarelli of the Aziz S. Atiya Middle East Library
at the University ofUtah.
For professionalism and efficiency beyond compare, I am again most
grateful to Beatrice Rehl at Cambridge University Press, and to her assistant,
xiii
xiv
Acknowledgments
Isabella Vitti. Thanks too go to two anonymous manuscript referees for

their insightful comments and to Brian MacDonald for invaluable editorial
assistance.
As always and for all of the usual reasons the full measure of my debt to
Katherine and Paul defies expression.
Abbreviations
AJA
AJP
AR
BASOR
BASP
BCH
BDB
American Journal of Archaeology

American Journal of Philology
Archaeological Reports
Bulletin of the American Schools of Oriental Research
Bulletin of the American Society of Papyrologists
Bulletin de correspondance hellnique
Hebrew and English Lexicon of the Old Testament (Brown,
Driver, and Briggs 1976)
CEG
Carmina epigraphica graeca (Hansen 19831989)
CIL
Corpus inscriptionum latinarum
CIS
Corpus inscriptionum semiticarum
CJ
Classical Journal
C Phil.
Classical Philology
CQ
Classical Quarterly
CR
Classical Review
DAA Dedications from the Athenian Acropolis (Raubitschek
1949)
Dietz

(Dietz 1836)
Dbner
Scholia graeca in Aristophanem (Dbner 1969)
EG
Epigrafia Greca (Guarducci 1967)
FD
Fouilles de Delphes
FGrH
Fragmente der griechischen Historiker (Jacoby 19541969)
FHG
Fragmenta historicorum graecorum (Mller 18411870)
Gr. Gr.
Griechische Grammatik (Schwyzer 1939)
xv
xvi
Abbreviations
Harv. Stud.
IC
ICS
IF
IG
IGA
JAOS
JHS
KAI
Harvard Studies in Classical Philology

Inscriptiones Creticae (Guarducci 19351950)
Les inscriptions chypriotes syllabiques (Masson 1983)
Indogermanische Forschungen
Inscriptiones graecae
Inscriptiones graecae antiquissimae (Roehl 1882)
Journal of the American Oriental Society
Journal of Hellenic Studies
Kanaanische und aramische Inschriften (Donner and
Rllig 19661969)
Kock
Comicorum Atticorum Fragmenta (Kock 1880)
Khn
Medicorum graecorum opera (Khn 19641965)
KZ
Kuhns Zeitschrift (= Zeitschrift fr vergleichende
Sprachforschung auf dem Gebiete der indogermanischen
Sprachen)
LIV
Lexicon der indogermanischen Verben (Rix 2001)
L-P
Poetarum Lesbiorum fragmenta (Lobel and Page 1955)
LSAG
Local Scripts of Archaic Greece (Jeffery 1961)
2
Local Scripts of Archaic Greece (Jeffery 1990)
LSAG
LSJ
Greek-English Lexicon (Liddell, Scott, and Jones 1996)
MSL
Mmoires de la Socit de Linguistique de Paris
MSS
Mnchener Studien zur Sprachwissenschaft
NRSV
Holy Bible, New Revised Standard Version
OLD
Oxford Latin Dictionary
PRU
Le Palais royal dUgarit (Schaeffer and Nougayrol 1955)
RS
Rpertoire dpigraphie smitique
SEG
Supplementum epigraphicum graecum
Walde-Pokorny Vergleichendes Wrterbuch der indogermanischen Sprachen
(Walde and Pokorny 19271930)
ZPE
Zeitschrift fr Papyrologie und Epigraphik
A n c i e n t Au t h o r s a n d Wo r k s
Aeschylus (Aesch.)
Choe.
PV
Supp.
Libation Bearers (Choephoroe)

Prometheus Bound (Prometheus Vinctus)
Suppliant Maidens
Abbreviations
xvii
Anthologia Palatina (Anth. Pal.; The Greek

Anthology)
Anthologia Planudea (Anth. Plan.; The Greek
Anthology)
Aristophanes (Ar.)
Eq.
Thesm.
Knights (Equites)
Women at the Thesmophoria (Thesmophoriazusae)
Bacchylides (Bacchyl.)
Epigr.
Epigrams
Callimachus (Callim.)
Epigr.
Epigrams
Clement of Alexandria
Protr.
Protrepticus
Cornutus (L. Annaeus)

Theol. Graec.

(Summary of the Traditions concerning
Greek Theology)
Dionysius of Halicarnassus
Comp. De compositione verborum
Erotian
Voc. Hippoc. col. Vocum Hippocraticarum collectio
Euripides
IT
Phoen.
Iphigenia Taurica
Phoenician Women (Phoenissae)
xviii
Abbreviations
Eustathius
Od.
Ad Odysseam
Galen (Gal.)
De anat. admin. De anatomicis administrationibus
De loc. aff. De locis affectis
Herodas (Herod.)
Hesiod (Hes.)
Op.
Theog.
Works and Days (Opera et dies)

Theogony
Hippocrates (Hippoc.)
Fist. De fistulis
Morb. De morbis
Nat. mul. De natura muliebri
Ulc. De ulceribus
VC De capitis vulneribus
Homer (Hom.)
Il.
Od.
Iliad
Odyssey
Homeric Hymn to Apollo (Hymn. Hom.Ap.)

Lucian
Apol.
Catapl.
Apologia
Cataplus
Nicander
Alex.
Ther.
Alexipharmaca
Theriaca
Abbreviations
Nonnus
Dion. Dionysiaca
Oxyrhynchus Papyri (P Oxy.)

Philostratus
Imag.
Imagines
Pindar (Pind.)
Isthm.
Isthmian Odes
Nem. NemeanOdes
Plato
Alc.
Resp.
Alcibiades
Respublica
Pliny
HN Naturalis historia
Scholia in Aristophanem (Schol. Ar.)

Thesm.
Women at the Thesmophoria (Thesmophoriazusae)
Scholia in Lycophronem (Schol. Lycoph.)

Scholia in Nicandrum (Schol. Nic.)
Ther.
Theriaca
Sophocles (Soph.)
Ant.
Trach.
Antigone
Women of Trachis (Trachiniae)
xix
xx
Abbreviations
Theophrastus
Hist. pl.
Historia plantarum
Xenophon
An.
Anabasis
1
Background
In 1983 the antiquities and rare book dealer H. P. Kraus of New York issued
Catalogue 165, in which were included, among other items, two copper plaques
inscribed with Greek alphabetic writing, listed as The Fayum Tablets. Their
provenience and date were given simply as Northern Egypt, eighth century
b. c . or earlier. The two plaques measure 215 by 135 millimeters (plaque 1)and
212 by 137 millimeters (plaque 2); both are approximately 1.3 millimeters thick.
The catalog further describes the plaques as: Inscribed on all four plate faces
(written surface 190 x 100mm.; one plate partially cleaned, both plates having
been covered by a layered structure of patina; small holes punched in each
corner). In a red moroccobox.
In 1988 these plaques were acquired by a Norwegian collector, Martin
Schyen, in whose collection they presently remain. The plaques in the
Schyen collection represent a subset of some larger set of such plaques. A
third member is housed in the Martin-von-Wagner-Museum of the JuliusMaximilians-Universitt Wrzburg, acquired by the museum in the gift of
Egyptian and Greek artifacts from the collection of Alexander Kiseleff in
1982. The existence of a fourth plaque has been reported, but its whereabouts
undisclosed.
In 1986 Alfred Heubeck, distinguished German classicist and linguist, published an analysis of the plaque from the Wrzburg museum. His analysis was
based on a careful visual examination of the plaque. Heubecks observations
were necessarily hampered, however, owing to the heavy patina that covers
a large portion of the Wrzburg plaque identical to the thick patina that
likewise obscures much of the surface of the two Schyen plaques. Even so,
Heubeck realized that the alphabet of the Wrzburg plaque was unique among
1
The Textualization of the Greek Alphabet
Greek alphabets in that it terminates with the letter tau: in other words, the
alphabet has no upsilon, it has none of the so-called supplemental consonant
symbols phi, chi, and psi and no omega. The Wrzburg alphabet is thus
coterminous with the Phoenician consonantal script, the source of the Greek
alphabet, which runs from alep (Greek alpha) to taw (Greek tau). Heubeck
also drew attention to the shape of alpha, lambda, and sigma, which he viewed
as being closer to later, rather than earlier, forms of these letters, and to iota
and nu, as being in allen griech. Alphabeten ohne exakte Parallele. Heubeck
proposed a date for the plaques of late ninth or early eighth centuryb c .1
In 1999 Martin Schyen sent the two copper plaques in his collection to
the J. Paul Getty Museum in Los Angeles for physical and chemical analysis in order to verify their authenticity. The plaques were there subjected to
a battery of tests, including binocular microscopic analysis, mettalography,
X-ray fluorescence spectrometry, X-ray diffractometry, X-ray radiography,
and environmental scanning electron microscopy. These tests were carried out under the direction of Dr. David Scott, then director of the Getty
Conservation Laboratory.2
In 2001 I contacted Dr. Irma Wehgartner of the Martin-von-WagnerMuseum regarding the possibility of examining the plaque held in that museums collection. The museum generously loaned the plaque to the Getty for
the same analytic treatments: the outcome of that examination was consistent
with the analysis of the Schyen plaques and revealed that the three plaques
had been cut from a single piece of copper. A summary of the analyses of the
three copper plaques appears in Chapter3, authored by David A. Scott.
A discovery that came to light in the process of examining the plaques in
the Getty laboratories was that, even though large portions of the plaque faces
cannot be read, and other portions read only with difficulty, or uncertainty,
being encrusted by this heavy patina, radiographic images of the plaques fully
reveal the letters with which they are engraved. A single radiograph displays
the abecedaria incised on each side of a plaque, front and back, one set superimposed upon the other.
Methodical and repeated examinations of the radiographs, and the plaque
surfaces where possible, revealed that the numerous abecedaria on the three
copper plaques preserve in the case of most letters not a single letter shape
but a set of variant letter shapes: in some instances the variation is subtle but
clearly observable, in other instances the variation is remarkably ostentatious.
The extent and degree of variation in some cases had not escaped Heubecks
attention. The treatment of the variant forms of the Greek letters of the copper
Background
plaques is the focus of the next chapter. The four sides of the two plaques in
the collection of Martin Schyen are identified as MS 1-1, MS 1-2, MS 2-1, and
MS 2-2. The two sides of the plaque in the Wrzburg museum are labeled W-1
and W-2. An X-ray image of each side of each of the three plaques is presented
between Chapters1 and2.
The find place of the plaques cannot at present be identified with more
precision than the Fayum, as specified in the Kraus catalog description cited
earlier. That description was provided to Kraus by the individual responsible
for introducing into the United States those two plaques that now reside in
the collection of Martin Schyen: that individual has been identified as the
late Dr.Aziz Suryal Atiya, a native of Egypt and a distinguished scholar of
Coptic and medieval studies, perhaps best known to broader audiences for
his Coptic Encyclopedia (published posthumously),3 a work that has received
new life in an expanded digital form from the Claremont Graduate University
School of Religion.4 More will be said shortly on Atiya and his apparent role
in bringing the plaques, ultimately, to scholarly attention.
However and whenever these copper (not bronze) documents came to rest
in the Fayum, I think it unlikely that they were crafted in that place. As the
reader will notice as she or he progresses through this work, there is a network
of affiliations that appear to draw the plaques into the world of the Greeks of
the eastern Mediterranean, especially East Ionic Greeks. Particularly seductive
are associations with the island of Samos a place with demonstrably strong
ties to Egypt (as will be discussed in Chapter6). The plaques were most likely
crafted for some specific cult purpose. With the performative alphabetic weaving to which I argue these plaques do attest, one might compare, for example,
the weaving of a peplos for Athena in the celebration of the Panathenaia, with
its accompanying rhapsodic performance of Homericepic.
No compelling reason presents itself for rejecting Heubecks dating of the
alphabet of the copper plaques, based on his examination of the Wrzburg
document. Recent finds require an earlier date for the origin of the Greek
alphabet than that advocated by many classicists throughout much of the
twentieth century, as do various considerations of the most likely scenario in
which that script was devised:5 late ninth century b c is reasonable. And the
analyses of the letters that is presented in Chapter2 fully allows such a date
for these materials, even if some of the letter shapes of the copper-plaque
abecedaria are better known from later periods. One could imagine that the
alpha-through-tau abecedaria that one finds engraved in these documents
are typical of that date but that the plaques themselves are of somewhat later
production; in other words, the scribes who produced the plaques were
recording an alphabet, in its variant avatars, of a more archaic type than one
used in their own day preserving a primitive Greek script that had been
preserved and transmitted to them in some ritual context perhaps on ritual
implements that have not survived, executed on a perishable medium such as
cloth or leather. If that should turn out to be the case, however, the crafting
of the plaques must be sufficiently early that the notion of alphabetic weaving is still a productive one at the time of their engraving, unless, of course,
this conceptualization of the production of the alphabet also itself became
fossilized.
It was the University of Utah professor Aziz Atiya (18981988), Copticist,
Islamicist, and medieval historian, a major collector of Egyptian epigraphic
materials, from whom H. P. Kraus acquired the two copper plaques that
were advertised in Catalogue 165 (1983) according to Martin Schyen, the
Norwegian collector who in turn purchased the plaques from Kraus.6 Atiya
was born in the village of El-Aysha, Egypt, located within the Nile Delta, later
moved to Zagazig, and was eventually schooled in Cairo. He attended university in Liverpool (19271931), subsequently earned a doctorate from London
(1933), and undertook postdoctoral work at the Universitt Bonn, Germany,
returning to Egypt in 1939, where he remained in academic employment until
1952 (1954?), departing the University of Alexandria under less than favorable circumstances, reportedly owing to his involvement with a Library of
Congress expedition to the Monastery of St. Catherine on Mount Sinai (1949
1950) in which he seems to have played a significant role not only as scholar
but as political facilitator.7 Atiya had been a Fulbright Fellow in the United
States during 19501951; after his departure from Egypt, he returned to the
United States. There followed a series of visiting appointments at American
universities: Michigan, Columbia, Union Seminary, Indiana, and Princeton
(including a year as a fellow of the Institute for Advanced Study). In 1959 he
joined the faculty at Utah.8
The Aziz S. Atiya Middle East Library at the University of Utah, a special collection within the universitys J. Willard Marriott Library, houses the
Atiya archives, to which I was given generous access during a visit to Salt
Lake City in 2011. Among Atiyas personal letters preserved in the archive
are numerous pieces of correspondence between Atiya and H. P. Kraus, dating from 1962 to 1980.9 References within the letters make it clear that the
archive does not contain the full set of all such correspondence. The earliest archived correspondence between the Kraus firm and Atiya is a 1962
Background
letter from the firm that must have accompanied a copy of H. P. Kraus
Catalogue 101, Liberal Arts Periodicals Works and Reference Works Including
Social and Political Science, Law, Business, Publications of Learned Societies,
Government Publications. The author of the letter draws attention to the
subject matter of the cataloged items as being of the sort that Atiya had
previously purchased for the University of Utah library and curiously refers
to the period of those purchases as a time when you [i.e., Atiya] were with
the University of Utah, as if he were no longer so. Atiya was affiliated with
Utah from 1959 until the end of his career. The mischaracterization in this
1962 missive may reflect Atiyas absence from the university in 19611962,
and subsequently, on buying trips10 to Egypt to purchase volumes for the
Middle East collection at Utah;11 in any event, it betrays previous contact
between Kraus and Atiya.
Much of the correspondence between the Kraus firm and Atiya has an air
of familiarity to it: both a personal and professional relationship must have
existed. In 1965 the Kraus Reprint Corporation published a revised edition of
Atiyas work The Crusade in the Later Middle Ages (first published in 1938 in
London by Methuen). In 1967 Atiya contributed an article to a Festschrift for
Krauss, Homage to a Bookman, edited by Hellmut Lehmann-Haupt, whom
Kraus called one of his two chief assistants:12 Atiyas contribution is an essay
on the Codex Arabicus, a unique tri-lingual quintuple palimpsest, which he
says he discovered on June 12, 1950, in the Sinai Monastery of St. Catherine.13
In correspondence dated December 8, 1978, Kraus writes in response to a
letter from a Mr. Olsen that Atiya had forwarded to him. Evidently Olsen
had something to sell, and Atiya was directing him to Kraus as a potential
buyer: Kraus tells Atiya that he is not collecting this kind of material. Olsen
is possibly to be identified with Fred Olsen (died 1986), an officer of Olin
Mathieson Chemical Corporation and patron of the Fred and Florence Olsen
Foundation of New Haven and Guilford, Connecticut, promoting Coptic and
Andean art among other activities, with whom Atiya first corresponded while
at the Institute for Advanced Study in 1959.14 The most recent of the archived
letters of correspondence with the Kraus firm is dated to 1980 and concerns
the republication of volumes on Eastern Christianity by the Kraus-Thomson
Organization Ltd. (a publishing arm that Kraus initiated in 1967 together with
the Canadian newspaper tycoon Roy H. Thomson [Lord Thomson of Fleet]);15
Atiyas A History of Eastern Christianity was republished by Kraus-Thomson,
carrying the publication date of 1980 (originally published by Methuen in
1968).
Atiya was a consummate collector. In an interview with Everett Cooley,

responding to a question about his collecting activities, Atiya states: I was
collecting all my life.16 The University of Utah Library boasts of holding the
largest Arabic papyrus, paper, and parchment collection in North America,
numbering more than seventeen hundred fragments (as of 2007): most of
these were a gift of Atiya to the library in 1975.17 In the Cooley interview, Atiya
describes a moment in which the continuation of the Utah Middle East Center
appeared to be in jeopardy. In order to save the center, Atiya says, he decided
to turn over to the university an amount of manuscripts in my possession as
well as a heap of Arabic papyri which looked like rubbish, because papyrus
looks like rubbish,18 thereby endowing in one motion a major international
research collection. These papyri, says Atiya, were my life, in a sense. These
papyri were collected all over the world, but in Egypt principally, and also in
Beirut and London, he says. Atiya continues: Everywhere where there was a
scrap of Arabic papyrus I pocketed it because Arabic papyri, unlike the Greek
and Coptic which were involved in Biblical material, were less in demand by
collectors.19 It is an interesting, and probably revealing, differential characterization, as we shallsee.
In the same interview, Atiya describes his method of acquisition:
I can tell you one thing, that when I went to Egypt, collections were lying idle
in vaults, which I invaded. When I came out I just spat mud. I swallowed so
much dust, sitting dust that had accumulated for a hundred years on these
books. I got all the treasures available in the market. It was after I made that
invasion in Egypt with my connection with book sellers and so on that the
other centers became aware and started looking for what I had left behind.
This collection, really, getting it, was in a sense a romance that needs a special
tape and we dont have time to do that sort of thing. Personally I had so many
pleasant ventures in my collecting; for instance, I saw a heap in a corner which
included something that I wanted very badly. So I would ask the sheikh, How
much do you want for that collection? He said, Nothing. Its rubbish. Take it.
I said, No, Ill give you ten pounds for it, ten pounds.20
The only commodity explicitly mentioned is books, but one must suspect that among the rubbish were the precious papyri (papyrus looks like
rubbish) and likely other treasures.
How was Atiya able to get these materials out of Egypt? A clue may perhaps be provided by comments that he offers in one of the Cooley interview
sessions as he discusses the microfilms that were made during the Library
Background
of Congress Sinai expedition. After describing his supervisory role in the

expedition, Atiya adds: I was instrumental in helping to export the material
untouched because of the sensitive microfilms. If the customs began to play
around with this material, it would have been calamitous. He goes on: I had
access to the minister of foreign affairs. Through him I was able to let the
microfilm consignments leave the country untouched.21
H. P. Kraus Catalogue 105, bearing the title A Collection of Papyri: Egyptian,
Greek, Coptic, Arabic, Showing the Development of Handwriting Mainly from
the Second Century B.C. to the Eighth Century A.D., contains an introduction authored by Father Theodore C. Petersen, who had been until 1948 a
professor of Coptic at the Catholic University of America. Kraus credits
Petersen with preparing the catalogue.22 The entry for Father Petersen in
Atiyas Coptic Encyclopedia includes a reference to Petersens Introduction
to Catalogue 105 and identifies the date of the catalogs publication as 1964.23
The catalog itself bears no publication date. In his autobiography, Kraus writes
of it, In 1961 I issued a catalogue of very fine classical papyri containing 140
items. A publication date of 1961 is certainly mistaken: a single buyer acted
quickly enough to procure the entire collection from Kraus, a purchase that
was announced in the New York Times on July 1, 196424 (and compare the
previously discussed Catalogue 101 that had appeared by 1962). In an article
treating one of these papyri, Emmel draws attention to Krauss 1961 dating of
the catalog, and he is likely correct when he suggests that Kraus had in mind
the date of his acquisition of the items.25
The buyer of the collection of Kraus Catalogue 105 was Edwin J. Beinecke,
and the catalogs contents were destined for the Beinecke Rare Book and
Manuscript Library at Yale.26 In his Introduction to the catalog, Father
Petersen writes that the papyri were until quite recently the private collection
of a renowned scholar in the field of papyrology. Somewhat similarly, Kraus
in his autobiography, after describing his sale of a Book of the Dead papyrus to the Swiss collector Martin Bodmer circa 1961, then mentions: Later I
came into the possession of a large, valuable collection of Greek and Coptic
papyri assembled by a well-known collector.27 This renowned scholar and
well-known collector who sold the papyri to Kraus was Aziz Atiya, according to Martin Schyen Kraus purportedly having revealed Atiyas identity
to Schyen when the Norwegian collector purchased from Kraus pieces that
had been part of the assemblage out of which came the Beinecke papyri: for
example, MS 244/01 (a Greek papyrus) and MS 245/07 (a Coptic liturgy on
a limestone fragment); see also MS 247 (a Greek funerary inscription on
linen that made its way from Kraus to Schyen via the collection of Richard
Linenthal). The timing of the appearance of the Beinecke papyri on the market undeniably, and neatly, correlates with Atiyas book-buying trips to Egypt
in 1961 and 1962 (and subsequently) and with the correspondence between
Atiya and Kraus, which was under way prior to the 1962 letter that resides in
the Atiya Archive, and can be taken as highly supportive, if not confirmatory,
evidence. It must have been on one such trip that Atiya stumbled across the
copper plaques.
W-1. X-radiograph of obverse face of Wrzburg plaque. Courtesy of Professor David A. Scott.
W-2. X-radiograph of reverse face of Wrzburg plaque. Courtesy of Professor David A. Scott.
MS 1-1. X-radiograph of obverse face of Martin Schyen plaque 1. Courtesy of

Professor David A. Scott.
MS 1-2. X-radiograph of reverse face of Martin Schyen plaque 1. Courtesy of

MS 2-1. X-radiograph of obverse face of Martin Schyen plaque 2. Courtesy of

MS 2-2. X-radiograph of reverse face of Martin Schyen plaque 2. Courtesy of

2
The Associative Structure of the Copper Plaques
, ,

And he sent him into Lycia, and gave him twisting signs,
many and life-destroying, written in a folding pinax.
Homer Iliad VI 168169 (on Proteus and the message he gave
Bellerophon to carry to Lycia)
2.0
i n t r o d u c t io n
In this chapter I examine in turn each of the twenty-two letters of the copperplaque (CP) alphabet. What such an examination makes readily apparent is
that the plaques display a remarkable set of morphological and typological
variants of many of the individual alphabetic characters. Each variant letter
type occurring in the abecedaria of the plaques is provided with a drawing of
its shape (shown sinistroverse [as with all examples reproduced from the copper plaques], which is the direction of writing in the plaques).
2.1
a lpha
The inaugural letter of the Greek alphabet occurs in four basic forms in the
copper plaques. None of the four is particularly primitive in appearance when
compared to the several examples of the sidelong alpha, a, scratched onto the
Dipylon oinochoe (IG I2 919; SEG XXX 46; CEG 432).1 This late eighth-century
15
16
graffito on a wine jug from Athens is commonly held to be the earliest Greek
inscription of substance.2 The letter is also found on a fragment of red pottery
from Euboian Pithekoussai, circa the mid-eighth century b c , the so-called
Lacco Ameno sherd3 if, in fact, the four letters preserved on the sherd are
Greek (rather than Phoenician).4 Like all other forms of archaic Greek alpha,
those of the copper plaques appear to stand at some evolutionary distance
from their Phoenician protoform a. Notice that the stance of this most primitive of Greek alpha shapes is reversed in comparison to the stance of the
Phoenician prototype.
One of the types of alpha occurring in the abecedaria of the plaques has
the following shape:
(1) Alpha-1
What I have designated as Alpha-1 might be judged the typologically earliest
of the four CP alpha types. But this judgment has significance only in relative
terms: again, none of the several types is especially close to the Phoenician
prototype; and to reiterate Greek alpha in most of its various local forms
shows considerable graphic separation from Phoenician alep. There is more,
however, that might be said about the relative typology of the CP alpha types,
and to this we will briefly turn after considering each of the forms.
The shape of CP Alpha-1 is not unique to the copper plaques, though
examples of this morphology are not abundant at the earliest period of Greek
alphabetic writing. Outside of the plaques, its earliest attestation perhaps
comes from Crete, a component of a verse inscription on a storage pithos
from Phaistos (SEG XXVI 1050):5

Erpetidm Paidophilsode
This [is the pithos] of Paidophila, [the wife?] of Erpetidamos/
This [is the pithos] of Erpetidamos, [the son] of Paidophila
The second alpha of is unmistakably of the CP Alpha-1 type

alpha type 8 in Jefferys catalog of letters;6 other forms in the inscription are
of Jefferys alpha type 4, that is, , here shown dextroverse (i.e., the crossbar
slants up toward the direction of writing). Variation in the morphology of
alpha within a single archaic inscription, as here in this Cretan specimen, is
not uncommon. The date assigned to the inscription is eighth century b c ,
making it the earliest known alphabetic inscription from Crete; more than
that, as Johnston notes in his revision of Local Scripts of Archaic Greece, the
17
text was inscribed before firing, and if Levi7 is right that the type cannot
be later than c. 700, this is one of the earliest datable inscriptions that we
have.8
The town of Selinous in Doric Sicily, founded by colonists from Megara
Hyblaia in the seventh century (and perhaps taking its alphabet ultimately
from Megara in Greece), provides a late seventh- or early sixth-century example of an alpha of the type of CP Alpha-1. The letter occurs in a graffito on a
Megarian cup from the necropolis:9
[] h
[E]pameinonos hakulix
[This is] the kylix of Epameinon
The first alpha is like that of CP Alpha-1; the second is, again, that of Jefferys
type4.
From the site of Naxos in Sicily the first Sicilian Greek colony, settled in
the second half of the eighth century by Euboians from Chalkis, with a contingent from the Aegean island of Naxos10 comes a late seventh-century
inscription on a marble block, a dedication to the goddess Enyo. The single
alpha occurring in the inscription is of the CP Alpha-1 type.11 The same sort
of alpha is attested in the alphabet of Chalkis, as in the mid-sixth-century
graffito from a vase found at Olympia: (Smonids m
anethken) Semonides dedicated me.12
From Lokroi Epizephyrioi, on the southeast coast of Italys toe, comes
another example. A graffito on a sherd, dated to circa the end of the seventh
century b c , preserves the ancient name of the goddess Cybele () and
at least one clear example of a CP Alpha-1 sort.13 The colonial alphabet of
Lokroi Epizephyrioi is like that of Ozolian Lokris in central Greece, being
similar to the alphabet of neighboring Phokis.
The type of CP Alpha-1 also occurs in archaic Attic script, Jeffery describing it as found at all periods, but rarely;14 Immerwahr also notes its rarity in
earliest periods.15 In the seventh century b c , one finds a similar type of alpha
in the graffiti from Mount Hymettos, southeast of Athens, though nearly all
examples are reversed in stance relative to that of CP Alpha-1; in other words,
the crossbar descends toward the direction of writing in the Hymettos materials (as in, e.g., H 147)16 but descends away from the direction of writing
in the copper plaques. Likewise, in the earliest inscribed materials from the
Athenian Agora, a similar alpha morphology occurs, but again one that is
typically the reverse of CP Alpha-1, though an alpha with the stance of the
18
copper-plaque type can be seen in a graffito ( [) on the handle of an

amphora, circa the sixth centuryb c .17
The same state of affairs characterizes the alphabet of Aigina (which most
closely resembles Attic and Cycladic alphabets)18 and that of Megara: alpha of
the type of CP Alpha-1 occurs in both stances. In the case of both alphabets,
Jeffery describes the stance which is like that of CP Alpha-1 as normal by the
end of the sixth century.19 The alphabet of Aiginas Cretan colony of Kydonia
preserves an alpha with the morphology of CP Alpha-1 in the fifth century.20
Fundamentally similar alphas with a similar distribution occur in the
alphabet of Phokis. Jeffery characterizes an alpha type with a stance like
that of CP Alpha-1 as most frequent; that with the reversed stance is not
uncommon.21 Both types can be seen co-occurring in the boustrophedon
inscription of a dedication to Athena and Hera, circa the first half of the sixth
century:22 the alphas are always turned in the same direction, regardless of the
direction of writing.
CP Alpha-1 also bears formal similarity to the eastern Ionic alpha of an
abecedarium inscribed on a cup from the Heraion of Samos, circa 660 b c .23
From the Heraion also, but a century later (ca. 570560 b c ), one of the forms
of alpha found in the inscription on the statue of the veiled Hera matches CP
Alpha-1 (the other form being that with reversed stance),24 as does an alpha
in a contemporary Samian inscription on a kouros dedicated to Apollo.25 On
neighboring Chios, the famous legal inscription on stone, also dated to circa
the first half of the sixth century b c , utilizes alpha of the type CP Alpha-1, as
well as (and more often) the comparable form in the reversed stance.26 The
same type alpha was in use in the early sixth century alphabet of Kolophon,
on the Ionic mainland, preserved in a graffito left by a Kolophonian mercenary at Abou Simbel (ca. 591 b c ); both stances that of CP Alpha-1 and its
reverse are again preserved in this graffito:
<> | |<>
Pa<m>bis o Kolophnios sun Psammata<i>
Pambis the Kolophonian with Psammatas
The first and second occurrences of alpha have the stance of CP Alpha-1; the
third, positioned at the right edge of the inscription and sandwiched between
the two lines, has the reversed stance.27 An alpha like that of CP Alpha-1 can
also be seen in sixth-century graffiti from the sanctuary of Apollo Milesios
at Naukratis,28 the Egyptian city founded by Ionic Miletos in the seventh
centuryb c .29
19
In the Doric eastern Aegean, the same type of alpha can be seen in graffiti
from Lindos on Rhodes of the first half of the sixth century b c .30 That alpha
having a stance that is the reverse of CP Alpha-1 is well represented at the
same period.31
A few words regarding this alpha that is fundamentally like that of CP
Alpha-1 but reversed in orientation are in order: Jeffery considers this what
we might term the non-CP stance to be the unmarked case, and the stance
like that of CP Alpha-1 to be the marked case. Indeed, in her discussion of the
previously mentioned alpha of Aigina, Megara, and Kydonia oriented in the
same direction as CP Alpha-1 she refers to it as the reversed alpha.32 The
unmarked stance (adopting Jefferys perspective) with the diagonal crossbar
descending toward the direction of writing is considerably more common
among the various local forms of alpha than the stance of the CP Alpha-1
with the crossbar descending away from the direction of writing as is readily
(if somewhat anecdotally) revealed by a perusal of the Table of Letters at the
end of Jefferys Local Scripts of Archaic Greece. Like the marked (i.e., the CP)
stance, the unmarked is also attested early, as in, for example, a late eighthcentury graffito on a sherd from Eretria in Euboia.33 The same symbol occurs
in the Phrygian alphabet of the mid-seventh centuryb c .34
Even in the abecedaria of the copper plaques, the unmarked stance can
be seen. There is a single occurrence of CP Alpha-1 in reversed stance: plaque
MS1-1, line6.
Yet a different alpha morphology is found in the abecedaria of the plaques,
herein denoted Alpha-2:
(2) Alpha-2
Alpha-2 commonly has a distinctive horizontal (or nearly so, rather than
markedly diagonal) crossbar stretching across the base of the letter. Its overall
shape varies from that of an isosceles triangle to that of a right triangle, giving
the letter a distinct delta-like appearance. In a few instances, however, the
base is perceptibly diagonal, giving the character the general shape of a sidelong delta (as, e.g., on plaque W-1, line1).
The archaic Greek alphabets provide scanty evidence for an alpha of the
CP Alpha-2 letter shape at an early period: an example appears in an alphabet on a sherd from Mount Hymettos (H165/184; seventh century b c ).35
Concerning the five well-preserved letter shapes in this fragment of an abecedarium, Langdon writes that the delta resembles a Boiotian or Euboian type
[i.e., R], but this is probably only the result of a slip of the stylus in the hand
20
of an inexperienced writer. The other letters in this line are also misshapen
and should probably be explained likewise.36 To etch the arcing delta of the
Boiotian or Euboian type would, however, seemingly require more, not
less, skill and experience, and one wonders if it is accurate to characterize the
other letters as misshapen: heterogeneity of letter shapes in the graffiti from
Hymettos is not unusual, as Langdon and other investigators have acknowledged, and as we shall see as we proceed through this investigation.
Examples of the CP Alpha-2 type can be found in the late archaic and early
classical periods. In his study of the Attic script, Immerwahr includes in his
chart of alpha forms one having the shape of an isosceles triangle.37 He notes
occurrences on red-figure pottery on an oinochoe by the Harrow Painter,
on an amphora by the Oionokles Painter, and on a psykter by Oltos.38
A third form of alpha occurring in the abecedaria of the plaques, Alpha-3,
has a shape fundamentally like that of Alpha-2, except that the front sloping
side (i.e., the side facing the direction of writing) extends diagonally downward below thebase:
(3) Alpha-3
As with Alpha-2, archaic alphabets are reticent to offer up forms of alpha
matching that of CP Alpha-3. An alpha with similar morphology occurs on
a bucchero fragment from Massarosa, circa the late seventh or early sixth
century b c . The stance of this letter from Etruria is the same as that of CP
Alpha-3 (i.e., with the extended stroke occurring at the front of the letter);
however, the body of the letter has the shape of a right triangle (with the right
angle on the front side), and the long stroke descends vertically, with no diagonal tilt.39
Other examples are provided by sixth-century materials. A right-triangular
alpha with a back stroke not front that extends vertically not diagonally
downward (and so much like the alpha from Massarosa, but with reversed
stance) occurs in the Attic alphabet: Jeffery notes its use circa 525500 b c .40
The same type of alpha occurs in the aforementioned legal text from Chios
(ca. the first half of the sixth century b c ), the one that also utilizes alpha of the
type CP Alpha-1, as well as (and more often) the comparable form in reversed
stance (i.e., the marked and unmarked varieties, respectively).41
In his chart of Attic letter forms, Immerwahr identifies an alpha closer in
shape to that of CP Alpha-3 with a diagonal stroke extending downward
beyond the base but, again, its stance is the reverse of that of the form of the
copper plaques.42 Immerwahr remarks that it is one of several forms of alpha
21
used on the Franois Vase (570560 b c ) an object that exhibits a wide variety of [alpha] shapes, including the marked and unmarked varieties of the
CP Alpha-1 type and that it also occurs on that set of vessels dubbed the
Little Master cups.43
In the Aegean, a similar alpha like CP Alpha-3 but in reversed stance is
found in an inscription from the vicinity of Amphipolis, a commemoration
composed for a Parian, Tokes, who was killed at Eion (first quarter of the fifth
century b c ). The alphabet is identified as that of either Paros or Thasos, both
Ionic Aegean scripts.44
The evidence, such as it is, appears to suggest that the stance of CP Alpha-3
is atypical: the reversed stance is the common orientation for an alpha of this
type. As with CP Alpha-1, the stance of Alpha-3 represents the marked case.
Whatever its significance, the agreement in the marked status of these two
forms of alpha in the copper plaques, relative to occurrences elsewhere, is of
interest.
In addition to the three forms of alpha described thus far, there occurs
within the abecedaria of the copper plaques still another alpha morphology.
This symbol incorporates two structural elements already encountered, but
utilizes them differently: (1) its shape is triangular (i.e., delta-like), like that of
CP Alpha-2 and Alpha-3; and (2) the letter is produced by elongating one of
the three strokes of the triangle, la Alpha-3 (and its reversed-stance variants
found in other archaic alphabets). In this instance, however, the lengthened
segment is not the front stroke of the triangle (or the back stroke in the case
of the reversed-stance variants) but the base of the triangle. The base extends
forward like a proboscis pointed toward the direction of writing:
(4) Alpha-4
The distribution of this symbol within the abecedaria of the three plaques
is limited, and the number of occurrences is low: it occurs four times in the
abecedaria of plaque W-2, where it is interspersed among occurrences of
Alpha-2 and Alpha-3, and one time on MS2-1. This form of alpha is, however,
quite distinctive and recurs, if not often, and for this reason I have classified
the morphology as a separate Alpha-4 type. A comparable morphology in
other archaic alphabets is wanting.
Setting aside the limited attestation of Alpha-4, the various forms of CP
alpha are broadly, if not evenly, distributed across the abecedaria of the six
plaque faces. The most common form of alpha on the plaques is, perhaps
unexpectedly, the delta-like Alpha-2, a morphology that is not well evidenced
22
elsewhere until late, though already attested in the seventh century. Within the
abecedaria of the plaques, Alpha-2 occurs a total of thirty-three times, slightly
more than the combined number of occurrences of Alpha-1 and Alpha-3, fifteen times and seventeen times, respectively. Alpha-1 is absent from both W-2
and MS1-2 but is the predominant form of alpha on the two sides of plaque
MS2, occurring eight times on MS2-1. The three forms of alpha are most
evenly distributed on MS2-2, showing four occurrences of Alpha-1, three of
Alpha-2, and five of Alpha-3.
The full distribution of the several forms of alpha across the six plaque
faces is presented as follows:
(5)Distribution of CP Alphatypes
W-1
W-2
MS 1-1
MS 1-2
MS 2-1
MS 2-2
TOTAL45
-1
2
0
1
0
8
4
15
-2
6
5
7
7
5
3
33
-3
3
2
2
4
1
5
17
-4
0
4
0
0
1
0
5
Accounting for the original morphology and morphological development

of alpha is not a neat affair. In his treatment of the earliest Greek alphabets
and their relationship to the Phoenician consonantal script, McCarter observes
that the reconstruction of the first alpha is problematic. It was apparently of
uncertain stance. The angle of the nose was probably rather narrow. Much more
than this cannot be said.46 Jeffery sets up a hypothetical intermediate form of
alpha (actually somewhat close to the alep of certain eighth-century Phoenician
inscriptions, such as that on the base of a statue of Astarte from Seville)47 to
allow for a transition between the Phoenician precursor and various local forms
of alpha (excluding that of the Dipylon oinochoe, a form that she imagines that
Greeks in Posideion took over from Phoenician script: a formal letter written
by a Greek from Posideion [?] or Athenian[s] who had visited there).48
While the uncertainty of the pathway by which the form of Phoenician
alep developed into archaic Greek alpha is tangential to the central concern of
this study, the co-occurrence at an early period of the several alpha morphologies identified here invites cautious reflection on the relationship of the four
23
types to one another and, hence, on an evolutionary pathway. One could well
imagine a linear progression for instance, that CP Alpha-3 was the developmental precursor of the triangular (i.e., delta-like) Alpha-2 and, further, that
Alpha-1 was itself the evolutionary precursor of Alpha-3: that is to say, Alpha-1
was rotated forward so that the diagonal crossbar became (nearly) horizontal, producing Alpha-3; a clipping of the resulting elongated front diagonal of
Alpha-3 then produced Alpha-2. The rotation that produced Alpha-3 would
be accompanied by an adjustment of the uppermost angle of the triangle to
keep the front side of the triangle, with its lengthy stroke, diagonal. Without
such an adjustment, the elongated stroke would be approximately vertical, as
in the case of the morphology of forms from Athens and Chios, noted earlier
(though with reversed stance, relative to the CP forms). But what then would
one do with Alpha-4 with its forward horizontal extension?
An alternative, perhaps more enlightening, interpretative approach would
be one that seeks to relate each of the several types of alpha found in the
copper plaques directly to a single proto-form. For example, if one were to
superimpose the morphologies of Alpha-2, Alpha-3, and Alpha-4 one upon
the other, the resulting composite form would look like the letter depicted
in(6):
(6) Proto-form Alpha
The formal starting point for such a morphology would likely not be the
early Greek sidelong alpha, attested on the Dipylon oinochoe, but rather an
eighth-century Phoenician alep of the type used on the jug belonging to nt
(so its five-letter inscription claims) from the temple of Astarte at Kition in
Cyprus (RS 1524; first half of the eighth century).49 A ninety-degree clockwise rotation of the symbol would provide a suitable pre-form, as depicted
in(7):
(7) Pre-form Alpha
The proposal of a ninety-degree rotation in the development of alpha is not

in itself unique to the interpretation being explored here. McCarter allows
that it is not impossible that early upright alpha is an unrotated modification
in detail of a primitive sidelong form but then contends that the weight of
the evidence is against this,50 though he seems to envision that it is a primitive sidelong alpha such as that of the Dipylon oinochoe that experienced the
rotation.51 In any event, ninety-degree rotation is evidenced in the history of
alep itself, as in the Canaanite inscription on the fragments of a bowl from
24
Qbur el-Walaydah, circa 1200 b c , about which Cross notes: The two aleps
are certain, though their A stance is paralleled only in the Beth-shemesh
Ostracon, and, of course, in the early Greek alphabet.52
Without speculating on scribal causes and motives (e.g., cognitive processing, graphemic streamlining, script interference [the graphic equivalent of
the linguistic phenomenon of speaking with an accent]), one can immediately discern that the clipping of the right horizontal and right diagonal of
the rotated form of (7) produces the reconstructed composite of (6), which of
course will then generate (diachronically) the individual graphemes Alpha-2,
Alpha-3, Alpha-4 upon which the reconstruction is based.
But now, what of Alpha-1? Alpha-1 is distinguished from Alpha-2/3 by
the presence of a diagonal rather than a horizontal crossbar. The morphology of Phoenician alep is saliently marked by the presence of diagonal
strokes (as is the primitive sidelong alpha of the Dipylon oinochoe). If the
early evolution of the Greek alphabet was a process that occurred in the context of a continued presence of and continued dialogue with the Phoenician
consonantal script, as must certainly be the case, then one could plausibly
propose that the adjusting of the crossbar of an early Greek alpha, such as
CP Alpha-2/3, to a diagonal position, as characteristic of Alpha-1, is consequent to that continued Greek-Phoenician interfacing. The unmarked
equivalents of CP Alpha-1 and Alpha-3 would have a similar history, beginning also with a form like the reconstructed composite symbol of (6), but
one having, in effect, a reversed-stance that is, a symbol generated from
the rotated form (7) by clipping the left (rather than right) horizontal and
diagonal.
2.2
be ta
We can divide the betas of the CP abecedaria into two fundamental types: the
first, Beta-1, is characterized by loops that do not meet at the midpoint of the
vertical leg (with varying degrees of separation small in some instances and
large in others):
(8) Beta-1
This is a known archaic form of beta, otherwise attested in the alphabet of
Attica in the seventh century b c ,53 as in the partial abecedarium on a plate
from Mount Hymettos (H 2, having pointed, however, rather than rounded,
loops).54
25
The alphabet used in the eighth-century (and later) rock graffiti from Thera
shows a somewhat similar form, though a typologically more primitive one,
having a small hook or open loop at the top.55 This Theran beta is given priority typologically in that it is morphologically close to Phoenician bet, except
inverted in stance.56 Some occurrences of CP Beta-1 are reminiscent of Theran
beta to the extent that the upper loop is noticeably smaller than the lower, as
in the abecedaria of MS2-2, lines 14 and20.
CP Beta-2, in contrast, shows loops that touch one another, though it displays several variant shapes. One type has more-rounded loops, a second
more-pointed loops (cf. the Hymettos beta of H 2 mentioned earlier); in both
cases the loops meet at about the midpoint of the vertical spine of the letter.
These two types with more-rounded and more-pointed loops are jointly
designated as CP Beta-2:
(9) Beta-2
The two varieties occur widely in local archaic scripts. Immerwahr notes that
both are found on Attic vases dating from the seventh century to the early
fifth.57 A seventh-century example of the pointed variety can be seen in the
abecedarium of the Hymettos sherd H165/184,58 mentioned earlier in the discussion of CP Alpha-2. The variety with more-rounded loops is attested early
and broadly: examples are to be found on a small vase from Eretria (FK9856),
second half of the eighth century b c ;59 on a loom weight from the Athenian
Agora (MC 907), late eighth or early seventh century;60 in the Boiotian verse
inscription that snakes across the thighs of a bronze statuette of a warrior,
dedicated to Apollo by one Mantiklos, circa the late eighth or early seventh
century;61 and in abecedaria from Etruria, preserving the script of the Euboian
Greeks from whom the Etruscans learned the alphabet, such as that inscribed
on an ivory writing tablet found in the necropolis of Marsiliana dAlbegna,
dated to the second quarter of the seventh centuryb c .62
A distinct type of CP connected-loops beta is characterized by having
one loop that intersects with the other (rather than both loops connecting to
the spine directly). Such a beta occurs in the abecedarium on the previously
mentioned cup from the sanctuary of Hera in Samos (eastern Ionic), dated
to circa 660 b c ,63 and another can be seen in one of the graffiti from Mount
Hymettos (H 192).64 McCarter is likely correct in seeing this form as intermediate between the open-loop beta of Thera, noted previously, and the
standard closed-loop forms of beta; he writes: The form resembles Theran
beta except that it is inverted, thus recalling the stance of Phoenician bet; the
26
lower stroke loops back not to the shaft but some distance forward on the
upper loop.65 This Samian type of beta also occurs on the copper plaques and
is designated Beta-2i:
(10) Beta-2i
The distribution of CP Beta-2i is limited: it occurs on plaques MS1-1 at line
15 and MS2-2 at lines 11 and 17, and probably on line 16 as well, based on the
X-ray examination, though the morphology of the symbol is close to Beta-2.
A less distinct example is found on plaque W-1, line11.
Within the abecedaria of the copper plaques, the far more common type of
Beta-2 is one that is characterized instead by the upper stroke looping down
onto the lower stroke; in other words, the stance is inverted when compared
to that of the Samian type of Beta-2 and, thus, displays the same orientation
as the Theran beta vis--vis the Phoenician precursor. This beta is herein designated as Beta-2ii:
(11) Beta-2ii
Outside of the copper plaques, a distinct example of this type of beta appears
in Etruscan letters from Veii, painted on an aryballos, dated to the late seventh century b c .66 CP Beta-2ii occurs most frequently on the two Schyen
plaques and is especially common on MS1-1 and1-2.
The most frequently occurring form of beta in the abecedaria of the copper
plaques is Beta-2 (with a total of twenty-eight occurrences), a frequency in
keeping with the data reported in Jefferys Table of Letters at the end of LSAG2,
where the same type of beta is presented as the most commonly occurring
typical form of the letter among the local Greek alphabets. Only slightly
less frequent in the copper plaques is Beta-2ii (twenty-three occurrences),
followed by Beta-1 (nineteen occurrences). As just noted, Beta-2i is found
five timesonly.
The full distribution of the several types of beta in the copper plaques is as
follows:
(12)Distribution of CP Betatypes
W-1
W-2
MS 1-1
-1
0
7
0
-2
4
4
3
-2i
1
0
1
-2ii
7
0
7
MS 1-2
MS 2-1
MS 2-2
TOTAL67
2.3
2
6
4
19
4
6
7
28
*
0
0
3
5
27
6
2
1
23
g a mm a
Throughout the copper plaques, gamma is morphologically quite consistent.

(13) Gamma
It is the shape of gamma characterized by Jeffery as the earliest form in most
places and is widely attested among local archaic alphabets.68 Its morphology, however, is typologically further removed from the Phoenician precursor, *gaml (Hebrew gimel), than is that of the early gamma of the shape
g.69 A variant with the crossbar angled upward (at an obtuse angle) is also
attested early,70 as in the aforementioned seventh-century abecedarium of the
Euboian-based Marsiliana writing tablet,71 though the form does not occur in
many of the other Etruscan abecedaria, which use instead the lunate gamma
(or a more angular variant). Guarducci summarizes:
I due tipi e , molto vicini al modello fenicio, si affermano quasi dovunque
nella Grecia. In alcuni luoghi per (p. es. Corinto e colonie, Megara, Sicione,
Elide, Arcadia, colonie euboiche) simpone il tipo a semirombo (c), poi addolcito in un semicerchio (1), tipo evidentemente nato dal desiderio di non confondere il gamma con il lambda o col pi ad uncino.72
The isosceles gamma (G) is regular in Attic script from earliest times;
Immerwahr observes that Ionic gamma [i.e., a gamma like that of the copper
plaques] first appears in some quantity in about 480460 b c and that the
question of the shape of [gamma] is bound up with the introduction of the
Ionic alphabet.73
Outside of the copper plaques, perhaps the earliest example of a gamma like
that of the plaques though reversed in stance is found in a partial abecedarium inscribed on the base of a conical oinochoe from Kyme, late eighth
or early seventh century b c ; the graffiti on this pot are particularly interesting, and we shall return to them further along.74 A similarly dated example of
a gamma having a morphology comparable to that of the copper plaques is
provided by the inscription on the Boiotian Mantiklos statuette.75
28
Several early examples come from Ionic areas. The abecedarium on the
cup from the Heraion of Samos (ca. 660 b c ; eastern Ionic) shows such a
gamma.76 A CP type of gamma is also found in early inscriptions from the
Ionic islands of Amorgos, Samothrace, and Tenos:77 as in the rock-carved epitaph of Deidamas (IG XII vii 442; from Amorgos but in Naxian script), first
half of the seventh century b c ,78 and the grave stele of Demainete (IG XII vii
141; from Amorgos),79 possibly mid- to late sixth century.80
2.4
d elta
Throughout the copper plaques, the shape of delta is generally consistent,

being that of an approximately equilateral triangle, not uncommonly tilting
slightly toward the direction of writing:
(14) Delta
The equilateral-triangle delta is a widely attested archaic form, though its earliest examples tend to have a base that is shorter than the two sloping sides (and
thus not fully equilateral). A delta with similarly compressed proportions also
occurs in the copper plaques, as in the abecedaria of W-2, line 11, and MS1-2,
line 3. Seventh-century occurrences of this type of horizontally compressed (i.e.,
isosceles) delta are found in, for example, the graffiti from Mount Hymettos;81
in the Naxian inscription on the base of a kouros from Delos, dedicated by
Euthykartides (SEG XXIII 504);82 in the rock-carved epitaph of Deidamas
from Amorgos;83 in the memorial inscription to Glaukos son of Leptines that
comes from Thasos (SEG XIV 565);84 and in a graffito etched into an amphora
from a childs grave on Thera (IG XII iii 986).85 The same delta morphology
shows up on one of the incised sherds from the Potters Quarter at Corinth86
sherds to which some specialists have assigned an eighth-century date.87 A
probable eighth-century example is provided by the inscription on a pithos
from Phaistos in Crete, discussed earlier in conjunction with CP Alpha-1.88
In this same era and in these same places, among yet others, a slightly different (horizontally compressed) isosceles delta is frequently found one
with a (nearly) verticalized sloping stroke on the side facing away from the
direction of writing. This variation in the back stroke and angle often creates
the appearance that the letter is leaning away from the direction of writing
(and at times the base of the triangle inclines slightly toward the direction of
writing), or it produces a right triangle. Such a delta occurs circa 750700b c
in the verse inscription on a skyphos from Pithekoussai, the so-called cup of
29
Nestor89 (CEG 454; for the text of the inscription, see the later discussion of
CP pi), and the damaged delta in the inscription of the still earlier Dipylon
oinochoe from Athens appears to be the same.90 Examples also occur in the
graffiti from Mount Hymettos in Attica.91 The occurrences of delta on the
Naxian inscription of the Nikandra statue from Delos92 (discussed later) tend
to be of this form. Sometimes these two variants of the horizontally compressed delta occur even within a single inscription, as in the epitaph of
Deidamas from Amorgos; in rock graffiti from Thera (eighth century b c and
later);93 and in the mid-seventh-century Cretan Dreros legal fragments.94
In some local Greek scripts there also occur forms of the compressed, isosceles-triangle delta that lean toward the direction of writing as the deltas of
the copper-plaque abecedaria often lean slightly toward the direction of writing. Again, this forward-inclined type of delta may co-occur with forms having
a comparatively upright stance. For example, such a delta is found in the Dreros
inscriptions from Crete, together with the two morphologies just noted.95
The typical CP delta one that is more fully equilateral is, however, also
seen elsewhere at an early period. On the basis of the rendering of the inscription carved into the now lost grave stele of Demandros, an example occurs in
late seventh- or early sixth-century Samos, the earliest Samian inscription on
stone from its appearance.96
Absent from the copper plaques are the common archaic forms of delta D
and D. Jeffery97 typifies theseas:
Essentially mainland forms (Argos, Elis, Lakonia, Euboia,98 Boiotia, Phokis,
Lokris, Achaia); they are not normal in the Aegean islands or in Asia Minor,
but occur once in Aiolis, in graffiti from Larisa.
2.5
epsil o n
Epsilon too has a relatively consistent appearance throughout the copper

plaques. The character is formed of a vertical spine with three horizontal bars
that meet the spine at a right angle. The spine continues below the lowest horizontal, typically producing a pronounced tail, though of variable length:
(15) Epsilon
Occasionally the tail flairs away from the direction of writing (i.e., to the
right) as it descends, as, for example, in the abecedaria of W-1, lines 12 and 17;
MS1-1, lines 1, 2, 5, 8, 14, and 16; MS1-2, line 17: this feature also characterizes
30
the spines of other CP letters at times and may be considered a scribal idiosyncrasy in many though perhaps not all cases. In some instances epsilon
tilts slightly in the direction of the writing (i.e., to the left), but the rightangled join of the horizontal bars is generally preserved (as in W-1, line 10),
or else the bars remain roughly horizontal in spite of the leftward lean of the
spine (as in W-1, line14).
This CP epsilon differs notably from most archaic forms of the letter to the
extent that outside of the copper plaques, early epsilon typically shows crossbars that slope downward toward the direction of writing, intersecting with
the spine at an acute angle. In this way, the typical archaic epsilon is typologically closer to its Phoenician precursor, he (having the idealized shape h) than
is the CP epsilon, though the he of the early tenth-century Ahiram sarcophagus and tomb inscriptions likewise shows fundamental right-angularity.99
An archaic tailed epsilon with vertical spine and right-angle crossbars is,
however, not unique to the copper plaques. An epsilon with approximate
right-angularity appears in the partial abecedary on the late eighth- or early
seventh-century conical oinochoe from Kyme mentioned in Section 2.3; the
spine of this epsilon extends slightly above the top crossbar a feature that is
not typical of the CP epsilon, though instances of such a spiked form do occur,
as in the abecedaria of plates W-1, line 4, and MS2-1, line 10.100 Similar to this
Kyme epsilon is that one found on an amphora from the Eretrian colony of
Methone ( 2425), also dated to the late eighth or early seventh century b c
(and characterized by a spine that sways markedly away from the direction
of writing at its lower end).101 Right-angle epsilons can also be found in the
seventh-century graffiti from Mount Hymettos in Attica: for example, at least
one of the two tailed epsilons in the fragment H 505, bearing the letters ,
has conspicuous right-angularity.102
Sixth-century examples can be identified in alphabets of the Peloponnese.
Two such forms appear in the proper name inscribed around the
rim of a bronze hydria from Laconia, dated to the early sixth century.103 Also
from Laconia, a circa late sixth-century bronze aryballos from the Menelaion
bears a dedication to Helen and Menelaus in which occur examples of tailed
epsilons with right-angular bars (as in the abecedaria of the copper plaques)
alongside tailed epsilons with sloping, acute-angular bars.104 The CP type of
epsilon also occurs in Arkadia in the late sixth century: Jeffery notes its use on
early coinage of Heraia and, from Melpea, its occurrence on a bronze statuette of a herdsman dedicated to Pan,105 bearing the earliest attested alphabetic
spelling of the gods name.106
31
In a few instances of CP epsilon, the tail is exceptionally short.107 Close

examination of X-rays reveals the presence of a stubby tail, almost imperceptible, on the epsilons of W-1, lines 7 and 16, and MS2-1, lines 6 and 21. Even
the shortest of tails is difficult to identify on the epsilons of W-2, lines 5 and
14, owing perhaps to the X-ray superimposition of strokes from the opposite side of the tablet. In these instances, the letter has an appearance that
approaches the shape of the developed classical style.108 Yet even this morphology is attested quite early: compare the several occurrences of epsilon on
a Protocorinthian aryballos from Kyme (IG XIV 865), circa the second quarter of the seventh century b c ; these characters show a pronounced forward
tilt, but the right-angular joins of the crossbars are preserved, and the tilt parallels that of the spines of various other letters in the inscription.109 Tail-less
right-angular epsilons also occur in the seventh-century graffiti from Mount
Hymettos, as in the inscriptions of, inter alia, H 130, perhaps preserving the
name of Heracles,110 and H 189: h he himself wrote [this].111
2.6
d ig a mm a
The digamma, or , of the copper plaques morphologically parallels epsilon,

being formed by a vertical spine and right-angle horizontal crossbars two,
rather than the three of epsilon:
(16) Digamma
The morphology of digamma, mutatis mutandis, parallels that of epsilon in
the abecedaria of the copper plaques. Thus, this character occasionally tilts
slightly toward the direction of writing, but the spine retains right-angled
joins to its two crossbars; or, alternatively, and more often, it preserves the
fundamental horizontal orientation of the crossbars. In a few instances the
bottom of the spine flairs to the right: for example, W-1, line 3; W-2, lines 3,
8, and 10; MS1-1, lines 1, 2, 4, 8, 10, and 14. Twice on plaque MS2-1 (lines 2
and10), a serif appears at the left end of the top crossbar, giving the character
a gamma-like appearance.
Outside of the copper plaques, archaic digamma commonly agrees with
epsilon in having crossbars that slope downward, meeting the spine of the
letter at an acute angle. It seems to have been assumed by some investigators
that (1) this type of digamma historically preceded the right-angular variety
of the copper plaques an archaic variety that is attested elsewhere and
(2) that the development of a more primitive acute-angular digamma into a
32
more innovative right-angular digamma followed, or paralleled, a comparable

evolution of epsilon (see, e.g., LSAG2 25). One cannot, however, convincingly
claim that the acute-angular digamma is typologically more primitive than
the right-angular form of the copper plaques, in that Greek digamma ,
that is in all of its variants is morphologically quite distant from its Semitic
namesake waw, the character that occupies the comparable position in the
Phoenician consonantal script. McCarter puts it plainly: Greek [] corresponds in form to no known Phoenician letter.112
The CP type of right-angular digamma is no less than the CP type of
epsilon attested quite early, however.113 The previously mentioned partial
abecedarium on a conical oinochoe from Kyme, late eighth or early seventh
century b c , displays not only an epsilon with approximate right-angularity
but a digamma with comparable morphology as well.114 For right-angularity,
compare the digamma on a skyphos from Eretrian Methone ( 2250), late
eighth or early seventh century, which, however, shows a spine that descends
at an angle, toward the direction of writing, and a top crossbar that bends
downward at its distal end.115 Particularly interesting is the use of the character in a dedicatory inscription running around the rim of a bronze lebes
from Boiotia, dated circa 700675 b c , a prize awarded at funeral games for
Ekpropos (IG I2 402). In this inscription, which will have bearing on my discussion of other letter forms from the copper plaques, right-angular digammas ( Isodikos [proper name]) are used side by side with acuteangular epsilons ( dedicated).116
The present state of the evidence does not obviate the possibility that the
right-angular digamma is older than the acute-angular digamma and that the
latter form, with its sloping crossbars, arose secondarily under the influence of
its alphabetic companion, acute-angular epsilon. Conversely, in some scribal
settings a primitive right-angle digamma may have motivated the developmental shift of acute-angle epsilon to right-angle epsilon.
Owing to the relatively early phonological loss of the glide [w] in several
Greek dialects, digamma is less robustly attested than many letters of the
Greek alphabet. I quote here Jefferys succinct summary of the occurrence of
the sound in the dialects and of the symbol in the epichoric alphabets:
In the Peloponnese, northern and central Greece (except Attica), Crete, Euboia
(through Boiotian influence?), [w] was normal in the dialect. In Attica and
Naxos it is attested in poetry only. In Rhodes, Thera, Melos, and Ionia it is
not attested, nor in the Cyclades except for Naxos. But the early abecedaria
33
of Attica and Amorgos, and the archaic Ionic abecedarium preserved in the
so-called Milesian alphabetic numeral system, all show that vau [i.e., ]
remained in its place in the row, as san and qoppa also appear to have done.117
Greek digamma, like Phoenician waw, spells a labiovelar glide /w/, a sound
that as just described eventually disappeared from all ancient Greek dialects (at least phonemically; on the survival of a nonphonemic [w], see under the
subsequent discussion of iota), and quite early in some cases. Oddly, in contrast to what we observed regarding the enigmatic morphology of digamma
vis--vis its Semitic precursor, Phoenician waw transparently provides the
raw graphic material for Greek upsilon, the symbol that spells the vocalic
counterpart of the glide /w/, that is, the vowel /u/. This letter upsilon was
the original supplemental character, being added directly to the end of the
Phoenician script, which terminates in taw, source of Greek tau. The alphabet
preserved in the copper plaques predates the appending of upsilon, unlike all
other known varieties of the Greek script, and so the synchronic alphabetic
system attested by the CP abecedaria must surely be one in which digamma
serves to spell both the consonant /w/ and the corresponding vowel/u/.
The use of digamma in place of upsilon for spelling /u/ (or its secondarily
fronted counterpart //) is an orthographic practice that is indeed occasionally
attested, notably in the alphabets of Argos and neighboring Kleonai, but elsewhere as well. From a sixth-century Kleonaian dedicatory inscription found
at Nemea (circa 560 b c ; the dedication is that of Aristis, the son of Pheidon,
who four times won the pankration at Nemea one of the earliest inscriptions found at Nemea)118 comes the form h for son.119 This inscription is also interesting for graphically distinguishing the eta-vowel from the
epsilon-vowel: while the latter is written with e, the former is spelled with the
Corinthian symbol B (which serves in the Corinthian alphabet to spell both
long- and short-e vowels), a graphic distinction that Jeffery identifies as characterizing the scripts of Kleonai, Phleious, and Tiryns.120 A mid-fifth-century
Argive121 inscription recording a treaty made under Argive auspices between
the Cretan cities of Knossos and Tylissos preserves the spelling [ for
, the name of the Dorian festival in honor of Hyacinthus.122 As we shall
see, this same inscription (SEG XI 316)is particularly important for informing
our understanding of another orthographic practice reflected in the copper
plaques. Digamma is also used at times to spell the second component of a
Vu diphthong (where V = vowel), as in fifth-century Locrian (i.e.,
with for ).123
34
Perhaps also germane to the claim that earliest Greek digamma was phonetically bivalent is the Pamphylian orthographic practice of using two
distinct forms of digamma in tandem. In addition to the symbol , the
Pamphylian alphabet includes a letter N (shown dextroverse); both characters are employed to spell w, or at least reflexes of earlier *w. The distribution
of the two symbols is in part complementary, and this distribution has hence
been taken to suggest some variation in the phonetic values of and N.124 The
origin of the Pamphylian digamma-character N is debated, though it seems
probable that, whatever its origin, the symbol is related to a form of digamma
in use on Crete, the letter (shown dextroverse), which occurs in lieu of at
Rhizenia, Axos, and Eleutherna.125
In a recent study of the relationship between the Greek and Phrygian
alphabets, Claude Brixhe has similarly taken note of the probable case that,
at a sufficiently early point in the history of the Greek and Phrygian scripts,
a single symbol must have been used to spell both the w-consonant and
u-vowel. Brixhe observes that il est vraisemblable que le ww tait primitivement utilis pour les voyelles /u(:)/ et la semi-consonne /w/.126 For Brixhe the
bivalent status of the symbol is bound up intimately with Phrygian-specific
developments in the spelling of the palatal glide /y/ and its vocalic counterpart /i/. The probability of a corresponding Greek practice will be germane to
my subsequent analysis of the orthographic system reflected in the abecedaria
of the copper plaques.
2.7
z e ta
The appearance of zeta in the copper plaques is an intriguing matter. Two

quite distinct forms of the character are used, one of which is a typical archaic
zeta, widely practically ubiquitously attested in the various local Greek
alphabets:
(17) Zeta-1
The second type, while no less familiar in appearance, is attested in no archaic
alphabet other than the one preserved in the copper plaques. It is in effect
the zeta of the classical era, though the stance of the CP symbol is given the
right-to-left (sinistroverse) direction of writing on the copper plaques the
reverse of that of the typical classical Z-shapedzeta:
(18) Zeta-2
35
For the sake of convenience, I will refer to this form of the letter as
Z-shaped, but it is important to bear in mind that its stance is the reverse of
later Greek and RomanZ.
This Z-shaped zeta occurs on all six plaque faces, though most frequently
on the two sides of MS2, where it is the predominant form of zeta on each.
At the time of his study of the Wrzburg plaque, Heubeck was aware that
the H. P. Kraus catalog reported the occurrence of Z-shaped zeta among the
characters on the New York plaques those plaques eventually acquired by
Martin Schyen. Owing to the heavy patina of the Wrzburg plaque, however,
Heubecks examination of that document failed to reveal the six occurrences
of the Z-shaped zeta inscribed thereon and found almost as frequently as
the archaic I-shaped zeta on side 2. And thus he wrote: Das hat die
Normalform; es ist bemerkenswert, da auf einer der New Yorker Tafeln sich
gelegentlich neben dieser Form auch das spter bliche klassische (Z) findet, das noch in allen archaischen Alphabeten fehlt.127
The distribution of Zeta-1 and Zeta-2 forms (the respective shorthand
denotations and Z will be used in the presentations of Chapter4) is summarized as follows:
(19)Distribution of CP Zetatypes
W-1
W-2
MS 1-1
MS 1-2
MS 2-1
MS 2-2
TOTAL
Zeta-1 ()
Zeta-2 ()
11
6
10
9
3
5
44
2
4
2
2
12
9
31
Included in this chart is the occurrence of Zeta-2 in the next to last line of
MS2-1, where the scribe appears to have superimposed Zeta-1 and Zeta-2 one
upon the other.
Earlier investigators have commonly supposed the classical Z-form zeta
(like CP Zeta-2, but with reversed stance) to be typologically later than the
archaic I-form zeta (the type of CP Zeta-1) and, thus, have held that the Z-form
arose secondarily. Jeffery, for example, in her general discussion of the archaic
Greek letters, simply states: The form remained close to the Phoenician. 128
36
Zayin, the Phoenician precursor of Greek zeta, is indeed morphologically

very close to that widespread archaic I-form zeta, consisting of two horizontal
crossbars at either end of a central vertical spine (z). In the Phoenician script
appearing on the Kition bowl from Cyprus (ca. 800 b c ), however (discovered
subsequent to Jefferys original publication of LSAG), zayin has a pronounced
Z-shape. In his study of the bowl and its enigmatic inscription, Coote writes
(of line1):
The zayin is an extraordinary anachronism. Though well established by the
end of the ninth century in Aramaic, in lapidary Phoenician the z-form does
not replace the I-form until the sixth century. It cannot be considered a mistake or a faulty yod since the phrase rz follows ml again in line 4 and appears
in line 3, and all three zayins are formed in the same manner.129
McCarter also calls attention to the uniqueness of the Phoenician zayin from
Kition and its similarity to instances of ninth-century Aramaic zayin;130 in his
discussion of the development of Greek zeta, he rightly observes:
The cursive option of z-form zayin existed at the time of the divergence of the
Greek alphabet. Much later z-form zeta in Greek may have been an independent development. It is possible, however, that the form was used in Greek
from the beginning in materials that have perished.131
The appearance of both I-form and Z-form zetas in the copper plaques offers
prima facie support for just such an early Greek acquisition of both forms of
Phoenician zayin.
The orientation of the zayin of the Kition bowl is like that of the copper
plaques and, hence, the reverse of that of the classical period zeta. That the
Z-shaped zeta should have reversed its position in its evolution to the typical
classical form would come as no surprise. Compare, for example, the variation in stances of alpha discussed in Section 2.1 and that which is common
among early forms of three- and four-stroke sigma.
2.8
e ta
There are both functional and structural issues at play here: two different etaforms occur in the abecedaria of the copper plaques; the distribution of the
two forms within the abecedaria suggests a synchronic bivalency of function
for each form. The two forms share an obvious graphic similarity. Among the
archaic Greek alphabets in which they are otherwise attested, however, these
37
two symbols typically diverge sharply from one another in their respective
phonetic values; moreover, one of these symbols in its own right shows a variation in phonetic value between local alphabets, and a bifurcation in phonetic
value even within individual alphabets. The first of these is the Greek character derived from the Phoenician letter het, representing in the Phoenician
language a voiceless pharyngeal fricative [].132
(20) Eta
This grapheme is typologically early and occurs widely among the various
epichoric Greek alphabets, where the shape is typically that of a rectangle with
greater vertical dimension than horizontal (i.e., taller than wide), h. At times
the symbol is found with additional horizontal crossbars,133 a variable that the
Greek letter shares with its Phoenician precursor.134 As illustrated in (20), the
CP forms of eta, in contrast, tend to be more squarish or to be rectangular
with greater horizontal dimension than vertical (i.e., wider than tall) though
not exclusively so: thus, tall thin etas occur, for example, in W-1, lines 1 and
9, and MS1-2, line 1. If, in fact, one compares lines 9, 11, and 12 of plaque face
W-1, one will see in these lines, respectively, a rectangular eta with greater
vertical dimension; a square eta; and a rectangular eta with greater horizontal
dimension. The more squarish CP eta is reminiscent of the het occurring in
the Phoenician script of the eleventh-century Nora fragment (CIS I 145).135
In some archaic Greek alphabets eta (whatever its particular graphic shape)
is used to represent a consonant;136 in others it represents a vowel; in still others it represents both a consonant and a vowel. To the first group belong the
alphabets of Attica, Euboia, Boiotia, Thessaly, Corinth, Megara, Argos, and
Laconia, among yet other locales: in these alphabets the symbol represents
the voiceless glottal fricative /h/ (the so-called spiritus asper). In contrast the
eta-symbol represents a vowel in alphabets of the Ionic Dodekapolis and of
Crete, as well in Aiolic graffiti from Larisa and Magnesia.137
More interesting are those alphabets in which the eta-symbol does double
duty, being assigned both a consonantal and a vocalic value. This is the case,
inter alia, in the alphabets of several of the Cycladic islands on which an Ionic
dialect was spoken. In the alphabet of Naxos, the eta-symbol h is used to spell
both the glottal fricative /h/ and the long vowel (which we might simply
denote as the eta-vowel, without delving into phonetic particularities) that
had developed secondarily from Common Greek * but it is not used to
spell the long vowel that continued earlier (i.e., Common Greek) *; for
spelling this inherited , the epsilon-symbol isused.
38
This Naxian practice is evidenced in the hexameters inscribed boustrophedon in three lines on the side of a statue depicting a female figure found
at Delos (ca. mid-seventh century b c ), the so-called Nikandra statue (SEG
XIX 507; mentioned in the discussion of delta).138 The lines are presented
below; h and transcribe h in its consonantal and vocalic uses, respectively.
In addition, , a variant form of h is used in the spelling of the consonantal
cluster /k+ s/ (i.e., the cluster more familiarly spelled by the xi-symbol x [in
dark blue alphabets]),139 also here transcribed per convention by h, as in
h, h, and hh:
h<>
| h, h ,
, | hh []
Nikandr m anethken h<e>kboli iokheairi
kr Deino|dik t Naksi, eksokhos aln,
Deinomenos de kasignt, | Phraks d alokhosn[un]
Nikandra dedicated me to the far-shooting goddess, arrows in her hand,140
Daughter of Deino|dikes the Naxian, standing out above others,
And sister of Deinomenes, | n[ow] the wife of Phraxos
Similar, if not identical, practices are seen elsewhere in the Ionic Aegean.
Jeffery offers summary observations (for ease of exposition, I have modified
her references to symbols and her transcription of the aspirate [i.e., the glottal
fricative /h/] in keeping with the practice used herein):141
In Naxos h / H was used (a) for h, (b) under certain circumstances for [as
described here]. By the early 5th c. [the Naxians] were using H for in all
circumstances, as well as for the aspirate.
Delos, Keos, and Syros show a confused use of h / H (a) for aspirate, (b) for
and sometimes for . [The latter confusion is also reflected in more recent
finds from Naxos.]142
Paros used h / H in the more common Ionic fashion for , and occasionally
also for the aspirate.
Siphnos used it for the aspirate ; possibly for (no examples as yet); not
for.
The earliest inscriptions to show the later form H appear to be Paros 28143 and
29,144 here conjecturally assigned to the middle and third quarter of the 6thc
39
A related vocalic practice characterizes the alphabet of early inscriptions from

Amorgos. Guarducci observes: Come a Nasso, anche qui ricorre luso di e
e di h per esprimere rispettivamente le lungo originario () e le lungo
secondario derivato da a lungo ().145 She here makes explicit reference to forms from the rock-carved epitaph of Deidamas (ca. 700650 b c ; IG
XII vi 442):146
| [] []
Didamani | Pugmas o patr [t]ond oik[on]
For Deidamas | Pygmas his father thistomb
To judge from a fifth-century inscription set up at Delphi, a related strategy

exists in the alphabetic system of Andros: Buck notes that in this inscription,
the eta-symbol and the epsilon-symbol are about equally common for spelling the -vowel that came from *, but that the epsilon-symbol alone is used
for spelling the inherited -vowel. The eta-symbol is used once for spelling /h/
(the spelling of which is omitted twenty-six times).147
The dual use of eta is not restricted to Ionic contexts, however. Still in the
Cyclades the archaic alphabet of the Doric island of Thera likewise uses
the eta-symbol h (and H) to spell both the consonant /h/ and the eta-vowel,
regardless of the historical source of that vowel.148 Farther east, Doric Rhodes
follows suit, using h (alongside H) for both the consonant and the eta-vowel149
(and at times for the short e-vowel as well).150 Knidos, Rhodess neighbor on
the Asian mainland, uses the symbol for the fricative /h/ and uses , a
modified form of the eta-symbol, for spelling the eta-vowel; this letter appears
to be the same as that one encountered in Naxian orthography (as noted),
used in the spelling of the /k + s/ cluster.151
There is evidence for a very early use of the symbol in the script of
Euboian colonists in southern Italy. The letter occurs in the two partial mirror-image abecedaria inscribed on the aforementioned Protocorinthian152
conical oinochoe from Kyme, circa the late eighth or early seventh century
B C .153 In both of the abecedaria, the symbol occurs in what must be the
intended position of eta, though the letter is placed mistakenly or otherwise between digamma and zeta (the last letter to be written in each of the
partial abecedaria); in other words, the positions of eta and zeta are reversed.
The value of in the abecedaria is perhaps /h/, the regular value of Euboian
eta. Though, could we possibly find here evidence of an early Euboian use
of the symbol for the eta-vowel? In the inscription that accompanies the
abecedaria on the oinochoe, the (different) eta-symbol h occurs with the
40
value /h/, if, as seems probable, the inscription is to be read as Greek.154 Such
a Euboian system a West Ionic alphabet would mirror that of Knidos, the
alphabet of which place is fundamentally Ionic.155
A recent epigraphic find has revealed a kindred practice in the orthographic system of the Sicilian city of Naxos, the colony of (West Ionic) Euboian
Chalkis and (Central Ionic) Aegean Naxos. This eta-variant occurs in the
late seventh-century Sicilian Naxian dedication to the goddess Enyo that was
mentioned in conjunction with my discussion of CP Alpha-1. In this inscription, as Guarducci has argued,156 the symbol appears to be used for the
glottal fricative /h/:157
[] | h | h[]
Lurako[s] | huros | Enhu[i]
As Guarducci proposes, [], Lyrakos, must be the name of the one

presenting the dedication; she interprets h [],158 the divine recipient, as a
spelling of the name of the war goddess Enyo. She is the deity whom Homer
knows as (ptoliporthos) city-sacker (Il. V 333) and as Ares
cohort (Il. V 592), and identified by Hesiod in his catalog of monsters as
(Enu krokopeplos) saffron-robed Enyo, a daughter of Phorkys
and Keto (Theogony 273). Guarducci interprets h as an ethnic adjective,
derived from the toponym Hyria, identifying it as a town of Paros that town,
she conjectures, mentioned by Plutarch (De exilio 602D) in conjunction with
a reference to Naxos.159 Yet one thinks also of Boiotian Hyria, located on the
Euboian Gulf, not far from Aulis according to Strabo (9.2.12), and so not far
from Chalkis that Hyria mentioned by Homer (Il. II 496)in the catalog of
Boiotian warriors who sailed againstTroy.
In her exposition of the inscription, Guarducci invokes an unpublished, mutilated graffito from a neighboring sanctuary (attributed to Ares
and Aphrodite) in which she believes to recognize another example of the
symbol .160 She also draws attention to the late seventh-century Aegean
Naxian inscription from Delos carved on the base of a kouros dedicated by
Euthykartides, mentioned in conjunction with my discussion of the delta of
the copper plaques;161 here, the eta-variant is not only used in the typical
Naxian way in spelling the /k + s/ consonant cluster but is also used, in lieu of
the expected Aegean Naxian h, to represent the glottal fricative /h/,162 as in the
inscription from Sicilian Naxos.163
As noted, a second symbol occurs in the position of eta in the abecedaria
of the copper plaques. Given the observations we have just made regarding
41
the use of eta-symbols in various Aegean alphabets and a related Sicilian

one it perhaps comes as no great surprise that the second CP eta-symbol
is a particular graphic avatar of the Greek letter xi, the symbol that represents
the consonantal sequence /k+s/:
(21) Xi
This same character of course also occurs, and predominantly so, in the alphabetic position of xi in the copper plaques. The CP scribes wrote the xi-symbol
in the eta-position of plaques W-1, lines 10 and 16; MS1-1, line 13; MS1-2,
line7; and MS2-1, line 21. Conversely, the CP eta at times occurs in the position of xi: W-1, lines 11 and 16; M-S 1-1, lines 15 and 18; MS1-2, line 6; and
MS2-2, line15.
The placement of the xiform in the etaposition and the etaform in the
xiposition within the alphabetic sequence results in individual abecedaria
having various marked permutations of the two characters. These can be
summarized in a series of sequences in which line numbers indicate those
lines on which the entire abecedarium occurs:
(22) Sequences of form and position for Xi and Eta

The formal sequence xi eta:

W-1, lines 1011

W-1, lines 1517

The formal sequence eta eta:

MS 1-1, lines 1415

MS 1-1, lines 1718

MS 1-2, lines 56

MS 2-2, lines 1415

The formal sequence xi xi:

MS 1-1, lines 1214

MS 1-2, lines 67

MS 2-1, lines2022
The CP xi-symbol (21) is of uncommon shape, though not otherwise unattested. In the archaic dark blue alphabets, the letter xi is typically formed of
three horizontal bars bisected by a central vertical spine (x). This dark blue
xi is formally close to its Phoenician precursor, samek (having the idealized
form s). In red alphabets xi commonly takes the shape ofc.164
A xi-symbol having the same shape as CP xi, , is known, however, from
the Euboian script preserved in various Etruscan abecedaria the alphabet
42
introduced to the south of Italy by colonists from Euboia who settled at

Pithekoussai and Kyme, beginning in the eighth century b c . Thus, the xisymbol occurs in (1) the abecedarium of the Marsiliana writing tablet, circa
the second quarter of the seventh century b c , the same abecedarium that we
encountered in our discussion of CP beta and gamma (and to which we shall
return several times below).165 Other examples occur in abecedaria (2) on a
bird-shaped bucchero bottle from Viterbo, circa the second half of the seventh century b c ;166 (3) on a bucchero amphora from Monte Acuto (Formello),
circa the last quarter of the seventh century b c ;167 (4) on a red-brown impasto
amphora from Veii, circa the last quarter of the seventh century b c ;168 (5) on a
bucchero bottle from Caere, circa the last quarter of the seventh century b c ;169
(6) on an aryballos from Veii, circa the late seventh century b c ;170 and (7) on
the wall of a tomb at Monteriggioni, in the vicinity of Siena, whereabouts now
unknown, perhaps sixth century b c .171
More significantly still the xi-symbol occurs in the alphabetic position
of eta in another of these Euboian-based Etruscan abecedaria precisely as on
the copper plaques. The inscription consists of only a partial abecedarium
running alpha through kappa hence the xi-position and the symbol that the
inscriber would have utilized for that position are absent. The object on which
these letters occur is a brown impasto goblet from Narce, dated circa the second quarter of the seventh century b c .172
About this Euboian alphabet and the several Etruscan abecedaria attesting
it, Jeffery writes:173
The contents of [Kymes] archaic tombs show that she flourished both commercially and artistically from the time of her foundation onwards. As is well
known, the most valuable evidence for her alphabet is that of the series of
abecedaria found at various sites in Etruria, which range in date from the early
seventh to the fifth century. Every letter-form in the row has its parallel in
the archaic inscriptions of either Kyme or Eretria, except the unique xi and
san. Whether it was the Etruscans, or the Kymeans, or even the people of
Chalkis and Eretria, who were responsible for the closing of the letter xi in
a square [i.e., ], it is impossible to decide; the type is as yet unknown elsewhere, but it occurs persistently in the Etruscan abecedaria.174
This enclosed xi () can, with the publication of the copper plaques, no

longer be said to be unknown elsewhere.
But there is still more that can now be said. Recent excavation of the sanctuary of Apollo Daphnephoros at the site of Eretria in Euboia has provided
43
numerous inscribed pieces of pottery, including a fragment of a monochrome

cup (FK78200) bearing three letters of uncertain writing-direction, the leftmost of which lies right along a break. In the figure that accompanies the
publication of the cup,175 the authors reproduce the letter sequence as ]g o
[ and transcribe as ][: in other words, identifying the leftmost symbol as
a broken pi, and thus the three characters as an alphabetically sequential set
, , , they interpret the three letters as a portion of an abecedarium
that runs around the rim of the vessel.176 This particular pot comes from a
disturbed context, though all of the datable pieces published along with this
inscription are assigned to the first or, more commonly, second half of the
eighth century b c ;177 and the authors suggest that this inscription may provide
the earliest example of the letter known to date.178
Yet even before the copper plaques and the Eretrian graffito came to light, a
symbol of the shape was indeed known. It occurs in the alphabetic tradition
of Argos where it is used, like the eta-symbol h (and H), to represent the glottal fricative /h/. Thus, , the enclosed xi-symbol of the copper plaques, is
found in this use on the mid-fifth-century Tanagra stele in the spelling of the
name of an Argive tribe, h.179 This monument bears the names of Argive
soldiers, divided by tribal affiliation, who died fighting with their Athenian
allies against the Spartan armies at Tanagra, in Boiotia, in 457 b c . Of the several names preserved in the fragments of the monument, this is one of two
that attests a spelling of /h/. In the case of the second name, []h[],
/h/ is spelled with the eta-symbol h.180 This Argive spelling variation formally
mirrors the graphemic alternation of h and in the eta-position of the CP
abecedaria, though the fluctuation is not necessarily functionally equivalent
in the two orthographic systems evidenced by the alphabetic remains of the
those systems.
Another archaic orthographic device demands our attention with regard to
this graphemic alternation of h and in the copper plaques. In the alphabet
of Aegean Naxos, as we saw earlier, eta (h) is used to represent both the consonant /h/ and that long vowel that had developed diachronically from long
vowel *. A variant form of this symbol one denuded of the middle crossbar,
is also used in the Naxian spelling of the consonantal sequence /k + s/.
In other words, in this alphabetic system I is the composite spelling used
to represent the sound sequence that is commonly spelled by the single letter
xi in many epichoric alphabets (typically x in the dark blue alphabets; c in
the red). However, Buck is almost certainly correct in interpreting the Naxian
symbol as having the value of /k + s/ in and of itself in other words, is
44
a xi-symbol and the spelling I as an example of pleonastic orthography

that is, a redundant spelling ks-s (i.e., xi-sigma) of the cluster /k + s/.181
As ready evidence for his interpretation of the Naxian forms h, and
so forth, as displaying redundant orthography, Buck invokes Rhodian pleonastic spellings of the type for (proper name; IG XII
i 709[ca. the third quarter of the sixth century b c ];182 compare the spelling
of the name for ' in IGA 482[early sixth century b c ])183
and for kylix184 (as in IG XII i 719).185 The same form with, mutatis mutandis, the same redundant spelling appears also to occur in a circa
eighth-century graffito etched onto a subgeometric cup from Rhodes, one of
the earliest examples of Greek alphabetic writing: (Korako
mi kulikss) I am the kylix of Korakos.186 In 1961, at the time of the publication
of LSAG, Jeffery could characterize the inscription as by its appearance, as
early as any inscription which we have, except the Dipylon oinochoe.187 The
same pleonastic spelling, , is also attested on a kylix from the Doric colony of Selinous in Sicily, circa the seventh/sixth century b c (SEG XXXI 838;
mentioned earlier, in Section2.1).
Pleonastic spellings of /k + s/ clusters are not, however, limited to Naxian,
Rhodian, and Selinountine orthography: Buck could have just as easily provided parallels from Ionic, Boiotian, Corinthian, Delphian,188 and so on. More
than that the corpus of Greek dialect inscriptions is replete with examples of
redundant spelling of consonant clusters that consist of the fricative /s/ (or its
voiced allophone [z]) plus a stop consonant (such as /p/, /k/, /d/) or nasal consonant (such as /m/). For example, the phonetic sequence [z + d], typically spelled
simply as , is at times spelled redundantly as ; in a completely parallel fashion, /s + t/ is at times spelled as , /s + k/ as , and so forth. Furthermore, in
addition to the spelling for /k + s/, one also encounters the spelling .189
Perhaps Buck chose to make direct reference to Rhodian practice in his
remarks on the redundant Naxian spelling I (i.e., ) of /k + s/ because of the
occurrence of the symbol , identical to the Naxian character, in the alphabet
of Knidos, a fellow member with Rhodes of the Doric-speaking community
of Asia Minor. Though Doric-speaking, Knidos has an alphabet having close
affinities to Ionic. As we noted earlier, the Knidian alphabet uses the Naxian
symbol for writing the eta-vowel, while the eta-symbol H is used to spell
the glottal fricative/h/.
In these several alphabets we see a confluence of orthographic devices
used for the representation of eta-sounds and xi-sounds. The evidence is both
inter-alphabetic and intra-alphabetic.
45
(23) Summary of these eta- and xi-sound orthographies

A. An enclosed xi-symbol , like that of the copper plaques, occurs
in a graffito from Eretria (i.e., in Euboia) and in several Euboianbased abecedaria from Etruria. In one of these, appears in the
alphabetic position ofeta.
B. This same symbol is used in the Argive alphabet as an alternative
to the eta-symbol h for spelling the consonant/h/.
C. The Naxian eta-symbol h (and ) spells both /h/ and long vowel
from *; similar or identical practices are evidenced in the orthographic systems of other Ionic Cycladic islands.
D. In the alphabets of Thera and Rhodes, h (and ) is used to spell both
/h/ and long vowel , regardless of the vowels historical source.
E. A Naxian variant letter is used in the spelling of the cluster /k
+ s/ (i.e., the xi consonantal sequence). The spelling is very likely
pleonastic; that is to say, the value of the symbol is /k + s/, and,
thus, Naxian like Euboian can be identified as a xi-symbol.
F. This Naxian xi-symbol is used in the Knidian alphabet to spell
the eta-vowel, while the eta-symbol spells/h/.
G. The symbol occurs in abecedaria from Euboian Kyme in the
alphabetic position ofeta.
H. The symbol also occurs in the alphabet of Sicilian Naxos where
it is used to spell /h/, as it does even in an inscription from Aegean
Naxos.
The values of these several avatars of the eta- and xi-symbols and the local
alphabets in which the symbols are used is charted in Figure2.1. An asterisk
following the designation of a local alphabet indicates that identification of
the value of the letters specified is based on alphabetic position, considered
together with alphabetic practice as otherwise attested. The of the graffito
from the Eretrian sanctuary of Apollo Daphnephoros is not included, as its
use, and thus its value, must be judged uncertain at the moment.
Such a well-documented and widely disseminated interplay of eta- and xisymbols in these archaic scripts must stem from some antecedent synchronic
orthographic tradition in a still earlier period of the Greek alphabet. While the
local alphabetic traditions garnered in the evidence cited here do not reveal
the use of the archaic eta-symbol h for expressing the composite value of /k+
s/, they do reveal (1) the converse presence of the enclosed xi-symbol in the
alphabetic position of eta, just as in the copper plaques; and (2) the use of an
46
Local Alphabet
Symbol
_________________________________________
(and/or H)
____________ ____________ ____________
Argive
Euboian (Kyme)*
Euboian (Etruria)*
Aegean Naxian
h/
h
h
ks / h
ks / h
h
Sicilian Naxian
Theran
h/
Rhodian
h/
Knidian
Figure2.1. Eta-symbols and Xi-symbols and their values.
eta-symbol (the formal variant ) for spelling not only the glottal fricative
/h/ and the long vowel regular cross-alphabetic, polyphonic values of eta
but for spelling the consonant sequence /k + s/ as well. The evidence for such
traditions suggests that the appearance of the xi-symbol in the eta-position,
as well as the appearance of the eta-symbol in the xi-position, in abecedaria
of the copper plaques is not due to scribal error or idiosyncrasy but is instead
the consequence of intentionality, reflecting ancient orthographic practices.
Further evidence of such intentionality and evidence of the raison dtre for
the playful homography revealed by these practices is provided, I will argue, by
what appears to be actual Greek linguistic data surviving on plaque MS2-2.
Perhaps more curious and interesting still is what we find in the next alphabetic position that of theta.
2.9
t he ta
The orthography of theta is again bound up with issues of form and function.
Formally, two varieties of the theta-symbol appear in the plaques; functionally, a duality of use occurs that parallels that of eta and xi (again, the latter
47
issue involves symbols that are characterized by graphic similarity but sharp
phonetic distinction).
The earliest forms of Greek theta are typically two being that of a circle that
encapsulates either (1) an x-shaped pair of strokes (i.e., ) or (2) a+-shaped
pair of strokes (i.e., ). Jeffery observes that in early inscriptions these two
forms appear to be used indifferently; by the late sixth century [] is the
more regularly used.190 Both forms have antecedents in Phoenician scripts,
which existed side by side in the ninth and early eighth centuries.191 In addition, however, the dotted form of theta (i.e., [on which, see subsequent
discussion]) that appears not rarely during the sixth century,192 becoming
commonplace by the mid-fifth century, is already attested on the Mantiklos
statuette from Thebes in the late eighth or early seventh century b c .193
By far, the most frequently used form of theta in the abecedaria of the copper plaques is the x-shaped variety . There is, nevertheless, a noticeable variation displayed among occurrences of theta on the plaques: in some instances,
theta is quite small relative to the size of most other characters; other instances
of the letter are much larger. The size of archaic theta is not typically a parameter by which the symbol is formally evaluated, though a variation in the size of
theta can indeed be observed among early epichoric alphabets. For instance,
the thetas of the early seventh-century Marsiliana writing tablet and various
other (seventh- and sixth-century) Etruscan abecedaria of Euboian origin are
comparatively large.194 In contrast, the thetas appearing, for example, on the
aforementioned Boiotian bronze lebes (a prize awarded at the funeral games
of Ekpropos)195 and that on the gravestone of Ankylion from Anaphe (SEG
XXVII 527),196 both also early seventh century, are relatively small. As size variation of the character is conspicuous in the abecedaria of the copper plaques,
and as such variation is not unknown in other early examples of Greek alphabetic writing, I have herein identified two CP theta-types, a smaller Theta-1
and a larger Theta-2. A third factor motivating the decision to distinguish
these two types will be addressed later.
(24) Theta-1
(25) Theta-2
All thetas in the plaques are so characterized. Discriminating between the two
types is in many instances a clear-cut process; in some cases, however, the distinction is less pronounced, and specifying the theta-type requires making a
value judgment.
48
In addition to the x-shaped theta , there are a few occurrences of the

+-shaped theta on the copper plaques. These several instances of +-shaped
theta show a size variation comparable to that of the x-shaped thetas and
have been accordingly marked as either Theta-1 (smaller) or Theta-2
(larger). The (less numerous) +-shaped thetas are to be found in the following positions:
(26)CP thetas

W-1, line 3 (Theta-2); line 9 (Theta-1); line 13 (Theta-1)

W-2, line 16 (Theta-2)

MS 1-2, line 5 (Theta-1)

MS 2-1, line 3 (Theta-1)
In a way that functionally parallels the previously discussed interchange
of eta- and xi-symbols in these plaques, the omicron-symbol at times appears
in the theta-position and, conversely, the theta-symbol appears in the position of omicron. Furthermore, CP omicron formally parallels CP theta to the
extent that omicron conspicuously occurs in both a small and a large variety.
These will be designated CP Omicron-1 and Omicron-2 respectively (again,
with value judgments required in the less-clear-cut instances):
(27) Omicron-1
(28) Omicron-2
Unlike the case of theta, however, the size of omicron in fact is a parameter utilized in discussion of archaic forms of that Greek letter. For example,
McCarter offers the following observation:
In the earliest scripts [omicron] displays certain interesting variations. In those
inscriptions where other letters tend to elongate greatly, omicron is regularly
quite small. In less elongated but equally ancient inscriptions omicron is
comparatively large.197
A somewhat similar size variation occurs occasionally in the case of omicrons

Phoenician precursor, ayin,which
undergoes little development in the course of the ninth and eighth centuries,
except to grow slightly smaller by the end of the ninth. In the Kition Bowl and
the Bal Lebanon inscriptions are occasional fluctuations to an extremely small
ayin; but this type is not used consistently in either inscription.198
49
The variation in the diameter of omicron (an already recognized parameter in the historical analysis of the character), a variation paralleling that
of theta in the copper plaques, taken together with alternation of theta and
omicron in CP abecedaria, provides the aforementioned third factor motivating the decision to distinguish formally two types of thetas occurring on
the plaques.
Consequent to the observed interchange of the theta-symbol and omicronsymbol in the abecedaria of the copper plaques, there exist three possible
marked permutations of alphabetic sequencing with regard to the form of
these symbols that is, omicron theta, theta theta, and omicron omicron.
Of these possibilities only the first two are attested in the various abecedaria
of the three copper plaques (line numbers again indicate those lines on which
the entire abecedarium occurs):
(29) Sequencing of form for theta and omicron

The formal sequence omicron theta:
W-1, lines 1517
MS 2-1, lines 2022
MS 2-2, lines 1214
MS 2-2, lines 1415

The formal sequence theta theta:
W-1, lines 12
MS 2-2, lines1112
The absence of the formal permutation omicron omicron among the abe
cedaria of the three known plaques is perhaps accidental.
The variant alphabetic sequencing involving theta and omicron is less common than that involving eta and xi. The former is attested in a total of six abe
cedaria on the copper plaques, the latter in a total of nine. Several instances
of the marked order of both sets of symbols co-occur within a single abecedarium (AB, E), or a marked order occurs in sequential abecedaria (CD):
(30)Alphabetic sequencing for theta and omicron
A. W-1, lines 1517: xi eta and omicrontheta
B. MS 2-1, lines 2022: xi xi and omicrontheta
C. MS 2-2, lines 1112: (xi199 and) thetatheta
D. MS 2-2, lines 1214: (unmarked eta xi and) omicrontheta
E. MS 2-2, lines 1415: eta eta and omicrontheta
50
Note that in each of these instances the scribe utilizes a different permutation
pattern in ordering the four symbols. Notice too that there is also a complementarity in the distribution of the two sets of marked permutations (the set
eta and xi and the set theta and omicron) to the extent that more than half
of the marked orderings of eta and xi occur on the two sides of plaque MS1,
whereas no instance of a marked ordering of theta and omicron is found on
either side of that plaque.
In the case of eta- and xi-symbols, we observed that there is strong independent evidence of archaic Greek orthographic practices that reflect the alternation of these symbols in the abecedaria of the copper plaques. Evidence for a
similar interchangeability of theta- and omicron-symbols also exists.
In this instance, Argos again figures prominently, and remarkably so. The
mason of the previously mentioned early fifth-century Tanagra stele on
which the CP enclosed xi-symbol is used to spell /h/ (alongside the use of
the eta-symbol h to spell /h/, paralleling the alternation of eta- and xi-symbols
in the eta-position in the copper plaques) appears to have also incised the
inscription on a stele set up in the Argive Heraion by four hieromnemones
(IG IV 517).200 In this latter inscription, the mason spells (iaromnamones) as , that is, using a theta-symbol (+-shaped theta
) to spell the second omicron-vowel of the word; his spelling thus parallels
the alternation of theta- and omicron-symbols in the omicron-position on the
copper plaques.
In addition to the theta- and omicron-symbols presented in (24)(26) and
(27)(28) and described in the accompanying discussions, there also occur in
the archaic alphabets a so-called dotted theta (noted already) and a dotted
omicron . The former is commonly said to be a graphic modification of the
primitive crisscrossed (i.e., +-shaped [] and x-shaped []) thetas,201 and certainly is so; nevertheless, dotted theta is attested quite early, being found in the
late eighth or early seventh century on the Boiotian Mantiklos statuette.202
Dotted omicron is yet more intriguing. The Semitic character ayin (eye)
provides the raw material for the Greek letter omicron. A dotted ayin is
attested in Phoenician scripts before the early tenth century203 and in still
earlier Canaanite scripts204 but the symbol is absent from Phoenician writing
thereafter. The occurrence of dotted omicron in eighth- and seventh-century
Greek alphabets as in that on an amphora fragment from Euboian Methone
( 2238; late eighth or early seventh century b c ),205 that of the rock graffiti from Thera,206 and that of the Euboian alphabet preserved in Etruscan
abecedaria207 has led some investigators, most notably Naveh, to argue that
51
the Greeks must have already adapted their alphabet from the Phoenician consonantal script in the late second millennium b c at a time when Phoenician
dotted ayin was still in use.208
Perhaps even more significant than the previously mentioned cases of early
dotted omicron, and ultimately uniquely revealing, is an orthographic practice preserved on a stele from the agora of Argos (ca. the mid-fifth century
b c ), inscribed with a treaty made between the Cretan cities of Knossos and
Tylissos (having Argive arbitrators) the same inscription that was mentioned
earlier in conjunction with the use of digamma for upsilon.209 This Argive
inscription is one in which co-occur dotted theta and dotted omicron, doing
so without exception throughout the forty-eight lines of Vollgraff s text of the
inscription,210 with the partial exception of the final four and one-half lines a
rider to the body of the decree,211 in which theta retains the dot but omicron
generally loses it (though the dot does appear to be faintly visible in some
instances). For example, (of victim; line 11)is spelled .
Vollgraff remarks: Lomikron, marqu dun point au milieu, est absolument
semblable au theta, ce qui peut tre considr comme un signe darchasme.212
Here, again, Argive spelling practice replicates the copper-plaque use of a single symbol in two alphabetic positions those of theta and omicron; moreover, as Vollgraff here suggests, this particular orthographic practice is not
unique to Argos, as we shall soonsee.
Subsequent to the discovery of the Argive stele, fragments of a text were
found at Tylissos that Vollgraff has identified as coming from a Cretan version
of the same treaty.213 The principal fragment, preserving in part or in whole
thirty-three lines of text, is unmistakably written with the Argive alphabet.
In this fragment, unlike the decree found in Argos, omicron and theta are
generally kept graphically distinct, being written as and , respectively;
though omicron / theta homography is not completely avoided: dotted
is used for theta in, for example, ] (for ] to cast [a
vote]; line 9), and for omicron in (for should approve;
line8).
Regarding the Tylissian fragment, Jeffery observes that it shows a curious
mixture of earlier and later forms,214 among which she notes the two nu-symbols N and n her 4 and 5, respectively stating that 5 is not normal even
in Attic before the third quarter of the fifth century. The fragment found in
Tylissos is thus reminiscent of the copper-plaque abecedaria both in its theta /
omicron homography specifically and in its use of both so-called earlier and
later forms generally.
52
We find the same theta / omicron homography at Kyrene in an early fifthcentury dedication on a stele from the Apollonion (SEG 78):
| h
Aiglanr m anethke | hontipatr dekatan
Aiglanor [the son] of Antipatros, dedicated me [as] atithe
On this stele, the third and fourth forms are spelled ,

with dotted theta and dotted omicron respectively.215 Chamoux judged this
dedication to Apollo to be one of the oldest inscriptions from Kyrene.216
Dobias-Lalous chronological analysis does not agree with Chamouxs assessment but acknowledges the inscription to be archaic and assigns it, as have
others, to the early fifth century.217
The North African site of Kyrene was settled by colonists from Thera (ca.
630 b c ) , 218 and it is the Theran alphabet that was used in that place in the
archaic period, though with probable Rhodian influence.219 The co-occurrence
of dotted-omicron and dotted-theta symbols within a single inscription is
also attested on Thera, as in the grave inscription of IG XII iii 814, in which
[-] - is spelled [ ] -. The use of dotted-omicron
for spelling the long omega-vowel, as here, as opposed to plain omicron
for spelling short o, is a common, though not consistent, practice of Theran
orthography.
A different graphic expression of the same ambiguous theta / omicron
spelling strategy is found in an inscription on one of the boulders lying above
the gymnasium at Thera. Here the form occurs for
(or possibly []).220 In this instance, a +-shaped theta-symbol is
used to spell both the theta-consonant and the omicron-vowel. This spelling
procedure is (in broad terms) functionally equivalent to the occurrence of
the formal sequence theta theta seen in the abecedaria of copper plaques
W-1 and MS2-2; this Theran use of the theta-symbol for the omicron-vowel
also agrees with the spelling found on the previously discussed stele from the
Argive Heraion.
***
In the preceding pages, evidence has been presented that reveals that the
homography of eta and xi and the homography of theta and omicron displayed
in abecedaria of the copper plaques are particular expressions of an archaic
orthographic phenomenon that is attested in various local Greek alphabets.
There is also evidence for the existence of a yet greater assimilatory propensity
53
in the archaic alphabetic tradition, one that graphically draws together these
two letter-sets eta / xi and theta / omicron.
The rim of a bronze lebes (ca. 600550 b c ) found at Delphi bears the dedicatory hexametric inscription (FD V3.29.271; CEG369):
h []
das me ho Deksip Puthd anethk[en]
das the [son] of Dexippos [who was sent] to Pytho221 dedicatedme
Perhaps the most interesting epigraphic feature of this brief inscription is the
use of theta- and omicron-symbols having a square, rather than circular, shape.
Theta is identical in form to the enclosed xi-symbol of the copper plaques,
, the symbol appearing at times in the eta-position on the plaques, as well
as in a Euboian-letter abecedarium from Etruria, and found in use as an etasymbol in Argive writing (representing /h/).222 Omicron has a shape close to
that of the eta-variant , used as a xi-symbol in Naxian script and as an etasymbol at Sicilian Naxos (spelling /h/; a use also seen in Aegean Naxian) and
at Knidos (spelling ). Two occurrences of square omicron on the lebes from
Delphi are a bit more squarish than the typical eta-variant , though one of
the instances of the latter form in the Enyo dedication from Sicilian Naxos
approaches the same shape; the third omicron from the Delphic inscription
is rectangular, but with its longer side oriented horizontally, rather than vertically, and thus bearing a particular similarity to the geometry of many of the
-symbols and h-symbols of the abecedaria of the copper plaques. I interpret
the varying box shapes carrying the functional load of eta-, xi- and theta-,
omicron-signifiers to be allographs, essentially in free variation, though perhaps locally determined.
In his publication of this inscription, Lerat refrains from offering a definitive statement regarding its origins. He remarks that it appears to belong to
the groupe occidental and makes explicit reference to a connection to the
Rhodian and Lakonian alphabets; indeed: Dans son ensemble, lalphabet
laconien concorde assez bien avec celui de notre inscription, mais il ne connat pas un signe daspiration analogue.223 For her part, Jeffery is quite confident about the Laconian provenance of the inscription.224 Others would
identify a Boiotian origin.225
The signe daspiration to which Lerat refers (setting this inscription
apart from the Lakonian alphabet) is yet another eta-variant, four-barred
eta that is to say, a form having two medial crossbars (as opposed to the
single medial crossbar of h). The sign is otherwise known but is limited in
54
its distribution and attestation. An example occurs in the inscription on a

lebes from Boiotia, circa the first quarter of the seventh century the prize
awarded at the funeral games of Ekpropos that was discussed earlier in conjunction with CP digamma.226 The same symbol can be seen in the Euboian
abecedarium on the Marsiliana writing tablet, second quarter of the seventh
century b c ,227 now encountered many times. A second Euboian occurrence is
found in a circa mid-sixth-century graffito on the handle of an oinochoe from
Eretria, discussed later in conjunction with my treatment of the letter pi in the
copper plaques.228
Jeffery suggests that the presence of this eta-variant on the lebes from
Delphi may simply be due to a scribal mistake in fact, she is inclined to dismiss all occurrences of four-barred eta as an error of writing in each case.229
This is patently unlikely and should be a warning of the dangers that lie in
casually dismissing the irregular forms data that depart from the norm
data that may well be especially valuable for that very reason. A precursor
for the Greek four-barred eta is provided by Phoenician four-barred het,
as in the eighth-century Bal Lebanon inscription from Cyprus (dedication
on a bronze bowl; CIS I 5; KAI 31)230 an inscription that Rhys Carpenter
had recognized as particularly important for the history of the Greek alphabet:231 Indeed, in the wording of Benjamin Sass, the Baal Lebanon letters
shapes and vertical stances were long recognized as resembling the oldest
Greek letters most.232 Four-barred het can in fact be found in still earlier periods of Phoenician writing: McCarter observes that two-, three-, and (infrequently) four-barred hets are known from the eleventh to the eighth century
in Phoenician inscriptions.233
Why do theta and omicron have the respective square forms and in this
inscription? La forme carre donne au thta et lomicron, doit tre interprte, nous lavons vu, comme un expdient du graveur aux prises avec une
matire rebelle,234 writes Lerat, though he appears to place little confidence
in such a for-ease-of-production explanation.235 One might be inclined to
accept this sort of interpretation were curved lines completely avoided in the
inscription (and they are not), as, for example, in a circa early fifth-century
dedicatory inscription on bronze from Epidauros (IG IV 1611): here a square
theta is used in conjunction with a diamond-shaped omicron , and all
letters show a pronounced rectilinearity.236 Yet even in such a case, one cannot eliminate the possibility that the avoidance of curves is fundamentally a
matter of graphic style. The alphabet of Sikyon shows circular omicron and
theta in seventh- and sixth-century inscriptions, as in a one-word graffito
55
from Delphi, Sikyonian (where is the s-symbol san), circa the

seventh century b c ,237 and in the long membership list on a bronze plaque
(SEG XXX 350), circa the late sixth century b c .238 But in the fifth century,
diamond-shaped theta and diamond-shaped omicron become predominant,
or, in Jefferys words, come into fashion.239
There are instances in which square theta and omicron unmistakably
co-occur with symbols executed with curvilinear strokes. This is the case
in the Barako abecedarium, where theta has the form (beside H-shaped
eta) and omicron has the form (noted earlier), but the loops of beta are
clearly curvilinear. Phi has a square head as well in this abecedarium.240 By
contrast, observes Langdon, two normally rectilinear letters in the abecedarium are made with curving strokes:241 digamma and sigma.
Another example we find square omicron used together with semicircular delta D in an inscription on a sixth-century Corinthian bronze helmet242
(that Jeffery identifies as written in Lakonian script).243 Similarly, the fragments
of a law code, circa the mid-sixth century B C , from Eleutherna preserve a
form of the Cretan alphabet in which theta and omicron are diamond-shaped;
yet some examples of omicron are egg-shaped instead, that is, produced with
curving strokes, and, correspondingly, rho varies between having an angular
head () and a rounded head (R
).244
Observations offered by Lang regarding the several forms of theta that
occur in fifth-century Athenian ostraca are pertinent to this line of inquiry,
revealing the extent of graphemic free-variation found in the ostraca:
There are on the early ostraka at least eight different forms: the crossbar theta
may have an upright cross or an X-cross; both crosses may occur in squares as
well as circles; either square or circle may be dotted; and a circle may have only
a vertical bar or only a horizontalbar.
Most illuminating here [in the spelling of for Themistocles]
are the 126 ostraka on which both thetas are preserved: on 96 (75%) the forms
[of theta] are the same, while on 30 (25%) the forms are different.245
Lang makes an interesting observation after examining the distribution of

forms of sigma in the ostraca that are other than the common forward-facing
three-stroke sigma of the early Attic alphabet:246 It is clear that for the most
part the vagaries of usage can be accounted for only on the grounds of individual preference.247 The viability of the observation seems self-evident.248
But to return to the bronze lebes from Delphi that I am here examining
the co-occurrence of square characters with rounded, just addressed, is also
56
found in the dedicatory inscription on this object. While both theta and omicron are square, the only other symbols that could be produced with curvilinear strokes three occurrences of delta are produced with curvilinear
strokes. In each instance delta has the shape D. Certainly the engravers choice
of the respective theta- and omicron- shapes and was not forced upon
him by his inadequacies as a graphic artist. These were symbols available to
him from the repertoire of archaic theta- and omicron-shapes, and he chose
to utilize them as appropriate options.
Lerat is certainly right in adjudging that linscription donne une impression de grande archasme.249 The earliest example of square theta that I
have been able to identify is found in a seventh-century graffito from Mount
Hymettos. While theta usually takes the form of either or, less commonly,
250 among the Hymettos graffiti, sherd H 217 preserves a square theta in the
recurring formula [- - ] he dedicated.251 The advent of square theta
must surely lie at a still earlier date; it is most unlikely to be the contribution to the Greek alphabet of this one seventh-century worshiper of Zeus on
Mount Hymettos.
Square theta and square omicron are additional participants in an archaic
playful homography a homography that we have already begun to see unfold.
Square theta shares a common graphic form with xi, as does eta. Square omicron shares a common graphic form with eta, as does xi. To put it in slightly
different, structural terms (having functional significance): the xi-grapheme
a second-half-of-the-alphabet symbol can serve not only in the etaposition but in the theta-position as well two contiguous letter-positions in
the first half of the alphabet. Conversely (if we assume for the moment that
is indeed originally an eta-symbol rather than a xi-symbol),252 then the
eta-grapheme a first-half-of-the-alphabet symbol can serve not only
in the xi-position but in the omicron-position as well two contiguous letterpositions in the second half of the alphabet. This arrangement is really quite
remarkable.
2.10
io ta
The iota of the copper plaques is not identical with any other iota found in
the epichoric Greek alphabets, though a somewhat similar iota can be seen in
an inscription from the island of Melos.253 While this copper-plaque character
is consistently formed of three separate strokes throughout the abecedaria of
the plaques, it displays clear morphological variation. Some examples consist
57
of a spine (vertical or leaning away from the direction of writing as the spine
extends upward) with a two-stroke head that tops the spine at, or somewhat
above, the midpoint of the vertical extension of the character: the head is
formed of a horizontal forward-projecting crossbar, from which a vertical
stroke extends upward at the distal end of the crossbar. In some instances, the
latter stroke leans slightly toward, or even away from, the direction of writing
(rather than being fully vertical). This form is designated CP Iota-1:
(31) Iota-1
With this form of the letter compare the several iota symbols found inscribed
on a column from Melos, the earliest Melian inscription, dating to the late
sixth century b c (SEG III 738).254 While the overall shape of this Melian iota
is similar to CP Iota-1, it differs in two regards: the symbol from Melos shows
greater right-angularity (i.e., the up and down strokes are executed vertically, or nearly so, making a right angle join to a midlevel horizontal); and the
Melian symbol faces in the opposite direction (i.e., the horizontal bar extends
away from the direction of writing). With regard to this orientation, note
that the gammas in this inscription are reversed in stance, as regularly in the
Melian alphabet.255
Other copper-plaque iotas are of a fundamentally similar form except that
the two-stroke head is composed of a marked diagonal (not horizontal) line
descending forward from the point at which it joins the spine; from its distal
end the second stroke either rises vertically or ascends diagonally toward the
direction of writing or, less frequently, leans slightly away from the direction
of writing. In some instances, the vertical expanse of the spine is coequal,
or almost coequal, with the total vertical dimension of the character. This
latter feature, together with the diagonal descent of the first stroke of the head,
makes this form of iota, CP Iota-2, conspicuously distinct from Iota-1:
(32) Iota-2
Of the two forms, Iota-2 is the more frequently occurring, outnumbering
instances of Iota-1 by a ratio of almost 2to1.
A few additional occurrences of iota depart somewhat from the two morphologies just described. Collectively they form a less homogeneous group:
while I would judge some to be conspicuously different from Iota-1 and Iota2, some are formally quite close to instances of Iota-2. What binds together
the several occurrences of this third set, CP Iota-3, is a diminution in the
size of the head: in some instances the most distinct the head has a very
58
narrow horizontal dimension; in others, it simply appears smaller relative to

the length of the spine:
(33) Iota-3
An odd state of affairs should be noted: some of the occurrences of CP
Iota-2 are morphologically quite close to forms of archaic Greek nu, such as,
inter alia, that of the Dipylon oinochoe or the Mantiklos statuette and, hence,
similar to the corresponding ancestral forms of Phoenician nun. Perhaps
more curious still some of the occurrences of Iota-1, with its horizontal
crossbar, are redolent of somewhat different forms of Phoenician nun, such as
the nun of the eighth-century Cypriot jug of nt (mentioned in our discussion
of alpha)256 and of the late ninth-century Kilamuwa inscription from Zenjirli
(KAI 25), written with Phoenician script but seemingly in the Aramaic language.257 In contrast, the forms of CP iota are morphologically quite distinct
from the forms of nu that occur on the copper plaques, as we shallsee.
The distribution of Iota-1, Iota-2, and Iota-3 forms on the six sides of the
copper plaques is summarized as follows:
(34)Distribution of CP Iotatypes
-1
-2
-3
W-1
W-2
MS 1-1
MS 1-2
MS 2-1
MS 2-2
TOTAL258
22
39
Outside of the copper-plaque abecedaria, two varieties of archaic iota are

typically distinguished: a straight form (5) and a crooked form, of which latter there are several allographs, such as three-stroke (I) and four-stroke ()
varieties. A three-stroke iota appears on the previously mentioned Dipylon
oinochoe from Athens,259 the wine jug preserving one of the earliest recovered
alphabetic Greek inscriptions (ca. 725 b c ), but a straight iota is otherwise used
in Attic script260 and is widely distributed outside of Attica. A four-stroke
character occurs in the Corinthian alphabet, as on the early Potters Quarter
sherds (ca. the late eighth century b c ), mentioned previously in the discussion
59
of CP delta.261 An S-shaped crooked iota () is used in the archaic alphabet of

Thera262 and commonly so in that of Crete.263
The positional precursor of Greek iota is the Phoenician letter yod, spelling
a palatal glide /y/, typically having the (idealized) shape y. For Jeffery there
can be no doubt that the crooked [iota] is the original form.264 McCarter
concurs that the three-stroke iota recalls the standard lapidary yod of late
ninth-century and subsequent Phoenician scripts.265 The straight iota (5) is
then interpreted as a secondary development effected to prevent confusion
between the graphically (nearly) identical crooked iota (I, , ) and sigma
(s, , I, ) in those alphabets that adopted sigma, rather than san (;), as
the letter of choice for spelling the fricative /s/.266 By this reasoning, the pair
{crooked iota, san} and the pair {straight iota, sigma} should occur virtually
in complementary distribution among the archaic alphabets.
There is sufficient difference between the Phoenician letters yod (y) and
shan (v) precursors of iota and sigma respectively to make one ponder why graphic ambiguity between archaic Greek iota and sigma should
have arisen to begin with. Though, to be sure, as we have just witnessed in
the preceding discussions of eta / xi and theta / omicron, users of the early
Greek alphabet were not shy about embracing graphemic similitude that
is, homography yet perhaps there were circumstances that contributed to
those homographies specifically (a point that we will take up later). The previously mentioned complementary distribution of iotas and s-symbols (i.e.,
{crooked iota, san} vs. {straight iota, sigma}) may hold generally, though it
is interesting that in that very early inscription of the Dipylon oinochoe, one
finds, in violation of that generalization, crooked (three-stroke) iota (unusual
for Attic script) used together with (three-stroke) sigma. Also, Johnston, in
his revision of LSAG, makes the point that the straight form [of iota] appears
in our earliest texts, Euboic colonies, Italy:267 his reference is to inscribed
fragments of storage amphorae from Pithekoussai (ca. 750730 b c ).268
Elsewhere he writes of this straight iota and of the terse Greek and of
their employment at Pithekoussai in such an early period in the history of
the Greek alphabet (mid-eighth century b c ): The importance of these letters
needs to be stressed: one is an addition to the basic Semitic series, the other,
iota, is a developed, simplified form of a Semitic model.269
A still earlier inscription written with alphabetic characters has come to
light since Johnstons revision of LSAG appeared270 five letters etched onto a
flask, discovered in the necropolis of Osteria dellOssa, dated to circa 775 b c
(or earlier).271 There is disagreement over the reading of the graffito, but that
60
reading which could perhaps be styled as most endorsed (and that of the editio
princeps)272 is . Ridgeway suggests a proper name or noun ()273
that is, the adjective good spinner (i.e., one who spins thread well),
as Watkins clarifies274 the form found in Pausanias 8.21.3:
,
,
.
Lukios de ln arkhaioteros tn hlikian, Dliois humnous kai allous
poisas kai es Eileithuian, eulinon te autn anakalei dlon hst(i)
peprmen(i) tn autn kai Kronou presbuteran phsin einai.
The Lycian Olen, a poet of a more ancient time who composed Delian hymns,
among others, one for Eileithyia, both proclaimed her good spinner clearly
equating her with Fate and said that she is older than Cronus.
The inscribed flask from Osteria dellOssa is part of a double burial, presumably that of husband (inhumed) and wife (cremated), with this pot being
conspicuously associated with the latter burial.275 Watkinss observations bear
rehearsing:
In the ancient Mediterranean world and elsewhere spinning belonged exclusively to and defined the world of women, and the pot inscribed eulin(os) occupied the place of honor among the grave goods of a married woman. Since
other female burials in Iron Age Osteria dellOsa were often accompanied by
spindle whorls and other accompanying grave goods termed by Bietti Sestieri
a weaving set,276 we may wonder whether the inscribed pot had rather a symbolic function by virtue of its message in the Iron Age Latial culture at this
particular time, if perhaps only in this particular family.
We will return to such matters directly, but for the present discussion what
is most important to note is that one of the symbols in this brief inscription,
regardless of its interpretation, in what appears now to be the very earliest
recovered example of Greek alphabetic letters, coming from the beginning of
the eighth century, is a straight iota.277
The iota of the copper plaques is unique among Greek alphabets: there is
not another iota like it. Ipso facto, one might be tempted to say that CP iota
corresponds in form to no known Phoenician letter. This sculpted quote
is appropriated from McCarters 1975 discussion of Greek digamma278 and
61
was cited in our examination of CP digamma. As we saw, digamma spells a

labiovelar glide /w/, as does Phoenician waw, with which digamma shares a
common alphabetic position and which is the source of this Greek letters
name that is, but Phoenician waw and Greek digamma do not share a
common graphic shape.
McCarters digamma dictum cannot, however, be applied so strictly to
CP iota. I have already called attention to a peculiar graphic correspondence between the iotas of the copper plaques and forms of Phoenician nun.
But could one seriously entertain the prospect that iota as preserved on the
plaques is actually modeled on Phoenician nun? Could any apparent similarity in shape be due to anything other than chance resemblance?
On the one side that Phoenician nun should have provided the graphic
raw material for Greek iota, as attested on the copper plaques, would seem
an a priori most unlikely scenario. On the other side CP iota unquestionably does not correspond in form to yod, the Phoenician precursor of iota
nor, for that matter, to any other Phoenician letter, excepting the aforementioned reminiscence involving nun. McCarters digamma dictum would thus
seem to hold also over CP iota at least in spirit if we are unwilling to
allow that a relationship of intentionality exists between the graphic shapes of
Phoenician nun and CP iota, and we affirm that there is no formal correspondence between the graphic shapes of Phoenician yod and CP iota. I will proceed with the working hypothesis that the digamma dictum does so hold over
CP iota in the sense described. That being said, we may still need, eventually,
to explore the suggestiveness of shape encompassing some forms of nun and
some forms of CPiota.
Like waw, Phoenician yod which shares a common alphabetic position with Greek iota and is held to be the source both of the typical Greek
grapheme (see Jefferys and McCarters remarks cited previously) and of its
name also spells a glide, the palatal glide /y/. Glides, generally speaking, are
the most vowel-like of consonants and, hence, are commonly also called
semi-vowels. In many of the worlds languages, [w] alternates with the vowel
[u] depending upon the phonetic contexts in which the sounds occur, as
does [y] with [i]. Apropos of such phonetic alternation, as we saw earlier, the
Phoenician letter waw, representing the Phoenician glide /w/, is transparently
the source of the Greek letter upsilon, representing the Greek vowel/u/.
Did the ancient Greek language, like the Phoenician, also possess a glide
/y/? Yes and no. In a prehistoric period, and thus long before the acquisition of the alphabet, Greek lost the phonemic glide /y/ that it had inherited
62
from its Indo-European ancestor in most of its phonetic contexts; in a few

instances, phonemic /y/ survived into the Mycenaean era and is spelled with
the Linear B syllabic script (as in the introductory particle jo- [from *yo-]),
but phonemic /y/ is not attested thereafter.279 By phonemic I mean to designate (in a traditional classic phonemic sense) a sound that fundamentally
has the capacity to effect a change in meaning by alternating with another
phoneme. English, for example, has such a phonemic /y/, which can, thus,
be used to vary meaning between pairs such as yet and wet, or yellow and mellow, as English also has phonemic /w/ and/m/.
After the loss of phonemic /y/, however, ancient Greek speech still possessed the sound [y] no longer phonemic, but a phonetic artifact of the kind
that arises spontaneously when the tongue moves from one articulatory position to another. This type of nonphonemic [y] also occurs in English, among
many other of the worlds languages, as in pronunciations of words such
as science (sai-y-ahnts) and fleeing (flee-y-ing); that is, in hiatus. In ancient
Greek, a phonetic [y] of this sort finds orthographic expression in some of
the local alphabetic traditions, and when it does, it is spelled with iota, as
in Pamphylian (i.e., y for through) and h (i.e., hy
for with the Hiaroi).280 Buck observes that the practice occurs frequently in early Argive inscriptions: consider the proper names and
,281 and the nominals for the goddess (SEG XXXVI 341)282 and
(title of officials; IG IV 506). Buck further notes that the practice
occurs occasionally elsewhere, invoking Ionic and Sikyonic forms, including the seventh-century graffito from Delphi mentioned earlier (in our discussion of theta and omicron), Sikyonian.283 In comparable fashion, the Greek Cypriot syllabary possesses the syllabic symbols ya, ye, and
yo, employed for spelling this nonphonemic [y], as in i-ya-te-ra-ne that is,
y (i.e., the accusative singular of physician; ICS 217.3). Compare
with that the Mycenaean spelling practice: the Linear B yV syllabic symbols
are used to spell not only phonemic /y/ (as noted already) but the articulatory-transitional, nonphonemic [y] as well, as in i-je-re-u for hy (i.e.,
priest; Am 821, Aq 218, Eo 247, etc.).
In the discussion of CP digamma, we noted that the primitive phoneme
/w/, the labiovelar glide, survived longer than did the palatal glide in Greek
being attested into the alphabetic period in some dialects. But the labiovelar
glide too has a dual existence: there is both a phonemic /w/ and a nonphonemic [w] that arises automatically as the articulatory organs move from one
position to another. English again provides a parallel, with phonemic /w/ in
63
wet (beside yet etc.) and transitional, nonphonemic [w] in pronunciations of,
for example, suicide (su-w-ah-side) and fruition (fru-w-ish-ahn).
Local Greek alphabets that retain digamma at times use it to spell this nonphonemic [w], as in Chalcidian and Eretrian (for two) here,
as in the English examples offered previously, [w] is a phonetic artifact of
articulatory movement from the first vowel to the second.284 Comparable phonetic occurrences of [w] occur in the syllabic script of the Cypriot Greeks, as
in the optative tu-wa-no-i (from * to give;285 ICS 217) this alongside
spellings of phonemic /w/, as in wa-na-kse (ICS 211 and 220)for (Attic
master). Both manifestations again occur in Mycenaean spelling: phonemic /w/ is found in di-we for (to Zeus; Fp 1, Tn 316, etc.), ultimately
from the Indo-European root *deiw- to shine, and phonetic [w] in ku-wa-no,
an oblique case form of (lapis lazuli; Ta 642, Ta714).
Could it simply be an accident that the two archaic symbols used for representing the glide sounds of ancient Greek digamma generally and the iota
found on the copper plaques uniquely bear no graphic similarity to their
Phoenician precursors?
Let us summarize what we have observed thus far regarding orthographic
representation of the two glide sounds of Greek. First the labiovelar (the
w-sound): the Greek letter digamma or an adaptation of the Phoenician
letter waw (representing the Phoenician sound /w/) was used to spell both
Greek phonemic /w/ and phonetic [w], as well as the vocalic counterpart of
/w/ the vowel /u/. The last-named practice is at times attested in various
local alphabets and, I have posited, is the regular orthographic practice at an
early stage of the Greek alphabet, as suggested by the abecedaria of the copper plaques, which lack the vowel letter upsilon (the original supplemental
character). As noted, Brixhe has similarly proposed that the use of digamma
to spell both the consonant /w/ and the vowel /u/ is an early Phrygian alphabetic practice.286
Alphabetic representation of the palatal glide (the y-sound) formally parallels representation of the labiovelar glide, though the two practices are not
functionally identical. This functional asymmetry is a consequence of the fact
that by the time of the creation of the Greek alphabet, the palatal glide had
lost the phonemic status that it still enjoyed in the Mycenaean era. Thus, the
Greek letter iota, an adaptation of the Phoenician letter yod (representing the
Phoenician sound /y/), was used to spell Greek phonetic [y] only phonemic /y/ no longer existing as well as being used, of course, to spell the corresponding vowel /i/. The use of iota for spelling phonetic [y] is attested in
64
local alphabets, as we have seen, and must also have been a regular practice,
I would posit, at a formative stage in the development of the Greek alphabet;
recall that the provision of a method for spelling phonetic [y] is a feature
incorporated into the orthographic strategies of both of the pre-alphabetic
Greek syllabic scripts Mycenaean Linear B and the Cypriot syllabary.287 The
alphabetic functional asymmetry involving representation of glides is schematized as follows:
(35) Early functional asymmetry of glide representation
Phoenician source letter
Greek derivative letter
Greek phonemic representation
Greek phonetic representation
waw
wau
/w/ and /u/
[w]
yod
iota
/i/
[y]
Cross-culturally and cross-linguistically, writing systems tend to operate

along phonemic lines. This is a well-recognized tendency, though it is not an
orthographic absolute. Thus the creators of the Greek alphabet themselves
deviated from the principle of phonemic spelling by utilizing the Phoenician
letter qop to represent a context-dependent variant (an allophone) of Greek
/k/ a phonetic avatar of this phoneme that arose when it occurred before
the u-vowel and before the o-vowel. The Greek letter used to spell this phonetic variant the letter called qoppa was, however, abandoned relatively
early in the history of the Greek alphabet (beginning to disappear after the
mid-sixth century b c ),288 and with its abandonment, the spelling of /k/ was
made in accordance with the previously mentioned tendency expectedly
phonemic.289 This modification, however, was not the first adaptation that
Greeks made to their Phoenician-derived alphabet that resulted in its having
greater phonemic transparency.
The developers of the Greek alphabet utilized the Phoenician consonantal
symbol yod (/y/) for spelling the Greek vowel phoneme /i/ and the Phoenician
consonantal symbol waw (/w/) for spelling the Greek vowel phoneme /u/ (two
instantiations of the Greek strategy of adapting a subset of the Phoenician
consonantal symbols for the spelling of vowels). This linguistically insightful adaptation of waw and yod using symbols for glides (/w/ and /y/) to
represent the corresponding vowels (/u/ and /i/) shows intentionality on
the part of the adapters, operating on the two Phoenician glide characters in
tandem. The Greeks also chose to adapt the Semitic glide character waw for
65
spelling the Greek phoneme /w/, thus both assigning a new vocalic value /u/
to the Phoenician symbol while simultaneously retaining its Semitic consonantal value (/w/). Additionally, this Greek symbol for phonemic /w/ could
be utilized for spelling nonphonemic [w] (a practice paralleled by both Linear
B and syllabic Cypriot spelling, as we have seen). This much of the historical
process, as described, is beyond question.
And so we return to the matter of the failure of Greek glide characters to
look like Phoenician letters. It must be the case, I would propose, that the
Greek character digamma, or wau, occurring in the alphabetic position of
Phoenician waw, bears no resemblance to waw, or to any other Phoenician
letter (the digamma dictum), because the adapters intentionally modified its
appearance consequent to its bivalency. That is to say, digamma was from
the very outset of the design of the Greek alphabet a bivalent symbol, having
both consonantal and vocalic values, and was graphically restyled to signal this
phonemic ambiguity this heavy functional load by using a non-Phoenician letter shape (a state of affairs that suggests ongoing interaction between
early users of the Greek alphabet and Phoenician speaker-writers).
At a still early period, however, digammas phonemic ambiguity was
resolved when the Greeks created the first of the supplemental characters
(those appended beyond that point at which the Phoenician script ends),
namely the vowel symbol called upsilon, to which the vocalic value /u/ was
then assigned (though vestiges of the use of digamma to spell /u/ would survive, as we have seen). This first of the supplementals, upsilon, is absent from
the abecedaria of the copper plaques: up to this point in the brief history of
their scholarly examination, the single most important contribution that the
copper plaques have made to the study of the Greek alphabet is, in my view,
the revelation that there indeed existed (as one likely should have suspected) a
form of the archaic Greek alphabet that was coterminous with the Phoenician
consonantal script that is, one lacking all supplemental characters. The supplemental vowel character upsilon unlike the consonantal character
preserves the shape of the Phoenician source-letter waw; and this is of course
in keeping with the Greek adapters practice of otherwise using graphically
unmodified Phoenician consonantal letters to spell Greek vowels.290
Perhaps the most intriguing artifact of the diachronic process whereby the
Greeks teased apart the original bivalency of digamma in keeping with the
cross-cultural tendency to spell phonemically is the previously mentioned
formal similarity that obtains between Greek upsilon and Phoenician waw
sharply contrasting with that graphic disparity which separates digamma
66
() from waw. In other words, Phoenician waw (/w/) was the starting point
for the new Greek symbol for /u/ just as it had been for the earlier-created
bivalent symbol representing both /w/ and /u/: the former (more recent) connection (between waw and upsilon) is revealed by both graphic similarity and
phonetic similarity; the latter (earlier) connection (between waw and )
is revealed by phonetic similarity, alphabetic position, and character name
but not by graphic similarity. The return of the Greeks to Phoenician waw
to obtain the raw material for upsilon, their unambiguous u-symbol, reveals
that this initial expansion of the Greek alphabet beyond the boundaries of
its Phoenician precursor took place in a cultural milieu in which there was
still a dynamic of Greek-Phoenician scriptic intercourse a social and intellectual context that we have seen hinted at already by our findings in this
investigation.
The iota of the copper plaques equally bears no resemblance to any
Phoenician character (in the sense proposed), I would further posit, because
it was generated by the Greek adapters with intentionality in concert with
their creation of digamma. As the Phoenician symbol for the labiovelar glide
/w/ (waw) was taken over as a bivalent character having both a consonantal
and a vocalic value /w/ and /u/ so the Phoenician symbol for the palatal
glide /y/ (yod) was adapted as a bivalent character, though the bivalency in
this case is of a different nature. Iota represented the phonemic vowel /i/ and
the nonphonemic consonant [y]; and because iota thus encodes no phonemic
ambiguity (unlike digamma), there was no need for a resolution comparable to that of the displacement of the phonemically distinct vocalic value of
digamma (i.e., /u/) to the end of the alphabet.
***
We began discussion of CP iota by noting that the numerous occurrences of
the letter in the copper plaques can be separated into three formal types, designated Iota-1, -2, and -3, and here illustrated again:
(36) Iota-1
(37) Iota-2
(38) Iota-3
There is a continuity of shape that binds together the three types. Does each
of the three represent a separate point along an evolutionary continuum? This
67
could be the case. And if it were, which of the three shapes should be recognized as the most primitive?
Iota-1 and/or Iota-2 could quite easily be the source of the archaic crooked
iotas of the various epichoric alphabets. A slight bending of the angles of
Iota-1 will produce the typical three-stroke iota (I).291 The addition of a fourth
stroke to the top of Iota-2 will produce something quite close to the mirror
image of a typical four-stroke iota (). Recall that both Jeffery and McCarter
have judged crooked iota to be more primitive than straight iota (5). Recall,
also, however, that the earliest attested examples of iota are of the straight
variety. In my own view and with all due respect to earlier scholarship on
the problem and to the scholarliness of those who have produced it archaic
crooked iota could be said to be morphologically similar to Phoenician yod
(idealized y) only by making the grossest of comparisons: they have in common sets of lines and angles. Indeed, impressionistically, many varieties of
Phoenician yod actually appear morphologically much closer to digamma
than to crooked iota.292
Iota-3, especially those examples that have a severely limited horizontal
dimension, is less reminiscent of the archaic crooked iotas. One might very
well suppose that CP Iota-3 could have been cropped of its diminished head,
or that the spine and head could have been melded into one, so as to provide
the equally archaic straight iota found in various epichoric alphabets: deriving the archaic straight iota from CP Iota-3 is undeniably more straightforward than deriving it by an arbitrary straightening of any one of the various
crooked iotas.
Might one want to see in Iota-3 the CP avatar of original bivalent iota,
one deliberately made distinct from any Phoenician precursor in parallel
with the creation of digamma as a signal of the characters bivalency? If
so, one might then ponder if perhaps Iota-1 and Iota-2 arose as the alphabet
continued to take shape in a milieu of ongoing Greek-Phoenician interaction the symbol acquiring morphological elements that are grossly but
certainly not distinctively more suggestive of the geometric components
of Phoenician yod. If so, this evolutionary course would somewhat anticipate the secondary creation of upsilon from the other Phoenician glide symbol, waw, in a context of continued Greek-Phoenician interaction. There
is, however, a complication that any such proposed scenario would need to
address: to wit, the aforementioned observed similarity between CP Iota 1/2
and Greek nu / Phoenician nun. For the moment, I will simply acknowledge
the complication.
68
Taking this a step further, one might then ponder if perhaps the evolving
CP iota variants were assigned some separate functional significance. Was
Iota-3, bearing no similarity to Phoenician yod, used to spell the nonphonemic consonant [y] (as non-Phoenician-like digamma was to be used to spell
only the phonemic consonant /w/ after the creation of upsilon), while Iota1/2 was assigned the task of spelling the phonemic vowel /i/ (as Phoenicianlike upsilon would be devised for spelling the phonemic vowel /u/)? If some
such functional difference existed one of nonphonemic [y] versus phonemic /i/ between iota variants, it did not survive;293 and we would very likely
be quite surprised had it survived, given the universal tendency of writing
systems to spell phonemically.
Such ponderings are admittedly quite speculative. It is interesting in
regard to these matters, however, that Claude Brixhe has recently suggested
a functional splitting of Phrygian iota into two forms along similar lines.294
Brixhe reasons that the Greek/Phrygian wau was originally bivalent (as
I have likewise proposed), being used for both the w-consonant and the
u-vowel, and that, in a parallel way, the Greek/Phrygian iota was originally
used for both the y-consonant and the i-vowel and, indeed, we have seen
that it continued to be used in that way among Greeks, spelling phonemic
/i/ and nonphonemic [y]. In contrast to Greek, however, Phrygian retained
a phoneme /y/. For spelling this phoneme, as well as for spelling nonphonemic [y] (occurring in Phrygian, as in Greek, as an artifact of articulatory
movement), the Phrygian alphabet employed a symbol i (or its reverse).295
Formerly, scholars had held that this symbol appeared only in the sixth
century b c , being a direct borrowing and modification of Phoenician yod
(y), and was thereafter limited in use to particular geographic areas. More
recently discovered evidence, argues Brixhe, suggests that the symbol i was
present already in the primitive Phrygian alphabet, likely serving to represent both /i/ and /y/. For Brixhe, Phrygian 5 that is, the Greek straight
iota was a derivative formation, introduced in order to spell /i/, leaving i
to represent /y/ and [y]296 the process effectively mirroring the proposed
ddoublement of Greek/Phrygian wau that produced separate symbols for
/w/ and /u/. The two envisaged processes would be different, I would note,
to the extent that in the latter case it is the transparently Phoenician symbol that serves to spell the vowel (/u/) and not the consonant (/w/); in the
former, the borrowed Phoenician symbol spells the Phrygian consonant
(/y/) and not the vowel (/i/).
2.11
69
ka ppa
The kappa morphology found in the abecedaria of the copper plaques is typical of forms of the letter that are otherwise attested in early Greek alphabets.
In her brief summary remarks on kappa, Jeffery notes that in archaic examples the transverse bars do not always join the vertical at the same point.297
Guarduccis slightly more expansive introductory comments make reference
to the same variation: La lettera greca sostanzialmente identica alla fenicia.
In alcune iscrizioni greche molto arcaiche si nota la tendenza ad innestare i
due tratti obliqui in due punti diversi del tratto verticale (k).298 To this chorus
we can add McCarters voice: Occasionally and sporadically, the two strokes
of the v do not meet the shaft at a single point, as, for example, in the script
of the Naxian statue dedicated by Nikandra.299 This form of kappa, one characterized by two diagonal arms that join the spine of the letter at different
points, occurs commonly on the copper plaques and is herein identified as
CP Kappa-1:
(39) Kappa-1
Outside of the copper plaques, some of the earliest attested examples of
kappa replicate the morphology of Kappa-1. A kappa of this type occurs in
a five-letter graffito inscribed on the sherd of a kotyle from Pithekoussai.300
Johnston dates the graffito to circa 720710 b c .301 If, however, Peruzzi302 and,
subsequently, Coldstream should be correct in reading the graffito to use both
of the s-characters, sigma and san, then one might well suspect an earlier date
or the preservation of a system of orthography of earlier date. Using the late
eighth-century inscription on the cup of Nestor from Pithekoussai as a comparandum, Coldstream judges that the graffito stands closer than the Nestor
inscription to the Phoenician prototypes, and may therefore be appreciably
earlier, or at any rate imply earlier knowledge of the alphabet.303
To this eighth-century Pithekoussai example, a roster of exempla of other
early occurrences of kappa of the CP Kappa-1 type can be appended. A kappa
of this type is found on a late eighth-century fragment of a boustrophedon
inscription on stone from the Athenian Acropolis (IG I2 484).304 Such a kappa
occurs in the Euboian-based Etruscan abecedaria on the Marsiliana writing tablet, second quarter of the seventh century b c ,305 and on an impasto
amphora from Veii, last quarter of the same century.306 An early seventhcentury graffito on a fragment of a clay box of Protocorinthian style from
70
Syracuse preserves a kappa of this type,307 as does the abecedarium inscribed

on the cup from the Heraion of Samos, circa 660 b c , on which an example of
CP Alpha-1 is found.308 This is also the type of kappa used in the mid-seventhcentury Naxian inscription on the Nikandra statue from Delos, discussed
previously in conjunction with the uses of the -symbol.309 Two instances
of the morphology also occur in the seventh-century graffiti from Mount
Hymettos: on sherds H 130 and H 217.310
Somewhat more common in the abecedaria of the copper plaques, however though perhaps a bit less so than the preceding observations might
prepare us to expect is that form of kappa designated CP Kappa-2 that is
characterized by arms that join the spine at a singlespot:
(40) Kappa-2
Outside of the copper plaques, this type of kappa likewise occurs at an early
period. One of the fragments of an amphora from Pithekoussai, circa the mideighth century b c , preserves a kappa with this morphology.311 Other occurrences can be seen in the eighth-century inscriptions on the cup of Nestor
from Pithekoussai312 and on the Dipylon oinochoe from Athens (though in
this case a small gap between the points at which the arms join the spine is
perceptible).313 It also occurs in the inscriptions on sherds from the Potters
Quarter in Corinth, circa the eighth century b c ;314 in graffiti on a skyphos
and an amphora from Eretrian Methone, both late eighth or early seventh
century b c ;315 on the late eighth- or early seventh-century Mantiklos statuette from Boiotia;316 and on the early seventh-century bronze lebes from
Boiotia, awarded as a prize at funeral games for Ekpropos.317 From the second
quarter of the seventh century come two examples on a candleholder from
Ithaka bearing the signature of the maker, 318 and that on the
Protocorinthian aryballos from Kyme319 that we encountered in the discussion of CP epsilon, the so-called aryballos of Tataie (and see the subsequent
discussion of tau). Other examples can be seen in several of the seventhcentury Euboian abecedaria from Etruria: those found on the impasto goblet
from Narce;320 on the bird-shaped bucchero bottle from Viterbo;321 and on a
bucchero bottle from Caere (discussed with Mu-2a).322 In addition, Langdon
notes that a kappa of this type the CP Kappa-2-type occurs in six of the
graffiti from Mount Hymettos in Attica.323
Recall that examples of kappa having the morphology of CP Kappa-1 also
occur on the Hymettos sherds. A similar fluctuation between morphologies
of the type of CP Kappa-1 (with arms joining the spine at different points)
71
and of CP Kappa-2 (with arms joining the spine at the same point) is found
among the occurrences of kappa in both the law code from Dreros in Crete
and the graffiti carved across the outcroppings of rock that fill the areas
above the gymnasium and around the temple of Apollo at the site of ancient
Thera. In the principal, and perhaps oldest, fragment from Dreros, circa the
mid-seventh century b c , Kappa-1 represents the predominant type, though
Kappa-2 occurs at least three times.324 In the rock inscriptions from Thera,
circa the end of the eighth century and later, one finds a kappa of the type of
CP Kappa-2 in, for example, the spelling of the name (IG XII iii
538). On the same outcrop of stone, the name appears, spelled with
a kappa of the CP Kappa-1 type (IG XII iii 539); the same morphology occurs,
inter alia, in IG XII iii 544: hh Tharumakhas [is] good.
In the latter inscription from Thera (IG XII iii 544), kappa is distinguished
by having a lower arm that arches perceptibly downward. In her treatment of
the alphabet of the island, Jeffery remarks on a particular characteristic
feature of the archaic Theran script being the habit of elaborately curling
the ends of iota and kappa.325 Theran iota can have a particularly serpentine
appearance,326 but the ends of kappa are probably better described as more
frequently bowed than curled.327 To illustrate the shapes she has in mind,
Jeffery provides the example of IG XII iii 573, where the kappa is like that of
previously mentioned IG XII iii 544, except that in 573 both arms of the letter arch the lower arm downward, the higher arm upward. Kappa with the
same type of bowed arms can be seen in the circa seventh-century law code
from the temple of Apollo Pythios at Gortyn.328
While iota has its own peculiarities in the copper plaques, as we have seen,
it does not show curling ends. Several occurrences of CP kappa, however,
display an arching of one or both arms like that seen in the Theran rock
inscriptions. On plaque face W-2, Kappa-1 in line 12 has two notably bowed
arms, while the two occurrences of Kappa-1 in lines 9 and 10 show a less pronounced bowing of both arms;329 the lower arm of Kappa-1 in line 4 is moderately bowed. On plaque face MS2-1, the Kappa-1 of line 12 has a lower arm
that shows a slight curve, as do both of the arms of the Kappa-2 of line18.
Other physical traits of the CP kappas can be noted. Both Kappa-1 and
Kappa-2 tend to have a vertical spine that exceeds in length the vertical
dimension of the two outstretched arms. McCarter, in his general comments
on Greek kappa, observes that such a geometry occurs often among the
archaic alphabets.330 In a few instances within the abecedaria of the copper
plaques, the two arms of kappa pass through the spine and join one another
72
on the right of (i.e., behind) the spine: namely in W-1, line 8; MS1-2, line
13; and MS2-1, line 21 (these are all classified as Kappa-2). The same morphology can be seen on sherds from Mount Hymettos H 136/142 and H
139/140.331
The distribution of Kappa-1 and Kappa-2 forms on the six faces of the copper plaques is presented in the following chart:
(41)Distribution of CP Kappatypes
W-1
W-2
MS 1-1
MS 1-2
MS 2-1
MS 2-2
TOTAL332
-1
1
7
2
6
2
0
18
-2
11
3
9
5
7
10
45
In one sense, however, the most conspicuous feature of the CP kappa is its
frequent omission, being absent from eleven of the abecedaria of the plaques.
No other letter comes close to being omitted so often except kappas neighbor, lambda, which is missing from ten of the abecedaria. Of these various
omissions, kappa and lambda are both missing five times, all on MS2 (four
times on MS2-1[lines 4, 10, 14, and 20], once on MS2-2[line 8]). Kappa and
lambda are inverted once on MS1-1 (line 1), and lambda and mu are inverted
once in conjunction with the omission of kappa on the same plaque and side
(line 12). Two of the dual omissions of kappa and lambda involve the omission
altogether of the sequence kappa-lambda-mu-nu (MS 2-1, lines 4 and 10); in
one instance the sequence kappa-lambda-mu is omitted (MS 2-1, line 20).333
The disruption in the alphabet at the point of kappa is curious, and we shall
return to it further along.
2.12
l a mb da
The lambda of the abecedaria of the copper plaques occurs in three forms,
typological gradients, all of which are otherwise known in the archaic period,
and at least two of which are broadly attested from earlyon:
(42) Lambda-1
73
(43) Lambda-2
(44) Lambda-3
Lambda-1 is typologically early, being in effect a 180-degree rotation (i.e.,
an inversion) of Phoenician angular (rather than rounded) lamed, such as
the type found in eighth-century Phoenician inscriptions from Carthage
and, most especially, Cyprus.334 A perusal of Jefferys and Guarduccis discussions and catalogs of local scripts reveals this type of lambda to be widely
distributed: use is attested in the archaic scripts of, inter alia, Thessaly, Phokis,
Ozolian Lokris, Aigina, Corinth, Sikyon, Phleious, the eastern Argolid, the
Ionian islands, Achaian and Doric colonies, and the Ionic and Doric Aegean
islands and in Ionic and Doric Asia Minor. In the eighth century it can be
found on the Dipylon oinochoe, atypical for Attic script;335 in rock graffiti
from Thera, as in IG XII iii 357, where we earlier saw a bowing-armed kappa
of the CP Kappa-1 type;336 in the graffito on the aforementioned subgeometric
cup from Rhodes, the kylix of Korakos;337 as well as on the eighth/seventhcentury sherds from the Potters Quarter at Corinth.338 In the seventh century,
it occurs, inter alia, on the grave stele of Ankylion from Anaphe339 and on the
Nikandra statue from Delos (Naxian script).340
Lambda-2 appears to be a diachronic variant of Lambda-1: Jeffery characterizes the latter type of lambda (the CP Lambda-1 type) as tending to develop
into the former (the CP Lambda-2 type) in the course of the late archaic
period,341 though a lambda of the CP Lambda-2 type is attested at least as early
as the beginning of the seventh century, in the inscription on the fragment of
a clay box from Syracuse mentioned in the discussion of CP Kappa-1.342 As
shown in the figure, the spine of Lambda-2 tends to lean toward the direction of writing that of Lambda-1 being more vertical and the hook-stroke
of Lambda-2 is typically somewhat longer than that of Lambda-1. The two
forms are of about equal occurrence in the copper plaques, Lambda-1 being
slightly more common. In some instances, discriminating between the two is
a judgment matter as their relationship to one another is, to some extent, one
of points on a gradient being, it would seem, a synchronic reflection of an
evolutionary pathway.
The character that I have designated as Lambda-3 occurs far less frequently in the copper plaques than the other two varieties of the letter
(about one-fourth as often), all examples being limited to Schyen plaque
2 (occurring on both sides). The shape of Lambda-3 approaches that of the
74
familiar inverted-V lambda (G), except that the stroke that faces the direction of writing tends to be somewhat shorter than its partner; Immerwahr
catalogs this letter and its mirror image that is, with shorter backstroke
(the stroke occurring away from the direction of writing) as variants of
inverted-V lambda.343 CP Lambda-3 unquestionably appears to be an evolutionary derivative of Lambda-2, produced by elongating the forestroke of
the latter, though not enough to produce a symmetrical character. Even so,
a lambda of similar form is attested early: an example of the mirror-image
of CP Lambda-3 occurs in the seventh-century graffito H 130 from Mount
Hymettos344 and likely that of H 182 as well.345 A lambda of the same morphology and having the same stance as CP Lambda-3 can be seen in that
late seventh-century inscription from Sicilian Naxos (the dedication to the
goddess Enyo discussed previously), which also preserves an alpha like that
of CP Alpha-1 and the symbol used for /h/.346
The graffiti from Atticas Mount Hymettos, dating through the seventh century b c and into the early sixth, and by now familiar to the reader of this
study, are interesting on several counts but especially for the variation in letter
forms that they show. Immerwahr notes that the letter-forms vary from very
archaic to more developed.347 He continues, observing thatthe
Hymettus graffiti, a remarkably complete seventh-century series, include dedications, divine ownership, abecedaria, drinking- and erotic inscriptions, and
perhaps owners names, forming altogether a typical collection of casual graffiti. The writers are proud of their skill, as is shown by the repeated reference
that so-and-so wrote it. Both Young and Langdon have suggested that
the alphabet was not yet fixed, but the irregularities are rather ordinary, and
it appears that the writers of these graffiti show a good deal of literacy, albeit
imperfect.
True to the invoked variation, the graffiti from Hymettos show not only a
lambda that is morphologically close to that of CP Lambda-3 but also, and
more frequently, the common archaic Attic lambda l (though it is not that
of the Dipylon oinochoe [= CP Lambda-1, as already noted]). This lambda is
essentially the inverted counterpart of CP Lambda-1 and, so, continues the
geometry of eighth-century Phoenician lamed. Its early distribution is also
quite broad. Guarducci summarizes:
In alcune localit del mondo greco lo spigolo del lambda rimase in basso: cos
in alcune localit della Creta arcaica (a Eltynia e a Praisos, spesso anche a
75
Dreros, talvolta a Cnosso e a Gortina); cos ad Eretria, a Calcide e nelle colonie

euboiche dItalia, nellAttica, nella Beozia, nella Locride Opunzia. Il lambda
con lo spigolo in basso, definito (non esattamente) calcidese, divenne poi,
attraverso Cuma, la elle (L) dellalfabeto latino.348
Regarding these two early orientations of lambda one with Phoenician stance
and the other inverted (like the forms of the copper plaques) McCarter
concludes:
The examples of both are numerous. Both are found in the early scripts of
Crete. There is therefore no sound basis for considering one to be derivative of
the other; and we cannot choose between them.349
The absence from the copper plaques of the common archaic lambda having
the stance of Phoenician lamed is noteworthy.
The distribution of the three forms of lambda that occur in the abecedaria
of the copper plaques is summarized as follows:
(45)Distribution of CP Lambdatypes
W-1
W-2
MS 1-1
MS 1-2
MS 2-1
MS 2-2
TOTAL350
-1
3
9
5
11
0
3
31
-2
11
0
6
2
5
3
27
-3
0
0
0
0
4
3
7
As already stated, lambda is notable in being absent from ten of the abece
daria of the plaques.
2.13
mu
The mu of the abecedaria of the copper plaques is a typologically early fivestroke mu. To that extent, the examples of CP mu are consistent; however,
four variant forms of this five-stroke mu can be identified, of which the most
commonly occurring is that one herein designated Mu-1:
(46) Mu-1
76
The spine (i.e., the elongated initial stroke) of CP Mu-1 slopes from the head
down toward the direction of writing, while the head itself is tilted upward. In
some of its occurrences, the head is strikingly elevated, as in, for example, W-1,
line 1: this shape recalls the later ninth-century Phoenician mem of the Nora
stone (CIS I 144; KAI 46),351 but especially that of the Gold Pendant inscription
(CIS I 6057; RS 5; KAI 73)from late eighth-century Carthage, with its elongated
spine.352 The typical CP Mu-1, with a somewhat less elevated head, is closer to the
eighth-century mem of the Seville statuette bearing an inscription to Astarte.353
Archaic Greek mu is of two fundamental types a five-stroke mu, as in
the copper plaques (and in morphological agreement with Phoenician mem),
and a graphically reduced four-stroke mu. Outside of the copper plaques,
the earliest examples of five-stroke mu occur in the Euboian alphabet as
attested both in Eretria and, still earlier, in the Euboian colonies of southern
Italy and in the Cretan alphabet, as also in the related scripts of the neighboring islands of Melos and Sikinos (but not that of Thera).354 Yet the form of
five-stroke mu in the earliest examples of these various alphabets commonly
differs from that of the typical CP Mu-1. A nearly exact match of CP Mu-1 is,
however, provided by the mu of the Euboian abecedarium preserved on the
writing tablet from the necropolis of Marsiliana dAlbegna;355 we have now
seen several instances in which this alphabet shows particular agreement with
forms attested in the abecedaria of the copper plaques.
The second most frequently occurring form of mu in the copper plaques,
designated Mu-1, is quite close to Mu-1, different only in that the head of the
letter is rotated downward so as to have a comparatively (if not always absolutely) horizontal orientation. Note that the spine of the letter continues to
descend forward (i.e., toward the direction of writing):
(47) Mu-1
The Greek letter is somewhat reminiscent of the Phoenician mem of the Kition
bowl from Cyprus (ca. 800 b c ).356 The elongated stroke forming the spine of
the Kition Phoenician character has a curved aspect not typical of CP Mu-1
though such a curvature does occur, as in the case of the Mu-1 of W-1, line 6,
where it is echoed by the curvature of the spine of the ensuing nu character.
Comparable forms of mu in archaic Greek inscriptions are difficult to identify though not impossible. Such a form does occur, and twice, in a graffito
preserved on fragments of an amphora from Pithekoussai, circa the mid-eighth
century b c one of the earliest known examples of Greek alphabetic writing.357
77
In the local alphabet of Melos, one finds a similar five-stroke mu with a horizontally oriented head and sloping spine; in this instance, however, the spine
slopes away from the direction of writing, as in the inscription on a fifthcentury grave stele (IG XII iii 1139).358 A mu much like the Melian character
can be seen earlier on neighboring Crete, occurring in the oldest alphabetic
writing at Gortyn legal texts inscribed on wall blocks and steps of the temple of Apollo Pythios,359 circa the end of the seventh century b c , or possibly
earlier.360 The leaning of the elongated stroke of five-stroke mu away from
the direction of writing is a harbinger of the form that would develop with
five strokes of more or less equal length (m); the letter is typologically later,
but an early avatar of it is attested already in the second quarter of the seventh century in a Euboian graffito from Kyme, that on the so-called aryballos
of Tataie,361 which has already twice attracted our attention (and see further
under tau). Mu with a shorter-than-typical first stroke (relative to the length
of the other four strokes) and generally having an appearance reminiscent of
this m-type occurs in the copper plaques on MS2-1 at line 14; cf. also the mu
of MS2-2 at line 20, having a longer initial stroke, but (otherwise) approaching an overall shape recalling a derivative mu of them-type.
The form of five-stroke mu that one typically encounters earliest in Greek
inscriptions (i.e., in inscriptions written in the Euboian alphabet362 and those
of Crete and environs)363 is similar to CP Mu-1 but differs from it in having a spine that is fundamentally vertical. Such a vertical-spined mu occurs,
for example, in the mid-eighth-century graffito from Pithekoussai just mentioned that one also preserving two examples of the (sloping-spined) CP
Mu-1 morphology.364 This sort of variation is consistent with what others have
observed: thus, Johnston, in his revision of LSAG, remarks that it is worth
noting the variety of letter-forms in the early graffiti from Pithekoussai.365
Other early examples of this vertical-spined five-stroke mu can be seen.
The symbol occurs in the eighth-century graffito of the cup of Nestor, also
from Euboian Pithekoussai.366 From the Eretrian colony of Methone comes
a Euboian skyphos (late eighth or early seventh century b c ) bearing a verse
inscription ( 2248) that displays a well-preserved mu of this morphology (the lower portion of the spine sweeps slightly away from the direction of writing).367 It occurs in certain of the previously mentioned Etruscan
abecedaria preserving Euboian script abecedaria that also attest the CP
enclosed xi-symbol as in those on the bucchero amphora from Formello
(ca. the last quarter of the seventh century)368 and that copied from the wall of
a tomb at Monteriggioni (perhaps sixth century b c ).369 It can be found in legal
78
inscriptions preserved on wall blocks from Eretria (Euboia) (ca. mid-sixth

century b c ).370 From the southern Aegean islands, examples can be seen in
the Cretan Dreros law code (ca. the mid-seventh century b c )371 and on a circa
seventh-century gravestone of Pasidikos from Sikinos (CEG 165).372
A mu of this type with approximately horizontal head and (near) vertical
spine also occurs in the abecedaria of the copper plaques; the morphology
is herein designatedMu-2:
(48) Mu-2
In contradistinction to other early Greek alphabets utilizing five-stroke mu,
this type of mu CP Mu-2 is found relatively infrequently among the abe
cedaria of the copper plaques. While CP Mu-1 occurs thirty-four times and
Mu-1 occurs twenty-nine times, there are only nine instances of CP Mu-2,
though considering the small number of occurrences, the distribution of this
type is remarkably broad and even, with a somewhat heavier concentration in
the Schyen plaques: the symbol is found once on each side of the Wrzburg
plaque, twice on each side of MS1, and once on MS2-1, twice on MS2-2.
The fourth variant of mu in the copper plaques is labeled Mu-2. Like Mu2, Mu-2 has a (near) vertical spine, but differs from it in having a head that
is oriented upward, sharing this feature withMu-1:
(49) Mu-2
This type of mu occurs only four times in the copper plaques once on W-2,
once on MS1-2, and twice on MS2-2. Its shape recalls those of the Phoenician
mem of the previously mentioned Bal Lebanon inscription from Cyprus373
and that of the Malta stele (CIS I 123; KAI 61A),374 both from circa the second
half of the eighth century b c . An Etruscan abecedarium, preserving Euboian
letters, on the bird-shaped bucchero bottle from Viterbo (ca. late seventh
century) shows a close match to Mu-2.375
The distribution of the four types of mu in the abecedaria of the copper
plaques is summarized in the following chart:
(50)Distribution of CP Mutypes
W-1
W-2
MS 1-1
-1
4
3
6
-1
9
6
4
-2
1
1
2
-2
0
1
0
MS 1-2
MS 2-1
MS 2-2
TOTAL
6
6
9
34
4
4
2
29
2
1
2
9
79
1
0
2
4
Before bringing this discussion of CP mu to a close, I would point out the

occurrence of a six-stroke form of Mu-1 on MS 1-2, line 8. Given the single
occurrence, ones natural inclination is perhaps to attribute the additional stroke
to scribal error, as indeed it could be though producing erroneous strokes in
copper with an engraving stylus must be a rather different thing than hastily
scratching off an additional tick with pen and ink. The added downward stroke is
about the same length as the final downward stroke of the CP four-stroke nu (see
section 2.14), the letter that next follows in the alphabet. A six-stroke mu is in fact
otherwise attested and, at a very early period, found on the bucchero bottle from
Etruscan Caere, dated circa the last quarter of the seventh century bc. Inscribed
around the base of the bottle is an abecedarium preserving Euboian letters, being
one of those aforementioned Etruscan abecedaria evidencing the enclosed xi of
the copper plaques.376 Within this abecedarium, both lambda and mu are illegible, the bottle being chipped at the point at which the two letters were inscribed.
However, the body of the bottle is encircled by rows of inscribed syllabic CV (consonant + vowel) sequences of the form C+i, C+a, C+u, and C+e; the letter mu
occurs four times within the syllabary (i.e., in the syllabic sequences m+i, m+a,
m+u, m+e), where it is consistently written with six strokes. The sixth stroke is
long, being approximately the same length as the initial stroke (i.e., the spine)
and so producing a character of symmetrical shape, unlike the six-stroke mu of
MS1-2, which has a short sixth stroke and thus retains CP mus asymmetry.377
2.14
nu
The nu of the copper plaques is idiosyncratic, though not completely unique.

Across the other early Greek alphabets, the shape of the letter nu is relatively
consistent. Thus, Jeffery writes, There are no marked local variations of nu.378
In his 1975 survey of the earliest Greek scripts, McCarter observes:
The oldest examples [of nu] are uniformly of three strokes. The third is elongated to form a stem. The length of this vertical stem is the only appreciable
early variable. It ranges from being only slightly longer than one of the strokes
of the head to being extremely long.379
80
In contrast to this typical parameter of slight variation between local forms

of nu, the most conspicuous distinguishing feature of CP nu setting it
apart from almost all other varieties of the character is its four strokes.
In gross terms, the nu of the copper plaques recalls a typical archaic fourstroke mu, such as those found in eighth-century Theran rock graffiti; in
the Naxian script of the Nikandra dedication from Delos (ca. 650 b c ); in
the epitaph of Deidamas from Amorgos (ca. 700650 b c ); and in graffiti
on the sherds from the Corinthian Potters Quarter (ca. 700 b c or earlier). A four-stroke nu is, however, otherwise attested being found in the
Euboian abecedarium and accompanying syllabary of the bucchero bottle
from Caere just discussed. In the fashion of the six-stroke mu of that same
object, the four-stroke nu differs from that of the copper plaques in having
a lengthened final stroke in effect a mirror image of the long first stroke,
producing, again, a symbol of symmetric shape, as opposed to the asymmetric CP four-stroke nu.380
The occurrences of the four-stroke nu of the copper plaques corporately
display that same set of variations that mark CP five-stroke mu. Consequently,
on the basis of the orientation of the head of the letter inclined upward
or horizontal and the disposition of the spine (i.e., the elongated stroke)
slanted forward or vertical nu can be classified in a way that parallels the
classification of mu in the copper plaques (again, with differences in some
cases being subtle):
(51) Nu-1
(52) Nu-1
(53) Nu-2
(54) Nu-2
The relative occurrence of these four forms also parallels that of the four types
of CP mu. Nu-1 occurs thirty-five times (Mu-1 thirty-four times) and Nu-1
twenty-seven times (Mu-1 twenty-nine times); Nu-2 occurs six times (Mu-2
nine times) and Nu-2 is found three times (Mu-2 four times). This close
81
correlation in numbers-of-occurrence reflects in part a certain CP scribal

propensity for producing morphologically homogeneous sequences of the
two letters, particularly of the sequence Mu -1 Nu-1 and Mu-1 Nu-1. The
number of instances of morphologically homogeneous sequences on each of
the copper plaques is as follows:
(55) Morphologically homogeneous sequences ofMuNu
Mu-1 Nu-1
W-1:
W-2:
MS 1-1:
MS 1-2:
MS 2-1:
MS 2-2:
TOTAL:
Mu-2 Nu-2
W-1:
W-2:
MS 1-1:
MS 1-2:
MS 2-1:
MS 2-2:
TOTAL:
3
1
4
5
6
3
22
Mu-1 Nu-1
W-1:
W-2:
MS 1-1:
MS 1-2:
MS 2-1:
MS 2-2:
TOTAL:
6
3
4
1
0
1
15
1
1
0
0
1
0
3
Mu-2 Nu-2
W-1:
W-2:
MS 1-1:
MS 1-2:
MS 2-1:
MS 2-2:
TOTAL:
0
0
0
1
0
1
2
Compare with the twenty-two occurrences of homogeneous Mu-1 Nu-1

sequences and fifteen occurrences of Mu-1 Nu-1 sequences, the appreciably
fewer occurrences of the heterogeneous sequences involving the same forms:
Mu-1 Nu-1 occurs nine times, Mu-1 Nu-1 is found eight times.
Most, though not all, of the possible morphologically heterogeneous
sequences of the two nasal characters do occur in the copper plaques; not
unexpectedly, the number of occurrences of some combinations is small. The
distribution of the three most frequently found heterogeneous sequences
Mu-1 Nu-1 (9 times), Mu-1 Nu-1 (8 times), and Mu-2 Nu-1 (4 times) is
shown in the followinglist:
82
(56) Most common morphologically heterogeneous sequences ofMuNu

Mu-1 Nu-1
W-1:
W-2:
MS 1-1:
MS 1-2:
MS 2-1:
MS 2-2:
TOTAL:
3
1
0
0
4
1
9
Mu-2 Nu-1
W-1:
W-2:
MS 1-1:
MS 1-2:
MS 2-1:
MS 2-2:
TOTAL:
0
0
1
2
0
1
4
Mu-1 Nu-1
W-1:
W-2:
MS 1-1:
MS 1-2:
MS 2-1:
MS 2-2:
TOTAL:
1
2
2
0
0
3
8
Other attested heterogeneous sequences are shown in the following list; each
combination has only a single occurrence.
(57)Additional morphologically heterogeneous sequences of Mu Nu
Mu-1 Nu-2: MS1-2 (1 time)
Mu-2 Nu-1: W-1 (1 time)
Mu-2 Nu-2: MS2-2 (1time)
Not attested are the possible sequences Mu-1 Nu-2, Mu-2 Nu-2, and Mu2 Nu-1.
And why does CP nu unlike its Phoenician precursor, nun, and unlike
other early forms of the Greek letter, save the Euboian script of the bottle from
Caere have four strokes, rather than three? A possible rationale might be
inferred from observations offered in the discussion of iota. Forms of threestroke CP Iota-1 and Iota-2 are unmistakably reminiscent of if not fully identical to some early forms of Greek nu, such as that used in the inscriptions of
83
the Attic Dipylon oinochoe (ca. the late eighth century b c ), the cup of Nestor
from Euboian Pithekoussai (ca. the late eighth century b c ), Theran rock graffiti (the eighth century b c and later), and the Dreros law code from Crete (ca.
the mid-seventh century b c ). One might imagine that the addition of a fourth
stroke to nu was motivated by an attempt to effect a distinction between nu
and the unique iota of the same alphabet an iota that was itself strategically
modified, I have suggested, because of its dual use ( la digamma) for representing both (phonemic) vocalic /i/ and (nonphonemic) consonantal [y]. Yet,
addition of a fourth stroke to nu moved that character into the graphic space
of san, creating other possible ambiguities, as we shallsee.
2.15
xi
The xi-symbol of the copper plaques was discussed in conjunction with the
treatment of eta. The reader is referred to that discussion, though the form of
CP xi is here illustrated again:
(58) Xi
2.16
omic r o n
Omicron was discussed together with theta. The CP omicrons are illustrated
again in (59) and(60):
(59) Omicron-1
(60) Omicron-2
2.17
pi
The letter pi occurs in four different forms on the copper plaques. In terms of
typological ordering, Pi-1 and Pi-2 are close to Phoenician precursors, while
Pi-3 appears to be an evolutionary intermediate between those two graphs, on
the one hand (closer to Pi-1), and the more advanced (i.e., typologically more
distant) Pi-4, on the other. Each of the four is otherwise known in the archaic
period: the assemblage of pi graphemes in the copper plaques echoes McCarters
observation that the forms of Greek pi show a subtle but distinct variety.381
The type herein denoted as Pi-1 has a certain jointedness or angularity
about the head, regularly joining the spine at an angle.
84
(61) Pi-1
In contrast, the head of Pi-2 has a more rounded shape overall, appearing as a
fluid continuation of the spine, rather than a join. It is this juncture with the
spine that is the principal criterion for distinguishing the Pi-1 and Pi-2 graphs
from one another, though we should note that Pi-1 shows a greater range of
allographs than doesPi-2.
(62) Pi-2
In their most discrete forms, CP Pi-1 and Pi-2 are clearly distinct; in some of
their occurrences, however, the distinction between the two is a fine one. In
his study of early Greek scripts, McCarter takes note of forms corresponding
to CP Pi-1 and Pi-2 and of respective Phoenician antecedents; regarding the
Semitic letter, he writes:
The standard lapidary pe of the ninth and eighth centuries provides a satisfactory prototype for the rounded Greek forms [i.e., of the type CP Pi-2]. The
Kition Bowl [from Cyprus, late ninth century b c ] may display the model for
the jointed form [i.e., of the type CP Pi-1].382
One might also compare the pe of the Nora stone from Sardinia (later ninth
century b c )383 with its slightly angular head, reminiscent of some instances of
CPPi-1.
Outside of the copper plaques, perhaps the most notable occurrences of
forms of pi matching CP Pi-1 and Pi-2 are to be found in the short hexametric
inscription of the Dipylon oinochoe.384 In this earliest Attic inscription of any
length, these two types of pi co-occur:
h
hos nun orkhstn pantn ataltata paizdei
who now among all dancers most exuberantly dances
A distinct Pi-1 form occurs in the spelling of , which is followed after

one intervening word by an equally distinct Pi-2 form at the beginning of
a variation to which others have drawn attention: thus, Immerwahr
writes: The Dipylon jug has (1) [= my CP Pi-2] and (2) [= my CP Pi-1].385
Similarly, distinct examples of the two forms also co-occur in the Dreros
legal inscription from Crete (ca. the mid-seventh century): thus, in the
phrase , to owe double, the first word is written with a pi of the
CP Pi-2 type, the second word with a pi of the CP Pi-1 type.386 Notice that in
the same lines of the law code from which this phrase is abstracted, a graphic
variation of a different sort is also attested: long vowel is now written with
85
the eta-symbol (h) and now with epsilon in the infinitive beside
to be.387
Other archaic examples of pi of the type of CP Pi-1 can be identified. Such
a letter can be seen in the seventh-century Mount Hymettos graffito H 553.388
A pi-symbol with this morphology also occurs on an inscribed skyphos from
Athens (P4663), circa the mid-seventh century b c ,389 and another in a graffito
on an oinochoe from the Euboian city of Eretria, circa the mid-sixth century
b c .390 The pi of an oracular inscription from Euboian Kyme,391 circa the early
sixth century, is fundamentally of the CP Pi-1 type,392 hinting at the more rectilinear form of CPPi-3.
Beyond the examples provided by the Dipylon oinochoe and the Dreros law
code, a pi like that of CP Pi-2 is found at an early period in several locales,
showing a broad geographic distribution. Such a pi occurs in the Cretan legal
text inscribed on the previously mentioned blocks and steps from the temple
of Apollo Pythios at Gortyn, circa the late seventh century b c ,393 as well as on a
column at Prinias, circa the early sixth century;394 and, still in the Doric Aegean,
the morphology is typical in the eighth-century (and later) rock graffiti of
Thera and yet other inscriptions from the island.395 Among the Ionic islands
of the Aegean, this form of pi is found in the early seventh-century Deidamas
epitaph from Amorgos396 and on the grave stele of Dame(as) from Naxos, second half of the seventh century.397 It is likewise used in the seventh century (ca.
650625) in an inscription from the sanctuary of Apollo Ptoios in Boiotia (IG
VII 2729), though more notable among the earliest Boiotian inscriptions is the
occurrence of a pi like that of CP Pi-3.398 In the Euboian abecedarium preserved
on the writing tablet from the necropolis of Marsiliana dAlbegna, circa the second quarter of the seventh century,399 and that on the bird-shaped bottle from
Viterbo, circa the late seventh century,400 the same type of rounded pi (= CP
Pi-2) occurs; the slight angularity with which each was executed is, however,
somewhat reminiscent of CP Pi-1 in certain of its occurrences.
Among the several Euboian scripts preserved in Italy, that one attested by
an abecedarium that occurs together with a set of syllabic spellings on the bucchero bottle from Caere (ca. 650600 BC)401 already discussed in conjunction
with my examination of CP five-stroke mu and four-stroke nu is of particular interest with regard to pi. Within the abecedarium, ringing the base of the
bottle, pi has a form similar to that seen in other of these Euboian abecedaria
from Etruria: the form is much like that of a typical CP Pi-1, but the head is
somewhat more curved, looping around and back toward the spine. More
interesting is the presentation of pi in the syllabic component of the inscription (i.e., in the syllabic sequences p+i, p+a, p+u, p+e) in which no fewer than
86
three variant forms of the grapheme pi appear. In writing the sequence p+i,
the scribe first wrote a pi formally close to that of the associated abecedarium,
but with a triangular head thus producing something quite similar to one
of the common forms of archaic Greek rho (approximately);402 the scribe
(or some subsequent scribe) then modified this character by writing over and
beneath it another pi, one essentially like (though not identical to) that of the
abecedarium. The scribe then, however, wrote the sequence p+a using a large
pi without a fully looped head403 in other words, one shaped fundamentally like a typical CP Pi-1. For writing p+u the scribe returned to utilizing a
pi close to that of the affiliated abecedarium (with fully looped head). Then,
however, to inscribe the syllabic sequence p+e, the scribe made use of a quite
different, distinctly angular pi,404 identical to CP Pi-3, to which we nowturn.
Pi-3, the third variant of pi found within the abecedaria of the copper
plaques, is, in effect, Pi-1 made fully rectilinear. The latters arcing head has
morphed into a descending diagonal joined to the top of the characters spine;
at the distal end of the diagonal is a small tick, which is either approximately
parallel to the vertical spine or angled diagonally back toward the spine. The
angular join of this tick to the descending arm is a prominent mark of Pi-3.405
(63) Pi-3
In some instances Pi-1 and Pi-3 are morphologically quite close: plaque MS1-1
displays several forms that appear to be transitional between discrete Pi-1 and
Pi-3 morphologies (see especially the occurrences of pi in lines 2, 6, and17).
This grapheme is otherwise attested, though less commonly than the types
of Pi-1 and Pi-2. Perhaps the earliest example is that provided by the fragmentary inscription running around the rim of a skyphos from Methone (
2253), the Euboian colony on the Thermaic Gulf (late eighth or early seventh
century b c ).406 Such a pi morphology can be seen in the archaic Boiotian
alphabet.407 For example, in the inscription on a fragment of a bronze lebes
from Thebes, that prize awarded at funeral games for Ekpropos which we
have now encountered several times,408 three examples of this type of pi
occur in the spelling of the commemorative phrase ; the letters
ofthe commemorative inscription (ca. 700675 b c ) are long and spindly and
thetick of each pi, angled back toward the spine, is commensurately long. In
the dedicatory inscription on the same lebes, however, the letters have shorter
stems, and the tick of the lone pi is correspondingly shorter, giving the letter
an appearance close to that of instances of Pi-3 in the copper plaques. Note
that a second, incomplete, commemorative inscription found on the same
87
lebes uses a pi like that of CP Pi-4; the letters forming the second commemorative inscription may have been added by a separate hand at a later date, it
has been suggested,409 though we shall see the co-occurrence of CP Pi-3 and
CP Pi-4 forms elsewhere.
A pi of the CP Pi-3 type also occurs in early Lakonian inscriptions.410 An
example can be seen in the dedication preserved on a fragment of a bronze
lebes found at Delphi:411 this is the very inscription discussed earlier in which
square theta and square omicron occur. The inscription is one of a set that
Jeffery groups together on the basis of certain common epigraphic features
(including this form of pi, though other forms of the letter occur among the
members of the set), cautiously dating the entire group of inscriptions to
circa 575550 b c .412 Her final observation on the set is of interest: In all these
inscriptions sigma varies between 1 and 2[i.e., I and S], and nu is never the
same twice even in the same inscription.413
The fourth avatar of pi in the copper plaques is formally similar to Pi-3 but
with the vestige of the arcing head of the letter now raised from the descending diagonal position of Pi-3 to a horizontal position. The shape of the character is thus much like that of the typical archaic gamma,414 but set off from it
by the downward-pointed tick that occurs at the distal end of the horizontal
stroke, joining the stroke at a right angle:415
(64) Pi-4
CP Pi-4, typologically farthest removed from the Phoenician precursor of pi,
echoes the most commonly occurring form of the letter in the various archaic
Greek alphabets.416 In spite of its typological remoteness from Phoenician
sources, however, this form of pi is attested at a very early period.
Of the four examples of pi occurring in the verse inscription of the eighthcentury cup of Nestor from Pithekoussai,417 at least one has the distinct rectilinear shape that is the hallmark of the morphology of CP Pi-4. The first line of
the inscription, composed in iambic trimeter,418 contains two instances of pi:
one within the third word, [], and one at the beginning of the fourth,
. The remaining two lines of the inscription are in dactylic hexameter;
the first of these likewise contains two occurrences of pi, the first at the end of
the second foot, the next at the end of the third: : []. Following the
restoration and translation of Watkinss 1976 publication, the inscription reads:
: [] : [] :
h <> : [] :
h h : [] :
88
Nestoros : e[st]i : eupot[on] : potrion

hos d a<n> tde pisi : potri[] : autikaknon
himeros hairsei : kalliste[pha]n : Aphrodits
Nestors cup is good to drinkfrom;
But he who drinks from this cup, forthwithhim
Will seize desire of fair-garlanded Aphrodite.
The morphology of the first pi of line 1 ([]) is partially obscured by a
break in the cup that occurs at its top edge, right where the looping head joins
the stem. Given the curvature of the head, the letter appears to be an instance
of the CP Pi-2 type;419 if not, then it should be classed as being of the CP Pi-1
type but showing particular morphological affinity for Pi-2. The second pi in
line 1 () is of the CP Pi-1 type, though approaching the angularity of
CP Pi-3. If the first pi is in fact of the sort CP Pi-2, then in this line from the
cup of Nestor, as in the graffito of the Dipylon oinochoe together two of the
earliest Greek inscriptions of any length both of the typologically earliest
forms of Greek pi are used in alternation. It is a curious coincidence. In line
2 of the inscription on the cup of Nestor, both instances of pi show the typologically advanced shape of CP Pi-4. The first of these () again, partially
obscured by a break may show some formal proclivity for CP Pi-3.420
Other early examples of the CP Pi-4 type can be readily identified. A pi
fundamentally of this sort but with a looping forestroke can be seen on a vase
fragment from Eretria (FK03663), second half of the eighth century b c .421
Several pi-symbols of this morphology can be seen in the seventh-century
graffiti from Mount Hymettos, as in H 189,422 H 1,423 H 169,424 and so forth.
The form is also found in seventh-century Boiotian inscriptions, including
a set of bronze lebetes found on the Acropolis in Athens (the examples of pi
occur in IG I2 402403 and 406),425 roughly contemporary with the previously discussed inscription on the lebes from Thebes (prize at the funeral
games for Ekpropos) that preserves examples of CP Pi-3.426 Other circa seventh-century occurrences include, inter alia, the inscription on a bronze aryballos (dedicated by Chalkodamas) from Argos (IG V i 231; CEG 363);427 the
boustrophedon Naxian inscription on the base of a kouros from Delos, dedicated by Euthykartides;428 the memorial inscription, also written boustrophedon, to Glaukos son of Leptines from Thasos;429 a Protocorinthian aryballos
from Megara Hyblaia, bearing a dipinto of five letters;430 and an aryballos of
Protocorinthian type from an uncertain Italian location, now in the Louvre,
with the dipinto [?].431
89
The distribution of the four morphological types of pi in the abecedaria of

the copper plaques is presented in the following chart:
(65)Distribution of CP Pitypes
W-1
W-2
MS 1-1
MS 1-2
MS 2-1
MS 2-2
TOTAL432
-1 -2 -3
1
2
0
1
3
3
4
2
3
3
6
0
3
6
2
1
4
0
13 23 8
-4
7
2
0
2
4
8
23
The four types of pi are broadly distributed across the six sides of the copper
plaques. At least one instance of each type occurs on each plaque face with
the exception of Pi-3, which is absent from W-1, MS1-2, and MS2-2, and of
Pi-4, absent from MS1-1. The former, Pi-3, is the least frequently occurring
form of pi on the plaques; the latter, Pi-4, is one of the two most commonly
occurring forms, the other being Pi-2. With regard to distribution of types of
pi, plaque side W-1 is the least heterogeneous, while its obverse, W-2, shows a
particularly even distribution of the four forms.
2.18
san
In her survey description of the Greek letter san, Jeffery notes only two forms
of the letter:
The Greek form of the letter, 1[i.e., ;], has no particular resemblance to the
sd of Phoenician formal lettering on stone, but might well be derived
from the cursive form as used on the painted sherds from Samaria . Like
mu, san develops to 2[i.e., ] in cursive writing by the late archaic period,
occasionally earlier.433
The san of the copper plaques which exists in three variant forms is identical to neither of Jefferys two fundamental types.
McCarter concurs with Jeffery concerning the typological distance
between the Phoenician and Greek symbols: Unfortunately, no good prototype for san exists among attested examples of early Phoenician sade.434 He
90
notes the early attestation of the symmetrical M-shaped san (equivalent to

Jefferys type 1): The two legs may be parallel, as, for example, in the san of
the Marsiliana tablet , or somewhat splayed, as in the san of the Dreros slab
from Crete.435 But McCarter also calls attention to an early asymmetrical type
of san: in the case of the just-mentioned splayed variety, he notesthat
the back leg may be somewhat shorter than the front leg. This is a feature
occasionally encountered in local scripts which used five-stroke mu.436 It suggests that the first san may have had one short leg, which was later lengthened upon the advent of four-stroke mu to avoid confusion. In later scripts,
after the M-form mu has developed, the two signs (san and mu) are often
indistinguishable.437
Such homography reverberates in the copper plaques where, as we shall see,

san and four-stroke nu (rather than mu) are often morphologically identical.
The three types of san found in the abecedaria of the copper plaques are
of the asymmetrical sort.438 Unlike the asymmetry of the forms noted by
McCarter, however, the asymmetry of CP san crucially involves the back leg
(that one opposite the direction of writing) being longer (not shorter) than
the front leg (that one proximate to the direction of writing). In addition,
one envisions, on the basis of these archaic forms of san, not so much the diachronic scenario of a symbol with an original short leg that would be lengthened so as to create greater perceptual distance between itself and four-stroke
mu, but the scenario of a symbol with an originally elongated leg upon which
symmetry would be imposed by bringing length and angular equivalence to
the two outer strokes though perhaps motivated by issues of homography.
In the case of San-1, the initial, lengthened stroke is vertical in its
orientation:
(66) San-1
Note that the three remaining strokes of CP San-1 are in almost all cases
approximately equal in length that is to say, to be more precise, that they
have an equivalent vertical dimension.439 In its overall morphology, San-1
echoes the shape of Nu-2 (or Nu-2).440 CP San-1 is virtually identical to the
typical four-stroke mu of many of the san-using archaic alphabets.
It is difficult to identify forms of san comparable to CP San-1 in the various
local scripts of archaic Greece; this is doubtless due in part to the relatively
sparse attestation of the character. We may begin by comparing the san in the
spelling of the name of the god of the North Wind, (IG XII iii 357),
91
inscribed beside one of the carved offering receptacles at Thera, perhaps late
eighth or early seventh century b c . The Theran san of the rock graffiti is typically symmetrical; in this case, however, the form is asymmetrical, having an
elongated back leg, as with CP San-1, though differing from most instances
of CP San-1 in that the length of the front leg of the Theran form somewhat
exceeds the vertical dimension of the two inner strokes.
Related forms may be provided by a set of names inscribed on blocks of
tufa from Delphi (second half of the sixth century b c ), blocks from a structure situated between the Theban Treasury and the Athenian Treasury.441 The
inscriptions would not have been visible when the blocks were positioned in
situ, leading to speculation that the names were either those of minor donors
or of masons, or both. The alphabet of the inscriptions may be that of Phokis,
though uncertainty in that regard has been expressed for at least some of the
names specifically those in which san is utilized to spell /s/, as sigma is the
typical sibilant symbol of Phokian orthography.442 There is, however, other
evidence for the use of san in Phokis;443 moreover, in his revision of LSAG,
Johnston affirms: It can now be accepted that san was used in Phokis, perhaps by way of import via Itea.444
Some of these spellings reveal a form of san similar, though not strictly
identical, to the type of CP San-1. The overall geometry of the symbols is
fundamentally the same, except that the Delphic examples have a broader
horizontal dimension and the elongated back leg is generally shorter, relative
to the other three strokes, than is the case with San-1 of the copper plaques.445
Little has been said thus far in my examination of the plaques concerning the
possible locale in which they were produced, though particular resemblances
to the Euboian alphabet have been seen to recur. The survival of a form of
san similar to the almost unique CP San-1 in what may be an early variety of
the Phokian alphabet could be of some relevance to the question of locale,
as Phokis quite possibly acquired its alphabet from Euboia via their mutual
neighbors, the Boiotians.446
One might consider the apparent example of the same type of san among
the painted letters on an oinochoe from Ithaka to be illusory. The symbol
occurs in a fragment of hexameter of early date, circa 700 b c , at least two lines
in length (CEG 453).447 The following phrase can be securely restored:
] []
Ks]enwos te philos kai p[is]tos etairos
Near and dear xenos and faithful companion-in-arms448
92
While the final san of ] provides a perfect match to CP San-1, its short
front leg lies along a crack in the pot that might possibly obscure a further
expanse of the leg; my inclination is to judge that this is not the case, but I
cannot be certain.449 Too little of the san-characters of and []
survives to identify their morphology. In contrast, the final san of
is fully visible;450 its shape is symmetrical (unlike the varieties of san found
in the copper plaques), having front and back legs of equal length. We have
of course seen variation in letter shapes in the eighth-century hexameters of
both the Dipylon oinochoe and Nestors cup; it is not impossible perhaps
not unexpected that two separate san morphologies one symmetric and
the other asymmetric, and matching a CP type of san are evidenced in this
phrase on the wine jug from Ithaka.
In addition to this phrase, two other words can be read that contain san:
h and .451 Unfortunately, however, both occur along a longitudinal
fissure: what survives of san in the latter form () is only the first two
strokes; in the former case (h), a portion of the third stroke remains as well.
In both instances, however, enough of the character is visible to reveal that the
orientation of the initial stroke (i.e., the back leg) is less vertical than in the
case of the sans of ] and , being angled away from the direction
of writing.
The picture that emerges is one of heterogeneity of san-shapes on the wine
jug from Ithaka. We should note that among other archaic elements of the
inscription, not found in later inscriptions from that place, is the lambda (6)
that Jeffery typifies as Euboian and Guarducci as calcidese.452
In contrast to San-1, the initial, long, stroke (the back leg) of the two
remaining types of CP san shows a diagonal orientation. That of San-2 leans
toward the direction of writing (opposite the direction seen in the two lastmentioned examples from Ithaka), with the result that this avatar of san echoes the form of Nu-1 (or Nu-1):453
(67) San-2
Among Greek alphabets, san of this shape appears to be unique to the copper plaques, and in the plaques it is the least-commonly encountered form
ofsan.
The elongated stroke of San-3 leans away from the direction of writing,
making the character distinct from CPnu:
(68) San-3
93
Sometimes, however, the angle of the diagonal slope of the long stroke is
very shallow, making the distinction between San-1 (with its vertical long
stroke) and San-3 quite a fine one.454 If the long stroke is clearly vertical at
the point at which it meets the head of the character, I have labeled it San-1.
However, judgments of verticality must at times make recourse to the vertical orientation of the legs of surrounding letters (pi and qoppa especially).455
In several instances, the elongated back stroke shows a distinctly lunate
shape. As the result is a long stroke with an overall rightward progression
(i.e., away from the direction of writing), I have also designated these forms
as San-3:456
(69) San-3 (lunate long stroke)
Distinguishing San-1 from this type of San-3 is at times complicated by the
tendency of one or more scribes to end vertical strokes in a rightward flourish; again, comparison with surrounding letters must be invoked.457
One searches practically in vain among the other archaic alphabets for a
san comparable in shape to the CP San-3 of (68). A dedicatory inscription
of the late seventh century b c written on a petite bronze aryballos in Argive
script (though reported to have been found in Sparta) does preserve a san
showing an elongated back leg that slopes diagonally away from the direction
of writing (IG V i 231; CEG 363).458 The character occurs as the final letter of
the spelling of the name of the dedicator, Chalkodamas:

Khalkodamans me anethke thiioin perikalles agalma
Chalkodamas dedicated me to the twin gods, a very beautifulgift
The back leg of the san is, however, only marginally longer than the front leg,
and both show a much greater vertical dimension than the two inner strokes,
making the overall (nearly symmetrical) morphology of the letter deviant
from that of the typical CP San-3. We should note, however, that there is a
single, anomalous, example of CP-San 3 (see W-1, line 12)that shows closer
though not exact agreement: the front leg of the copper-plaque character is
still somewhat shorter than that of the Argive. The second occurrence of san
in the Chalkodamas dedication (in ) is yet more symmetrical and,
thus, further removed from CP San-3.
Perhaps it is worth mentioning another example of archaic san, a closer
match to CP San-3 than is the Chalkodamas san, except that it is flipped around.
The letter occurs in an inscription found on the base of statues of Kleobis and
94
Biton from Delphi (SEG XXXV 479; ca. the late seventh/early sixth century
b c ).459 The name of the sculptor (partially obliterated) is engraved in Argive
letters (from right to left):
[?] h
[Poly?]mds epoie hargeios
[Poly?]medes the Argive made[me]
The san of h is the mirror image of CP San-3, having a single elongated diagonal stroke on the side that faces the direction of writing, the left
side (i.e., having a long front leg [rather than long back leg] that extends
away from the body of the character), with the remaining three strokes being
short, each of roughly the same length. The san of - likewise has a long
stroke on the left, but the orientation of the stroke is vertical rather than diagonal. This letter is thus approximately the mirror image of CP San-1, though
the opposing outside stroke (on the right, the back leg) is somewhat longer,
relative to the two inner strokes, than is typical for the corresponding stroke
of the CP character.
And what about the CP San-3 (69) with the lunate long back leg? No exact
morphological parallel has presented itself outside of the copper plaques.
However, a san with a long back leg, having a lunate shape much like examples of CP San-3, is attested in one of the Euboian-based abecedaria from
Etruria. The earlier-mentioned late seventh-century bucchero amphora from
Formello is incised with two abecedaria:460 in one of these, san has a long
curving back leg; the front leg, somewhat less elongated, is noncurving, as
are the vertical stems of other letters in the abecedarium. The wedge-form
inner strokes of this san are much shorter than the two outer strokes, and in
this way the morphology of the letter is distinct from that of the typical san
of the copper plaques. In the second abecedarium on the amphora, san has
the familiar -shape (i.e., with noncurving, elongated outer legs and short
wedge-form inner strokes), again departing from typical CP morphology.
Note, however, that the morphological relationship between the two sans of
the Formello abecedaria is, mutatis mutandis, the same as that between CP
lunate San-3 and CP San-1.
In the search for morphological parallels to the lunate san of the copper
plaques, the san of a short inscription on a cup of subgeometric style found
in the precinct of the Argive Heraion is worth noting. Jeffery tentatively identifies the alphabet of this early inscription (SEG XI 306; late eighth or early
seventh century b c ) as that of neighboring Kleonai,461 chiefly on the basis of
95
the B-shaped eta symbol (B); the graffito reads B (khos emi), apparently meaning I am a khous, where a khous would be a local unit of measure, smaller than the Attic khous.462 The san character has, relative to other
vertical lines in the inscription, a pronounced windswept appearance,463 with
the lower portion of the long and gangling back leg sweeping away from the
direction of writing. The back leg is thus somewhat reminiscent of that of
some of the CP San-3 lunate letters; it differs from CP san, however, in having
a lengthened, bending front leg as well (though the front leg is shorter than
the back).464
The san of the copper plaques, in all three of the varieties here identified,
thus can lay claim to very few counterparts among the local alphabets of archaic
Greece. We began our discussion of san by rehearsing earlier investigators
remarks regarding the wide typological divide that separates Phoenician sade
and archaic Greek san. Given the differences between the san morphologies
of the copper plaques and those of most other archaic alphabets, the question
of the typological relationship between CP san and Phoenician sade naturally
presents itself.
What we find when we compare CP san with the Greek letters Semitic
prototype, sade, is a much closer typological match. The gross shape of San-1
matches that of Phoenician forms of the eighth century b c (typically x),
though reversed.465 The relationship is not precisely one of a mirror image,
however. In the generating of a symbol having the morphology of CP San-1,
the spine of the Phoenician letter has as it were been removed from the
front side of the character and reattached on the back side so that the threestroke head retains its shape: that is, given the orientation x, the shape of
[upward diagonal + downward diagonal + upward diagonal]. The example of
San-1 on line 1 of plaque W-2 has a strikingly Phoenician appearance, mutatis
mutandis, as does that of MS1-2, line12.
If we allow for this operation, a Phoenician protoform of CP San-2 and
CP-San 3 also presents itself. This prototype is evidenced in the inscription
cut into the Gold Pendant from Carthage, dated circa 700 b c , that I invoked
in the preceding discussion of CP mu; here sade is formed with a spine that
leans diagonally inward beneath the head of the letter, and so descends diagonally away from the direction of writing.466 Transferring the spine to the back
side of the letter while keeping its inward orientation produces a symbol having the morphology of CP San-2. Alternatively, transferring the spine to the
back side of the letter while retaining its diagonal orientation away from the
direction of writing produces a symbol having the morphology of CP San-3.
96
The motivation for the removal and transverse reattaching of the spine is
not transparent, given the potential for homography with four-stroke nu
and four-stroke mu outside of the copper plaques that the operation created. It is almost as if the earliest Greek users of the alphabet were flirting
with ambiguity. If so, this would prove to be a dalliance from which some
later Greeks would retreat in imposing a symmetrical morphology upon san,
distinguishing it from mu, only to have that remedy undone, as the morphology of mu advanced toward that ofsan.
Of the three types of san occurring in the copper plaques, San-1, the type
having a vertically oriented long stroke, is by far the most common, occurring
thirty-nine times (found most prominently, relative to the other two types, in
W-2 and MS1-1). San-2, with its long stroke angled toward the direction of writing, occurs least often, being found only in the abecedaria of MS1-1, MS2-1, and
MS2-2, for a total of eleven occurrences. The number of instances of San-3 is
almost the mean of the previous two, with twenty-two occurrences. Both San-1
and San-3 are broadly distributed throughout the abecedaria of the six plaque
surfaces, though MS1-1 attests no instances of San-3. The distribution of the
three types San-1, San-2, and San-3 is summarized in the following chart:
(70)Distribution of CP Santypes
W-1
W-2
MS 1-1
MS 1-2
MS 2-1
MS 2-2
TOTAL467
-1
5
7
9
-2
0
0
2
-3
6
2
0
6
7
5
39
0
5
4
11
6
3
5
22
The particularly even distribution of the three types in the abecedaria of

plaque MS2-2 is conspicuous. Line 10 of MS2-2 contains an ambiguous sign:
the letter is either San-2 or Nu-1;468 the letter occurs out of its proper alphabetic position, and it is this isolation that occasions the ambiguity. If it should
in fact be San-2, then the distribution of the three forms of san in the abecedaria of MS2-2 would then be a perfectly equivalent 5:5:5. As we shall see,
MS2-2 is in a sense the most unusual, most interesting, and most revealing of
the six plaque faces.
2.19
97
q o ppa
Archaic Greek qoppa is typically subcategorized on the basis of the juncture

of the circular head of the letter and the spine that supports it. In the one case,
the spine joins the circle with little or no penetration (the type eventually
generalized);469 in the other, the spine passes through the head, effectively
bifurcating it. As McCarter discusses, each type has its own Phoenician precursor.470 The qoppa of the copper plaques is consistently of the formertype:
(71) Qoppa
CP Qoppa shows some small variation in the size of the circular head471 and
in the length of the spine but is otherwise quite constant throughout the abe
cedaria of the plaques. McCarters survey of the earliest Greek alphabets suggests a broader distribution of this type of qoppa that is, the type with an
undivided head.472 A notable characteristic of CP Qoppa is the length of its
spine, which tends to be somewhat longer than that of many occurrences of
archaic Greek qoppa.473 As the illustration shows, the spine of qoppa in the
abecedaria of the copper plaques is typically vertical; in some cases, though,
as the spine descends from the head, it curves away from the direction of
writing.474
Outside of the copper plaques, perhaps the earliest example of a comparable form of qoppa that is, one having an undivided head and long spine is
found in the script of the early seventh-century grave stele of Ankylion from
the Aegean island of Anaphe (mentioned in conjunction with the earlier discussions of theta and lambda).475 Others occur on a sherd from the Athenian
Agora (P4664)476 and on the grave stele of Pasidikos () from the
Cycladic island of Sikinos (invoked in the discussion of CP mu),477 both
seventh-century objects. Also from the Aegean the mid-seventh-century
inscription meandering across the leg of the Nikandra statue from Delos preserves (in Naxian script) a qoppa of similar shape.478 Only slightly later circa
650600 b c are examples of this type of qoppa occurring in Euboian-based
abecedaria on bucchero pottery from Etruria: namely, the bucchero amphora
from Monte Acuto (Formello),479 and the bottle from Caere bearing a syllabic presentation of letters.480 From the end of the seventh century, a similar
qoppa appears in the Euboian abecedarium on the bird-shaped bottle from
Viterbo481 and in names etched on exported Attic SOS amphorae, such as
<> (from Gela) and (from Caere).482
98
2.20
r ho
Archaic Greek rho typically shows either a rounded head ([dextroverse] r)

or an angular head ([dextroverse] ). The two are widely distributed: either
the former or the latter is the normal form everywhere in the early period,
observes Jeffery;483 and in her catalog of alphabets she indicates an early cooccurrence of the two in the scripts of Attica, Euboia, Boiotia, the Ionic and
Doric islands of the Aegean, and the Doric Hexapolis. Each of the two has a
Phoenician prototype: Early rho displays the same formal variation between
the old, round-nosed re (Phoenician r) and triangle-head re which characterize the Phoenician form early in the eighth century, writes McCarter.484
While the eighth-century Dipylon oinochoe from Athens485 and the midseventh-century Euboian abecedarium of the Marsiliana writing tablet486
both evidence angle-headed rho, the rounded variety also occurs in early
alphabetic writing from Attica and in other inscriptions from Etruria. Both
rounded and angular types occur, for example, in the seventh-century graffiti
from Mount Hymettos;487 and rounded varieties appear in the syllabic presentation of letters on the seventh-century bucchero vase from Caere (reversed in
stance), while an angular form is found in the accompanying abecedarium.488
Both varieties also appear in the eighth-century verse inscription of Nestors
cup from Pithekoussai (the rounded plainly in , though the spine is
damaged; an angular form in ).489 Both types show up in the rock
graffiti of circa eighth-century Thera: the rounded, for example, on IG XII iii
536 (), the angular on IG XII iii 539 (). Both are to be found
on the inscribed blocks from the temple of Apollo Pythios in Gortyn, Crete
(late seventh century b c or earlier; the angular, for example, in and
the rounded in ).490 And soon.
The head of the rho of the copper plaques is of the rounded variety; there are
no unambiguous examples of the triangular-shaped head so common among
the local archaic alphabets, with the possible exception of the rho of plaque
W-2, line 15, and that of MS1-2, line 13. There is, nonetheless, a variation in
the shape of the head one broadly attested among the abecedaria of the copper plaques. Most often, the spine of the letter abuts the right extremity of the
head in such a way as to flatten as it were that side of the globular head.
This morphological trait is typical of the rho with rounded head that occurs
in the various epichoric Greek alphabets, and also of its Phoenician precursor; the Semitic symbol is, after all, in origin a depiction of a human head in
profile. The rho of this shape in the copper plaques is designated Rho-1:
99
(72) Rho-1
CP Rho-1 accounts for forty-six of the seventy-six occurrences of the letter in
the abecedaria of the plaques.
Rho-2, the second variety of rho in the copper plaques, differs from Rho-1
in having a more bulbous head that is, one not flattened along the expanse
of its border with the spine (though see further for a caveat). Its shape is more
like that of a balloon attached to a stick or a qoppa with its head rotated to
the side:491
(73) Rho-2
In at least some of its occurrences, the character appears to consist of one continuous stroke (though this is not to suggest that the symbol was etched without the stylus being lifted from the plaque face): see, for example, plaque W-1,
line 11, and, especially, MS1-1, line 6. The spine of CP rho is at times slightly
shorter than that of its neighbor qoppa, and, as illustrated in (72) and (73), its
orientation is often predominantly vertical. However, as with qoppa, in some
cases the spine, as it descends from the head, shows a perceptible bend or
curve away from the direction of writing.492
This second avatar of rho in the copper plaques Rho-2, with its more
balloonish head is rarely encountered among other archaic alphabets
from Greece. A morphological match shows up on a Doric island of the
Dodecanese, on a sherd from the precinct of Apollo at Kalymna, probably of
eighth-century date.493 Scratched onto the sherd are various letters, some, at
least, clearly identifiable as Greek letters, including a rho with a circular head
of the Rho-2 type of the copper plaques.494 The alphabet of Kalymna shows a
curious similarity to that of Argos. Jeffery observes:
We can only say that Kalymna has produced sherds ranking among the earliest Greek inscriptions which we have; and these sherds may show that the
Argive local script came not from the same source as the Corinthian, but from
Kalymna or else some other, unidentified place whence Kalymna and Argos
both took their writing.495
Another example is provided by a fragmentary graffito from Mount Hymettos

(H 74; seventh century b c );496 the head of this rho is only loosely coiled, and
the character is formed of a single continuous stroke (and in this regard is
reminiscent of examples of CP Rho-2 just noted).497
100
While the familiar archaic rho having a triangular head may be all but
absent from the abecedaria of the copper plaques, some occurrences of the
symbol both Rho-1 and Rho-2 do exhibit a head marked by a degree of
angularity. Instances of this sort can be found, for example, at W-1, line 18
(Rho-1); W-2, lines 6 (Rho-1), 8 (Rho-2), and 14 (Rho-1); MS1-2, lines 6 (Rho2) and 17 (Rho-1); MS2-1, lines 4 (Rho-1) and 6 (Rho-1); MS2-2, lines 3 (Rho-1)
and 15 (Rho-1). One may be naturally inclined to attribute whatever angularity
these forms exhibit to the difficulty of executing curved lines with a stylus on
copper and indeed this is likely the cause of certain aspects of their individual morphologies. However, these occurrences of rho consistently share one
feature in common: the etching of a horizontal stroke directly to the spine
precipitously cuts off the lower portion of the curvinghead:
(74) Rho with horizontal stroke
While a comparable character again appears to be uncommon among archaic
Greek alphabets, a rho having a similar morphology can be seen in one of the
graffiti from Mount Hymettos.498 More interesting, and likely of much greater
significance the feature is typical of Phoenician r (i.e., re) from the tenth
through the eighth century B C .499
Rho-1 and Rho-2 occur with almost equal frequency in the abecedaria of
plaques W-2 and MS1-1 (with ratios of 6:5 and 7:5 respectively). The distribution of the characters is displayed in the following chart:
(75)Distribution of CP Rhotypes
W-1
W-2
MS 1-1
MS 1-2
MS 2-1
MS 2-2
TOTAL500
-1
4
6
7
8
10
11
46
-2
8
5
5
4
5
3
30
As can be seen, on plaque W-1, Rho-2 is twice as common as Rho-1 (8:4); conversely, Rho-1 is twice as common as Rho-2 on both MS1-2 (8:4) and MS2-1
(10:5). The greatest disparity occurs on MS2-2, where Rho-1 outnumbers Rho2, eleven occurrences to three.
2.21
101
sigm a
Regarding the distribution of the major morphological types of archaic sigma

among the local Greek alphabets, Jeffery generalizesthus:
The letter was stood on end by the Greeks; and the earliest writers apparently had not a fixed idea of how many strokes it should contain. The fourstroked 1[i.e., 1] is normal in early Samian and Milesian , Rhodian, Parian,
Arkadian, Boiotian; the three-stroked 2[i.e., ] in Naxian, Euboic (4[i.e., ] in
Rhegion), Attic, Thessalian, Elean, in the Ionic of Kolophon and Smyrna: but
the use of one form or of the other is never invariable.501
In contrast to Jefferys concluding observation that the use of one form or

of the other is never invariable the abecedaria of the copper plaques display notable invariability in the form of sigma utilized.502 CP sigma is a typical
four-stroke sigma, the Phoenician an (Hebrew in),503 c, stood onend:
(76) Sigma
Allowing for the difference in stance, the typological proximity of four-stroke
sigma to Phoenician an provides prima facie evidence for it being the earliest
form of the Greek letter. What is perhaps the earliest known example of this
letter, four-stroke sigma, comes from Pithekoussai, one of five letters found on
the previously mentioned fragment of a skyphos, dated to the second half of
the eighth century b c ; the stance of the character is, however, reversed (i.e., s
in a sinistroverse inscription).504 Aigina provides a late eighth-century example of four-stroke sigma (not reversed), found in a dipinto on a votive sherd
recovered in the area of the temple of Apollo.505
Seventh-century examples of four-stroke sigma are provided by the graffiti
from Mount Hymettos in Attica, as in the dedicatory inscription of H 163,506
where it co-occurs (although in reversed stance in this instance) with three-stroke
sigma, the more frequent form found in the Hymettos graffiti.507 Among several
other occurrences of four-stroke sigma from Hymettos is that of H 232,508 occurring in the intriguing dedication (Smioi Di) to Zeus Semios, where
is certainly a derivative of sign: Langdon interprets this otherwise
unattested epithet as a reference to (cloud-gathering) Zeus,509 the giver of weather
signs.510 In her review of Langdons edition of the Hymettos graffiti, Jeffery offers a
particularly insightful comment in this regard, worth rehearsing in full:511
The interesting problem remains why was so much inscribed ware offered,
so early, to this agricultural deity? L. suggests that Attica had only recently
102
become literate, and that the inscriptions were, in a sense, the dedication:
an inscribed pot was still an impressive object. He may be right; at least the
quantity of graffiti suggests that, in the worshippers view, Zeus Semios liked
to see inscriptions. The categories (above)512 could imply that, though some
pots were inscribed on the spot, others were brought along because they were
inscribed; one is even tempted to wonder whether Zeus epithet was thought
now, by some simple minds, to include these alphabetic , as well as
weather-signs. Finally, L. warns us that fine, decorated ware was, naturally,
rarely scratched over with graffiti; we should not expect to find so much writing on the sherds from a major sanctuary at this date. The Hymettos sherds
simply demonstrate anew how freely the alphabet spread through the levels of
Greek society, once it arrived.513
Returning to seventh-century examples of four-stroke sigma the symbol

is found in the first quarter of the century in the inscription on the Mantiklos
statuette from Boiotia, where instances of four-stroke sigma co-occur with
an example of five-stroke sigma (the so-called long sigma),514 an avatar of the
letter that is associated with the Lakonian alphabet,515 though also attested at
Smyrna in Ionia516 (a form of long sigma appears on the aforementioned sherd
from Pithekoussai of ca. 740 b c as well).517 Four-stroke sigma also occurs on
seventh-century bronze lebetes from Boiotia,518 co-occurring (in reverse
stance) with an instance of three-stroke sigma on the lebes awarded at the
funeral games of Ekpropos.519 The Naxian inscription on the Nikandra statue
from Delos preserves two occurrences of four-stroke sigma in reverse stance,
co-occurring with several examples of three-stroke sigma:520 Jefferys generalization regarding the distribution of three- and four-stroke forms of sigma
among the alphabets of the Ionic islands of the Aegean is that Naxos used the
three-stroke, and Paros and Keos the four-stroke, while Thasos, Amorgos,
and Delos used both.521 Four-stroke sigma also occurs in some of the seventhcentury Euboian-based abecedaria of Etruria, as in the two abecedaria on the
bucchero amphora from Formello, with one example showing forward stance,
the other reversed. Beyond the two abecedaria, additional letters appear on
the vase and among these are other four-stroke sigmas, some facing one direction, others turned opposite.522 Four-stroke sigmas also appear in both the
abecedarium and the syllabic presentation of letters on the bucchero bottle
from Caere.523 These Etruscan examples of four-stroke sigma are characterized by rounded, rather than sharply angled, joints; a similar morphology
occurs elsewhere, as in the Ionic Dodekapolis.524
103
Given sigmas variability within and across the local archaic Greek alphabets, the occurrence of a uniform type in the copper plaques where, as we
have by now observed many times over, a high degree of graphemic variation,
subtle or flagrant or something in between, presents itself almost takes us
by surprise. This absence of sigma variation may very well be noteworthy. It
is certainly curious.
2.22
tau
Archaic Greek tau, on the other hand, shows little variability among the
epichoric alphabets. The tau of the copper plaques is morphologically consistent with what one finds elsewhere:
(77) Tau
Tau allographs can be identified, both in the copper-plaque abecedaria and in
various local scripts of archaic Greece. Within the copper plaques, however,
the variants are so few and, often, the variation so subtle that variant lettertypes will not be individually designated but will simply be noted in the following discussions.
Jeffery speaks of the typical form t as universal in Greek except for variations like 2[i.e., t] in badly-written graffiti; she continues:
No marked changes in the shape took place; evidently the Greek instinct for
symmetry was satisfied with it from the start. In Etruscan, however, the form
2 [i.e., t] is often found, and we can trace this back to the Etruscans early
model abecedaria.525
Such forms, with the crossbar descending toward the direction of writing, do
indeed occur in the seventh-century alphabets from Etruria: examples can
be seen in the abecedarium of the Marsiliana writing tablet,526 in that of the
bird-shaped bottle from Viterbo,527 in both of the abecedaria of the Formello
amphora,528 in the partial abecedarium on a red-brown impasto amphora
from Veii,529 and in both the abecedarium (here with the crossbar descending
away from the direction of writing) and the syllabic presentation of letters on
the bucchero bottle from Caere.530
But in Italy this morphology is not limited to Euboian-derived Etruscan
abecedaria. An example can be seen in a late eighth-century graffito of four
letters () found on an amphora of local manufacture from Pithekoussai.531
At least one of the instances of tau in the inscription of Nestors cup, also from
104
Pithekoussai, is unmistakably of this sort (in line two, []), having a

crossbar that descends at an angle equivalent to that of the descending bars
of the epsilon that stands next to it.532 Another, less pronounced, is found in
a graffito on a small Protocorinthian aryballos from Kyme, the so-called aryballos of Tataie (IG XIV 865; ca. 675650 B C ):533
h
Tatais mi lkuthos hos d an me klephsei thuphlosestai
I am the lekythos of Tataie; whoever steals me will goblind
The bar of the first tau is horizontal, that of the second tilts forward slightly;
the third tau is different still: while the crossbar joins the spine at a right angle,
the entire letter tilts forward precipitously in a manner that exceeds any tilt
shown by the characters that surround it. Distinct sixth-century examples of
tau with forward-leaning crossbar are provided by funerary inscriptions from
Kyme, such as that of Kritoboule (IG XIV 869; ca. the early sixth century)534
and that of Lenos (IG XIV 871; ca. the last quarter of the century).535
Even in the Greek heartland, forms of tau with the tilted crossbar occur
and occur in at least some contexts that one may be hesitant to relegate to
the status of badly-written graffiti. From the Athenian Acropolis comes a
fragment of inscribed stone, circa the late eighth century, perhaps preserving
hexameter (IG I2 484): the second line can be restored as ][, where
the crossbar of tau has a downward tilt, though the angle is somewhat less
acute than that of the bars of the ensuing epsilon (with a tiny omicron intervening).536 The same can be said of one of the instances of tau in the painted
inscription on the oinochoe from Ithaka (ca. 700 b c ) discussed previously
that tau occurring in the sequence ].537 Occurrences can be seen in ostraca from the Athenian Agora, such as P 4664 (late eighth or early seventh
century b c ), <>.538
Tau with forward-tilting crossbar is also found in the copper plaques,
though forms with a conspicuous, unambiguous tilt are few in number and
limited to two of the plaque surfaces. Examples of such a tau occur on plaques
MS1-2, line 3, and MS2-1, lines 1, 4, and 7. Less pronounced is the tilt of the
crossbar of the tau of W-1, line 8. Less marked still and of uncertain morphological significance are the occurrences of MS1-1, line 3; and MS2-1, lines 3,
6, and 15 (though compare the conspicuous examples of lines 1, 4, and 7 just
mentioned).539
The form of Phoenician taw to which Greek tau is typologically closest
is the elongated +-form of the early eighth-century and later,540 that is,
105
forms such as t, or more upright varieties such as those seen in Phoenician

Cyprus541 the red slip jug of nt542 and the Bal Lebanon inscription543 and
on the Malta stele.544 The Phoenician character possesses a spine that continues upward through the crossbar; and this is indeed a characteristic of some
forms of archaic Greek tau; thus, Guarducci observes:
Nelle iscrizioni greche pi antiche il tau conserva un aspetto pi conforme
a quello della seconda fra le due varianti fenicie sopra indicate [i.e. the form
I have noted above]: il tratto verticale emerge dal tratto orizzontale. Poi si
giunge al tipo t, che resta, attraverso I secoli, invariato. Qualche volta, specie
nelle epigrafi meno diligentemente scritte, il taglio si presenta obliquo (T).545
In some instances of Greek tau, the extension of the spine above the crossbar
is so slight as to leave one suspecting that the penetration may be only the
artifact of a less-than-efficacious attempt to produce a tau having t morphology; however, in other instances, as Guarducci advocates, the Greek symbol
appears to be unquestionably echoing the penetration characteristic of the
spine of Phoenician taw. The tau of the Etruscan abecedarium inscribed
around the base of the bucchero bottle from Caere, for example, is of this
type,546 as is that of the early sixth-century funerary inscription of Kritoboule
from Kyme547 already mentioned (and, as noted, both also have a tilting crossbar). At least one occurrence of tau in the inscription of Nestors cup from
eighth-century Euboian Pithekoussai is of the extended-spine variety (in line
one, ).548 Immerwahr finds three different varieties of tau co-occurring in the inscription of the Dipylon oinochoe, from eighth-century Athens:
in addition to the unmarked tau, he identifies in the graffito both the type
with a tilting crossbar (the type discussed earlier) and the type with a spine
extending upward beyond the crossbar.549
Examples of tau with a spine that continues upward through the crossbar also occur in the copper plaques, but, as with the variety having a tilting
crossbar, well-defined instances of the morphology are few: MS1-2, lines 11
12, and 15. In addition, a nick of the spine is visible above the crossbar in the
occurrences of tau in W-1, lines 10 and 17; MS2-1, lines 10, 13, and 16; and
MS2-2, line4.550
***
Great care has been exercised in identifying the forms of each of the letters of
the CP abecedaria and the occurrence of parallel forms in local archaic Greek
scripts or the absence thereof. What I would like to do next is to examine
106
each plaque face individually, to discover the sequences of letter forms that are
used to generate the several abecedaria appearing on each face, and to identify
the various irregularities that occur in the fabricating of these abecedaria,
and other noteworthy phenomena (see Chapter 4). Before doing so, however, a chapter authored by Professor David A. Scott of the Cotsen Institute
at UCLA and the J. Paul Getty Museum in Malibu will be presented, in which
the author presents the results of chemical and physical analyses of the copper
plaques results that demonstrate the antiquity of the documents.
3
Physical and Chemical Examination
of the Copper Plaques
David A.Scott
,

, .
Many, many are the things I left behind on this fools journey
to this place; yet from here others gold and ruddy copper
and women fair of waist and grayishiron
shall I bear away what I took by lot at least.
Homer Iliad IX 364367 (spoken by Achilles,
as he prepares to depart the Greeks atTroy)
3.0
i n t r o du c tio n
There are seven important techniques employed for the technical examination of these plaques. These are visual examination with a binocular bench
microscope (BBM), scanning electron microscopy with environmental
capabilities allowing sample to be entered into the machine without coating
(ESEM), X-radiography (XR), X-ray fluorescence analysis (XRF), X-ray diffraction analysis (XRD), electron probe microanalysis (EPMA), and optical
metallography(OM).
A complete description of these techniques is beyond the scope of the present chapter. Suffice it to say that careful optical examination of an artifact with
a bench microscope, illuminated with a fiber optic light source, using magnifications from 4x to 40x, is an essential part of the study of the surface features
107
108
of the plaques. ESEM allows the imaging of samples using electrons, either in
secondary electron mode or in back-scattered electron mode, to allow imaging of very fine detail; X-radiography can be used to examine the internal
structure of an object in much the same way as a hospital radiograph is used
to examine internal parts of the body, nondestructively. X-ray fluorescence,
which is also primarily employed as a nondestructive technique, is a method
of chemical analysis that employs a beam of X-rays to energize a tiny area of
the surface of the object. The object then emits a characteristic set of X-rays
that is dependent on the chemical constituents of the artifact placed in the
path of the beam. It is primarily used for surface analytical studies and cannot
be relied upon to provide good-quality quantitative data; this is the province
of the EPMA, where comparison with freshly acquired elemental standards
enables polished sections of the metal to be analyzed with good precision for
a variety of elements. For copper alloy artifacts, we usually analyze for the following: copper, lead, tin, zinc, antimony, arsenic, nickel, silver, gold, manganese, and bismuth. X-ray diffraction can be used nondestructively, or on tiny
samples smaller than a pinhead, and uses a focused beam of X-rays to impinge
on the artifact, this time providing information on the diffracted X-rays from
the specimen or object, revealing the crystal structure of the components,
rather than their elemental composition per se. Metallographic examination
is the only destructive part of the procedure employed on the plaques but is
essential; it involves the removal of a small sample, about 1.5mm3 in volume,
which is then mounted in a synthetic resin and ground and polished to a mirror finish on silicon carbide and diamond laps. This mounted sample can then
be etched in selective reagents to reveal the nature of the microstructure of
the artifact being studied. The microstructural examination provides details
concerning the extent of casting or working, the extent of corrosion, the interface between the patina and the remaining metal, and the type of alloy or its
thermal history. The mounted sample can then be used for EPMA analyses
after careful repolishing, to obtain quantitative elementaldata.
The three plaques examined are referred to in this chapter as Greek Plaque
I, from the Schyen collection, the first plaque to be studied; Greek Plaque II,
the second from the Schyen collection; and Greek Plaque III, the plaque in
the collection in the Martin-von-Wagner-Museum in Wrzburg. It is the lastnamed plaque that created difficulties in the art-historical critique of these
plaques as it was pronounced to be a forgery by scientists at the Doerner
Institut in Munich, a conservation facility associated with the Bayerischen
Staatsgemldesammlungen, centrally dedicated to caring for the paintings
Physical and Chemical Examination of the Copper Plaques
109
housed in that collection (fourteenth century to contemporary) and offering

for-fee examination of art for the declared purpose of detecting forgeries.1 The
reasons for this conclusion were based on the fact that the plaque was made
in pure copper, with the presumption that it should have been made in a tin
bronze; that the patina was composed essentially of paratacamite, one of the
basic copper chlorides,2 whereas the typical buried patina component is malachite; and that the symbols on the plaque suggested a version of the Greek
alphabet unlike any single known epichoric alphabet.
3 . 1
e l e c t r o n p r o be mic r oa na lysis
Examination of these potential problems will help us to address the important questions of how the plaques were made and what they were made from.
During the Bronze Age a number of different copper alloys were in use, the
most important of which were the copper-arsenic alloys of the early Bronze
Age, followed by a transition to tin bronze by the middle Bronze Age.3 The
addition of tin to copper hardens the alloy, changes the color to a pale gold,
and lowers the melting point for casting operations, all of which are beneficial
changes in property compared with pure copper. But are these characteristics
required for the purpose of the copper plaques? X-ray fluorescence analysis
of the plaques reveals that they are, indeed, made in pure copper, but there
are many other examples of artifacts made in pure copper from the Bronze
Age, so how much of a logistical difficulty does this create? The polished sections cut from each of the edges of the plaques, which will be described later,
were used to obtain EPMA data, which are given in Tables 3.13.3 (for all
figures and tables associated with this chapter, the reader is referred to the
books dedicated Web site: www.cambridge.org/9781107028111). The EPMA
data confirm the overall conclusion derived from the XRF study that the
plaques are made in quite pure copper. Some tiny inclusions present in the
copper are rich in antimony, nickel, and lead, which are all elements associated with copper deposits. Notice in all of the analyses in these three tables
that the same features in the composition appear. All contain some nickel and
no tin; have low arsenic content, only present here in a few tiny inclusions;
and have low sulfur content. None of the inclusions in the metal appear to be
of sulfide minerals, suggesting that the smelted copper was produced from
oxidized copper ore bodies and not from complex sulfides. The presence of
small amounts of nickel in copper is a common feature of early Greek copper. For example, Liritzis notes that the lowest nickel levels in copper from
110
the Aegean are observed from the Troad and the Cyclades, where the levels
may reach approximately 0.4percent.4 Some copper with antimony content
is also found from the Troad.5 Levels of antimony in the Cyclades are low and
are present in artifacts with 9193 percent copper, that is, those which are
already alloyed. Unalloyed coppers in the Cyclades have no antimony, and so
the antimony comes from the alloying element. Cretan artifacts have a fairly
consistent level of antimony around 0.12percent whether alloyed or not. As
regards iron content, Liritzis notes that values can be scattered.6 The levels seen
in Tables3.13.3 are consistent with the use of high-quality copper ore, and
the tiny amounts found are comparable with data from Crete, where 0.01percent iron is an average, and Greece, where a majority of artifacts analyzed has
no iron content.
3 . 2
bi n o c u l a r be n c h mic r o s c o p y
One of the most important aspects of the scientific examination of the plaques
is what might be called the technical connoisseurship of the patina and the corrosion. BBM examination revealed a complex array of different patina variations on the surfaces of the three plaques. For example, Figure3.1 shows a letter
on Plaque I that is preserved in a cuprite surface, with the depressions in the
metal where the copper has been inscribed, partially infilled with light green
corrosion products. To the right-hand side of this photomicrograph can be seen
a long, thin sliver of wood and next to it some small flakes of charcoal. Probing
the lettering with a microscalpel reveals further cuprite under the letters themselves, showing the alphabet is preserved entirely within a layer of corrosion,
principally cuprite, rather than within the metal itself. This feature is usually
associated with the natural corrosion of the copper alloy during burial.
The patina has many variations, such as that shown in Figure3.2 for another
region of Plaque I. Here, only a small, darker cuprite layer can be seen overlaid with malachite in the form of small, hard concretions, and which incorporate some soil minerals and rounded quartz grains cemented to the malachite surface. Malachite frequently assumes a botryoidal morphology; there is
a slight patterning to the surface as if it may have been in contact with a textile
or other organic material.
Figure 3.3 presents an unusual appearance for a copper patina with
squashed-looking bluish regions on a light green-blue corrosion, which is
principally of malachite. The bluish regions may be of azurite and indicative
of local variations in burial conditions over the surface of the plaques.
111
The presence of azurite within the corrosion crust is confirmed by the typical dark blue crystalline particles that are interspersed within the malachite
crust (see Figure3.4). The very dark azurite may also represent a slight patterning over the surface of one side of Plaque II, which has the appearance of
a netting or textile imprint with which the plaque may have been in contact.
3 . 3
x - r adio g r aph y
X-radiography was carried out at 90110 Kv, 30 mA, with a variety of timed
exposures to ensure that the lettering was revealed clearly. An overall view
of one such X-radiograph is shown in Figure 3.5. Magnified views of these
X-radiographs enable the lettering to be read by Professor Woodard on both
sides of the plaques, as they are all inscribed on the obverse and reverse
surfaces.
A remarkable fact revealed by the X-radiographic examination of the
plaques is that each of them shows an identical pattern of mottling in a morphology that continues along the length of each plaque. Each plaque shows
the same kind and distribution of markings in the X-radiograph, which is a
unique fingerprint of the fabrication technology used in the production of
the plaques.
There are three ways in which large flat sheets of copper alloy could have
been produced from an initially smelted copper product. First, they could be
cast by pouring the alloy into a shallow mold. This, however, is very difficult
to control and has been rarely used for sheet production. Second, the alloy
could be heated and passed through rollers, which would squeeze the copper
out into a flat sheet, being effectively hammered and annealed in one operation, a process that is called hot-working. Third, the copper could be hammered while cold and annealed in the fire to shape, followed by further cycles
of hammering and annealing to shape the required sheet. The X-radiographs
prove that the third method, the hammering and annealing process, has been
used to make the copper sheets. Not only that, but a considerable degree of
forethought must have gone into the process, because the initial shape of
the hammered copper sheet would have to have been the length of all of the
plaques combined. This long sheet was then cut, probably with a chisel, into
the required shapes for the plaques we are examining here. After cutting,
edges could be rounded and holes cut through the sheets at each corner. The
X-radiographs show that this is the case, as the same pattern of hammering
marks, evidenced by the mottling revealed on the X-radiographs, is common
112
to each. Pure copper was used for this purpose, not only because it is so much
easier to hammer out into a very long sheet but also because the hardness of
the final product is quite soft and could be easily engraved or cut with a chisel.
For comparison we can consider the microhardness of a 5percent tin bronze
when hammered to shape. On the Vickers scale, the tin bronze might reach
140Hv, while the pure copper cannot get any harder than about 90Hv. This
means that a well-hardened bronze chisel with a more substantial amount
of tin, for example, from about 810percent tin content, could reach a hardness of about 180Hv and could easily be used to cut lettering into the copper
plaques. There are two basic methods of recording information for letters or
symbols in copper: the lettering can be made by either repouss or engraving.
In repouss, the copper is displaced by a blow from a sharp edge, resulting in
the lettering being forced into the copper surface, displacing the surrounding
metal in the process. This technique was employed, for example, in the fabrication of the copper Dead Sea Scroll, which incidentally is also made in pure
copper. This is not the technique used on the plaques, where engraving of the
copper surface removes a V-shaped channel in the copper surface and there is
no displacement of metal in the process. The approach to formation of these
letters is therefore dependent to some extent on the thickness of the copper
sheet employed for the plaques and is therefore an integral part of the fabrication process in terms of the choices to bemade.
In the case of the Dead Sea Scroll, a copper sheet was chosen, somewhat less
than 1mm in thickness, which can be easily deformed under pressure for the
repouss lettering employed on that object. In the case of the copper plaques,
a copper sheet a little more than 1mm in thickness was made, which would
be hard to deform by repouss, but would be ideal for engraving, especially as
the copper plaques are engraved on both sides of the sheet, which would not
be possible if repouss had been employed, as the displacement of the copper
would seriously interfere with the aim to produce an inscription on both sides
of the plaque. We therefore see the rationale behind the decision to make the
plaques in copper and to produce them in a certain preferred thickness.
How do we know that these sheets were worked (hammered) and
annealed? Apart from the indirect evidence derived from the X-radiograph,
we can determine the extent of working from a metallographic examination
of the copper plaques, which entails the removal of a small sample of each of
the three plaques from a suitable area of one edge. The sample was cut with
a fine-toothed jewelers saw, which avoids any heating of the sample while
it is removed. A V-shaped section was cut from the plaques, approximately
113
1 x 1.2mm at the widest point, and mounted in a Buehler epoxide resin, followed by grinding on silicon carbide papers of grit sizes 240, 400, 600, and
800, polishing on diamond laps of grit size 6 micron and then 1 micron, followed by etching in alcoholic ferric chloride.7
3 . 4
o ptic a l meta l l o g r aph y
The aim of this metallographic examination is to determine the extent of corrosion of the plaques, the nature of the copper grains that make up the sheet,
and the extent of the deformation of the copper. Figure3.6 illustrates the cross
section taken through Plaque I and shows that the grains are worked and
annealed to shape with twin lines in the copper crystals that are perfectly
straight. When copper alloys are annealed, they recrystallize with twin planes
that are completely straight, and if subsequently cold-worked again, these
twin lines then become themselves deformed.
Repeated annealing recrystallizes the metal, and the twin planes in the copper crystals are once again straight. Another photomicrograph of the recrystallized grains from Plaque III is shown in Figure3.7.
We therefore know that the plaques were worked and annealed to shape,
with annealing being the final working process. This is generally the preferred
fabrication technology employed in the manufacture of hammered copper
alloys, because leaving them in a hardened condition risks embrittlement
and lack of ductility, and soft copper is in any case required here in order to
engrave the plaque with hardened bronze tools.
3 . 5
the pati na
What kind of patina is present on the plaques, including Plaque III, which has
been called into question? The normal sequence of corrosion products that
we find on buried copper alloys is, first, a cuprous oxide crust of red or redorange color overlaid with an outer patina or corrosion crust of a basic copper
carbonate, usually malachite. This malachite crust can be found in a variety of
different forms, from botryoidal to acicular, from fibrous to massive, banded
to spherical, and can be a variety of shades of green. Incorporation of soil
minerals or associated organic remains can then be cemented into this outer
malachite crust and may be partially replaced or infiltrated with copper salts
from the corroding plaque, which can help to preserve traces of associated
organic materials, which would otherwise undergo rapid biodeterioration in
114
burial. The corrosion crust on the Greek Plaque I can be seen in Figure3.8,
which is an unetched view taken under crossed polar illumination.
This reveals the dichroic colors of the minerals under examination; thus
cuprite appears red, and the oxidized copper salts, the basic copper carbonates, and chlorides will appear various shades of green. The cross section
shows a complex morphology of patina constituents in which discrete crystalline growth of some microconstituents has taken place. The cuprite corrosion layer is of a substantial thickness, being about 20 microns thick, and
contiguous with the remaining metallic grains, a feature usually associated
only with genuinely old copper alloys, because getting the cuprite crust to
grow to a thickness of more than a couple of microns or so is very difficult
indeed. A natural patina, formed slowly in the soil, should possess a substantial cuprite crust overgrown by a cupric crust, which is often of malachite. The
depth of corrosion of the plaques is even more substantial on Plaque II, as can
be seen in the photomicrograph shown in Figure3.9.
The depth of penetration of the red, crystalline cuprite layer into the copper
grains revealed by Figure3.9 is generally taken as one indication of authenticity, because we have evidence for this kind of event only from natural corrosion processes over long periods of time. Figure3.10 shows the polarized light
metallograph of Plaque III, which also shows an extensive corrosion crust,
but one rather different in extent to that seen in PlaqueII.
The inclusions within the three copper plaques are all small circular globules of cuprite, and there are no sulfide inclusions visible at all. These cuprite
inclusions are not corrosion products but primary cuprite from the smelting
and melting stages of the original ingot of copper from which the plaques
were formed. These cuprite inclusions show a stationary cross under polarized light, as shown in Figure3.11, which is a consequence of the optical indicatrix for cubic minerals such as cuprite. Because the particles are monocrystalline, the polarized view of them is able to discern this feature of the primary
cuprite, which is a characteristic usually associated only with copper extracted
from high-quality oxidized or carbonate ores of copper.
The predominance of malachite in copper alloy patinas assumes that soil
groundwater solutions containing carbonate or bicarbonate solutions predominate in the environment in which the plaques were buried. This is invariably
the case, for example, in Europe or parts of Asia, where the extent of periodic
rainfall is enough to ensure that the groundwaters are continually replenished
with carbonated solutions promoting the formation of a malachite patina over
the copper artifact, but this is not necessarily the case with burial in drier soils,
115
such as those which might be found in parts of Cyprus, Greece, or Egypt,

possible locations where the plaques may have been originally buried.
Bronzes from the Middle East or Northern Africa may be buried in soils
where chloride ion concentrations can locally reach levels that are substantially higher than those found in more temperate zones. Under these circumstances, the patina that develops may incorporate substantial amounts of
chloride ions, resulting in a patina that is composed of basic copper chlorides,
such as atacamite or paratacamite. These kinds of compounds are also associated with nantokite, cuprous chloride, which is the principal agent of the
decay process known as bronze disease, in which this chloride-containing
compound reacts with oxygen and moisture to form one of the copper trihydroxychlorides, such as paratacamite. The existence of these compounds can
be rationalized on the basis of the Pourbaix diagrams, which relate the pH
of the natural environment with the electrochemical potential of the milieu
in which the object is buried. We can see what should occur in the presence
of soil carbonate-containing groundwater solutions by regarding Figure3.12
as simply a kind of map for the specific burial parameters assumed by the
concentrations of copper ions, carbonate ions, and chloride ions indicated at
the top of Figure3.12. At typical soil pH levels between 6 and 8, we can see
that the stable mineral species are malachite and atacamite. Malachite occurs
if the chloride ion concentration is low, and as the chloride content rises, the
malachite is replaced by atacamite.
These diagrams always show atacamite as the principal chloride-containing species, as this form of the copper trihydroxychlorides is the most stable
from the thermodynamic standpoint, but in practice a variety of compounds
may be found.8
This model helps to explain that, given the right kind of chloride ion concentration in the soil, the overall patina of a buried copper alloy may very
well be a phase such as atacamite or paratacamite. In fact, analogous types
of patina have been reported by Scott and Dodd in their discussion of the
corrosion of an inlaid Egyptian bronze statuette of the god Osiris,9 which
was covered with a thick and voluminous patina of atacamite over a cuprite
underlayer. Atacamite is known as a patina constituent,10 but is unusual in
forming the complete patina of an object. Not only do we have clear evidence
to link all of the plaques together, but there is a perfectly reasonable explanation for the presence of the paratacamite patina (actually a mixture of paratacamite and malachite) that the Munich lab cited as a reason for questioning
the authenticity of PlaqueIII.
116
How are the letters carved in the copper surface able to preserve the exact
impression of the letters as if they were still present within copper metal?
This we can explain as part of the growth process that occurs when copper
interacts with oxygen and moisture to form slowly the cuprite layer. In this
process of corrosion, copper atoms move outward and interact with oxygen
as the anions move into the metal. The formation of this cuprite layer grows
in such a way that it mimics some of the lattice characteristics of the original
copper surface in which the letters were carved. This form of growth is called
epitaxial, and there is therefore an epitactic relationship between the original
carved letters and the cuprite that replaces the copper metal. One of the consequences of this type of growth is that the corroded surface is a pseudomorphic replica of the original, preserving intact the original form of the letters
within the corrosion crust, not within remaining metal, as is clearly seen in
Figure3.13.
3 . 6
s c a n n i n g e l e c t r o n mic r o s c o p y
Some of the letters are easier to read on the X-radiograph than on the plaques
themselves because a further series of corrosion events has obscured the epitactic growth and filled in the surface details with another layer of corrosion
and soil minerals, which are firmly cemented to the surface of the plaques
in some areas. The principal of pseudomorphic replacement can extend to
associated materials from the burial context, such as textiles, wood, feathers,
bone, or leather. Figure3.14 shows part of a pseudomorph of a woody cellular structure that has been preserved by replacement with copper corrosion
products, usually malachite, from Plaque I. The impressively preserved cellular structure shown in the ESEM scanning electron photomicrograph of
Figure3.14 is a remnant structure of the spiral thickening along tracheid cells
in hardwoods that wind along the inner surfaces of the tracheid.
During burial, the organic matrix of material buried with the plaques may
be replaced, in part or in whole, by copper salts. This process of mineralization involves the combination or replacement of the organic matrix with
an inorganic one.11 In positive replacement, copper (or other) ions penetrate
the fiber of organic materials and coordinate with the organic matrix. These
coordination sites create additional nucleation areas that attract more copper ions. Corrosion products then gradually replace the fiber or wood as it
decays, forming a positive cast composed primarily of copper corrosion products, typically the basic chlorides or carbonates. Negative casts form when
117
the corrosion products deposit on the surface of the fiber, which then decays
and leaves residual copper corrosion products as a negative impression of the
original material. Because copper ions are strongly biocidal, their impregnation of the organic material may preserve it from biodeterioration during
burial, and indeed small fragments of undecayed wood and charcoal are to
be seen within the corrosion layers of the plaques. A view of a different area
of the malachite crust on Plaque I with collapsed coils of copper corrosion
products resulting from the spiral thickening of the hardwood cells is shown
in Figure3.15.
In general terms, we take the presence of mineralized wood fragments on
the surfaces of the plaques as further evidence for the authentic nature of the
burial deposits.
3 . 7
x - r ay diff r ac tio n a na lysis
X-ray power diffraction was used to determine the mineral phases present on
the plaques. These were found to be of malachite and cuprite; or malachite,
atacamite, and cuprite; or paratacamite, malachite, and cuprite. Two typical
X-ray diffractograms are shown in Figures3.16 and 3.17, where malachite and
atacamite, and malachite and cuprite are found to be present.
3 . 8
c o n c lu sio n s
In authenticity studies it is important to pursue all lines of evidence, which

has been attempted here with the resources available. The type, extent, and
penetration of the corrosion into the copper grains are all perfectly acceptable
for being due to long-term burial. In fact, the bright-red crystalline cuprite
in some depth occurring within the patina is usually taken as very strong
evidence for authenticity in combination with other technical lines of evidence. The composition of the copper and the nature of the inclusions within
the copper are typical for copper extracted from high-quality ores, and the
presence of tiny cuprite inclusions are commonly found only in early copper.
The presence of mineralized wood remains is a feature often associated with
burials in which the organic materials have slowly been replaced by copper
corrosion products, and the only known way to fake these is to glue or attach
genuine old fragments of such mineralization crusts to the surfaces of other
objects, but this clearly has not been carried out in the present case and is
another marker of authenticity. The hammering marks on the X-radiographs
118
not only clearly link all of the plaques together as having been made from
the same sheet of copper but also show the handmade nature of the sheets
employed and the fact that they formed an integral part of a very special
commission.
The difference in the corrosion products found on Plaque III, which are
predominantly basic copper chlorides as compared with the often complex
patina of Plaques I and II, shows that the burial environment of the three
plaques was subject to local variation in burial conditions, with higher chloride ion concentration in some. There is some evidence for the presence
of textile patterning on the surface of Plaque II, where tiny dark corrosion
products are patterned along the surface. The presence of the well-crystalline azurite particles within the malachite crust is another indication of the
authenticity of the patina and is due to local variations in the carbon dioxide
concentration in the groundwaters during burial, perhaps influenced by the
presence of organic materials associated with the plaques.
The lettering of the plaques can now be interpreted here in the knowledge
that the results of the scientific connoisseurship are strongly supportive of the
authenticity of the plaques and that their authenticity has been established
beyond any reasonable doubt.
4
The Syntagmatic Structure of the Copper Plaques

.
And having joined the yoked letter [consonantal] with the unyoked [vocalic],
Cadmus will teach the mysterious rites of well-spoken language.
Nonnos, Dionysiaca 41.381382 (a prophecy revealed by Harmonia
for Cypris from a golden plaque [pinax], when Cypris had
interrupted Harmonia from her weaving)
4.0
i n t r o du c t io n
On the following pages, the right-to-left sequence of letter types appearing on

each of the plaque faces is presented line by line. Thus the reader will here find
a complete transcription of the abecedaria. The presence of what might be construed as an anomaly has been signaled chiefly in two ways: (1) by underlining the
symbols involved or (2) by placing the affected characters within parentheses.
Underlining indicates a departure from regular alphabetic order: typically
this results from the metathesis of two letters or sequences of letters or the
insertion of one or more letters within an abecedarium so as to disrupt expected
alphabetic order. For example, the transcription of plaque W-1, line18
(1) T -1 -1 -4 O-1 N-2 -1 M-2 K-2 I-2-1
indicates that the alphabetic order of lambda and mu has been reversed, while
that of W-2, line5
(2) M-1 M-1 -1 K-1 I-3 -1/2 H Z FE
signals the insertion of an additional mu in the symbol sequence.
119
120
Included among the occurrences of underlined characters are those instances

in which the xi-symbol occurs in the eta-position, and vice versa, and those
instances in which the omicron-symbol occurs in the theta-position, and vice
versa. Thus, the underlining of letters in the transcription of W-1, line16
(3) -2 -1 -1 -1 -2 -2 -2
denotes the metathesis of the xi-omicron letter sequence and the eta-theta
letter sequence.
Parentheses are used to mark the absence of a letter or sequence of
letters from their expected alphabetic position. For example, MS 2-1, line
10, is transcribed as follows:
(4) F E (B) A-4 T -1/2 -1 -4 O-1 (N MK)
The parentheses indicate that the sequence of kappa through nu has been
deleted from the abecedarium that comes to an end in this line and that beta
has been deleted from the abecedarium that begins in thisline.
I refer to these departures from expected alphabetic order as anomalies
and not as mistakes. Scribal errors quite possibly occur (people make mistakes), but, as will become clear, or so I will argue, the common culprit in the
commission of many of the anomalies was intention, not error.
In addition to underlining and parentheses, I have once used a double slash
// to indicate the anomalous absence of the greater part of an entire alphabetic
sequence: the transgression can be found in MS 2-2, line 10. The gap is attended
by a disordering of symbols on either side. The single occurrence of such a radical departure from the scribal repetition of abecedarium after abecedarium (if
not always fully canonical) makes for a lapsus calami that is so conspicuous that
one immediately asks why. I say single occurrence, though a few lines after
(line 16), a different sort of disruption a near complete abandonment of alphabetic order can be seen (and is not marked in the transcriptions). These lines
will occupy our attention in Chapter5.
4.1
na r r ow o rt ho g r aph ic t r ansc r i p t io ns o f
t he c o ppe r p l aqu es
As in the synoptic presentation of the distribution of zeta characters in

Chapter2, in the following transcriptions, the symbol is used for the I-form
zeta (i.e., CP Zeta-1) and the symbol is used for the reverse-stance Z-form
zeta (i.e., CP Zeta-2). The transcriptions of other letters are also as specified in
the discussions of Chapter2.
121
Wrzburg 1
-1 -1 -2 -2 -1 -2 F -2ii-2
-2ii -3 -2 -3 -4 -1 -1 -1 -2 -2 -1-1
O-1 N-1 M-1 -2 K-1 -1 -2 H F E -2/
-1 H F B-2/ -1/3 T -2 -1-2
-2ii A-1 T -2 -3 -4 O-1 N-1 M-1 -2 K-2I-2
-4 O-2 N-1 M-1 -1 K-2 I-1 -2 H Z FE
-2 H F E -2ii A-2 T -1-3
-2ii A-3 T -1 -1 -4 O-1 N-1 M-1 -2 -2I-2
O-1 N-1 M-1 -2 K-2 I-2 -1 H FE
-2 F E -2ii A-2 T -2 -3 -2
-2i -2 -2 -1 -4 -2 -1 -1 -2 ()-2
-3 -1 -2 N-1 M-1 -2 K-2 I-1(?) -1 HFE
-2 K-2 I-1 -1 (H) Z F E B-2 A-3 T -1
F E -2 -1/3 -2 ( O ) N-1M-1
-2 -1 -2 -3 -2 -1 -1 -2 () -2 -2
-2 -1 -1 -1 -2 -2 -2
H F E B-2 A-2 T -2 -1-4
T -1 -1 -4 O-1 N-2 -1 M-2 K-2 I-2-1
M-1 -2 K-2 I-2 -2 H F E -2iiA-2
10
15
122
Wrzburg 2
E B-1 A-2 T -2 -1 -4 O-1 N-1
-1/3 -4 O-1 N-2 M-2 -1 K-1 I-2 -2HF
-2 H F E B-1 A-4 T -2
B-1 A-2 T -2 -3 -2 O-1 N-1 M-2 () K-1I-2
M-1 M-1 -1 K-1 I-3 -1/2 H Z F E
F E B-1 A-3 T -1 -1 ()O-2
-1 -2 -1 -1 -1 -1 -1 -2 -2
-1 -2 F -1 -4 -2
A-3 T -1 -1 -1 O-2 N-1 M-1 -1K-1
-1 M-1 -1 K-1 I-2 -2 H Z F E B-1
I-3 -2 (H) F E B-1 A-2 T -1 () -2O-1
A-4 T -1 -1 -3 O-1 N-1 M-1 -1K-1
N-1 M-1 -1 K-2 I-2/3 -2 H F E -2
F E B-2 A-2 T -1 -1 -3-1
-1 -1 -3 -1 -1 -1 -1 -2 -2 -2
K-2 I-3 -2 H Z F E B-2 A-2T
B-2 A-4 T -2 -3 -2 O-1 N-1 M-1-1
10
15
123
Martin Schyen 11
O-1 N-1 M-1 K-2 -2 I-2 -1 H FE
H F E -2ii A-2 T -1 -1-1/3
-3 -1 -1 -2 -1 -1 -1 -1 -1 -2-1
-1 -1 -1 -2 -2 -1 -2ii
-2 (H) F E -2ii A-1/3 T -1 -1 -1O-1
-2ii A-1 T -2 -1 -1/3 O-1 N-1 M-1 -2 K-2I-2
O-1 N-1 M-1 -2 K-2 I-2 -1 H FE
H F E -2ii A-2 T -1 -2-3
-2 -2 -1 -2 -2 -1 -1 -1 -2 -3-1
N-1 M-1 -2 K-2 I-2 -1 H F E -2ii
H Z F E B-2 A-2/3 T -2 -2 -3O-1
B-2 A-2 T -1 -1 -3 O-1 N-1 -2 M-1 (K) I-3-1
N-1 M-1 -1 K-1 I-1 -2 Z FE
-2ii -2 -2 -1 -1-2
-2i -2 -1 -1 -1 -1 -1 -2 () -2 -2/3 -2
O-1 N-1 M-1 -2 K-2 I-1 -2 H FE
-2 H F E B-2 A-2 T -2 -1-1/3
B-2(?) A-3 T -1 -1/3 -1 O-1 H N-1 M-2 -1 K-2I-2
10
15
124
Martin Schyen 12
-2(?) O-1 N-2 M-1 -1 K-2 I-1 -1 H FE
-1 -2 -2 () F -2ii -3 -1 -3-2
E -2ii A-2 T -1 -3 -2 O-1 N-2 M-2-1
-1 -1 -1 O-1 N-1 M-1 -1 K-1 I-2 -1HF
N-1 M-1 -1 K-1 I-2 -1 H F E -2ii A-3T
F E B-1 A-3 T -2 -3 O-1-1H
T -2 -1 -2 O-1 M-1 M-1 -1 K-2 I-3 -1()
N-1 M-1 -1 K-1 I-2 -2 H F E -2iiA-3
F -2 -2 -2 -1 -2-1
-1 -3 -4 O-1 N-1 M-1 -1 K-1 I-1 -2 H
M-1 -1 K-1/2 I-2 -1 H Z F E B-2A-2T
E B-1 A-2 T -1 -1 -2 O-1N-2
-1 -1 -1 -1 -1 -1 -2 -2 -2 -2F
M-2 -1 K-2 I-2 -2 H F E -2ii A-2T
E B-2 A-2 T -2 -3 -4 O-1 N-1 -2
-1 () O-1 N-1 M-1 -1 K-1 I-1 -2HZF
-1 K-2 I-3 -2 H F E () B-2 A-2 T-1
E -2ii A-2 T -3 -1 ( O ) N-1M-2
10
15
125
Martin Schyen 21
E B-2 A-1 T -2 -1 -4O-1
-1 -4 O-1 N-1 M-1 -2 K-2 I-2 -1HZF
-1 -1 () F -2 -2 -1
E B-2 A-1 T -1 -1 -2 O-1 (N MK)
-2 -1 O-1 N-1 M-1 -3 K-2 I-2 -2 HZF
K-1 I-1 -1 H Z F E () B-2 A-1 T -1
-1 -1 -2 -2 -3 -1 -1 -1-3
-2 -2 O-1 N-1 M-1 -3 (K) I-2 -2 HZFE
I-1 -1 H Z F E -2ii A-3 T -1
F E (B) A-4 T -1/2 -1 -4 O-1 (N M K)
-1 -1 -1 -2 O-2 N-1 M-1 -2 K-2 I-1 -2H
-2 K-1 I-1 -2 H Z F E B-1 A-2(?)(?)
-2ii -1 -2 -1 -2 -1 -1-1
-2 -2 O-1 N-2 M-2 ( ) I-1 -1 HZFE
K-2 I-1 -2 H Z F H B-2 A-1 T -1
E B-1 A-2 T -2 -1 -1 O-1 N-1 M-1-3
-1 -3 -2 O-1 N-1 M-1 -2 K-2 I-1 -1HF
N-1 M-1 () K-2 I-2 -2 H Z F E B-2 A-1T
F E B-1 A-2 T -1 -3 -4O-1
() B-1 A-1 T -2 -2 -1 O-1 () N-1 (M K) I-1 -1H
-3 -2 N-1 M-1 -2 K-2 I-2 O-2 ZFE
H Z F E B-1 A-2 T -1-3
10
15
20
126
Martin Schyen 22
A-1 T -2 -3 -4 O-1 (N) M-1 -2/3 K-2 I-1-2
-1 -1 -1 -3 -2 -1 -1 F (?)-2
-1 H F E B-2/ A-3 T -1 -1-2
-2ii A-3 T -1 -2 -4 O-1 () N-2 M-2 -3 K-2I-2
-3 -1 O-1 N-1 M-1 -3 K-2 I-2 -2 H Z FE
K-2 I-2 -1 H Z F E B-2 A-3 T -1
B-2 A-2/3 T -1 -3 -4 O-1 N-1 M-2-2
-1 -1 -2 O-1 N-1 -1 ( K) I-3 -2 HFE
N-2 M-1 () K-2 I-2 -2 H Z F E B-1 A-3T
F -2 -1 // -2/N-1 B-2 (A) T -2 -1 -4 O-1
K-2 M-1 (I) -2 (H) F E -2i A-2 T -1
-1 -3 -2 -1 -4 -2 N-1 M-1-2
-1 -2 -4 -1 N-1 M-1 -1 (K) I-2 O-1/2 HZFE
N-1 M-1 -1 (K) -2 O-2 H Z F E () B-1 A-1/3T
I(?) F E B-2 A-1 T -1 -2 -4 -1H
I(?) -2i A-1 H Z F E -1HM-2
N-1 M-1 -2 K-2 I-2 -1 H Z F E -2iA-2/4
F E B-2 A-2 T -1 -3 -2O-1
-1 -1 -2 -1 -1 () -2 -1 -2 -2
N-2 M-2 (?) K-2 I-3 -2 H Z F E B-1 A-2
F E B-2 A-1 T -1 -3 -4O-1
10
15
20
127
4.2
b r oa d o rt ho g r aph ic t r ansc r i p t io ns
o f t he c o ppe r p l aqu es
The transcription of the copper plates is repeated here but in a different form.
Unlike the preceding transcription, this transcription is broad that is, the
sequence of letters forming each abecedarium is presented without specification of morphological type. The reason for the repetition is twofold: (1) to
reveal clearly the distribution and the frequency of anomalies within the corpus of the plaques, without the business of narrow morphological parsing;
and (2) to provide explicit annotation of various phenomena.
Wrzburg 1

F

O N M K H F E

H F B T
5
A T O N MKI

O N M K I H Z FE

H F E A T

A T O N MI
O N M K I H FE
10
F E A T

()

N M K I HFE

K I (H) Z F E B A T

F E ( O ) N M

15
()

H F E B A T
T O N M K I

M K I H F E A
forO
repeated
forH
K deleted
Hfor
H deleted
, O, , deleted
and inverted
deleted
for
forH
Ofor
Hfor
forO
and M inverted
128
One immediately notices that the single most common anomaly on side 1
of the Wrzburg plaque is that of the collective alternation of (1) eta- and
xi-symbols and (2) theta- and omicron-symbols the phenomenon discussed
at some length in Chapter2 and which I have chosen, for the time being, to
label as anomalous because of the very striking nature of the out-of-place
occurrence of these letters. In addition to these eta / xi and theta / omicron
alternations, both eta and xi, as well as omicron, are deleted from the abe
cedarium that begins on line 13 and ends on line 14. Most of the remaining
anomalies entail letters that fall within the alphabetic range that extends from
kappa to san (of which xi and omicron are, of course, also a subset), excepting
nu: __ .
Wrzburg 2

E B A T ON

O N M K I HZF

H Z F E B A T

B A T O N M () K I
5
M M K I H Z F E

F E B A T () O

Z

F

A T O NMK
10
M K I H Z F EB
I (H) Z F E B A T () O

A T O NMK

N M K I H Z F E

F E B A T
15
Z

K I H Z F E B AT

B A T O NM
deleted
Mfor
deleted
deleted
H deleted
and inverted
Of the six plaque faces, side 2 of the plaque from the Wrzburg Museum
shows the fewest number of anomalies. Most entail the deletion of some letter,
including eta; deletion is, as we have just seen, also prominently present among
the anomalies of the obverse side of this plaque. The remaining anomalies
129
involve, yet again, an alphabetic range one that overlaps the sequential set
of symbols just noted (i.e., __ ): in this instance the letters
involved extend from lambda through rho, though with (curiously) xi and
omicron (conspicuously involved in the alternation anomalies of side 1)not
participating: __ __ .
Martin Schyen 11
O N M K I H Z F E and inverted

H Z F E A T

Z
5
(H) Z F E A T O H deleted

A T O N MKI
O N M K I H Z FE

H Z F E A T

10
N M K I H F E

H Z F E B A T O

B A T O N M (K) I
K deleted

and inverted

N M K I Z F E
for

Z
15
()
deleted

for
O N M K I H Z FE

H Z F E B A T

B A T O H N M K I
for
On the obverse side of plaque 1 from the Schyen collection, eta and xi again
figure prominently, accounting for half of the noted anomalies, but not solely
in the form of alternations: the xi-symbol occurs in the eta-position once,
the eta-symbol in the xi-position twice, and eta is deleted once. The remaining anomalies, once more, involve symbols that belong to an alphabetic
sequence: in this instance . This alphabetic sequence constitutes a
proper subset of the range of letters participating in anomalies of plaque W-1
(i.e., __ ) and overlaps with the corresponding set of W-2
(i.e., __ __ ).
130
Martin Schyen 12

(?) O N M K I H F E

() F

E A T O NM

O N M K I HF
5
N M K I H Z F E AT

F E B A T O H

T O M M K I (Z)

N M K I H Z F E A

F
10
O N M K IHZ

M K I H Z F E BAT

E Z B A T O N

ZF

M K I H Z F E AT
15
E BAT ON

() O N M K I H Z F

K I H Z F E () B A T

E A T ( O ) N M

deformed and repeated?

deleted
for
and inverted
deleted
for
for
inserted
inserted
deleted
deleted
, , deleted
, , and inverted
131
The reverse side of Schyen plaque 1 in large measure follows suit with regard
to which letters participate in anomalies, as I have been terming these deviant
phenomena. The xi-symbol occurs in the eta-position once, the eta-symbol
in the xi-position once. Three anomalies, however, involve the letter that precedes eta in the alphabet, namely zeta, which is twice deleted and once inserted
out of alphabetic order. With the exception of the deletion of gamma in line 17,
all of the remaining anomalies, of one sort or another, involve letters that yet
again are members of a range of characters, and in this case a continuous
alphabetic sequence (i.e., with no internal members excepted). On the three
previously examined plaque surfaces, the similarly involved alphabetic ranges
begin at, or almost at, the beginning of the second half of the Greek alphabet.1
In this instance, the range of affected characters is lambda through rho , ,
, , , , , , and so begins exactly with the second half of the alphabet
(as is also the case with the corresponding set on plaque faceW-2).
A word about line 1. The line begins with a gamma that is of somewhat
unusual form in at least two respects: it has a diminished vertical dimension and a crossbar that extends to the right of the spine (rather than joining
flush with the spine). This much is clear; but in addition, the X-ray image
shows what could be interpreted as a crossbar at the bottom of the spine as
well one that extends almost as far to the right of the spine as to the left
and which descends diagonally toward the direction of writing. One might
well suspect, however, that this could be merely a scratch on the plaque and
not an intentionally inscribed second crossbar, and for that reason I have
not traced it on the enhanced image of the plaque (see the annotated X-ray
on the associated Web site: www.cambridge.org/9781107028111) or noted the
letter as an anomaly in the transcription of the plaque. The final letter of line
1 appears to be a pi, but one of greatly reduced size; line 2 then begins with
another pi, one of typical shape and dimensions. It may be worth noting
that both gamma and pi are involved in anomalies at the end of the plaque:
pi is deleted from both the third from last line and the last line (in the latter
case in conjunction with other anomalies) and gamma from the intervening, second from last,line.
132
Martin Schyen 21

E B A T O

O N M K I HZF

() F

E B A T O (N M K)
5
O N M K I HZF

K I H Z F E () B A T

O N M (K) I H Z F E
I H Z F E A T
10
F E (B) A T O (N M K)

O N M K I H

K I H Z F E B A (?)

O N M ( ) I H Z F E
15
K I H Z F H B A T

E B A T O NM

O N M K I HF

N M () K I H Z F E B A T

F E B A T O
20
() B A T O () N (M K) I H

N M K I O Z F E

H Z F E B A T
deleted
, , , deleted
deleted
deleted
, , , deleted
deleted
for
deformed?
and deleted
for
deleted
, , deleted
deleted
deleted
for
for
for
133
While the number of anomalies on side 1 of the second Schyen plaque is

greater than the number seen on any single plaque face thus far (just as the
number of lines on this plaque face exceeds the number on any of the four
already examined), the types of anomalies are familiar, in large part replicating those that we have previously encountered. Thus, the xi-symbol occurs
in the eta-position once, the omicron-symbol in the theta-position once,
and the theta-symbol in the omicron-position once; in addition eta is once
deleted, as is xi. Curious, though not marked as an anomaly in the transcription, is the space that the engraver left after the line-initial sequence of etatheta and before the ensuing iota in line 20 (a line that is otherwise anomalous). Interesting, potentially for (at least ultimately) phonetic reasons, is the
use of the eta-symbol in the place of epsilon in line 15, bringing to mind the
East Ionic practice (discussed previously) of using both symbols, epsilon and
eta, to spell the long mid-front vowel . Also now familiar indeed practically expected is the participation in anomalies (primarily deletions) of an
alphabetic range of sequential letters; more than that, the sequence of letters
involved is in keeping with what we have previously witnessed, being drawn
from the beginning of the second half of the alphabet though starting with
kappa, the last letter of the first half (as in the case of the comparable letter
ranges of plaques W-1 and MS 1-1): .
Among the few remaining anomalies is the deletion of gamma twice, perhaps of significance in light of observations offered regarding deletion and possible deformation of gamma on preceding plaque face MS 1-2. Note the apparent deformation of tau at the beginning of line 12 of this plaque face.2 As pointed
out in my discussion of CP zeta in Chapter 2, the engraver appears to have
superimposed a Zeta-1 (the I-form zeta) and Zeta-2 (the reversed-Z-form zeta)
in line 21, just before the sequencing of xi and omicron in lieu of eta and theta.
134
Martin Schyen 22

A T O (N) M K I

F (?)

H F E B A T

A T O () N M K I
5
O N M K I H ZFE

K I H Z F E B A T

B A T O NM

O N ( K) I H F E

N M () K I H Z F E B A T
10
F // /N B (A) T O

K M (I) (H) F E A T

N M

N M (K) I O H Z F E

N M (K) O H Z F E () B A T

15 I(?) F E B A T H

I(?) A H Z F E H M

N M K I H Z F E A

F E B A T O

()

20
N M (?) K I H Z F E B A

F E B A T O
N deleted
deformed?
deleted
K, deleted
deleted
A deleted
N or inserted
E deleted (14 letters)
and F inserted
and inserted
deleted
deleted
M inserted
for
for
K deleted
for
deleted
for
forI
deleted
for
for
punctuation?
punctuation?
shifted forward
deleted
punctuation? / deformed?/
I for?
135
The reverse side of the second Schyen plaque not only shows the greatest
number of anomalies, as the term is being here used, but is far and away the
most interesting and revealing of the six plaque faces. We find the same processes of anomaly, involving the same letters, that are displayed on the previously examined five surfaces. Thus, the eta-symbol is used in the xi-position
once, the omicron-symbol in the theta-position twice, and the theta-symbol
in the omicron-position three times; in addition, eta and xi are each deleted
once. The participation in anomalies of letters forming an alphabetic sequence
again presents itself. In this case the range extends from eta through omicron
( ): at one end of the range are eta and theta, at the other
end xi and omicron the letters involved in the familiar transpositions here
serving to bookend the sequence; in between are five letters, each of which is
deleted on at least one occasion, in one position or another, from the abecedaria of this plaqueface.
What looks to be a deformed gamma appears in line 2. I pointed out earlier (without marginal annotation) that a gamma of unusual shape appears
in line 1 of MS 1-2. The modification of gamma in the present instance (MS
2-2, line 2)is crucially different: here the crossbar is conspicuously diagonal,
declining toward the direction of writing, giving the letter a lambda-like morphology. One could envisage that the engraver has here accidentally executed
a malformed symbol or, alternatively, that he has intentionally introduced a
lambda morphology into the gamma-position, taking advantage (so to speak)
of the two letters formal similarity,3 in much the same way that eta and theta
are repeatedly substituted for the graphically similar xi and omicron, and vice
versa. These are instances of symbol substitution that are, in effect, I would
propose, sanctioned by morphological verisimilitude. With regard to gamma
anomalies, note that the letter is once deleted from an abecedarium on this
plaque face, as we also saw occurring on MS 1-2 and MS2-1.
Lines 10 and 11 show the most extensive alphabetic deviance that we have
seen thus far in any single abecedaric stretch on these plaques. The first eight
letters of line 10 complete the abecedarium that begins early in line 9: this
abecedarium is unremarkable except for the deletion of lambda in line 9. The
remainder of line 10 consists of an additional eight letters letters that do not
unfold in regular alphabetic sequence, though traces of such a sequence can
be seen. Alpha is missing from the initial position of this deviant abecedarium, though beta and gamma do then occur. Prior to the ensuing delta (the
expected next letter in alphabetic order), however, there is inserted a letter of
136
ambiguous morphology: its shape matches that of both Nu-1 and San-2 (see
the discussion of the two letters in Chapter2 regarding their potential isomorphism). Following delta, there is a hiatus of fourteen letters (marked in
the transcription by //), extending from epsilon through san inclusively, after
which the scribe returns to alphabetic sequencing, engraving qoppa and then
rho. After rho, however, alphabetic order is once again disrupted by the insertion of a lambda followed by a digamma. But yet again, following the lambdadigamma anaptyxis, the scribe returns to alphabetic sequencing, engraving
sigma, the letter that follows in the alphabet after rho, only to disrupt immediately the alphabetic order again by inserting a xi and then a theta. Following
the anomalous xi and theta, the scribe turns back to alphabetic sequencing,
engraving a tau, the letter following sigma in alphabetic order, and in so doing
brings this abecedarium to an end. Thus what we find in lines 10 and 11 is an
abecedarium interrupted by a series of alternating deletions and insertions,
here rewritten left to right:
(5) (A) N/ // F T
4.3
a l phabe t ic va r iat io n i n t he c o ppe r

p l aq u es
My ideas concerning what it is that is being played out in these (and other)
lines and, especially, in early Greek alphabetic practice more generally will
occupy much of the remainder of this work, beginning with some preliminary considerations in the next chapter, but I would note at this point that
one possible interpretation that of a massive, even laughable, scribal error,
or series of errors would seem to offer an a priori unlikely and unsatisfactory interpretative scenario. The scribes of the copper plaques, including the
scribe(s) of MS 2-2, are clearly capable of producing canonical abecedaria
(i.e., respecting alphabetic order), as they do time and again. Are we to imagine that the engraver has here been struck by some sort of temporary scribal
delirium? Even a casual examination of this aberrant abecedarium would
reveal possible (playful) patterns of regularity (if not an explanation for their
existence) within the irregularity (as already suggested by my exposition of
letter sequencing). For example, if in (5) we were mentally to jump the first
inserted character forward over delta, and were to view that ambiguous character as a san, and were to jump sigma and tau backward so as to clear the
two sets of dual anaptyctic symbols, we would recognize a coherent alphabetic
137
coda: san, qoppa, rho, sigma, tau. We would then be left, of course, with what,
if any, sense could or should be made out of the now mentally stranded four
anaptyctic letters: lambda, digamma, xi, and theta.
More intriguing still are lines 15 and, especially, 16 of MS 2-2. The first eight
symbols of line 15 provide the final portion of the abecedarium that begins
early in line 14; a new abecedarium then begins (about mid-line 15), running
from alpha through digamma, with no anomalies displayed up to digamma,
the next to last letter of line 15. The ensuing character the final symbol of the
line is not, however, the letter that next follows digamma in the alphabetic
order, namely zeta. Instead, what we find is a symbol that is formally close to
some occurrences of Iota-2, though of somewhat diminished size, and having a third stroke (the ascending vertical stroke proximate to the direction of
writing) that is shorter than that typical for Iota-2. In addition, while line 15
ascends slightly from right to left across the plaque face, this final character is
conspicuous in that regard, almost having the appearance of a superscript.
Fully half of the symbols inscribed on line 16 do not occur in alphabetic
order, or anything that would appear to be an algorithm of alphabetic order.
Prima facie inspection of the line would suggest an arbitrary (i.e., aperiodic)
distribution of the first seven symbols: mu, eta, lambda, xi, sigma, epsilon,
lambda. Following these seven there occur three characters that do subscribe
to alphabetic order (i.e., periodic distribution) digamma, zeta, eta but these
are followed in turn not by a graphemic continuation of this order but by a
new alphabetic sequence, this time with four symbols alpha, beta, gamma,
delta. Thus, the first fourteen symbols of line 16 align themselves according to
the following sequence, here rewritten from left to right:
(6) F
The final symbol of line 16 is similar, though not identical, to the final symbol
of the preceding line, 15, and thus displays morphological likeness to some
forms of iota, but in this case appears closest to Iota-1 (not Iota-2). The symbol shows, however, greater right-angularity than is typical of Iota-1 and, as
with the corresponding symbol at the end of the preceding line, appears to be
somewhat vertically displaced, having a slight superscript orientation.
The alphabetic sequence that begins at the end of line 16, alpha, beta, gamma,
delta, situated before the curious line-end symbol, does not continue at the beginning of line 17, as would typically be the case. Instead, line 17 begins with a new
alphabetic sequence alpha, beta, gamma, and so on that runs uneventfully,
uninterruptedly all the way to tau, terminating at about the midpoint of line18.
138
And what is to be made of the two different iota-like symbols occurring

at the ends of lines 15 and 16? They are certainly made the more conspicuous
by the unique structure of the line that they bracket (line 16). Is it possible
that these symbols are here functioning, in effect, as markers of punctuation as signals of the pronounced deviance of a line of at least partially arbitrarily ordered graphemes floating within a veritable orthographic ocean of
abecedaria? A symbol very close to that one terminating line 15 can be seen
once more only somewhat larger in the abecedarium that begins at the
start of line 20. Here the symbol is used in the position of lambda in an otherwise unremarkable abecedarium, and in this context presents itself (self-evidently) as potentially a deformed lambda, having a shape somewhat intermediate between Lambda-1 and Lambda-2, but with a vertical stroke extending
upward from the distal end of the descending hook-stroke, thus giving the
symbol its structural affinity for CPiota.
I would offer one final observation before bringing this portion of my study
to a close. On plaque face MS 2-2, we have found two lines (or line sequences)
that deviate appreciably from the norm: (1) lines 10 and 11, preserving an
abecedarium interrupted by alternating deletions and insertions:
(7) Lines 10/11: (A) N/ // < F> T
and, more especially, (2) line 16, significantly characterized by an arbitrary
sequencing of graphemes.
(8) Line 16: < F>
Both graphemic sequences share in common the letter cluster lambda +
digamma ( F), set off within angled brackets in (7) and (8). In lines 10/11
the cluster represents a chunk of inserted material interrupting a periodic
sequence of graphemes. In line 16 it represents the final member of an arbitrary aperiodic sequence of graphemes (lambda) plus the initial member
(digamma) of what has the prima facie appearance of being a brief periodic
expression of alphabetic sequencing. The coincidental occurrence is notable
and may, or may not, be of some significance.
A second feature that the two line-sets share in common is the overt expression of the periodic sequence alpha, beta, gamma, delta the onset of the
alphabet (or, more precisely, of the abecedarium). In 10/11 this primal sequence
is made less overt by an alternation consisting of a deletion process (alpha is not
spelled) and an insertion process4 (nu or san is interposed [or, actually, it would
seem, san is forwarded from the latter portion of the line and interposed]) but
139
the primal sequence is made more conspicuous is foregrounded by the radical break in periodic sequencing that occurs immediately after delta a chasm
into which a sequence of fourteen letters tumbles. In line 16, the same primal
periodic sequence appears, this time at the end of the line (rather than at the
beginning [two salient positions]) again foregrounded in this instance by
its position before the curious line-end marker and by the abrupt cessation of
the periodic order with delta, marked by the immediate beginning of a new,
pristine, abecedarium at the start of the nextline.
The next chapter will begin where this one has ended, with MS 2-2, line16.
5
Langue et criture
Langue et criture sont deux systmes de signes distincts;

lunique raison dtre du second est de reprsenter le premier.
Ferdinand de Saussure, Cours de linguistique gnrale
5.0
At the end of Chapter 4, I directed the readers attention to two line-sets

from plaque face MS 2-2 that depart markedly from other lines that is, abe
cedaria of the copper plaques: lines 1011 and, most especially, line 16, examined together with the preceding line 15. It is the second of these to which I
would like to turn my attention. The left-to-right transcription of MS 2-2, line
16 is here repeated:
(1) F
If the line ends with a clear presentation of what I referred to in Chapter4
as the primal periodic sequence alpha, beta, gamma, delta it begins with
what is, from an abecedaric perspective, an arbitrary string of graphemes, one
that continues for at least half theline.
5.1
a r b i t r a r i n e ss : pa rt 1
Arbitrariness of sound sequence is the hallmark of language, not of the fixed

canonical display of the ordered symbols of an orthographic system here, an
abecedarium. The realization of the arbitrariness of the linguistic sign, both its
140
Langue et criture
141
phonic and its conceptual components, marked the laying of the foundation
of twentieth-century synchronic linguistics (and its derivative disciplines);
the mason of the foundation was Ferdinand de Saussure.1 In the Cours de linguistique gnrale, Saussure offers examples drawn from French and German
to illustrate the concept.
There is no internal connexion, for example, between the idea of sister and
the French sequence of sounds s--r which acts as its signal. The same idea
might as well be represented by any other sequence of sounds. This is demonstrated by differences between languages, and even by the existence of different
languages. The signification of ox has as its signal [French] b--f on one side
of the frontier, but [German] o-k-s (Ochs) on the other side. (trans. Harris)2
Further along in the Cours, in discussing the variability of the linguistic sign
over time, Saussure observes:
A language is situated socially and chronologically by reference to a certain
community and a certain period of time. No one can alter it in any particular.
On the other hand, the fact that its signs are arbitrary implies theoretically a
freedom to establish any connexion whatsoever between sounds and ideas.
The result is that each of the two elements joined together in the linguistic
sign retains its own independence to an unparalleled extent. Consequently a
language alters, or rather evolves, under the influence of all factors which may
affect either sounds or meanings. Evolution is inevitable. (trans. Harris)3
On the Protocorinthian aryballos from Kyme (the aryballos of Tataie circa

the second quarter of the seventh century b c ), discussed in my treatment of
CP tau, ones sees an arbitrary string of boldly etched graphemes that includes
the sequence:
(2)
One familiar with the linguistic signs of ancient Greek can with little difficulty parse this string of orthographic symbols to reveal an argument structure with verbal and nominal components, including a sentential clause with
pronominal subject and object:
(3)
One can rewrite this parsed orthographic string in a slightly more familiar
form (more familiar to those acquainted with the linguistic structures of
142
ancient Greek), and even write it yet again, using a different, but related, set
of orthographic symbols:
(4) h
hos d an me klephsei thuphlosestai
Each unit in these parsed strings representing particles, pronouns, nouns,
verbs is an orthographic expression of an arbitrary sequence of sounds (the
phonic component of a linguistic sign). With each there is arbitrarily affiliated a conceptual element. Corporately these individual conceptual components express, fundamentally, the notion expressed by the following English
structure:
(5) Whoever steals me will go blind.
There is no necessary connection between the several individual phonic components of (4) and their associated ideas.4 In the same way, there is no necessary connection between the arbitrary strings of sound of which (5) is an
orthographic expression and their affiliated ideas. Greek h
and English whoever steals me will go blind each represents an
argument structure consisting of arbitrary phonic signals, with which is arbitrarily associated the (approximately) common idea Whoever steals me will
go blind. There could exist, and do, other arbitrary phonic signals that express
the same fundamental notion: thus a French investigator can invest the idea
expressed by Greek h in the phonic string
transcribed as qui me drobera deviendra aveugle;5 an Italian in chiunque mi
ruber sar cieco;6 a German in wer mich stiehlt, wird blind werden.7
The language of which h is an arbitrary
phonic string can be, as per Saussure, situated socially and chronologically
by reference to a certain community and a certain period of time. That community was the body of Greek speakers of the south Italian city and Euboian
colony of Kyme; that period of time was broadly the archaic Greek era, narrowly the earlier seventh century b c . No member of that community could
alter the language; no member of that community could alter the relationship
that existed, and exists, between this string of arbitrary phonic signals and the
affiliated arbitrary conceptual elements. A language is beyond the control of
its speakers. Saussure writes concerning le signifiant (the signifying component [the signal in Harriss translation] of the linguistic sign):
Speakers are not consulted about its choice. Once the language has selected a
signal, it cannot be freely replaced by another. There appears to be something
Langue et criture
143
rather contradictory about this. It is a kind of linguistic Hobsons choice. What

can be chosen is already determined in advance. No individual is able, even if
he wished, to modify in any way a choice already established in the language.
Nor can the linguistic community exercise its authority to change even a single
word. The community, as much as the individual, is bound to its language.8
A writing system is, at least in part, also an arbitrary phenomenon. In the

Cours, Saussure explicitly addresses the arbitrariness of that other system of
signs, writing [lcriture]:
1. The signs used in writing are arbitrary. The letter t, for instance, has no connexion with the sound it denotes.
2. The values of the letters are purely negative and differential. So the same
individual may write t in variant forms. The one essential thing is that
his t should be distinct from his l, his d, etc.
3. Values in writing are solely based on contrasts within a fixed system, having a determinate number of letters. This feature, although not the same as
2 above, is closely connected with it; for both 2 and 3 follow from 1. Since
the written sign is arbitrary, its form is of little importance; or rather, is of
importance only within certain limits imposed by the system.
4. The actual mode of inscription is irrelevant, because it does not affect the
system. (trans. Harris)9
While a writing system is certainly also a system of signs, as Saussure

observes, the nature of its arbitrariness is, or at least can be, somewhat different from that of language. Given the medium at hand, in order to express the
arbitrary phonic string h , or hos d an me
klephsei phulos estai, a phonic string belonging to the language of the Greekspeaking community of Kyme in the earlier seventh century b c , we must make
recourse to orthographic symbols of one system or another. In other words,
we are here using orthographic symbols (graphemes) to give graphic expression to arbitrary sound sequences (a relationship of which Saussure was, of
course, well aware).10 Graphemes encode sounds; they map language onto a
medium. In that sense a graphemic string is also arbitrary it is derivatively
arbitrary in that it maps arbitrary phonic strings onto paper, papyrus, leather,
clay, stone, bronze, copper, or whatever the medium of choice mightbe.
This observation does not nullify Saussures points 2 and 3. The latter
the claim that values in writing are solely based on contrasts within a fixed
system, having a determinate number of letters holds, but also for derivative reasons: a grapheme encodes (maps) a sound; the linguistic value of the
144
individual sounds of a language are based solely upon contrasts within a fixed
phonological system having a determinate number of distinct consonants and
vowels (i.e., what we can refer to as phonemes); ergo, the value of a grapheme also is based solely on contrasts. Regarding such derivative symbolic
semiotic systems, of which writing is a prime example,11 Roman Jakobson
observes (and drawing our attention to the fact that even a writing system can
have its own derivative symbolic expression):
The relationship between the verbal pattern and the other types of signs may
be taken as a starting principle for their grouping. One variety of semiotic systems consists of diverse substitutes for spoken language. Such is writing, which
is both ontogenetically and phylogenetically a secondary and optional acquisition [emphasis is my own] as compared with the all-human oral speech.12
In the relation between graphic and phonological entities, the former always
functions as a signans and the latter as a signatum. On the other hand, written language, often underrated by linguists, deserves an autonomous scientific
analysis with due respect to the particular characters of writing and reading
(see Derrida 1967, 1968). The transfer of speech into whistles or drumbeats
offers another example of a substitutive system, while Morse code exhibits a
second-order substitution: its dots and dashes are a signans which stands for
the ordinary alphabet as their signatum (see Sapir 1921:20, 1949:7).13
Ideally then, in an alphabetic system, such as the ancient Greek alphabet the
first alphabet and the ultimate source of all the worlds alphabets14 there
would be a one-to-one mapping relationship between grapheme and phoneme. In other words, the set of graphic symbols would be coequal with and
determined by the set of contrastive sounds that the graphic symbols map
(i.e., from which set they are derived).
In actuality, however, this ideal one-to-one relationship typically does not
exist in writing systems. Let us consider two cases that exemplify how this is
so. In the Greek alphabet, the grapheme (alpha) is used to encode the value
of two phonemically distinct vowels, // and //.15 In the case of the Arkadian
dialect, for example, spells (i.e., is the graphic encoding of) both the
present middle indicative and present middle subjunctive forms of the thirdperson singular verb to stand: the two modal forms differ phonically, however, in that the penultimate vowel of the indicative is //, whereas that of
the subjunctive is //.16 The grapheme (iota) is used, likewise, to encode the
value of two phonemically distinct vowels, // and //; hence, both the present
and the imperfect tense forms of the first-person plural of the verb meaning
Langue et criture
145
we approach as suppliants / we were approaching as suppliants are spelled

, though the phonic string that has the former affiliated conceptual
component (present tense) begins with //, and the phonic string having the
latter affiliated conceptual component (imperfect tense) begins with //. As
these examples show, the number of distinct contrastive graphic symbols can
be fewer than the number of distinct, arbitrary contrastive sounds of the language which that writing system graphically encodes.
The ratio of graphemes to sounds can be far smaller than one-to-one in
some writing systems, as in Linear B, the syllabic script of the Mycenaean
Greeks. For example, the Linear B syllabic symbol , romanized as ka, could
potentially represent at least six contrastively different phonic strings of
the Mycenaean language: /k/, /k/, /kh/, /kh/, /g/, and /g/. In addition,
because of the Linear B scribal strategy of orthographically deleting many
members of consonant clusters,17 the same symbol could provide the graphic
encoding of phonic strings such as /rk/, /lk/, /skh/, /skh/, /rg/, /g/, to
cite but a few of the logical possibilities.
The converse of the condition exemplified by the several preceding examples also occurs in writing systems: namely, two graphically contrastive symbols can be assigned the role of encoding a single contrastive sound. For
reasons that have conventionally been deemed to be purely phonetic, having
nothing to do with the arbitrary system of contrastive sounds (phonemes) utilized by the Greek language, the consonant /k/, standing in a contrastive relationship with each of the other consonants and each of the vowels of Greek,
is realized in a phonetically variant way when it stands before a back vowel
compared with when it stands in any other phonetic context.18 This phonetic
variation finds an orthographic expression in early Greek alphabetic writing: the default version of /k/ is spelled (kappa) while the variant occurring
before a back vowel is spelled (qoppa). In this case, the set of graphic symbols exceeds the set of contrastive sounds and encodes differences that are not
ones of an arbitrary, contrastive nature, but differences that are of a motivated
phonetic nature.19 This procedure tends to be avoided by writing systems; in
other words, spelling tends to be phonemic, as discussed in Chapter2.
A variant expression of the same phenomenon in the Greek alphabet is provided by the orthographic symbols (xi, having the graphic shape in the socalled red alphabets)20 and (psi).21 Neither of these symbols ( [or ] and
) provides the graphic encoding of any single member of the arbitrary set of
sounds that make up the fixed inventory of Greek consonants (in their common
usage that is here being described). Instead, each symbol redundantly provides
146
the graphic encoding of a two-member consonantal string, /ks/ and /ps/ respectively: I say redundantly as in each instance there exist graphic symbols that
are used to encode each member of the string individually. The graphic symbol
(kappa) encodes the contrastive sound /k/, (pi) the sound /p/, and either
(sigma) or (san), depending upon the alphabet, the sound /s/.22
Notice that here, in the instance of the graphic signs xi and psi, we have a
case not of derivative arbitrariness of a writing system only but of primary
(or immediate) arbitrariness as well. The fixed Greek inventory of consonant
sounds did not contain among its members a unitary sound /ks/ and a unitary sound /ps/ for which individual graphic symbols were required in order
to map those sounds onto a medium; /ks/ is a string composed of two of the
unitary units of the Greek sound system, as is /ps/. In other words, the occurrence of the graphic symbols xi and psi in the alphabetic writing system of the
Greeks is not solely a consequence of individual contrasts within the fixed
phonological system of the Greek language. The choice to utilize available
graphic raw material in order to incorporate a sign for the consonantal string
/ks/ within the Greek alphabet was an arbitrary decision made by the Greek
adapters of the Phoenician consonantal script. Though the decision was culturally motivated, I have argued elsewhere, by the occurrence of comparable
<ksV> syllabic symbols in the Greek Cypriot syllabary a writing system in
which I judge the Greek adaptors of the Phoenician script to have been already
literate (and a writing system within which such symbols were required for
proper functioning of syllabic spelling). The decision to incorporate a graphic
sign for the parallel /ps/ string was subsequently made by persons responsible
for extending the Greek alphabet beyond its original Phoenician-set boundaries through the appending of supplemental characters.
5.2
d i st i n c t i v e n e ss a n d am b ig u i t y
In light of the observations offered in the preceding section, and also of the
epigraphic practices that we encountered in Chapter 2, Saussures point 2,
repeated here, requires consideration aswell:
The values of the letters are purely negative and differential. So the same individual may write t in variant forms. The one essential thing is that his t
should be distinct from his l, his d,etc.
Phonemic distinctiveness is a sine qua non for the successful operation of

language; graphemic distinctiveness might then seem, in principle, to be a
Langue et criture
147
necessary characteristic of a writing system. Some degree of ambiguity in this

regard can be and is tolerated in the practice of writing, however. In the case
of an alphabetic writing system, one that graphically encodes the individual
sounds of a languages phonological system, a writer-readers knowledge of
the language that is being encoded permits the language to be written in such
a way that graphemic distinctiveness can be minimized. Certainly in contemporary Western societies utilizing Latin alphabets this is a familiar phenomenon and seems to be especially prevalent among the most literate members
of society.23 Thus, a t may very well end up looking like a d or an l, or like an i
or an e, and still the writing system can function effectively.24 This is because
language is primary and writing is derivative. Writing is but a shadow of language, and at times the shadow may be more a shade and less a silhouette.
Or, as Saussure put it in the Cours, lcriture voile la vue de la langue: elle
nest pas un vtement, mais un travestissement.25 The sort of ambiguity here
described, however, is a function of the practice of orthography rather than
of the design of orthography. The Latin alphabets of contemporary Western
societies, all stemming from the Greek alphabet, were designed with generally
high, though varying, degrees of graphic distinctiveness built in.26 Yet, Roman
Jakobson reminds us, as a rule, systems of writing are fairly complicated and
cannot be reduced to a limited number of distinct visual oppositions.27
The sort of linguistically nondisruptive graphic ambiguity that arises in
practice in the foregoing alphabets is from a process perspective akin to
ambiguities present in other writing systems by design. In the Phoenician
script from which the Greek alphabet was formed, and in other, related
Semitic scripts, all having a common origin in the Proto-Canaanite writing
system, there is effectively no representation of vowel sounds.28 The reader of
such scripts must recognize the graphically encoded arbitrary phonic strings
of the language on the basis of the consonantal portion of those strings alone:
the vowels must be supplied from the readers knowledge of the inventory of
the linguistic signs of that language. This is not the case of to paraphrase
Saussure t being graphically identical to d or l (that condition which point
2 states that it is essential to avoid), but, more radically, of any given vowel
encoding being identical to any other vowel encoding, being a zero-encoding
in eachcase.
A burden equally great, if not greater but certainly not a debilitating
burden for native speakers is placed on the reader of the Mycenaean Greek
syllabic script. I noted in the discussion of Saussures closely correlated
point 3 that the contrastive consonant sounds /k/, /kh/, /g/ are graphically
148
ambiguous in Linear B spelling as are, moreover, /t/ and /th/ (though /d/
is kept distinct); as are /p/, /ph/, and (presumably) /b/; as are /kw/, /kwh/,
and /gw/; as are /r/ and /l/. The Linear B script also operates with extensive
graphic nondistinctiveness in the encoding of the vowel components of the
arbitrary strings that compose the phonic components of the linguistic signs
of Mycenaean Greek. The contrastively long and short vowels (// and //;
// and //; // and //; // and //; // and //) are not differentiated from
one another neither when the vowel occurs as a monophthong nor when
it occurs as the first element of a diphthong. In addition, diphthongs are
only partially distinguished from monophthongs: the second element of the
diphthongs /au/, /eu/, and /ou/ typically receive graphic expression through
use of the symbol <u>, but the i-component of /ai/, /ei/, and /oi/ only sometimes is overtly expressed (using <i>). Word-initial /h/ (the spiritus asper) is
not recorded; many of the consonants in consonant clusters receive no overt
graphic expression.29
A writing system that graphically assimilates much of the phonological distinctiveness of the language that it records, as Linear B does, can still operate
effectively for native speakers of that language. Such speakers bring to the task
of writing and reading an intimate, even automatic, knowledge of the language
behind the script its lexicon (i.e., the repository of linguistic signs and any
affiliated specifications regarding their use), its phonological redundancies,
its permitted patterns of arbitrary phonic strings, its morphology and syntax.
The task of effective reading may, however, require the reader to draw upon
nonlinguistic structures discourse and societal contexts, and cultural sensitivities to a degree greater than that required in the reading of a more phonologically distinct alphabetic script. For the modern investigator of Linear B,
who did not acquire Mycenaean Greek at her or his mothers knee, the script
can be far less user-friendly, and the act of unlocking its linguistic contents an
exercise in puzzle solving. Witness the great disagreements that have erupted
in the wake of the discovery of new Mycenaean tablets from Thebes.30 Had
a literate Mycenaean been unearthed with the tablets, the controversy could
have been settled in short order.
A similar confounding in the encoding of distinctive sounds obtains in
the practice of syllabic spelling among the Cypriot Greeks. In the Cypriot
syllabary there is no graphic distinction made between the contrastive consonant sounds /k/, /kh/, /g/, nor between /t/, /th/, /d/, nor between /p/, /ph/, /b/.
As with the Mycenaean syllabary, the Cypriot system does not differentiate
between the contrastively long and short vowels (// and //; // and //; //
Langue et criture
149
and //; // and //; // and //). Diphthongs are, however, distinguished, and
there is a more complete representation of consonants in sequence.31
In the instances of the two ancient Greek syllabic scripts, the strict parallel
to Saussures point 2 would be one that states that, for example, the syllabic
symbol Ka (where K graphically encodes any of the three contrastive consonants /k/, /kh/, or /g/) could be of any form in effect; what is essential is that it
is graphically distinct from the syllabic symbol Ke or Pa (where P graphically
encodes any of the three contrastive consonants /p/, /ph/, or /b/), and so forth.
Yet the sort of graphic distinctiveness that Saussure envisions as essential in
the point-2 example of the alphabetic symbol t is itself compromised in the
design of the Greek syllabaries with their extensive phonemic ambiguity.
Both the Phoenician consonantal script the source of the Greek alphabet
and the two pre-alphabetic Greek writing systems of Linear B and the Cypriot
syllabary arranged with the alphabetalong a historical continuum of Greek
literacy, and informing the design of the Greek alphabet, I have argued32 are
writing systems characterized by a high degree of phonemic ambiguity, by a
failure to encode graphic distinctiveness of contrastive sounds.33 These are
vagaries by design; and though a writing system so designed may be said to
be linguistically impoverished, the system nonetheless works (i.e., performs
its intended communicative function). At the start of this section, I invoked
the case of alphabetic writing in highly literate modern societies in which a t
may be indistinguishable from a d, and so forth. These are vagaries that arise
by execution; yet the system, though performatively impoverished, carries out
its function. In each instance, the system works for the same reason: writing
is not language; writing encodes language, and effective writing needs only to
encode sufficient linguistic representation to reveal to a linguistically knowledgeable reader the encoded linguistic structures. Language licenses, up to
a point, an orthographic disregard for linguistic detail; language permits a
graphemic playfulness.
In Chapter2, I drew the readers attention to multiple instances of graphic
ambiguity homography was the term used in archaic Greek alphabetic
writing, as evidenced both within and beyond the abecedaria of the copper
plaques. Thus, we saw that the symbol h (and/or its variant form ) could
be used to represent both a consonant the glottal consonant /h/ and a
vowel the long vowel in the local alphabets of Naxos, Thera, and Rhodes.
The symbol represents both /h/ and the consonantal sequence /ks/ in the
Naxian alphabet, just as the symbol stands in both the eta- and xi-positions of Euboian-based alphabets of Etruria. The -symbol similarly occurs
150
in both the (vocalic[/consonantal])34 eta-position and the (consonantal) xiposition in the copper plaques as does the h-symbol.35 In the case of the use
of h to spell both a consonant and a vowel and of the use of to spell both a
single consonant and a consonantal sequence in various epichoric alphabets,
the ambiguity involved is unquestionably by design. I propose that a synchronic practice of homography that is fossilized as the variable positioning
of in the abecedaria of the copper plaques and of Etruria is no less a matter
of vagary by design.
In Chapter2, we also saw that within the copper-plaque abecedaria, the
(consonantal) theta-symbol () can be found in both the theta-position and
the (vocalic) omicron-position. Conversely, the (vocalic) omicron-symbol
occurs both in the omicron-position and in the (consonantal) theta-position.
This phenomenon parallels that of the alternation of h- and -symbols and
must also be an instance of ambiguity by design. Supporting evidence for this
view is provided by the wide-scale use of a single symbol for both theta and
omicron in local alphabetic practice, most impressively so in the recurring use
of the symbol O for both consonant and vowel in the long Argive inscription
preserving a treaty between Knossos and Tylissos. Recall that the mason executing this inscription also uses elsewhere both the xi-symbol and the etasymbol h to spell /h/ (and so parallels the alternation of eta- and xi-symbols
in the eta-position in the copper plaques).
The list of homographies encountered in Chapter2 could be expanded. For
the present, however, this should be sufficient.
5.3
a r b i t r a r i n e ss : pa rt 2
Such homographies vagaries by design focus our attention on the seeming arbitrariness of assigning values to graphic symbols: the same symbol can
represent a single consonant, a sequence of consonants, or a vowel. That being
the case, we can see (again) that there is a sense in which the values assigned
to graphemes can also be said to be primarily arbitrary and not only derivatively arbitrary. This observation then brings us back to Saussures point1:
The signs used in writing are arbitrary. The letter t, for instance, has no connexion with the sound it denotes.
Specifically with regard to the Greek alphabet, can one say that the letter
tau () has no connection with the sound it denotes? The choice made
by those persons responsible for the creation of the Greek alphabet to use
Langue et criture
151
tau for graphically encoding the contrastive sound /t/ was motivated by the
phonic characteristics of the Phoenician donor symbol taw, which likewise
spelled the sound /t/. If the Greek letter tau is arbitrarily connected with the
consonant sound /t/, the relationship would best be identified as a form of
what Saussure termed relative arbitrariness (larbitraire relatif) to the extent
that it is neither historically nor linguistically fully (or intrinsically) arbitrary
(larbitraire absolu), but motivated by a preexisting, recurring use of, mutatis
mutandis, the same symbol for the same phonic representation.36
5.3.1 greek writing

Many of the consonantal symbols of the Greek alphabet show a similar relative arbitrariness, being taken over from the Phoenician script in form and
in function. We saw in Chapter 2, however, that the Greek adapters could
preserve a Phoenician consonantal function while eliminating the associated
Phoenician form, as in the Greek use of digamma () that is, wau [] to
represent /w/; even though wau occurs in the same alphabetic position as its
Phoenician antecedent waw, which likewise encodes /w/, the Greek grapheme does not continue the form the morphology of Phoenician waw.
Given this state of affairs, the arbitrariness of the relationship of this Greek
symbol to its sound value could perhaps be said to be enhanced, relative to
that of tau and other Greek characters. But, of course, the situation is more
complex than this, as we have alreadyseen.
Recall that the Phoenician consonantal symbol waw is also the source of
the Greek vocalic letter upsilon. This particular use of waw marks one of several instances in which Greek adapters reshape the function of a Phoenician
symbol. This process is notably (though not exclusively) associated with the
creation of the Greek vowel characters from Phoenician consonantal characters. In Chapter2, I argued that, at the inception of the Greek alphabet, wau
(digamma) must have been created as a bivalent symbol, linked to two contrastive Greek sounds, the consonant /w/ and the vowel /u/. I also suggested
that, in a fundamentally parallel fashion, the Greek adapters modified the
Phoenician letter yod, representing the consonant /y/, to produce a bivalent
symbol iota that was used to spell both the nonphonemic (i.e., noncontrastive) Greek consonant [y] and the contrastive Greek vowel /i/. In all archaic
Greek alphabets, the form of Greek wau does not continue that of Phoenician
waw, as just noted; similarly, the form of iota in the abecedaria of the copper
plaques does not continue that of Phoenician yod (and, I suggested, should
152
perhaps be seen as the morphological source of both archaic straight and

crooked iota). Moreover, in the adaptation process, the function of the Greek
symbols has been expanded in comparison to that of their Phoenician precursors: there has been a change in form and a partial change in function as well.
Given the change in both form and function vis--vis the Phoenician letters,
the connection of these Greek symbols to the sounds they encode is clearly
one of intrinsic arbitrariness (see the following paragraphs for more on this)
yet there remains some element of relative arbitrariness: the Greek symbols
stand in the same alphabetic position as the Phoenician symbols; they retain
the Phoenician symbols Semitic names; and the change in function shows
phonetic motivation (as the phonetic alternation of w ~ u and y ~ i is common
cross-linguistically).
Still at an early stage in the history of the Greek alphabet, Greek adapters
appended the letter upsilon to the end of the Phoenician-provided portion
of the script, creating the first of the supplemental characters. This is an
instance of the Greek adapters retaining the form of a Phoenician symbol
(waw) but abandoning its function: the Greek symbol has the associated
phonic value of the vowel /u/, while the Phoenician represents the consonant /w/. Aside from iota an originally bivalent character, as I have just
(again) suggested, showing change of form and (partially) of function
the remaining vowel letters created by the Greek adapters follow the same
pattern as upsilon: form is substantially preserved; function is completely
changed. These historical relationships are summarized in the following
chart:
(6)
Phoenician
_________________________
Form
Name
Sound
Greek
_________________________
Form
Name
Sound
alep
alpha
and
he
epsilon
and
ayin
omicron
and
The Phoenician symbols all spell consonants: represents a glottal stop, h a

voiceless glottal fricative, and a voiced pharyngeal fricative.37 In the Greek
column, and are used as cover symbols, each representing two distinct
vowels of Greek the long vowel inherited from Proto-Indo-European and
Langue et criture
153
the (contrastively different) long vowel that developed secondarily within

Greek. These long vowels would eventually be orthographically distinguished
from one another and from the affiliated short vowels.
I would view the Greek relationships of letter shapes to sound values that
are set out in the historical chart (6) as being of an arbitrariness that extends
beyond that of earliest, bivalent, wau and iota. While the Greek letters retain
the alphabetic position that their antecedents occupied in the Phoenician
script (so that their presence within the alphabetic system is historically motivated), they are used for phonic ends that are completely foreign to the consonantal system of the Phoenicians. This is a radical reassignment of graphic
material by the Greek adapters one that abandons not only the historical
values of the Phoenician characters but also the dynamic of Phoenician writing. The changes are both local and systemic. The Greek adapters denude
the Phoenician script of its particular vagaries by design (any given vowel
encoding being identical to any other vowel encoding, a zero-encoding in
each case) and replace those with the vocalic vagaries of the Greek syllabaries, both Mycenaean and Cypriot (vowel encoding that fails to distinguish
between contrastively long and short vowel phonemes). The resulting system
is closer to that of the Cypriot script, in which both vocalic components of
diphthongs are regularly given graphic expression.
This level of arbitrariness is not limited to vowels, however. For example,
the Greek adapters of the Phoenician script assigned a profoundly different
value to the Phoenician letter samek. The Semitic character spells the fricative
[s], but upon borrowing the symbol, the Greek adapters arbitrarily assigned
to the symbol the value of two contrastive consonants occurring in sequence
/k/ + /s/. The Greek letter, (or [xi]), and especially its Roman successor, x,
are so familiar to many users of the modern Roman alphabet that its presence
in that script scarcely seems out of place. But this change introduced by the
Greek adapters also marks a sharp departure from the modus operandi of
the Semitic consonantal script38 not to mention being a, seemingly, singularly peculiar choice for those individuals designing this new Greek segmental writing system in which a symbol for /k/ and a symbol for /s/ were also
included. We see here again that the newly crafted alphabetic system departs
systemically from its Phoenician model to follow the practice of the Cypriot
syllabary, in which system symbols with similar value ksV syllabic symbols occur and do so, I have argued, out of necessity, being required for
proper functioning of the syllabary.39
154
5.3.2 semitic and egyptian writing

And what of the symbols of the Phoenician consonantal script itself? Is their
own relationship to the sounds that they encode one of an intrinsically arbitrary matching of graphic shape to phonic value or is there also here some
element of relative arbitrariness? In order to address these questions properly,
one must look beyond the advent of Phoenician writing to a more distant
time. And so, VictorHugo:
Have you ever noticed that the letter Y is a picturesque letter open to countless
interpretations? A tree is in the shape of a Y; the fork of two roads forms a Y;
two rivers flow together in a Y; the head of a donkey or that of an ox is in the
shape of a Y; the stem of a glass is Y-shaped; a lily on its stalk is a Y; a man who
prays to the heavens raises his arms in the shape ofaY.
Besides, this observation can be applied to all aspects of what constitutes
basic human writing. All that is to be found in the demotic language is there
because it was put there by hieratic. The hieroglyph is the essential root of the
written character. All letters began as signs, and all signs began as images.
Human society, the world, and the whole of humankind is to be found in
the alphabet.40
In this excerpt from his travel writings, the French romantic looked to the
Egyptian hieroglyphic script (an object of fascination for romantics [and postromantics alike]) as the primeval source of alphabetic writing and of all that
it subsumes. Hugos thoughts were penned in 1839, in the decade following
that of the decipherment of the Rosetta Stone by Jean-Franois Champollion,
himself a child of the Romantic era [as] shows in his prose as well as in his
tumultuous life.41
In addition to Greek alphabetic letters, the Ptolemaic Stone preserves two
forms of ancient Egyptian script: the better-known hieroglyphic and the
more utilitarian demotic. As Hugo seems to suggest, demotic evolved, ultimately, from hieroglyphic, but stands some distance from it; demotic is a
very cursive script, almost wholly lacking in iconicity and replete with ligatures, abbreviations and other orthographic peculiarities, making it difficult
to read and virtually impossible to transcribe meaningfully into any kind of
hieroglyphic original.42 Not represented on the Rosetta Stone is the third
Egyptian script that Hugo names, hieratic, a modified, cursive version of the
highly iconographic hieroglyphic script, used primarily for nonmonumental
purposes. The hieroglyphic and hieratic scripts are of almost equal antiquity
Langue et criture
155
and, hence, were used to record the Egyptian language as it evolved over a
span of time that can be measured in millennia. The demotic script, in contrast, appears only relatively late in the history of the Egyptian language and
was used for recording that stage of the language likewise labeled Demotic,
beginning circa the seventh century b c . With the appearance of the demotic
script, the use of hieratic script began to wither; however, hieroglyphic writing, first evidenced in the late fourth millennium b c , continued (witness its
use on the early second-century b c Rosetta Stone). The last recorded use of
the hieroglyphic script is dated to a d August 24, 394 (an affiliated graffito
written in demotic script gives the date), found in a temple on the Nile island
of Philae in southern Egypt: The hieroglyphs are crude compared with most
of the texts from the site and a romantically inclined visitor can imagine that
they had been carved when the light was fading the light of knowledge.43
The demotic script would itself succumb to the Greek-based Coptic alphabet
by the middle of the fifth century a d ; demotic, like hieroglyphic, is also last
attested at Philae.44
There is yet another script that was spawned by the iconographic orthography of the Egyptian peoples though not one used for recording the Egyptian
language. The earliest evidence for this script is now dated to circa 1850 b c
(its origin likely to be somewhat earlier) and comes from the desert site of the
Wadi el-Hl, lying west of Luxor along the road that led from ancient Thebes
to Hou, cutting a diagonal swath across the region enclosed by the elbow bend
of the Nile at Qena.45 The script shows the creative influence not only of hieroglyphic forms but of hieratic as well.46 This writing system was employed by
peoples from southwest Asia very likely serving in one capacity or another
with Egyptian military contingents at and around the Wadi el-Hl47 and was
used for writing their Semitic language.
Before the discovery of the Semitic writings at the Wadi el-Hl in
1994/1995,48 the very same script was known from inscriptions found at
and around Serabit el-Khadim in the Sinai and accordingly, the script has
long carried the moniker of Proto-Sinaitic.49 Most of these materials first
came to light during Sir William Flinders Petries 1904/1905 excavations at
Serabit el-Khadim; the inscriptions are the handiwork of Semitic-speaking
Asians engaged in or otherwise supporting the Egyptian turquoise mining
operations in the region. A date of circa 1500 b c was earlier commonly
assigned to the Serabit inscriptions, but recent work has tended to push
the date back; Darnell etal. suggest that a date close to that of the materials from the Wadi el-Hl may be appropriate: It is now not unlikely that
156
at least some of the Serabit material (if not all!) dates to the Middle
Kingdom period.50
The visually elegant but operationally complex writing system of the
Egyptians consisted of both logograms (symbols representing entire words)
and phonograms (symbols having phonetic values):51 The Egyptian hieroglyphs constitute a variable set of graphemes, ranging from about 1000 in
the Old Kingdom (third millennium b c ) down to approximately 750 in the
classical language (second millennium b c ), then increasing to many thousands during the Ptolemaic and Roman rule in Egypt, from the third century
b c to the second century a d .52 Functionally, the phonographic symbols are
of three types. Each type encodes only consonants; vowels do not receive
explicit orthographic expression: there are (1) 24 monoconsonantal symbols, each representing, as the name reveals, only a single consonant, such
as n, spelling m; (2) fewer than 100 biconsonantal symbols, spelling two consonants (that may be interrupted phonetically by a vowel [i.e., reflecting a phonetic sequence -CVC-] but the vowel is not spelled), such as M, for sw; and
triconsonantal symbols (fewer than the number of biconsonantal symbols,
greater than the number of monoconsonantal), encoding a sequence of three
consonants (with the same qualification [i.e., concerning vowels] applying),
such as F, spelling nfr. Typically the biconsonantal and triconsonantal phonograms are accompanied by other (most frequently monoconsonantal) phonograms, serving as phonetic complements, disambiguating potentially ambiguous
spellings: for example nfr good can be spelled as Fp, where p is the monoconsonantal symbol for r. Notice that the numbers of phonograms cited here
make it clear that the majority of Egyptian writing symbols are logograms.
The Proto-Sinaitic script using the term broadly so as to place beneath
this rubric both those earlier-discovered materials from the Sinai and the
more recent finds from the Wadi el-Hl was created by exploiting one of the
conceptual components of the Egyptian writing system. The exploitation was
effected for the purpose of recording the language of Semitic peoples present
in Egypt during the early second millennium b c , peoples with no (known)
preexisting tradition of literacy. That one conceptual component exploited
was the monoconsonantal though it was not a matter of simply taking over
the Egyptian monoconsonantal symbols. Instead, the adapters identified some
number of Egyptian iconographic symbols, drawing on both the hieroglyphic
and hieratic scripts, and assigned to those symbols a phonetic, consonantal,
value on the basis of iconic association: the process is linked to both of the
arbitrary elements of the Saussurian linguistic sign phonic (signifiant) and
conceptual (signifi).
Langue et criture
157
Without ascribing to the adapters any particular historical process or methodology, the adaptation must have involved essentially the following two steps.
(1) An iconic symbol of the Egyptian writing system (criture) and a corresponding Semitic conceptual structure (signifi) were identified, the one with
the other. (2) The initial consonantal sound of the affiliated Semitic phonic
structure (signifiant) was associatively linked to the Egyptian iconic symbol (in
accordance with the so-called acrophonic principle). For any given consonantal symbol so derived, the steps could have been ordered either as (1) followed
by (2) or as (2) followed by (1) (i.e., a sound value may have been first targeted
and the phonic component of an appropriate donor sign identified). The overall historical event by which the Proto-Sinaitic script was generated would have
likely involved back and forth movements in both directions, adjustments and
readjustments, with the monoconsonantal component of the Egyptian writing
system always serving, at some level of consciousness and with some degree of
intentionality on the part of the adapters, as a guiding metric.
As an example, consider the iconic Egyptian symbol q, a logogram representing a courtyard house.53 The finds from the Wadi el-Hl have clearly
revealed this Egyptian symbol to be the source of one of the Proto-Sinaitic
characters.54 The adapters paired this Egyptian courtyard house-symbol
with a Semitic conceptual structure having the sense house; the associated
Semitic phonic structure was an ancestral form of the later West Semitic bet
(also meaning house). Abstracting from this phonic structure its initial consonant, the adapters linked that consonant, /b/, with the iconic symbol q and
thus created a monoconsonantal symbol with the valueb.
In this way the adapters designed a set of symbols, iconic in shape and
phonetic in function, related in shape and function through the mechanism
of the linguistic sign. This earliest of West Semitic writing systems passed out
of Egypt to take root in southwest Asia. Scattered examples of its use there
(and beyond) are attested between the seventeenth and twelfth centuries
b c , with the script in its Asian setting being assigned the denotation ProtoCanaanite.55 By the eleventh century b c , the Proto-Canaanite script had progressively evolved into the linear (i.e., far less iconic) consonantal script utilized by speakers of Phoenician that same script that would be adapted by
Greeks and transformed into a vowel-spelling alphabet three centuries later,
if not somewhat earlier.
***
Though it may not be immediately obvious, there is some insight to be
gained by comparing a more recent adaptation of the monoconsonantal
158
portion of the ancient Egyptian writing system. In this case, a linkage is

made not via the linguistic sign (langue) but via equation with symbols of
another writing system (criture). At tourist sites in Egypt, but also and
more familiarly at fairs and craft shows across North America and Europe,
one finds booths set up at which vendors will, for a fee, write your name in
Egyptian hieroglyphs. Similarly, museum gift shops now often sell kits that
provide the museum visitor with the equipment and expertise to accomplish the task at home, or they provide such a service online for virtualmuseum visitors. The prospect of giving graphic expression to ones own
name utilizing one of humankinds most elegant and exotic writing systems,
one shrouded in the mysteries of ancient Egypt, has a certain undeniable
appeal. For accomplishing the task, vendors and do-it-yourselfers principally utilize only one small component of the complex Egyptian system
the set of twenty-four monoconsonantal symbols the very same symbols
adapted for writing a West Semitic language within Egypt early in the second millennium b c . In this particular case, as noted, adaptation involves
making graphemic matches synchronically aligning ancient Egyptian
monoconsonantal symbols with modern alphabetic symbols (themselves
the diachronic reflexes of Egyptian iconic graphemes) used for the spelling
of modern European languages.
The University of Pennsylvania Museum of Archaeology and Anthropology
has a Web site at which one can generate the spelling of ones name in
Egyptian.56 If, for example, one inputs the name Chuck a name consisting
of a phonic string of three sounds, a vowel bracketed by two consonants [k],
but requiring five letters for ordinary English orthographic rendering the
output is a five-letter string of Egyptian monoconsonantal symbols sqjs%.
The phonetic value of the Egyptian placenta symbol s is [x] the velar
fricative of, for instance, the German word Buch book57 but the algorithm
of the Penn Museum Web site is one that adapts the symbol for transcribing
Latin c, without regard to phonetic considerations: thus, both the c that is
used as the first component of the English ch digraph, commonly spelling, as
here, the palato-alveolar affricate //, and the so-called silent c that appears in
the ck digraph, spelling the velar stop /k/ are spelled with s.
The q symbol of Egyptianized Chuck we have already encountered: this is
the courtyard house symbol adapted by Semitic speakers to produce a symbol for /b/ (and the ultimate ancestor of the Latin letter b). As a symbol of the
monoconsonantal portion of the Egyptian writing system, its phonetic value
Langue et criture
159
is /h/;58 and it is adapted by the Penn algorithm to transcribe Latin h, appearing here in the English digraph ch. The single initial consonant of Chuck that
is, // is thus spelled using two separate Egyptian monoconsonantal symbols, neither of which carries its Egyptian phonetic value (and the Egyptian
monoconsonantal symbols transcribe two Latin letters, c and h, that do not
here have the phonetic values that they typically carry when used singly in
English orthography).
For representing the // vowel of Chuck, the quail-chick symbol j is
employed; its value in the monoconsonantal system of Egyptian is /w/. The
algorithmic output is an interesting one in that it reflects a consonantal-w
/ vocalic-u alternation: the symbol for the Egyptian w was selected by the
Penn designers to transcribe Latin u appearing to be a cross-linguistically phonetically motivated choice (see my remarks on w / u alternation
in the discussions of digamma and iota in Chapter2). Note, however, that
here the transcription is n o t phonetically sensitive: the Egyptian symbol
j, spelling Egyptian /w/, is used to transcribe the Latin grapheme u, not
the high back vowel /u/; the vowel of Chuck, written u, is the low central
vowel//.
The final consonant of Egyptianized Chuck is spelled with the basketwith-handle symbol %, having the phonetic value /k/ in the monoconsonantal component of the Egyptian system.59 The symbol is here used to transcribe
the Latin letter k, which, in this word, also encodes the sound /k/, though, as
noted earlier, the Latin symbol is actually rendering the second component of
the English-orthographic ck digraph that spells /k/. While there is an element
of phonetic motivation in the use of % here, consider that the Penn algorithm
also uses this basket-with-handle symbol to Egyptianize, for example, the first
letter of the name Knightly, where the Latin grapheme k encodes neither the
sound [k] nor any other sound.
From this brief discussion it can be clearly seen that the adaptation of the
Egyptian monoconsonantal system utilized by the Penn Museum program is
orthographically based. That of the West Semitic adapters early in the second
millennium b c was phonetically based. In other words, the two conversions
are transacted upon two fundamentally different systems. The Penn scriptic
adaptation targets a writing system (criture) and, in so doing, replicates the
orthographic idiosyncrasies of that preexisting writing system. The ancient
Semitic scriptic adaptation targeted language (langue) a language that
appears to have had no writing system of its own and in so doing replicated
160
in its graphemic inventory the arbitrary phonic consonantal units of that

language. If the illustration is lighthearted, the distinction it illustrates is significant. A system of writing and a language are two separate beasts.
To return, again, to Saussures point1:
The signs used in writing are arbitrary. The letter t, for instance, has no connexion with the sound it denotes.
A claim of abject arbitrariness in the shapes of the alphabetic letters is somewhat compromised by the diachronic linkages that bind them, via Canaanite
writing, to their Egyptian ancestors. Thus, Roman t is descended from Greek
tau (T), in turn descended from Phoenician taw (), compare Hebrew tw,
meaning mark,60 symbol descended from the Proto-Sinaitic mark sign (C)
which is clearly attested in the new finds from the Wadi El-Hl,61 and is well
known among the materials from Serabit el-Khadim, as in the name of the
goddess blt, Baalat. At each stage, mutatis mutandis, the symbol gives written
expression to the sound /t/. The symbols Proto-Sinaitic origin lies in the adaptation of an iconic Egyptian symbol, likely the symbol , the crossed-planks
sign.62 The process was, as we saw, one in which each component of the linguistic sign phonic and conceptual was crucially involved: the Semitic adapters
paired this geometrically suggestive Egyptian symbol with a Semiticconceptual structure having the sense mark; the associated West Semitic phonic
structure was the ancestor to the later-attested signifiant taw, which acrophonically provided the value /t/ to the Proto-Sinaitic graphemeC.
If there were any arbitrariness in play here, it would be that of an arbitrary
selection of one particular stylization (conventionalized shape) of an iconographic symbol over other possible stylizations. Though beyond that, one may
also identify a form of the derivative arbitrariness that we have seen to be prevalent in the design of writing systems: in this instance, however, that arbitrariness is not only phonic in nature but semantic as well. It does not merely follow derivatively from that is, replicate the intrinsic arbitrariness of a set of
linguistically significant, contrastive sounds (the consonantal inventory of the
form of West Semitic present in Egypt in the early second millennium b c ), but
is based upon the arbitrariness of the linguistic sign itself, graphically capturing elements of both the phonic and conceptual components of signs. There is
a certain (narrow and anachronistic) sense, thus, in which the romantic Victor
Hugos remarks are validated: The hieroglyph is the essential root of the written character. All letters began as signs, and all signs began as images.63
Langue et criture
5.4
161
al p ha b e t ic o r d e r
The preceding chapter ended and the present chapter began with references
to the alphabetic order, to the abecedarium that primal periodic sequencing of letters that is an integral component of alphabetic traditions. It is an
element of such writing systems that is typically mastered by the alphabet
learner before the full panoply of the letter shapes themselves (and thus there
is a sense in which it has an existence distinct from that of its constituent
graphemes). Are we to see in that ordering an example of Saussures arbitraire
absolu the kind of arbitrariness that is fundamental to a linguistic system?
In order to approach an answer to that question, we must again adopt the
diachronicview.
The earliest-known attestations of the alphabetic order come from SyriaPalestine. A Proto-Canaanite abecedarium appears on an ostracon found at
Izbet arah (east of Tel Aviv), dated to circa 1200 b c .64 Earlier still, however, are abecedaria preserving the letter order of the Ugaritic consonantal
script,65 a script attested circa 13001190 b c . The writing system of the Syrian
kingdom of Ugarit (modern Ras Shamra) is clearly descended from the ProtoSinaitic system but differs from related Semitic scripts in having cuneiform
letter shapes. In addition, the Ugaritic system encompasses a total of twentyseven consonantal symbols, as opposed to the twenty-two of Phoenician; the
Ugaritic system is thus commonly dubbed the long consonantal script.66 The
cause of the greater length of the Ugaritic script is phonological: the Ugaritic
language possessed a larger number of consonant phonemes than did related
(well-attested) West Semitic languages, such as Phoenician; the Ugaritic writing system, being derivatively related to the Ugaritic language, accordingly
possessed a greater number of consonant graphemes. It is generally agreed that
the shorter length of the Phoenician script is a consequence of the diachronic
loss of consonantal phonemes from the Phoenician language, resulting in the
excision of consonant graphemes from an ancestral Canaanite writing system67
having a greater graphemic inventory.68 As Pardee putsit,
The order of the long alphabet, as illustrated by the Ugaritic abecedaries, makes
the hypothesis plausible that the short alphabet is a simplification of the longer
one: that is, certain sign forms were dropped from usage when not needed
(probably in the process of adoption from one language to another, rather
than by alphabet reform in a given community), for alphabet usage tends to
be conservative.69
162
The conservative or, perhaps, persistent nature of the periodic order

of alphabetic characters (the abecedarium) is actually quite extraordinary.
In a pair of intriguing articles penned in the late 1980s, W. C. Watt focused
closely on the letter order of early West Semitic scripts.70 One of the ancillary
observations that he makes in the course of his second study concerns the
alphabet that he acronymically designates as the SEER alphabet:
Ignoring a more or less constant evolution of letter shapes over the 3,500years
of their use, then, and a few other vicissitudes, we may speak with impunity
of the alphabet held in common by us and the Canaanites and the roughly
140 generations between as a single writing system one we might call the
Semitic/Euboean/Etruscan/Roman alphabet, or SEER alphabet for short.
It is a remarkable fact that during most and perhaps all of its history the letters of the SEER alphabet have been passed on in the form of a canonical list,
an ABC or abecedarium, in which those letters were ordered almost exactly
as they are today: A, B, C, D and so on. (Of the letters that have persisted
continuously in the SEER abecedarium since its known beginning at around
1400 b c [now shown to be earlier: RDW], not one has changed its place in
the list.) It would of course be quite possible for one people to adopt anothers alphabet without also adopting their traditional letter list, their ABC.71
This said, we now note that the original SEER letter-list ended at T, though
in most descendent alphabets more letters were added later. Properly speaking, then, the SEER abecedarium is coextensive with what we could and will
call the common alphabetic core (A through T), even though almost all
abecedaria also sport a kind of tail or coda (in the English alphabet, U
through Z).72
This alphabetic core , , at is, rigid, stable, bracing. When this

archaic core is extended by the appending of graphemes to the coda such
as the Greek supplementals , , , 73 the extended periodic structure
conforms to the rigidity of thecore.
This last observation focuses our attention on a phenomenon of writing
systems specifically, in the case at hand, the Greek alphabet, its ancestors,
and its offspring that my discussion has thus far assumed, without explicitly
addressing. Namely, there exists the possibility of an individual, or some set
of individuals, creating a new writing system or modifying an existing writing system. Thus, we have seen how some person or persons of southwest
Asian extraction, speaking a West Semitic language, living in an Egyptian
setting, perhaps in partnership with an Egyptian scribe, or scribes, devised
Langue et criture
163
a consonantal script by adapting Egyptian iconic symbols and exploiting the

model of Egyptian monoconsonantal orthography. We have seen how a Greek
individual or group, undoubtedly living in a Phoenician context, adapted that
West Semitic consonantal script by converting superfluous consonantal symbols into vocalic symbols, as well as making yet other modifications, creating a
core Greek alphabet that would be extended by the appending of additional
characters again, the work of individuals or sets of individuals. This script,
the Greek alphabet, would in turn be adapted in Italy for and by Etruscans
and Indo-Europeans; and in Anatolia for and by Phrygians, Lydians, Lycians,
and Carians; and in Africa for and by Copts. From Greece and Italy the alphabet would be carried deep into Europe, and beyond, modified, for example,
for use in recording Germanic the runic alphabets (futharks) devised by
unknown persons, and the Gothic alphabet created by the Arianist Visigothic
bishop Wulfila in the fourth century a d and adapted for recording Slavic by
the brothers Cyril and Methodius in the ninth century and Armenian by the
cleric Mesrop in the fifth century (according to tradition) and so on. I will
return to this matter of an individuals (and individuals) ability to create and
change a writing system but will first turn to one other consideration regarding the periodic structure of the alphabet.
The linear alpha-through-tau structure (with eventual appendages) of the
Greek alphabet (and of its ancestors and descendants) is akin to a fundamental property of language namely, syntagmatic structure. Again, Saussure as
presented in the Cours:
Words as used in discourse, strung together one after another, enter into relations based on the linear character of languages. Linearity precludes the possibility of uttering two words simultaneously. They must be arranged consecutively in spoken sequence. Combinations based on sequentiality may be called
syntagmas. The syntagma invariably comprises two or more consecutive units:
for example, re-lire (re-read), contre tous (against all), la vie humaine (the life
of man), Dieu est bon (God is good), sil fait beau temps, nous sortirons (if its
fine, well go out). In its place in a syntagma, any unit acquires its value simply
in opposition to what precedes, or to what follows, or to both. (trans. Harris)74
The periodic order of letters is a linear structure, composed of units

arranged consecutively in spoken sequence, a combination based on
sequentiality a syntagma. In the case of the Greek alphabet or, perhaps
I should say, of that continuum of scripts that Watt labels as SEER it is an
existentially fixed syntagma, permitting no variation in sequencing.
164
In this way, the syntagmatic structure of the alphabet is fundamentally

different from that of a natural language, even one characterized by a high
degree of word-order restrictiveness. On the one hand, the systemic disruption
that would arise by varying the place of units in a French phrase such as la vie
humaine, or a sentence such as, sil fait beau temps, nous sortirons would not
seem, prima facie, to be replicated by varying the periodic order of symbols in
the SEER abecedarium. Indeed, the Arabian consonantal scripts,75 of common
origin with Ugaritic and Phoenician scripts, and certain of the descendants
of the Greek/Latin alphabet, notably the runic futharks,76 deviate, at the time
of their earliest attestation, appreciably from the periodic order of the core of
the Canaanite and Greek (etal.) scripts though these deviant descendants
then in turn display their own stable and unvarying periodic order (i.e., form
different syntagmata). The SEER orthographic system, of which the individual
graphemes constitute fundamental components, would function equally well
with an abecedarium of deviant order. On the other hand, for a given orthographic system consonantal or alphabetic such deviant orders, in effect,
simply do not exist. The graphemic units of the alphabet form an alphabetic
syntagma,77 but the alphabet (unlike the French language, continuing the preceding comparison) lacks a productive syntax the syntax and syntagmatics
of a language being asymmetrically related, per Saussure:
Syntax as generally understood nowadays the theory of word sequences
falls under syntagmatics, since word sequences invariably involve at least
two units spatially ordered in the same way. All syntagmatic facts are not to
be classed as syntax, but all syntactic facts belong to syntagmatics. (trans.
Harris)78
The syntax of French or English, or Phoenician, or Greek is productive

to the extent that it is capable of generating a variety of syntagmatic structures all such structures that the language permits. The syntax of the alphabet, unlike that of these languages, is nonproductive that is, it is frozen in
form: the syntax of the alphabet is the syntagma of the alphabet.
To this extent, the syntagmatic structure of the alphabet, an orthographic
system, is reminiscent of that of the idioms of a linguistic system. In the discussion of syntagmatic relations in the Cours, Saussure (along with his editors)79
draws attention to idiomatic constructions:80
There are, in the first place, a large number of expressions. these are readymade phrases, absolutely invariable in usage, in which it may even require
Langue et criture
165
reflection to distinguish the constituent parts: e.g. quoi bon? (whats the
use?), allons donc! (come along!). The same is true, although not to the same
extent, for expressions like prendre la mouche (to take offence), forcer la main
quelquun (to force someones hand), rompre une lance (to break a lance),
avoir mal la tte (to have a headache), force de (by dint of ), que vous en
semble? (what do you think of it?), pas nest besoin de (no need to ),
etc. These are idiomatic expressions involving oddities of meaning or syntax.
These oddities are not improvised, but are handed down by tradition. (trans.
Harris)81
Like these ready-made phrases [les locutions toutes faites], absolutely invariable in usage [auxquelles lusage interdit de rien changer], the inflexible syntagmatic structure of the SEER abecedarium was handed down by tradition.
The flat syntax of the Greek abecedarium, may, however, not always have
been so. In his two studies of alphabetic order in the Canaanite script and
its descendants, Watt presents tightly reasoned arguments in favor of the
early existence of a matrix abecedarium.82 Somewhat like others before
him,83 Watt detects certain patterns in the linear arrangement of the letters
of the Canaanite abecedarium. But quite unlike those others, Watt is able to
abstract from that linear arrangement a recurring phonological periodicity
(statistically improbable of being the product of chance): essentially a twodimensional structure consisting of rows and columns a, one might say,
periodic table of the alphabetic elements.
The structure of the matrix, Watt argues, is best preserved by the long
Ugaritic script what he labels as the Ras Shamra Matrix. That is to say,
the Ugaritic matrix is primitive, as opposed to the Phoenician the Byblos
Matrix which is derivative: Overall then, the Ras Shamra Matrix, understood as applying also to whatever system was ancestral to both Ugaritic and
Phoenician, reduces to the Byblos Matrix of Phoenician.84
For Watt, the arrangement of symbols within the matrix was determined
by the principle of maximal separation, whose force is to segregate similar
sounds, as much as possible.85 The segregation occurs between columns (i.e.,
serially along the horizontal axis of the matrix), and to an extent within several subcolumns that Watt identifies. In the next chapter I will return to the
form of Watts matrix for a somewhat closer look, but for the present, its gross
columnar structure can be summarized as follows. The initial two columns
contrast graphemes for sounds produced at the margins of the oral cavity
back and front (laryngeals/pharyngeals; labials and interdentals) and the
166
last two columns contrast graphemes for sounds articulated at slightly less
extreme positions back and front (velars; dento-alveolars) while graphemes for more centrally articulated sounds (alveolars; palatals) are assigned
contrastively (and metaphorically?) to the middle column.
Watt argues that the principle of maximum separation also applies along
the vertical dimension that is, within a column, between its constituent
rows. Within the vertical domain of the matrix, it is a contrast in voicing
(i.e., the presence or absence of vocal cord vibration) that achieves maximal
separation.86
The matrix thus has a horizontal structure consisting of the following
sequence of places of articulation:
(7) Extreme BackExtreme Front MiddleBack Front
Six rows (i.e., six periodic series of graphemes) appear beneath the columnar
headings, with the Ugaritic form of alep (i.e., Greek alpha) being located in
the first row of the Extreme Back column, and the Ugaritic form of taw (i.e.,
Greek tau) being located in the sixth row of the Front column.87
The structure of the matrix, Watt conjectures, may have served a fundamental pedagogical purpose:
I will propose that the Matrix was designed to obey that principle [of maximal
separation], scanning both vertically and horizontally, because it was designed
to be recited both vertically and horizontally, and because it was expected
that such recitations would be eased and/or that monitoring them would be
eased if highly-similar (hence easily confused) sounds were as distant from
each other as possible.88
The two-dimensional design that Watt perceives to be evidenced in the

Ugaritic (and, generally, by extrapolation, the Canaanite) abecedarium is
consequent to and an expression of a principle of maximum separation of
graphemes a separation based not on their graphic form but on their phonic (articulatory) value. This orthographic principle itself, one could argue, is
derivative in nature, mimicking a well-known cross-linguistic tendency (or
principle, one might say). The set of vowel phonemes occurring in a given
language tends to distribute itself through vocalic space in such a way as to
maximize the perceptual and articulatory distance between the vowels.89 Thus,
if a language has only three vowel phonemes (and very few languages, if any,90
have fewer than three vowels), those three vowels will typically be /i/ (high
Langue et criture
167
front), /u/ (high back), and /a/ (low central) the three maximally separated
points on the classic vocalic triangle.91 As the number of vowel phonemes
increases,92 the distribution tends still to be one in which vowels are dispersed
maximally throughout phonetic space.93
In groundbreaking work on the acoustic components (distinctive features)
of individual vowel and consonant sounds, Roman Jakobson observed that
the maximally separated vowels of the vocalic triangle are distinguished by
oppositions occurring along two axes axes defined by the acoustic properties of compact-diffuse94 and grave-acute.95 Jakobson further showed that
from an acoustic perspective this fundamental /u/-/i/-/a/ vowel triangle is
closely paralleled by the widely occurring set of stop consonant phonemes
/p/, /t/, /k/. The consonant phonemes likewise plot a triangular arrangement
as points along the axes defined by the features compact-diffuse and graveacute features that are (nearly) universal in occurrence: The oppositions
compact-diffuse and grave-acute are the only two features which belong
to vocalism and/or consonantism in all languages of the world; but in the
whole ensemble of the worlds languages, with a small handful of exceptions,
these oppositions usually pertain both to the vocalic and to the consonantal systems.96 Regarding the similarity between the /u/-/i/-/a/ vocalic and
/p/-/t/-/k/ consonantal configurations, Jakobson writes:
The vocalic and consonantal patterns show a few striking symmetrical traits:
in each pattern the division into grave and acute opposites is much more frequent, even nearly universal, within the set of diffuse consonants and vowels
than within the compact ones. Thus, the triangular scheme is widespread both
in vocalism and in consonantism; note the parallelism between the two triangles [see (8)] with two perpendicular axes: grave-acute on the horizontal
and compact-diffuse on the vertical.97
(8)
VOWEL
Compact
Diffuse
a
u
i
Grave Acute
The triangles (with their two oppositional axes) to which Jakobson here makes
reference98 have the following forms:
CONSONANT
and
k
p
t
Grave Acute
Compact
Diffuse
168
The fundamental principle (the principle of maximum separation) that

Watt sees to be at work in the structuring of the Canaanite abecedarium into a
matrix would likely be one secondary to a phenomenon of phonological (i.e.,
linguistic) segregation finding expression in the maximal spacing of vowels
around the periphery of the vocalic triangle. This distribution is a function of
binary oppositions involving two acoustic features that are nearly universal,
if not universal, in occurrence; the same features in binary opposition play a
similar role in consonantal structure. This is not necessarily to suggest that
the separation of columns and rows in the matrix is based on acoustic features (grave-acute, compact-diffuse, or otherwise).99 It is to suggest, rather,
that a cognitive adaptation of a (near) universal structure of human language
may be in play: a discriminatory principle of maximal graphemic spacing qua
binary phonetic featural distinctions. As Jakobson points out, the two axes,
grave-acute and compact-diffuse, have long been seen to have other cognitive
correlations:
Two prominent psychologists, Carl Stumpf (18481936), who was the first to
say that one cannot assess psychological problems without approaching the
question of the search for structural laws (1907:61ff.), and his great inspirer
in psycho-acoustics, Wolfgang Khler (18871967), renowned particularly for
his synthetic Gestaltist work (see Khler 1929), revealed the role played in the
vocalic system by these two axes the U-I process and the A-process (in
Stumpf s terms). Both scholars invoked a phenomenological, adequate coordination of these two dimensions with the corresponding dimensions within
other sensory domains and particularly within the system of colors.
The further development of the acoustic inquiry into sound shape and of
its linguistic interpretation has permitted the expansion of the idea of the two
axes onto the whole framework of language, as well as the comparison of their
treatment in the vocalic and consonantal subsystems, a comparison taking
into account both convergences and divergences (Struktureigentmlichkeiten)
in these two opposite subsystems (see Jakobson [1968]).100
In the words of Balzac: Tout est bilatral dans le domaine de la pense. Les
ides sont binaires. Janus est le mythe de la critique et le symbole du gnie. Il
ny a que Dieu de triangulaire.101
The two-dimensional structure of the Canaanite abecedarium as revealed
by Watts matrix serial rows of graphemes discriminated by maximally separated columnar categories turns our attention back to Saussure. In addition
to syntagmatic structure, Saussure describes a second fundamental structural
Langue et criture
169
property of language, paradigmatic structure, or, in Saussures terms, les

rapports associatifs.102 While syntagmatic connections constitute in presentia relations, associative, or paradigmatic, connections constitute in absentia
relations framed in the Cours in terms of words or smaller morphological
units that can substitute for one another as members of une srie mnmonique virtuelleles groupes forms par association mentale.103 Such associative sets can be narrowly or broadly delimited (the following translation is
that of Harris; I have modified line spacing for expository reasons):
For the mind grasps the nature of the relations involved in each case, and
thus creates as many associative series as there are different relations. In enseignement (teaching), enseigner (to teach), enseignons ([we] teach), etc., there
is a common element in all the terms, i.e. the stem enseign- (teach-).
But the word enseignement also belongs to another series based upon a different common element, the suffix -ment: e.g. enseignement (teaching), armement (armament), changement (change),etc.
The association may also be based just on similarity of significations, as in
enseignement (teaching), instruction (instruction), apprentissage (apprenticeship), ducation (education),etc.
Similarly, it may be based just on similarity of sound patterns,104 e.g. the final
syllables of enseignement (teaching) and justement (precisely).
So sometimes there is a double associative link based on form and meaning,
but in other cases just one associative link based on form or meaning alone.
Any word can evoke in the mind whatever is capable of being associated with
it in some way or other.105
Syntagmatic and associative relations are at the core of all language structure and function. This is a phenomenon on which Jakobson wrote recurringly, with characteristic insight and clarity: Any linguistic sign involves two
modes of arrangement: combination [and] selection106 (where combination denotes the syntagmatic axis and selection the associative axis). These
processes are at work within the phonological domain, no less than within the
morpho-syntactic:
Did you say pig or fig? said the Cat. I said pig, replied Alice (Carroll 1866:
chap.6). In this peculiar utterance the feline addressee attempts to recapture
a linguistic choice made by the addresser. In the common code of the Cat and
Alice (spoken English), the difference between a stop and a continuant, other
things being equal, may change the meaning of the message. Alice had used
170
the distinctive feature stop versus continuant, rejecting the latter and choosing
the former of the two opposites; and in the same act of speech she combined
this solution with certain other simultaneous features, using the gravity and
the tenseness of /p/ in contradistinction to the acuteness of /t/ and to the laxness of /b/. Thus all these attributes have been combined into a bundle of distinctive features, the so-called phoneme. The phoneme /p/ was then followed
by the phonemes // and /g/, themselves bundles of simultaneously produced
distinctive features. Hence the concurrence of simultaneous entities [associative axis] and the concatenation of successive entities [syntagmatic axis] are the
two ways in which we speakers combine linguistic constituents.
As Jakobson goes on to demonstrate in the essay from which these lines are
drawn, Two Aspects of Language and Two Types of Aphasic Disturbances,
these two structural processes one of concurrence, one of concatenation
are fundamental to a variety of cognitive phenomena, including language dissolution: The varieties of aphasia are numerous and diverse, but all of them
lie between two polar types. Every form of aphasic disturbance consists
in some impairment, more or less severe, of the faculty either for selection
and substitution or for combination and contexture.107
Thus, on one side of the aphasic spectrum are conditions in which impairment lies along the associative axis: patients are unable to substitute one linguistic form for a similar, or otherwise alternative, linguistic form but remain
sensitive and responsive to linear context. In severe cases, a patients utterances may be reduced to strings of words that carry a high functional load as
syntactic operators pronominal forms, grammatical particles, conjunctions,
and so forth. As Jakobson saw, individuals afflicted with such aphasic conditions suffer from the impairment of metalinguistic capabilities: The aphasic
defect in the capacity of naming is properly a loss of metalanguage.108 The
ability to interpret metaphorically is reduced (or lost), but metonymy, in contrast, is commonly utilized in compensating for the loss of word selectivity.109
On the other side of the spectrum lie those forms of aphasia that are characterized by impairments associated with the syntagmatic axis. Patients suffering
with these conditions, the contiguity disorder, experience contextural deficiencies a degeneration of syntactic capabilities. The most resilient elements
in the speech of persons afflicted with the similarity disorder described in the
preceding paragraph are, thus, the most fragile components of the utterances
of those suffering with the contiguity disorder: Word order becomes chaotic;
the ties of grammatical coordination and subordination, whether concord
Langue et criture
171
or government, are dissolved.110 In this instance, metaphorical expression is

commonplace, while metonymy is foreign to the condition.111
The polar opposition of metaphor and metonymy (of similarity and contiguity) that is displayed on the aphasic spectrum, Jakobson goes on to point out,
is of course to be found elsewhere, among other language-related phenomena,
particularly in the crafting of language as art. Thus, in Russian lyrical songs,
for example, metaphoric constructions predominate, while in the heroic epics
the metonymic way is preponderant.112 Similarly, metaphor tends to predominate in the poetic traditions of Romanticism and Symbolism, while it is the
predominance of metonymy which underlies and actually predetermines the
so-called Realist trend, which belongs to an intermediary stage between the
decline of Romanticism and the rise of Symbolism and is opposed to both.113
Similar oppositions, he continues, can be seen in the history of painting and
cinema, in Freudian interpretation of dreams, in Frazerean analysis of sympathetic magic: The dichotomy discussed here appears to be of primal significance and consequence for all verbal behavior and for human behavior
generally.114
Similarity and contiguity associative and syntagmatic axes are no
less to be seen in the verbal behavior that finds expression in the repeating abecedaria of the copper plaques: contiguity in the linear syntax of the
alpha to tau ordering of the letters of the abecedarium, as we observed earlier,
and similarity in the variant forms of letters that substitute one for another
within the many abecedaria of the plaques. These variants were the object
of detailed investigation in Chapter2, and it is these variants substitutive
graphemic expressions on the axis of similarity (the metaphoric axis) that
are principally responsible for the making of almost each occurrence of the
CP abecedarium a unique structure, as is made clear by the examinations of
Chapter4.
5.5
la n g uag e a n d n o n - la n g uag e
While the abecedaria of the copper plaques share with other verbal behaviors, and other human behaviors generally, to co-opt Jakobson,115 the dichotomous structures of contiguity and similarity, what I hope has been made
clear by the discussions of this chapter is that language and writing are two
distinct phenomena in other words, that writing (criture) is not language
(langue), that language is not writing. Language is primary; writing is derivative. Havelock himself made the distinction cleanly:
172
The biological-historical fact is that homo sapiens is a species which uses oral
speech, manufactured by the mouth, to communicate. That is his definition.
He is not, by definition, a writer or reader. In short, reading man, as opposed
to speaking man, is not biologically determined. He wears the appearance of a
recent historical accident, and the same can be said of whatever written symbols he may choose to use.116
Beyond points made in the discussion of writing as a derivative phenomenon that I have set out in the preceding pages, consider too that writing
does not share those fundamental characteristics of language which Saussure
highlights in the passage cited at the outset of this chapter observations
that surround his remarks on the arbitrariness of the linguistic sign. Thus, he
notes, regarding language (langue), that personne ne peut rien y changer;
more fully:
In order to emphasize that a language is nothing other than a social institution, Whitney117 quite rightly insisted upon the arbitrary character of linguistic
signs. In so doing, he pointed linguistics in the right direction. But he did not
go far enough. For he failed to see that this arbitrary character fundamentally distinguishes languages [Saussure: la langue] from all other institutions.
This can be seen in the way in which a language evolves. The process is highly
complex. A language is situated socially and chronologically by reference to
a certain community and a certain period of time. No one can alter it in any
particular. (trans. Harris)118
While a language (langue) cannot be consciously altered, the potentiality most certainly exists for a writing system to be intentionally and radically modified even by a single individual. The history of humanity gives
evidence of this event occurring many times over: we have encountered several instances of it in this chapter. In the early second millennium b c , some
person or persons exploited a single component of iconic Egyptian script to
create a fully phonetic, monoconsonantal script, attested in its Proto-Sinaitic
and Proto-Canaanite forms. Some person or persons of Ugarit modified this
script later in the second millennium by replacing its Egyptian-mimicking (or
inspired) graphic-forms with Mesopotamian-like cuneiform symbols, and
by adding three syllabic characters to its inventory of consonantal letters. At
some point in the history of Canaanite consonantal writing some person or
persons modified the established periodic order of letters, or otherwise canonized two distinct letter orders that typical of west Semitic scripts, and that
Langue et criture
173
typical of south Semitic scripts (though also attested in Syria-Palestine).119 By

the early first millennium b c , some person or persons, probably living in or
otherwise affiliated with Cyprus and already literate, adapted the Phoenician
form of Canaanite writing for spelling a non-Semitic language namely,
Greek and in the process utilized unneeded Phoenician consonantal graphemes to produce a set of vowel characters. Subsequently, this script underwent various local adjustments to produce the attested epichoric alphabets of
archaic Greece. And so on and on, alterations in the script have continued to
be designed.
Though language (langue) cannot be volitionally altered, it is, on the other
hand, as Saussure points out, continually in a state of evolutionary flux: Cette
volution est fatale. Just as no individual or individuals can alter langue, no
individual or individuals can hold back the tide of language change. But while
language is thus from a diachronic perspective fundamentally unstable, writing systems tend to be relatively stable diachronically, in spite of the fact that
individuals possess the potentiality to alter scripts in a radical fashion when
the need, or urge, arises. In this way, langue and criture are precise mirror
images, the one being the complementary reflection of the other.
The diachronic stability of writing is something that we have already
encountered. Recall Watts observations regarding what he denotes as the
SEER alphabet: Ignoring a more or less constant evolution of letter shapes
over the 3,500years of their use, then, and a few other vicissitudes, we may
speak with impunity of the alphabet held in common by us and the Canaanites
and the roughly 140 generations between as a single writing system.120 Letter
shapes do of course evolve (as all elements of human institutions do) a notable example from the history of the alphabet being the Phoenician linear conventionalization of the more-pictorial Canaanite symbols and, what is more,
letter shapes can be graphically modified to yield synchronically alternating
character sets, often for principally aesthetic, sometimes utilitarian, reasons
one thinks immediately of Roman capital and uncial; of Carolingian minuscule and Gothic hands; of Japanese hiragana and katakana; of Times New
Roman and Courier; and soon.
However, the evolution of letter shapes may be minimized in effect eliminated over long expanses of time. Early in this chapter I drew the readers attention to the inscription on the seventh-century aryballos of Tataie
from Kyme and to how a French and an Italian scholar can invest the idea
expressed by the Greek language recorded on that pot in the phonic strings
transcribed qui me drobera deviendra aveugle121 and chiunque mi ruber sar
174
cieco, respectively. The alphabetic symbols here used in the two transcriptions, French and Italian, are identical. When speakers of the two Romance
languages here transcribed visit Roman historic sites, they encounter graffiti
and monumental inscriptions written with alphabetic symbols. Those Roman
symbols are closely related if not fully identical to the alphabetic symbols
used to produce the cited transcriptions, and most such visitors, if literate in
their own languages, would have no difficulty in giving oral expression to the
Roman inscriptions. They would not be able, however, to comprehend the
meaning of the inscriptions unless they have learned the Latin language. In
other words, while the language (langue) of Roman inscriptions has evolved
into distinctly different languages French and Italian mutually unintelligible (when one compares the literary standards), the graphic device with
which the language of the inscriptions is recorded (criture) has remained
comparatively unchanged. The differences between the two systems language and criture is yet more striking when one expands this exercise so
as to include all of the members of the Romance language family.122 The differences grow, in a certain sense, more significant still when we introduce
English- and German-speaking scholars and tourists into the scenario, whose
West Germanic languages are written using the same Roman script but whose
genetic affiliations with Latin language are realized only by climbing nearly to
the top of the Indo-European tree and lookingdown.
Saussure was, of course, well aware that writing (criture) is not language
(langue). The comparison that he offers of written signs to linguistic signs vis-vis arbitrariness, discussed earlier in the chapter, must not be construed as
confusion on Saussures part regarding the langue-criture distinction. In the
Cours he declares:
A language and its written form constitute two separate systems of signs. The
sole reason for the existence of the latter is to represent the former. The object
of study in linguistics is not a combination of the written word and the spoken
word. The spoken word alone constitutes that object. But the written word is
so intimately connected with the spoken word it represents that it manages to
usurp the principal role. As much or even more importance is given to this
representation of the vocal sign as to the vocal sign itself. It is rather as if people believed that in order to find out what a person looks like it is better to
study his photograph than his face. (trans. Harris)123
And let us recall what we have already seen Saussure to judge regarding writings relationship to language: lcriture voile la vue de la langue: elle nest pas
un vtement, mais un travestissement.124
Langue et criture
175
In a recent article on the introduction of the alphabet into archaic Greece,

the author of the study writes:
Those who are fully literate (particularly in more than one language) rarely
make the mistake of confusing language with script. The two represent quite
different and essentially independent and separable types of systems. What
does seem likely, however, is that particularly in illiterate contexts or contexts
where literacy is very limited, language and script do have the potential to
become more closely identified, with the signs in a real sense becoming the
visible symbol of the language, even (or perhaps particularly) if the viewer
cannot actually read them.125
Certainly language and script do have the potential of being very closely
identified, and the discussions heretofore presented within this chapter
demonstrate that language and script indeed represent quite different and
independent systems (in effect the author has paraphrased Saussures
remarks quoted here). However, this authors opening premise that those
who are fully literate (particularly in more than one language) rarely make
the mistake of confusing language with script is just as certainly misplaced:
nothing, in fact, could be farther from the truth; one need merely consider
Nonnoss remarks rehearsed at the beginning of Chapter 4. As a linguistic
educator, I have repeatedly found that bright and highly literate undergraduates have considerable difficulty in grasping the difference between phoneme
and grapheme between sound and letter when first exposed to phonological concepts, often attempting to negotiate those concepts in terms of conventional English orthographic units.126 The same problem, mutatis mutandis,
seems not infrequently to beset classics graduate students who for the first
time study Greek and Latin as objects of comparative linguistic examination.
All too often it appears that the distinction between phoneme and grapheme langue and criture is never really grasped in an intellectually and
experientially meaningful way. It is a curious cognitive roadblock, yet one
well attested within the history of linguistic study one to which Saussure
could already draw attention in his lectures that underlie the Cours.
Following upon his previously cited remarks regarding the confusion of
the two separate systems of language and writing on en vient donner
autant et plus dimportance la reprsentation du signe vocal qu ce signe
lui-mme Saussure notes: Cette illusion a exist de tout temps, et les opinions courantes quon colporte sur la langue en sont entaches. He continues:
The first linguists were misled in this way, as the humanists had been before
them. Even Bopp does not distinguish clearly between letters [la lettre] and
176
sounds [le son]. Reading Bopp, we might think that language is inseparable
from its alphabet. His immediate successors fell into the same trap. The spelling th for the fricative misled Grimm into believing not only that this was
a double consonant, but also that it was an aspirate stop. Hence the place he
assigns to it in his Law of Consonantal Mutation or Lautverschiebung.
Even nowadays educated people confuse the language with its spelling: Gaston
Deschamps said of Berthelot that he had saved the French language from
ruin because he had opposed spelling reforms. (trans. Harris)127
Bopp is, of course, Franz Bopp (17911867), author of the multivolume tome
Vergleichende Grammatik des Sanskrit, Zend, Griechischen, Lateinischen,
Litthauischen, Altslawischen, Gotischen und Deutschen, credited with first
demonstrating the relatedness of various Indo-European languages and
holder of the first chair of Indo-European studies (the chair of Orientalische
Literatur und Allgemeine Sprachkunde at Berlin, created in 1821).128 Grimm is
Jacob Grimm (17851863), and his Lautverschiebung the formulation probably
most readily recognized in the English-speaking world by the title Grimms
Law. Gaston Deschamps (n en 1861), polygraphe franais trs connu la fin
du sicle dernier [i.e., the nineteenth century], avait affirm en 1908, parlant
lAcadmie de P. E. M. Berthelot (18271907), que le savant stait oppos le
ruine de la langue franaise en se prononant contre les tentatives de rforme
de lorthographie entreprises par les autorits franaises entre 1901 et 1905.129
Berthelot is the distinguished French chemist and historian of alchemy, Pierre
Eugne Marcellin Berthelot (18271907). The student who fails to comprehend the divide between langue and criture is in good company.
***
The abecedaria of the copper plaques belong to the realm of criture, not of
langue. Yet in at least one instance, the steadfast periodic order of the archaic
Greek abecedarium is abandoned for a seemingly arbitrary sequence of graphemes, orthographic units that encode the phonological units of the Greek language. The arbitrary sequencing of phonological units is the hallmark of the
realm of langue. What one finds in one (perhaps two) lines of the copper
plaques is language text and to this preservation of linguistic utterance I
nowturn.
6
Of Styluses and Withes
,
Once they set out the warp, its the beginning of aweb.
6.0
Callimachus, fr.520
The copper plaques compel one to ask the a priori question of why. Why produce string after string of the alphabetic letters in their periodic order? The
answer to this question, I will argue, lies within the line or lines of the
copper plaque that preserve(s) langue an arbitrary sequence of graphemes
encoding an arbitrary sequence of phonemes rather than criture alone.
I began Chapter5 by setting out the transcription of MS 2-2, line 16, rewritten from left to right, but further consideration of that line had to be postponed until the relationship between langue and criture could be carefully
examined and the two systems teased apart. Here, again, is the transcription:
(1) F
One familiar with the linguistic signs of ancient Greek can parse the string
of orthographic symbols inscribed on the aryballos of Tataie from seventhcentury Kyme (2) to reveal an argument structure (3), as we did in Chapter5:
(2)
(3)
One can likewise parse with little difficulty the string of MS 2-2, line 16 (here
repeated):
177
178
(4) F
The following, I will argue, is the correct parsing:
(5) F
Rewriting this orthographic string in a more familiar form will make the
argument structure somewhat more obvious:
(6)
Ml se luzdabgd
Each constituent of the argument structure will be examined inturn.
6.1
( m l )
In the discussions of Chapter2, the reader saw that in the abecedaria of the
copper plaques a recurring anomaly is the placement of the xi-symbol in
the eta-position and the placement of the eta-symbol h in the xi-position. It
was further seen that this practice and the shapes it involves are revealed to be
one expression of a broadly attested confluence of orthographic devices used
for the representation of eta-sounds and xi-sounds. I would propose that the
graphemic sequence that is, the spelling h (i.e., ) in line 16 of
MS 2-2 reflects not merely a slotting-in of the xi-symbol in the eta-position
of the abecedarium but, more than that, reveals an actual synchronic orthographic use of the xi-symbol for spelling the eta-vowel . Thus, following this
interpretation, one finds in the spelling of this word, h, both the etasymbol h and the xi-symbol used for the graphic encoding of the -vowel.
Such a utilization of varying letters for representing the -vowel is of course
well attested among the epichoric Greek alphabets, as discussed in Chapter2,
and, as we saw there, entails the use of symbols with xi affiliations. The use of
diverse symbols for spelling the -vowel is grounded, fundamentally, in the
dual diachronic source of such vowels: one set of -vowels is descended from
inherited *; the other set of -vowels arises by a change of inherited * to .
This orthographic process has its origins, as we saw, in Ionic alphabets.
If we accept the parsing of this line of CP 22 to reveal the occurrence of
an arbitrary phonic string that is encoded by the graphemic sequence
(ml), can this phonic string be matched to an arbitrary conceptual element so as to yield an otherwise attested linguistic sign of the ancient Greek
language?Yes.
179
The feminine noun (ml) occurs frequently in the Ionic literary dialect of the Hippocratic corpus and other medical texts: the term is used to
denote a medical instrument. From its described uses, one envisions a long
pointed implement; its typical use is that of a probe. A can be used,
for example, to probe a head wound, to determine what sort of damage a
weapon has inflicted on a cranial bone (Hippocrates, VC 10.10), or to check
the depth of the cut made when trephining a cranial bone (VC 21.23, 34); it
is used to open the cervical orifice (Hippoc., Nat. mul. 37.9) and to probe a
fistula (Hippoc., Fist. 5); a tin (and thus pliable) (ml) with an eye can
be outfitted with a thread and passed into a fistula and back out of the anus,
the thread detached (and the removed) and the ends of the thread tied
together so as to ligate the wall of the fistula (Fist. 4). Hippocrates mentions
the use of a (ml) in the surgical treatment of cutaneous ulcers: for some
ulcers, an incision should be made against a (ml) (Ulc. 10); in some
instances, a lesion is to be lacerated and blood removed by applying pressure
with a (ml) (Ulc. 24). Physicians insert the instrument presumably
of a blunted variety (cf. the [amphimlon] also called a
[dipurnon] mentioned later) into the throat of a patient to induce vomiting (noted in a scholion on Aristophanes Knights 1150[Scholia in equites 1150b
(Jones and Wilson), and see later comments]). A notched can be used to
place a cord around a nasal polyp for removal (Hippoc., Morb. 2.35).
The term occurs frequently in Galens medical works. The sense is again
typically probe. Galen also writes of a probe called a (mltis)1 and of
a spatulate probe called a (spathoml).2 In addition he explicitly
reveals that a (ml) serves not only for exploring, leveraging, guiding,
and gagging but for lacerating as well. Thus, he describes a procedure involving the cutting of a membrane, in which, he writes, the surgeon should not
use a (ml) but should instead employ the (onux), an instrument
that attaches to a surgeons finger (Khn 8.55.6[De loc. aff.]).
In addition to (mltis) and (spathoml), several morphologically complex forms of (ml) are attested in Greek medical
texts. Galen (Khn 19.85.6) attributes to Hippocrates the term
(apurnoml) or (apuroml), that is, a (ml) lacking any
bulbous enlargement ( [purn] stone, knob) on the end (and seemingly
the same as a [mltis]). Both Galen (Khn 19.69.7) and the grammarian Erotian, who compiled a Hippocratic lexicon in the first century a d (Voc.
Hippoc. col. 51.5), report for Hippocrates the term (agkurouml),
denoting a (ml) having a hook ( [agkura]). There is a thin
180
form of the instrument, a (leptoml; see Dietz 47.11), and a wide

form, a (platuml), the latter attested in the work of the physician
Antyllus (second century a d ), cited by Oribasius (44.19.2.3; fourth century).
From the noun (ml) a denominative verb (mlo) is formed,
meaning to probe, as in Hippocrates, Morb. 1.6.11, where it is used of probing
the cranium to discover if bone is broken. Complex forms of the denominative
verb are attested with similar meaning. Thus, (katamlo) is used
of inserting a probe (Hesychius 1266), as for inducing gagging (Ar. Eq. 1150;
Suda 649). According to Photius (Lexicon 305), (katamlo)
was used proverbially of conjecturing about, puzzling out ( [tekmairomai]) (adla) secret things. Similarly, (promlo) is
used by Hippocrates of the act of probing a fistula prior to lancing it (Fist.5).
6.1.1
i ( sm i l )
and
( m l )
In Greek medical literature, a commonly occurring term denoting an implement used for cutting is (smil). Hippocrates knows the word, writing
of lacerating the nose with a (smil) in order to gain access to a nasal
polyp (Morb. 2.36.3); in the Hippocratic corpus, though, the form is not often
found. In the works of later medical authors such as, inter alia, Galen (especially), Oribasius, and Paul of Aegina (seventh century), the term occurs with
frequency in the same sense. (smil) remains the Modern Greek word
for scalpel.
The form can be found with the same denotation outside of the medical
corpus proper. Lucian, the second-century literatus, seems all too familiar with
procedures coupling the incising (smil) with the burning
(kautrion diapuron) red-hot cauterizing iron and how one must
be silent and bear up (Ap. 2.11), just as when ones ideals and literary expressions are being cut and burned, purportedly for ones own benefit. In one of
his Aesopic fables (1.98.13), Valerius Babrius (ca. second century a d ) tells of a
lion whose claws were removed (hupo smils) with a smil.
While within the medical corpus the simplex form (smil) is predominantly used to denote a cutting instrument, there is evidence that the
- (smil-) morph can carry a somewhat different semantic sense, at least in
a complex formation. Galen, among other medical authors, knows of a type
of probe that has a protuberance at either end:3 he calls such an instrument a
(dipurnon), from - (di-) two + (purn) stone, knob; compare the previously mentioned -- (a-puro-ml), a probe (-
181
[-ml]) without (- [a-]) any such enlargement. Alternatively, Galen notes,

this probe can be called an - (amphi-smilon; Khn 2.581.9[De anat.
admin.]);4 in contrast, in the fourth century, Oribasius (7.14.5.2, 4), citing the
second-century physician Antyllus (perhaps slightly later than Galen), knows
the probe as an - [amphi-mlon]: Oribasius thus uses the expected
morphology for a probe word (i.e., amphi-ml-on versus amphi-smil-on).
Elsewhere within the same work (Khn 2.574.810[De anat. admin.]), Galen
rehearses various tools that fall beneath the rubric of (elasmata): he
names as the members of this set several of the probes that we have already
encountered (spathoml; though here actually in a slightly different form, discussed later), (dipurnon), (mltis) and
incorporates within the list (between the first and second of these three items)
the tool called the (amphismil). In light of its occurrence within
this set and the earlier identification of (amphismilon) as a type
of probe, an (amphismil), must likewise be identified as a probing
instrument; hence, in these complex forms, the formant -- (-smil-) shares
with (ml) the denotation probe.5 We observed earlier that, conversely,
(ml) can, as (smil) typically does, denote an implement that
serves to cut. The conceptual range of each of the two terms, (ml) and
(smil), shares an area of overlapping semantics.
Perhaps even more interesting, the first item appearing in Galens set of
(elasmata) is (smil). One might interpret the intent of Galens
presentation to be a division of the enumerated (elasmata) into two
subsets: (1) the cutting implement (smil), followed by (2) the several
probes. It is possible, however, that here the simplex term (smil) is
being used to convey the sense probe as well, as in the complex forms already
noted. The evidence for this interpretation is provided by a passage from the
same work (Khn 2.711.1214[De anat. admin.]), in which Galen, discussing
examination of brain morphology, instructs the anatomist to probe one of
the cranial cavities with either a (dipurnon), a (smil), or
a (spathoml): whichever instrument is chosen, it is to be used
to stabilize the tissues from within while the overlying structure is dissected
from without. Clearly here the arbitrary phonic string (smil) is being
used to identify an instrument that is functionally equivalent to the probes
called (dipurnon) and (spathoml). The semantics of
(smil) and (ml) againblur.
The blurring may also be morphological. In the last passage cited (Khn
2.711.1214[De anat. admin.]), there is a concatenation of the terms
182
(dipurnon), (smil), and (spathoml). In the passage of the

preceding paragraph (Khn 2.574.810 [De anat. admin.]), as we saw, this
triad recurs, along with (mltis) and (amphismil), as
examples of (elasmata). However, as already noted parenthetically,
the commonly occurring name of the spatulate probe, - (spathoml), here (line 9)actually has a slightly different form: it appears as (spatho-mil), a variant that is attested in three other places within the
Galenic corpus.6 How are we to make sense of this? One could suggest that
this form (i.e., - [spatho-mil]) is a variant that arose secondary to
sound change: in other words, the sound of the Greek eta-vowel had begun
(though only begun, to judge by the inscriptional evidence), at least in some
contexts, to approach that of the iota-vowel by Galens second-century a d
date;7 the blurring at the edges of the cognitive components of the linguistic
signs (ml) and (smil) could have resulted in a remodeling of the
phonic component of the linguistic sign - (spatho-ml) to produce a free variant - (spatho-mil).8 Add to that two instances of
Galens use of a simplex (mil; Khn 11.300.16; 18.1.478.16). Alternatively,
one could propose that these variants reflect vestiges of some antecedent state
of morphology, a potentiality to which I shall return. Either way, such variability testifies to the morphological fluidity of the forms involved.
6.1.1.1 m i (smil) as s t y lu s a n d e n g r ave r . This word (smil)
has several other senses, each denoting, as with the terms principal medical
usage, a tool for incising. Aristophanes, for example, knows the as an
implement of writing. In his Women at the Thesmophoria (Thesmophoriazusae)
(776784), when his character identified as Euripides In-Law has been captured by Athenian women celebrating the Thesmophoria (the autumn festival
of Demeter and Persephone), the comic playwright parodies a scene from
Euripides lost tragedy, the Palamedes (as he parodies Euripides himself): as
Oeax, the brother of Palamedes one of the figures whom the Greeks credited with the invention of the alphabet had inscribed a message on oars,
relating Palamedes scandalous death at Troy, and thrown the oars into the
sea so they would float to Euboia and carry news of the death to their father
Nauplius (Schol. Ar. Thesm. 771[Dbner 269]), so the trapped In-Law will
write ( [graph]; Thesm. 771, 774)a call for help on wooden votive tablets and throw them here and there, as if they would float away and so bring
aid. As he tries to inscribe these plaques, the In-Law cries in a state of aporia
(Thesm. 776781):
,
.
, ,
,
. ,
.
kheiresemai,
egkheirein erg(i) khr porim(i).
age d, pinakn ksestn deltoi,
deksasthe smils holkous,
krukas emn mokhthn. oimoi,
touti to hr mokhthron.
Oh my hands,
Youve got to work to find a wayout.
Come on! tablets of smooth-polished plaques,
Accept the etchings of the stylus,
Be the messengers of my desperate straits! Blast it,
This rho sucks!
183
780
780
780
As a medium for inscribing his message, In-Law is (re-)using votive plaques

that are lying or hanging in the sanctuary of the Thesmophorium. Such plaques,
(pinakes [plural]), with or without an inscription, were commonly
installed in sanctuaries to commemorate vows made to a deity (see Schol. Ar.
Thesm. 773[Dbner 269]):9 as such, they are (agalmata; see Thesm.
773), sacred offerings honoring the divine.10 The term (pinaks [singular])
is also used to denote a writing tablet, such as that one in which Bellerophons
(lugra smata), so-called baneful signs, were written, to be
delivered to the king of Lycia (Hom., Il. VI 168169). Aristophanes usage in
line 778 captures both senses of (pinaks) votive plaque and writing
tablet.
For inscribing his plaques, In-Law presumably uses the altars sacrificial
knife, with which he has just sacrificed a wineskin wrapped in babys garb
(Thesm. 688758), a scene that parodies Telephuss abduction of Agamemnons
infant son Orestes in Euripides tragedy Telephus.11 In-Laws use of a sacrificial knife for a stylus denoted (smil) is clearly intended to add to
the comic effect of the scene. Sommerstein, like others before him, has called
attention to Aristophanes use of Euripidean language in these parodies of the
184
Telephus12 and Palamedes. He characterizes the latter as a Euripidean pastiche

with comic elements incorporated.13 Regarding (ksests), translated in
line 778 of the quoted passage as smooth-polished, Sommerstein notes that
the word is a favorite adjective with Euripides which he may apply to any
artifact of wood (IT 111, a cult-image; Phoen. 1179, the rungs of a ladder) or of
dressed stone.14 One might suspect that (smil), stylus, in the ensuing
line, denoting the implement with which In-Law inscribes the smooth-polished plaques, is preserved from the parodied scene in Euripides Palamedes,
in which Oeax inscribes the message of the treacherous execution of his
brother, the inventor of the alphabet: in other words, Aristophanes appropriation of the lexeme (smil) with the comically marked referent sacrificial knife continues Euripides linguistic choices. In any event, Aristophanes
must be punning on the customary senses of (smil).
The Suda entry for (smil; 742) similarly identifies this implement as a tool for inscribing. We learn from the same ancient lexicon that
the bits of debris dislodged by the scoring of the stylus are called
(smileumata; 741) or (aposmileumata; 3535). Hesychius
( 1255)likewise records the former term and its sense, as does a scholion on
Aristophanes Frogs (819).
In an epigram from the Greek Anthology (7.429), attributed to Alcaeus of
Mytilene, (smil) is used to denote the instrument used for etching letters (and smithing an enigma) into a gravestone:
,

.
;
.
,
;
.
,
, .
Dizdmai kata thumon, hotou kharin ha paroditis
dissaki phi mounon gramma lelogkhe petros
laotupois smilais kekolammenon. ara gunaiki
ta(i) khthoni keuthomena(i) Khilias n onoma?
touto gar aggellei koruphoumenos eis hen arithmos.
10
185
to men eis orthan atrapon ouk emolen,

ha d oiktron naiousa tod rion epleto Phidis?
nun Sphiggos griphous Oidipos ephrasaman.
ainetos houk dissoio kamn ainigma tupoio,
pheggos men ksunetois, aksunetois d erebos. 10
I wonder why this rock along the road

Was fated to have phi its only sign twice
Engraved with stonecutter styluses. Was thewoman
Who lies here hidden in the ground named Chilias?15
For summed up, the letter proclaims the number. 5
Or have I strayed from the straight path
Was she who dwells in this piteous tomb Phidis?16
Now I an Oedipus have pondered the riddles of the Sphinx.
Praises for the man who wrought an enigma from two symbols,
Light for the wise, for the foolish darkness. 10
An adjectival derivative is preserved in another epigram from the same
collection (Anth. Pal. 7.411; Dioscorides). Here Aeschylus is described, in
comparison to Thespis, as | ,
(ho m smileuta kharaksas | grammata, kheimarr(i) d hoia
katardomena) he who inscribed not letters etched ( [smileuta]), but
swept along, as it were, by a torrent.
In an epigram attributed to Julian, prefect of Egypt (Anth. Pal. 6.67), a
(smil) is presented as one component of a writing kit dedicated to Hermes
by its aged owner, Philodemus, along with lead ( [molibdos]), rule
( [kann]), porous sharpening stone ( [laas]), and ink (
[melan]) with reed pens ( [kalamoi])17 the last-named being qualified by the haunting description (mustria phns
andromes) mysterious devices of the human voice. As a part of this assemblage, the term (smil) almost certainly identifies an inscribing tool, to
judge from the fundamentally parallel structure of a second epigram attributed to Julian. The subject of this (second) epigram, Anth. Pal. 6.68, is likewise
the dedication of writing implements to Hermes by a geriatric Philodemus,
and these tools match item for item those named in Anth. Pal. 6.67. In the second epigram, Anth. Pal. 6.68, however, the poet, in lieu of naming the
(smil), substitutes the couplet:
,

186
Dekhnuso kai gluptra sidreon, h(i) thrasusArs

sun Mousais idin dke diaktorin
Receive as well the iron engraver, to which wildAres
And Muses both gave its own function.
The engraving tool lies within the purview of Ares and the Muses alike
recording instrument of the epitaph of the battle dead, and of inspired poetic
speech more generally. Both of these references are likely subsumed beneath
the identifying phrase with which the (smil) is named in our former
epigram by Julian, Anth. Pal. 6.67: (smils oksutomos
kopis) sharp-cutting blade of the (smil). (kopis), derived from
(kopt) to strike (with a weapon), is a word that commonly denotes a
blade used as a weapon: its use here to identify a writing implement hints at
an ambiguity that is teased apart in the latter epigram, Anth. Pal.6.68.
On the other hand, the form (smila), occurring in the epigrams of
Anth. Pal. 6.62 (attributed to Philippus of Thessalonica) and 6.295 (attributed
to Phanias), is used to denote a penknife that is used to sharpen reed pens.
This (smila) is called (donakn akrobeln
gluphida) knife for sharp-tipped reeds (6.62.2) and (donakogluphon) reed-cutting (6.295.2). The Suda entry for (smil; 742)recapitulates the distinction.
6.1.1.2 m i ( smil) as c u t t i n g t o o l . The phonic string (smil)
can be semantically linked with notions of other tools that etch, gouge, and
remove. These fall fundamentally into two groups: tools affiliated with metal
smithing and tools for leatherwork. A (smil) can be used to score a
metal bolt-pin, as attested twice by Aeneas Tacticus in his Poliorcetica (18.5
and 18.16; fourth century bc ): note that in the earliest surviving manuscript of
the work (Codex Laurentianus Graecus LV 4), parent of the remaining manuscripts, the form both times appears as (sml). In an anonymous epigram of the Greek Anthology (Anth. Plan. 15B), the poet writes of an enchained
Satyr, compelled to work at the forge of Hephaestus, wielding a (smil)
in his metalworking labors, as Hephaestus fashions a shield for Achilles.
More intriguing the Alexandrian poet Callimachus preserves the word
(smil) in his literary Ionic of Iambus 12 (line 66). According to the Milan
Diegesis, Callimachus presented this poem as a gift to the infant daughter of
a friend, one Leon: the present was offered at the celebration of the childs
amphidromia, an infancy ritual entailing presentation of the neonate and a
187
circumambulation of the household hearth.18 Callimachus likens his own gift

of this song (Iambus 12)on the occasion of the amphidromia of Leons daughter to a poetic gift presented by Apollo19 to Hebe, daughter of Zeus and Hera,
following her birth.20 Apollos gift of a song is contrasted with other divine
gifts made to Hebe, splendid objects presented by various deities, including
gifts wrought and offered by the artisans Athena and Hephaestus, gifts that do
not compare favorably with the poetic composition of Apollo. In contrast to
Apollos song which, again, the poet is equating to his own (lines 6567)
,
,

65
Tn Athnais de kai hetern dosin,

kaiper eu smil(i)sin kribmenn,
ho pros phoiten amaursei khronos
65
The gift of Athena and of others,

although exactingly made with chisels,
time, as it passes on, will diminish.
65
In line 66, (smil(i)sin) is commonly interpreted to denote chisels,

as I have it translated here. The occurrence of (smil(i)sin) in line
66 has been taken as support for the filling out of the lacunar [ as
(gluph(i)) by carving in line 27, as the poet sings of gifts brought to the infant
Hebe by Athena: [] [ | (polla tekhnenta
poik[il]a gl[ | paikhnia) many playthings skillfully and variously wrought [by
carving] gifts that pale in comparison with Apollos song.21
Though self-styled as simply a birthday gift for the infant child of a friend,
the language of the poem is decidedly metaphorically charged in the praise
of poetry. In his Hymn to Apollo, Callimachus refers to the god in his role as
singer (lines 4244), and then goes on to describe Apollos fabrication of the
horned altar at Delos; but in that description, in addition to the vocabulary of
construction, Callimachus makes use of the verbs (huphain) weave
and (plek) plait: In his portrayal of the young god building the altar
of bone at Delos, Callimachus chooses some metaphors that have a deliberately ambiguous quality, metaphors from earlier epic and choral poetry that
are also used of the composition of song.22 Compare Aetia fr. 26.5, where he
writes of (muthon huphainomenon) woven muthoi, that is
woven words (in at least an archaic sense, words that accomplish something
188
in their very production)23 The epic and choral uses of (huphain)

weave and (plek) as metaphors of poetic composition, particularly as
used in conjunction with (muthos), are matters to which we will soon
turn to consider in some detail. Notice that in the opening fragmentary lines
of Iambus 12, Callimachus writes (line 5) (nthousai) spinning, perhaps in invoking the Muses, with (nth) to spin employed metaphorically like (huphain) to weave of poetic production.24
Might it be possible that the products of the chiseling that the poet
Callimachus intends in the phrasing of line 66,
(smil[i]sin kribmenn) exactingly made with (smilai), are engraved
letters in keeping with many usages of (smil) for stylus that we have
already encountered? We will see that Callimachus uses (gluphanoi)
as the term for chisels that carve images of gods. Is Apollos superior gift,
his poetic composition and so Callimachuss own song and superior gift
being set apart from what is metaphorically conceptualized as letters etched
in metal, however precisely executed? One is reminded of Callimachuss professed loathing of epic; in Epigram 30, for example, he writes:
(ekhthair to poima to kuklikon) I despise the Cyclic
poem.25 In his retort to his literary detractors, whom he addresses as the
Telchines, invoking the fabled metalworkers of Greek tradition (on whom see
Section 6.2.7), critics who upbraid him for the slenderness of his compositions and the absence of epic themes, he writes (Aetia fr. 1.16): []
[] ([donides] hde melikhr[o]terai) little nightingales
are all the sweeter; and (lines 2124):

, []
. . ] ,
, ]
Kai gar hote prtiston emois epi delton ethka
gounasin, A[po]lln eipen ho moi Lukios:
. . ] aoide, to men thuos hotti pakhiston
threpsai, t] Mousan d gathe leptalen
For when first I set a tablet uponmy
knees, Lycian Apollo said this tome:
. . poet, the sacrificial victimyou
should fatten up, but, friend, make the Muse slim
189
One may wonder if in his Iambus 12 Callimachus contrastively invokes epic

by the metaphor of letters cut in cold metal with (smilai) productions
that will corrode over time unlike his own poetic offerings.
Plato (Resp. I 353a) knows the (smil) and the (makhaira)
dagger as blades that one could use for trimming a vine, but neither is the
optimal instrument (which would be a [drepanon]).26 He writes too
of a (skutotomos) leather cutter, shoemaker using a
(smil), among other cutting tools, for his work (Alc. 129c). Lucian attests
the same usage in his Cataplus (15): the shoemaker Micyllus, upon reaching
the Netherworld, tells Clotho that at the moment of his departure from the
world of the living he tossed aside the leather and blade, (smil), that he
was using to make a shoe. Later in the same work (Catapl. 20.17), Micyllus,
urged by Hermes to show some sorrow as Charon ferries him to Hades shore,
makes an exhibition of grieving, asking what will become of his (smil)
and his (kenttrion) awl. The word seems similarly used by
Herodas, when his shoemaker Cerdon of Mimiambus 7 makes reference to
a (smil; line 119). Note also the diminutive (smilion), which
Lucian places in the mouth of his cobbler Micyllus at Gallus 26; and compare
the spelling (smlion) in P Oxy.326.
6.1.2 the common origin of

and ( m l )
i ( sm i l )
I believe that we can identify the Greek terms (ml) and (smil)
as having a common origin from a well-known Proto-Indo-European root
meaning to carve, to work with a sharp tool, which I would reconstruct as
*(s)meh2-, with an extended root *(s)meh2y-.27 The Greek forms are then built
from this root using a frequently occurring suffix -lo- or -leh2-. Indo-European
roots that alternate between forms with an initial *s- and without this initial
*s-, the so-called s-mobile, are common, and at times reflexes of both variants
can be seen within single languages. A few examples of such a phenomenon
from the lexicon of ancient Greek are shown in the following chart:
(7)
roof
holm oak
bindweed
to rub with ointment
moon
190
The proposed Indo-European root *(s)meh2(y)- can be identified as the

source of several words attested in Germanic languages, such as the following
(Walde-Pokorny 2:686):28
(8) Gothic aiza-smia copper smith
Old Norse smir worker in wood and metal
Old High German smid; Old Frisian smith; Old English smi smith
Greek (smil) has been conventionally included in this group (as in
Chantraine 1968:10271028).
The same root, without s-mobile, is likely also the source of forms such as
the following (Walde-Pokorny 2:222):29
(9) Gothic maitan to cut, to hew
Old Norse meitill chisel
Old High German meizil chisel
Compare too the following:
(10)Old Irish mel hornless, bald, shorn
Old Church Slavic md ore and Russian md copper
One of the synchronic phonological processes of the Proto-Indo-European
language involves a metathesis (positional exchange) of a laryngeal consonant (i.e., *h1, *h2, *h3) and an immediately ensuing glide consonant (i.e., *y,
*w) when this sequence itself occurs between two consonants. In other words
(where H designates laryngeal; G glide; and C consonant):
(11) */CHGC/ */CGHC/
If, however, the phonemic sequence */HG/ is not snugged between two consonants, as in, for example, the concatenation */CeHGC/, then metathesis
does not occur.30 Such variant strings as vowel-less */CHGC/ versus voweled
*/CeHGC/ often (though not solely) arise as a result of the Indo-European
morphological process of ablaut (or apophony): the regular alternation of
morphemes between (1) the so-called e-grade (marked by the presence of an
e-vowel), as in the Greek present-tense form - (leip-) I leave; (2) the
o-grade (marked by the presence of an o-vowel), as in the Greek perfect-tense
form -- (le-loip-a) I have left; and (3) the -grade (i.e., zero-grade;
marked by the absence of either the e-vowel or the o-vowel), as in the Greek
aorist-tense form -- (e-lip-on) I left.31
191
The result of Proto-Indo-European laryngeal-glide metathesis is that

descendant Indo-European languages show reflexes of both metathesized
and nonmetathesized sequences. There is also some indication that at least in
Hittite the metathesis remained for a while a productive synchronic process.32
In either of the resulting phonemic structures (i.e., metathesized */CGHC/ or
nonmetathesized */CeHGC/), the glide remains positioned between two consonants and will thus be expressed phonetically as the corresponding vowel
(i.e., */y/ as *[i] and */w/ as *[u]).33
Let us consider a few examples drawn from Melchert, Rix, and Watkins:
(12) Proto-Indo-European *seh2w- to pour; to rain34
PIE */sh2i-sh2w-V-/ Hittite i-(u)wanzi they sprinkle (reduplicated present; without metathesis, due to the V that follows the HG [h2w] cluster)
PIE */sh2w-C-/ */swh2-C-/ Hittite uai (s)he pours (with metathesis to GH [wh2], due to C appearing on either side of the HG cluster)
(13) Proto-Indo-European *peh2wr fire35
PIE * peh2wr Hittite paur fire (without metathesis of HG [h2w], as
the sequence follows the root vowel)
PIE */ph2wr/ */pwh2r/36 Greek (pr); Umbrian pir fire (with
metathesis to GH [wh2] in the interconsonantal context)
By regular and well-established sound changes, Greek [ml] would
develop from a nonmetathesized form of the proposed Indo-European root,
having e-grade of the root, and [smil] would develop from a metathesized form of the root, having -grade; specifically:
(14)Nonmetathesized root:
*meh2-leh2- [ml]
Metathesized extended root with s-mobile:
*smh2y-leh2- *smyh2-leh2- *smih2-leh2- [smil]
An exact parallel is offered by developments of the Proto-Indo-European
root meaning to drink *peh3- / *peh3y- in Sanskrit, with comparable Greek
reflexes as shown:
(15)Nonmetathesized root:
*peh3-ti p-ti (s)he drinks
cf. Aiolic Greek - (p-n) I drink
Metathesized extended root:
*ph3y-to- *pyh3-to- *pih3-to- pta- drunk
cf. Attic Greek - (p-n) I drink
192
6.2
*
( l u z d )
(ml) then, I would argue, can be reasonably identified in line 16 of the

copper plaque MS 2-2 as a nominal denoting an implement for incising and
displacing specifically in this context, for engraving: in other words, a stylus.
Inflected in the nominative case, (ml) must occupy the role of subject
in the argument structure (ml se luzd abgd).
(luzd), I would propose, is the verb of the predicate of this argument. Given
the attested word roots of ancient Greek, the set of morphologically possible
forms that could provide a stem for a verb (luzd) is quite small surprisingly small, in fact. When semantic considerations are added, the set of
possibilities, I will claim, can be effectively narrowed toone.
The Proto-Indo-European root *leug- (e-grade) / *lug- (-grade) means to
bend. There are several evolutionary reflexes of the root, which tend to be
concentrated within the so-called centum37 wing of the Indo-European family; examples include the following:38
(16)Latin luctr to wrestle
Latin luxus dislocated, sprained
Gothic ga-lukan to close
Old Irish fo-long- to endure
Probablyalso
(17)Lithuanian lgnas flexible
6.2.1
( lu g o s )
and plaiting / weaving
Greek presents the nominal reflex (lugos), fundamentally denoting a

flexible branch or shoot that is, a withe. (lugos) occurs three times
in Homeric epic. In Book 11 of the Iliad, we learn that Achilles had once captured Priams sons Isus and Antiphus as they had tended their sheep on the
shoulders of Mount Ida, and that he had bound the brothers
(moskhoisi lugoisi) with pliant withes (Il. XI 105). In Book 10 of the Odyssey,
Odysseus relates how he had made a binding with which to tie the legs of a
stag he had killed on Circes island (Od. x 166168):
,
, ,

193
Autar eg spasamn rhpas te lugouste,

peisma d, hoson t orguian, eustrephes amphoterthen
pleksamenos sunedsa podas deinoio pelrou
Then I broke off some brush and withes
and a cord, about six feet long, plaited well from one end to theother
I wove, and I bound the hooves of that giant of a beast.
The third Homeric example, and perhaps the most interesting, comes from
the ninth book of the Odyssey. Odysseus and his men are trapped within the
cave of the Cyclops Polyphemus; he schemes how he and his surviving companions might be able to elude the now-blinded giant and exit from his cave
(Od. ix 420429):
, , 420

.
.
, , 425
,
,
, ,

Autar eg bouleuon, hops okh arista genoito, 420
ei tin hetairoisin thanatou lusin d emoi aut(i)
heuroimn; pantas de dolous kai mtin huphainon
hs te peri psukhs; mega gar kakon egguthenen.
hde de moi kata thumon arist phaineto boul.
arsenes oies san eutrephees, dasumalloi, 425
kaloi te megaloi te, iodnephes eiros ekhontes;
tous aken suneergon eustrepheessi lugoisin,
t(i)s epi Kuklps heude pelr, athemistia eids,
suntreis ainumenos.
Then I pondered how things might turn out for the very best, 420
if I could figure out some dodge from death for my comrades-in-arms
and for myself; and all kinds of deceptions and cunning Iwove,
as life hung in the balance; for a great evil was athand.
And in my heart this plan appeared to be thebest.
194
There his well-fed rams were, thickly fleeced, 425

and beautiful as big, with wool dark like violets;
these I quietly bound together with well-plaited withes,
the bed on which the Cyclops slept monster, knowing only wrong
taking the rams by threes.
What do we learn of (lugos) from Homeric epic? The term denotes

a plant limb that is flexible, even soft enough for bedding, but strong when
its strands are interwoven. Formulaically it appears in conjunction with the
adjective (eustrephs) well-plaited: an adjective that Homer also
employs to describe (1) sheeps wool plaited to make a sling ( [sphendon], Il. XIII 599; cf. XIII 716), and which can secondarily serve for bandaging; (2) a bowstring ( [neur], Il. XV 463); (3) a length of ships line
( [hoplon], Od. xiv 346);39 and (4) a lyre string of sheeps gut (
[enteron oios], Od. xxi408).
In both of the passages in which (eustrephs) well-plaited and
(lugos) withe are conceptually and formulaically linked, the phrasing
elicits the semantics of weaving. In my translation of Odyssey x 167168, I
render (plek) as weave (rather than plait): this is not solely for stylistic reasons; the choice of wording follows from and is legitimated by the
semantic relationship between (plek) and (huphain). Nagy
notes that weaving is a specialized form of interlacing or plaiting. The Greek
equivalent of plait is plek. Whereas plaiting is the process of joining basically
two sets of whatever element, such as thread, in an over-under-over-under
pattern, weaving superimposes a frame on this process.40 Nagy continues,
observing, with regard to Greek vocabulary of weaving (the bracketed phrase
is myown):
We may note some important semantic convergences and divergences in the
Greek. First of all, huphain weave is a specialized kind of plek plait, but
there are contexts where plek can be used as a synonym of huphain. Also, the
process of uniting, by way of weaving, the warp with the weft is described as
sumplok [a nominal derived from the verb plek] in Plato Politicus 281a.41
Examples of the synonymous use of (plek) and (huphain)

are noted in the 1996 study of Scheid and Svenbro. Citing instances of the
interchangeability of the verbs from Plato42 and Theocritus, they observe:
Plekein is used instead of huphainein when one wishes to emphasize the aspect
of interweaving.43 A further cited example is provided by a comparative look
195
at lines from the lyric poets Pindar and Bacchylides: in his sixth Olympian
Ode (lines 8687), Pindar writes of weaving [plekn] a hymn with rich
designs [poikilon humnon] for fighting men;44 compare with this phrasing,
words penned by Bacchylides in his fifth Victory Ode (lines 814) about
which Scheid and Svenbro observe:45
From his island of Keos, the poet asks King Hieron of Syracuse to forget his
troubles: Turn your thoughts this way: with the help of the deep-girded Graces
your guest-friend has woven a song of praise [huphanas humnon] and sends it
from the sacred island to your distinguished city. Whether this is an etymological figure or not,46 Bacchylides considers that he wove his hymn.47
In the work of these two choral poets, the (plek humnon) of

one equates to the (huphain humnon) of the other to weave
a hymn.
In our second passage from the Odyssey in which (eustrephs)
well-plaited and (lugos) withe co-occur, the conceptual linkage with
weaving is yet more overt. Odysseus ponders how best to save his comrades
and himself from the fate of death. His desperate deliberating is depicted
by the poet as the weaving ( [huphain)]) of all kinds of deceptions
( [dolous]) and cunning ( [mtin]) (Od. ix 422). The scheme that
eventuates from this cognitive weaving finds expression in the physical plaiting of (lugoi): withes are woven to make tethers that will bind together
three rams, beneath which Odysseus and his men will attempt to make an
escape. The process echoes and reminds the epic audience of Penelopes iterative weaving of a robe, by which means she will attempt to escape the suitors. For Odysseus to weave [dolos]) and [mtis] in the cave of the
Cyclops is to plait (lugoi), as for Penelope to weave ( [huphain)])
with deception ( [dol(i)]) (Od. ii 106; see also ii 93) and to threadwind ( [tolupeu]) deceptions ( [dolous]) (Od. xix 137) in
her halls, glorying in her cunning ( [mtis]) (Od. xix 326), is to weave
a great web ( [huphainein megan histon)]) (Od.
xix 139; see also ii 94, 104)on her loom.48
6.2.2 ( lu g o s ) as a synonym for a

( ag n o s )
The nominal (lugos) appears in one of the fragments of the archaic lyric
poet Anacreon, an Ionian from Teos who took up residence on Samos at the
196
court of Polycrates relocating to Pisistratid Athens following Polycrates

death. In the two lines of fragment 7, the poet sings of his acquaintance
Megistes:
<>
.
<ho> Megists d ho philophrn deka d mnes epeite
stephanoutai te lugi kai truga pinei melidea.
The kind-spirited Megistes for ten monthsnow
garlands himself with withes and drinks honey-sweetmust.
Since antiquity, the intent of the poet and the significance of Megistes garlanding himself with withes has occasioned a good deal of speculation.
For example, in his Deipnosophistae, Athenaeus (second/third century a d )
includes a lengthy treatment of Anacreons lines and of (lugoi; see especially 15.671d674b), and of garlands generally. In part, these and related discussions, especially among modern commentators,49 have been consequent
to the Greek use of the lexeme (lugos) as a synonym for (agnos), a
term that denotes a willow-like tree, Vitex Agnus-castus, the so-called chastetree. Regarded as an antaphrodisiac, parts of the tree are reported to have covered the bedding of Athenian women during the Thesmophoria (Pliny HN
24.59; Dioscorides 1.103; Galen [Khn 11.808]); Aelian 9.26; Eustathius Od. i
356; Schol. Nic. Ther. 71a71b).
A mention of (agnos) occurs in the Homeric Hymn to Hermes, line
410.50 We will return to this passage below, but for the moment perhaps
it will be sufficient to note that it is here used of (kartera
desma) that is, of strong bonds made of this plant material. In the text of
the hymn there is no mention made of (lugos) and no otherwise overt
association of (agnos) with (lugos). Some investigators have,
however, noted this passage in conjunction with lines 1314 of the Homeric
Hymn to Dionysus:51 pirates have captured young Dionysus, taken him on
board ship, and bound him (desmois argaleoisi) with
painful bonds (7.12):
,

Ton d ouk iskhane desma, lugoi d apo tlos epipton
kheirn depodn.
197
But the bonds did not hold him, and the withes fell awayfrom
his hands and hisfeet.
More than this, some textual critics have conjectured that the text of the Hymn
to Hermes should be emended by insertion of a line containing (lugos),
following 409, one equivalent to line 13 of the Hymn to Dionysus.52
The earliest-attested explicit references to the synonymy of (lugos)
and (agnos) appear to date to the first century a d . Of agnus castus,
Pliny HN 24.59 writes: Graeci lygon vocant, alias agnon (the Greeks call it
lugos, or else agnos). From the same century, and in Greek language, are those
references found in two medical texts: Pedanius Dioscorides Materia medica (1.103), a treatise on pharmaceuticals; and Erotians Hippocratic lexicon,
Vocum Hippocraticarum collectio (57.78). In the former, Dioscorides equates
the two terms, beginning the relevant entry by identifying the plant material
as (agnos lugos) agnos or lugos. In Erotians lexicon weread:
. .
agnou : lugou. esti de onoma phutou. hs kai Nikandros en Thriakoislegn
Agnos : lugos. It is the name of a plant. As Nicander also says in his Theriaca
Both within the extant Hippocratic corpus and in Nicanders (third/second-century bc ) didactic poem, the Theriaca, a treatment in hexameters of
poisonous pests and remedies for their venoms, one does indeed encounter the terms (lugos) and (agnos). In Hippocrates De mulierum
affectibus, (lugos) occurs at 78.116 and (agnos) two lines later
(78.118): each material is specified as one ingredient in two different concoctions, among several such concoctions, said to be effective in expelling afterbirth. Nicander mentions the former term at line 63 and the latter at line 71:
both occur as members of a larger set of botanicals that one can place by
ones bedding when sleeping outdoors to drive away snakes (lines 5779); for
example, in the lines that separate the references to (lugos) and
(agnos), Nicander also makes mention of hulwort ( [polion]), vipers
bugloss ( [ekhieion]), marjoram ( [origanon]), wormwood
( [habrotonon]), tufted thyme ( [herpullos]), and fleabane ( [konuzda]).53
At least two observations regarding these references readily present themselves. In each instance, (lugos) is used to denote a specific plant material, rather than being linked to a (more) generic referent withe, as appears to
be its usage in Homer. Second, while (lugos) and (agnos) co-occur
198
within the space of a few lines in these works of Hippocrates and Nicander,
it seems that the two terms are being used to name different plant materials.
These two observations then lead immediately to a third, which, no doubt,
has not escaped the readers attention: namely, whenever and however
(lugos) came to be used to denote a particular botanical item and whenever
and however it entered into a relationship of synonymy with (agnos),
these developments are secondary: (lugos) is one member of that set of
linguistic signs affiliated with conceptual notions of bending, plaiting well
attested among the Centum component of Indo-European languages, and
perhaps within Baltic as well, descended from a common ancestral IndoEuropean root meaning to bend.
6.2.3 on samos: part1

Before leaving this general discussion of the usages of Greek (lugos),
there is one additional matter to which we should pay some attention. There
are particular etiological associations of (lugos) with Hera on the Ionic
island of Samos. Citing the Samian historian Menodotus (FGrH 541 F 1), who
lived in perhaps the third century bc , Athenaeus (15.672a673b) tells the following etiological tale. Admete, daughter of the Argive king Eurystheus (for
whom Heracles performed the Twelve Labors), fled to Samos where, thankful
for her deliverance, she devoted herself to caring for Heras temple. But the
people of Argos hired some Tyrrhenian mercenaries ( [l(i)strikoi])
to steal the goddesss image, (bretas), from her temple, believing that
this would turn the Samians against Admete. Greek (bretas) is a term
typically denoting a wooden image of a deity,54 and perhaps just a plank in this
instance, as we shall see. The Tyrrhenians raided the temple and carried the
image to their ship but found that, inexplicably, their ship would not move.
Believing that the ship was being stayed by some divine power, they offloaded
the stolen image onto the shore and, after depositing sacrificial cakes beside
the image, put out to sea withoutit.
At dawn, when the theft of the (bretas) from the temple was discovered, a search for it began. When the non-Greek inhabitants of the
island, Carians, discovered it on the shore, in their barbaric imaginings,
they thought that the image had walked away from the temple; they placed
it on some platform ( [thrakion]) made of (lugos) and pulled
the longest of the branches ( [kladoi]) around the image to secure it.
Admete found the image, untied the withes, and returned it to its proper
199
place. From that time on, the image is said to have been carried annually to
the shore where it was purified and cakes were offered to it: the festival was
given the name Tonaia, from (tonos) band, cord.55 The Carians consulted an oracle regarding what their response should be to these events and
were told by Apollo that they, from that time on, should make their wreaths
of (lugos), binding their own heads with the same branches with which
they had tied the goddess.
But this is not all. Pausanias (7.4.47), in his description of the Ionic island
of Samos, writes that the Samians believe that Hera was born there, by the
Imbrasos River, about which Frazer observes in his commentary on Pausanias:
The stream flows about 400 paces to the east of the temple of Hera. Its banks
are fringed with oleanders and agnus castus.56 Similarly, Kyrieleis has more
recently described the latter vegetation: (Vitex agnus castus), a tree or bush
common to the Imbrasos plain.57 Pausanias continues, further specifying the
locale of the goddesss birth (7.4.45,7):
.

.
,
.
, ,
,
.
Kai hupo t(i) lug(i) t(i) en t(i) Hrai(i) kat eme eti pephukuia(i). einai d oun
to hieron touto en tois malista arkhaion oukh hkista an tis kai epi t(i) agalmati
tekmairoito; esti gar d andros ergon Aigintou Smilidos tou Eukleidou. houtos
ho Smilis estin hlikian kata Daidalon, dokss de ouk es to ison aphiketo;
.
ho de Smilis, hoti m para Samious kai es tn leian, par allous ge oudenas
phaneros estin apodmsas; es toutous de aphiketo, kai to agalma en Sam(i) ts
Hras ho poisas estin houtos.
And [she was born] under the lugos which in my own day still grew in the
Heraion. This sanctuary is extremely old, as one might judge not least from
the agalma; for it [the agalma] is the work of an Aiginetan man, Smilis the son
of Eucleides. This Smilis belongs to the time of Daedalus, but did not acquire
a reputation equal tohis.
.
200
And this Smilis it is not clear that he went abroad to any place other than
among the Samians and the Eleans. But to these he did come, and he is that
one who crafted the agalma of Hera in Samos.
6.2.4
( lu g o s )
and the oldesttrees
A couple of additional points should be made. This tree under which the
people of Samos say that Hera was born, a (lugos), is elsewhere said
by Pausanias to be the oldest of all trees. In describing the environs of the
Arkadian town of Kaphyai, he writes (8.23.4) that a little way beyond the town
there is a spring, and by this spring there grows a large and lovely
(platanos) plane tree. The locals name this tree, and the associated spring,
Menelas, as they claim the tree to have been planted there by the Spartan
king Menelaus at the time he was gathering an army to lead to Troy. Pausanias
continues (8.23.5):

,
,

.
Ei de Hellnn tois logois hepomenon katarithmsasthai dei me hoposa dendra
sa eti kai tethlota leipetai, presbutaton men h lugos estin autn h en t(i)
Samin pephukuia hier(i) Hras, meta de autn h en Ddn(i) drus kai elaia
te h en akropolei kai h para Dliois; trita de heneka arkhaiottos nemoien an
t(i) daphn(i) t(i) para sphisin hoi Suroi; tn de alln h platanos estin haut
palaiotaton.
And if I follow the account ( [logoi]) of the Greeks in computing how
many such trees remain alive and flourishing the oldest of these is the
(lugos) which grows in the Samian sanctuary of Hera; and after it, the oak
( [drus]) in Dodona, the olive tree ( [elaia]) on the Acropolis and
that one at Delos. Third in antiquity the Syrians would consider to be the laurel
( [daphn]) found in their own homeland. Of the remaining, this plane
tree is the most ancient.
201
This is not the only list of most ancient living trees to survive from classical antiquity. Thus, Theophrastus,58 identifying a partially overlapping set,59
writes (Hist. pl. 4.13.2):

.
Tn de makrobiotta marturousin epi ge tinn kai hmern kai agrin kai hai
paradedomenai phmai para tn muthologn.
And the longevity of some trees, both cultivated and wild, is evidenced by the
reports passed along from the muthologoi ().
The trees of Pausaniass list, like those of Theophrastuss, are trees rooted in
mythic past in sacred time and in sacred space. Their stories are transmitted as a part of that body of tradition that Theophrastus assigns to
muthologoi () a term in some sense both redundant and oxymoronic only clumsily translated into English as (the near transliteration)
mythologists. Pausaniass third-most-ancient tree, the laurel ( [daphn])
of the Syrians, is that of the morphed nymph Daphne, changed into the laurel as she fled from Apollo, reportedly seen by Apollonius of Tyana at the
temple of Apollo at Daphne ( [to hieron
tou Daphnaiou Apollnos), near Antioch, (Philostratus, The Life of Apollonius
of Tyana 1.16). The olive tree ( [elaia]) on the Acropolis sprang up at
the dawn of Athenian history as Athena trumped Poseidon in the contest for
divine supremacy of the city. The oak of Dodona mentioned by Pausanias
is the oracular tree of Epirus (northwestern Greece) held to be the most
ancient oracle ( [mantion]) in Greece: to this oak we shall return.

To return to Samos it was previously reported that the stump of Heras
(lugos) had been identified by H. Walter and his German team (19631964) in
an area in front of the Samian Heraion, where earlier, in 19321933, Buschor
had excavated, though not completely:
Near his Altar III and half under water, the stump of a tree was found which,
since it lies on the axis of the earliest temple, is to be regarded as a relic of
Heras sacred Chaste Tree. This area was particularly rich in Geometric and
202
Archaic terracottas, while the presence of Mycenaean sherds carries the cult
back into the second millennium.60
This is now known, however, not to be the anticipated agnus castus. In

Archaeology in Greece 198889, Catling reports:
The tree stump found by H. Walter in 1963 (AR 196364, 24)thought to be
Heras agnus castus, was exposed and re-examined, where it was found not, in
fact, to be an agnus castus but a juniperus. A 14C determination placed the tree
between 750450 b. c ., while a dendrochronological test showed it had had a
life-span of no more than 80years. These results exclude the possibility that
the stump could belong to the tree referred to by Pausanias (VII 4.4). It also
suggests that the altar was sited without reference to the sacred tree.61
This tree lying on the axis of the earliest temple,62 as it does thus belongs
perhaps to a Greece approximately as early as that of Homer and so chronologically abutting the earliest example of alphabetic writing found at Samos (ca. 650
bc ) and perhaps dating to the very beginnings of the temple itself.63 At the other
end of the dating span, the tree would be no later than the Greece of Pindar. An
equation with Pausaniass (lugos), his most ancient tree in Greece among
those still living in his own day, is excluded by some hundreds of years.
Yet, one wonders if this of necessity means that the juniperus and altar
could not have been sited in some mutually respective fashion.64 We noted
that the conceptual linkage of the lexeme (lugos) with the notion agnus
castus is clearly secondary (certainly vis--vis its Indo-European etymon,
but likely within its documented Greek history as well) and that the earliestattested explicitly synonymous uses of (lugos) and (agnos) appear
to occur only in the first century a d . In any event, whatever the source of the
affiliation of the Greek phonic string [lugos] with the Samian Heraion, an
affiliation giving rise to or otherwise bound up with etiological accounts, the
sacred (lugos) of the Hera of Samos whatever botanical species was
earliest construed with that phonetic string in Samian tradition beneath
which the goddess was declared to have been born, is, like the oracular oak of
Dodona, the olive of Athens, and the laurel of Antioch, a tree whose origins
must be situated in the time and space of (muthos).
6.2.6
( sa n i s )
and
( sm i l i s )
There was not always a sculpted image of the goddess, it is said, in this reportedly deeply ancient sacred space of Samos. Citing the Samian historian Aethlius
203
(fifth/fourth century bc ), Clement of Alexandria (Protr. 4.40 [= Aethlius

fr. 1, FHG 4:287]), the learned Christian polemicist of the second and third
centuries a d , writes that the (agalma) of Hera in her sanctuary was
at first ( [proteron]) a (sanis) much as the (agalma)
of Artemis at Ikaros was of unworked wood and that of Cithaironian Hera
in Thespiai was a wooden pillar a tree trunk ( [premnon]) that had
been felled ( [ekkekommenon]). The lexeme (sanis), with
which Clement identifies Samian Heras earliest (agalma), fundamentally denotes a board or plank: Homer applies the word to doors (e.g.,
Od. ii 344)and to flooring (e.g., Od. xxi 51). In the plural, it can be used to
denote painted panels, pictures65 and commonly writing tablets, as in
Euripides Alcestis, where one reads (lines 966969)that:

,

oude ti pharmakon
Thr(i)ssais en sanisin,tas
Orpheia kategrapsen
grus
and theres no remedy
in Thracian tablets,that
the voice of Orpheus
wrote out66
Callimachus, that Alexandrian Greek who played a role in our discussion
of (smil), had prior to Clement recorded the same tradition. In Book
4 of his Aetia (fr. 100), the poet addresses Heras image at Samos and writes
of the ancient custom ( [tethmos dnaios]) in accordance with
which (gluphann aksoos stha sanis) you were
a (sanis) not carved by chisels ( [gluphanoi]).
Back to St. Clement who further specifies (following his Samian historian) that it was later ( [husteron]), in the time when Procles ruled
Samos, that the Heraions (agalma) of a (sanis) was replaced
by one that was (andriantoeides) having the form of a statue
(i.e., human-shaped). A scholion on Callimachuss aetion (Diegesis in Aetia)
notes in agreement that the wooden image ( [ksoanon]) of the goddess
204
acquired a human likeness when Procles was king. This is presumably the
Procles who is said to have led the Bronze Age Ionic settlement of Samos (see
Pausanias 7.4.1), another event lying in the time of (muthos).
A few lines further along, Clement, citing another Samian historian,
Olympichus (before second century B C ; fr. 1, FHG 4:466),adds:

to de en Sam(i) ts Hras ksoanon smil(i) t(i) Smilidos tou Eukleidou pepoisthai Olumpikhos en Samiakois historei
and the wooden image ( [ksoanon]) of Hera in Samos was made with a smil,
that of Smilis, the son of Eucleides, says Olympichus in his Samian History.
This, in any event, is the text as one commonly finds it cited: the particular
form rehearsed here comes from Farnell 1907. It is an emended text, based
on an emendation proposed by C. G. Cobet in Mnemosyne 10 (1861). The
manuscript tradition shows confusion: what one finds in the manuscripts,
rather than the sculptors name, Smilis, is a set of variants preserving forms
of (smil), that word denoting an engraving instrument that occupied
our attention earlier in this chapter.67 The editorial motivation to emend the
text by incorporation of the name Smilis is a transparent one: the sculptor of
the Samian image is named by Pausanias as Smilis the son of Eucleides; and
Clements lines that precede and follow this one analogously name specific
artists and the local divine images that they carved.
Thus, if one accepts the textual emendation, Clement echoes Pausaniass
report of the local tradition that the (agalma) of Hera in her temple
at Samos was crafted by Smilis, son of Eucleides. The poet Callimachus, however, knows a different tradition: he describes the Samian image of Hera that
came subsequent to the ancient (sanis) as the
(Skelmion ergon euksoon) well-carved work of Skelmis. This disparity was
not overlooked in antiquity: a scholiast on Pausanias 7.4.4 notes that while
Pausanias identifies the sculptor of Heras Samian (agalma) as the work
of Smilis, son of Eucleides, Callimachus says Skelmis instead of Smilis.
6.2.7
( sm i l i s )
and
i ( sm i l )
What of the sculptor Smilis, who according to the tradition preserved by

Pausanias had produced the (agalma) image of Hera in her
Samian sanctuary in the time of Daedalus?
205
Some investigators have tried to identify Smilis as a historical figure, such

as Frazer, following earlier work: The date of the sculptor Smilis is uncertain. That he was an historical person there is no real reason to doubt, though
Pausanias represents him as a contemporary of the fabulous Daedalus.68
Frazer goes on to summarize the views of Brunn, Overbeck, and Furtwngler,
offering dates of circa the second quarter of the sixth century bc , or earlier,
and entertaining the prospect that he was not Aiginetan at all, but a native
Samian, since there was enmity between Samos and Aigina.
The rescue attempts, which seem contorted and artificial, are little different
in essence from the now discredited claim that the tree stump found at Heras
altar was that of the mythic (lugos) of Samos: efforts to historicize the
ahistorical, to retemporalize and relocate elements that belong to the time and
space of (muthos). The tradition that Pausanias preserves is quite clearly
one that situates the itinerant craftsman, dubbed Smilis, in mythic time in
the era of Daedalus, legendary artisan par excellence.69 That name Smilis
is morphologically too highly suggestive to be ignored: the text of Clement
itself attesting variant forms of (smil), edited by supplying
(Smilis) whatever its original wording, should tip us off and lead us to
consider the relationship of the craftsmans name to the name of the crafting
implement (Smilis) to (smil). Bechtel, in fact, had already in
the early twentieth century associated the name with the instrument, listing it
in his collection of Greek names beneath the rubric Handwerkszeug proper
names built on words for hand tools.70 Without etymological discussion, he
presents (Smilis) alongside the name (Smiln), known from an
inscription from Thasos.71
A plausible etymology for the name of the sculptor (Smilis) readily
presents itself. Sanskrit attests root noun compounds72 (i.e., compounds having a root noun as their second member) such as vtrahn-, properly, smiting
the resistance,73 hence slaying the enemy; the form is commonly used in the
Vedas as an epithet, chiefly of Indra, Vtrahan, to identify the heroic slayer
of the dragon that bears the name Vtra. Other Sanskrit examples include
havird- eating the oblation, avavd- knowing horses, vedavd- knowing
the Vedas; Burrow, from whom several of these examples are drawn, points
out that parallel compound constructions are found in other Indo-European
languages, as in Greek (bouplks) ox goad (striking the ox); Latin
fidicen lyre player (playing the lyre), artifex artisan (making art).74 The list
could be lengthened.75
The name assigned to the sculptor of Heras (agalma) in the Samian
tradition that Pausanias preserves the artisan (Smilis) who also,
206
writes Pausanias, carved the crude images of the Horae in Heras temple at
Olympia is a nominal, I propose, that parallels in morphology the Sanskrit
compounds such as avavd- (knowing horses) and vedavd- (knowing the
Vedas). The initial member of the compound is built from a form of that
stem that we have encountered as the feminine noun (smil), denoting an etching and carving tool, confounded in the manuscript tradition of
Clement of Alexandria with the name of the craftsman (Smilis). The
final member of the compound is the Greek cognate of the Sanskrit root
noun vid- (as in avavd- and vedavd-), that is - (wid-), later - (id-)
with the loss of the glide /w/ (on which, see Chapter2). Both Sanskrit vidand Greek - (wid-) are -grade descendants of the Proto-Indo-European
root *weid- to see. In Greek, the -grade of the root also appears, perhaps
more familiarly, in forms such as Homers first-person plural finite verb
(idmen) we know (Sanskrit vidm, Gothic witum); the o-grade can be seen
in the first-person singular of the same verb, (oida) I know (Sanskrit
vda, Gothic wait).76
Behind the proper name - (Smil-is) having an oblique stem -(Smil-id- [as in the genitive -- (Smil-id-os)]), there lies a compound
nominal meaning knowing the (smil), an adjective referencing one
skilled in the use of the instrument. Given the productive Greek pattern, the
compound would likely be built with a thematic-stem form of the athematic
feminine noun (smil): that is, - (smilo-):77 compare, for example,
- (hulo-tomos) cutter of wood from (hul) wood; -
(diko-logos) an advocate (i.e., one who speaks for right) from (dik)
right.78 The resulting compound *-- (*Smilo-wid-) would become,
with loss of /w/, *-- (*Smilo-id-), which then by elision of the thematic
vowel of the first member of the compound would produce the attested stem
-- (Smil-id-), nominative singular - (Smil-is).79
Regarding elision in the formation of compounds, one reads that when
the second part [of the compound] once began with or , the final vowel of
the first part is not elided but remains or is contracted with the following.80
Dialectal exceptions to this generalization of lack of elision are, however, well
attested. Thus, *- (*dmio-wergos), skilled worker, craftsman (literally one who works for the people) gives, with loss of /w/, Homeric (dmio-ergos), the term that Homer uses for a wandering skilled craftsman, of which he names five types: prophets, healers, woodworkers, bards
(Od. xvii 383385), and heralds (Od. xix 135).81 While the Homeric form does
not show elision (, [dmio-ergos]), comparable Doric dialect
207
forms from Kyrene (the colony of Thera) and the eastern Aegean island sites
of Astypalaea (on Cos) and Nisyros (just south of Cos) do show elision of the
final (thematic) vowel of the first member, - (dmi-ergos).
This word is also built using the o-grade root of the second member of
the compound: in other words, *- (*dmio-wrgos). Doric,
Northwest Greek, and Arkadian all show reflexes that have undergone elision,
- (dmi-orgos). The Ionic form of this o-grade variant follows suit,
showing o, [dmi-orgos], as in the Central Ionic dialect of the island
of Amorgos and in the East Ionic of the island of Samos itself.82
That in the East Ionic dialect of Samos *-- (*smilo-wid-) had likewise become -- (smil-id-), nominative singular - (smil-is), via a
process of /w/-loss followed by elision seems most likely. One could, however, also advocate a process of analogy for the formation. Buck notes that
after the analogy of names containing inherited -stems arose also forms
like [Arkhi-lokhos], [Arkhi-damos], etc. in various
dialects,83 including those of Rhodes, Cos, Nisyros Samoss Doric neighbors and also Elean, the Northwest Greek dialect of that other place to
which Pausanias reports Smilis as having traveled. Notice that there exists a
compound adjective - (smili-gluphoi) meaning chiseling and used
to modify (tekhnai) arts, skills; about this form, Chantraine remarks:
la finale -- du premier terme ne peut gure tre interprte comme l de la
loi de Caland84 comme semble faire Schwyzer, Gr. Gr. 1,448 ; il est analogique
ou issu de (smilion),85 the latter being the diminutive denoting scalpel, shoemakers knife, pen knife that we encountered previously.
Concerning the proposed form *-- (*Smilo-wid-), a morphologically and semantically similar, if not identical, compound adjective is well
attested in archaic Greek: - (n-is [from *n-wid-s]; nominative), - (n-id-os; genitive) developing from an earlier stem *- typically
glossed as lacking knowledge, lacking skill that is, in effect, knowing
nothing in some area of expertise.86
Thus, at Iliad VII 198 Ajax is speaking, preparing to duel with Hector,
declaring that no warrior will make him flee by force or by skill (i.e., knowhow: [idrei], earlier, *- [wid-rei]):

epei oud eme nda g houts
elpomai en Salamini genesthai te traphemente
208
since I am pretty sure that I wasnot

born and raised knowing nothing in Salamis
As Kirk paraphrases, he is no ignoramus in the arts of war.87
In Odyssey viii, Odysseus, having declined a request to take part in the
Phaeacian games, is taunted by Euryalus, who likens Odysseus to a greedy
merchant rather than an athlete. Angered, Odysseus returns insult for insult
and adds (line 179): (eg d ou ns aethln) I myself do
not know nothing of athletic competition. Like Ajax, Odysseus possesses the
requisite know-how he is skilled in the arts of the contest.
In the Homeric Hymn to Hermes, the young Hermes plays a song with his
new musical creation, the lyre, and then, as he presents the instrument to
Apollo, tells him (lines 482488):
.
,
,
,

,
.
485
. hos tis anautn

tekhn(i) kai sophi(i) dedamenos eksereein(i),
phtheggomen pantoia no(i) kharienta didaskei,
rheia sunthei(i)sin athuromen malak(i)sin,
ergasin pheugousa dupathon; hos de kenautn
ns en to prton epizdaphels ereein(i)
maps auts ken epeita metora te thrulizdoi.
485
. to whomever asks of thelyre

with artistry and skill those talents honed her voice replies
and gives all kinds of things that please themind,
lithely singing with mellow intimacy,
485
fleeing painstaking labor; but whoever asksher
in a fury knowing nothing from the start-does so in vain, and wavering she breaks the notes.
And in the Homeric Hymn to Demeter, the term is used again to name those
who are without knowing. The disguised goddess of grain has become the
209
nurse of the child Demophon, son of the king and queen of Eleusis. Each night
she buries the child deep within live coals in a transformative process that will
make him immortal; but the process is cut short when one night Demophons
mother, Metaneira, happens to see him in the fire and hysterically turns on
the goddess. An enraged Demeter shrieks in response (256258):

.
Ndes anthrpoi kai aphradmones out agathoio
aisan eperkhomenou prognmenai oute kakoio;
kai su gar aphradi(i)si te(i)s nkeston aasths.
Humans! unskilled and without sense to foreknow
a divine-decreed destiny of coming good orill;
You too in your senselessness are incurably impaired.
The semantics of the verb with which the last line here cited ends, (aa),
a verb of epic, typically entail an encumbrance upon the mind or some otherwise cognitive impairment. The use of the verb here seems clearly to echo
deductively the initial word of the first line ( [ndes]): Metaneiras personal deficiencies are those of the human race generally. The one who knows
nothing in the realm of perceiving divine will is one who lacks a certain nimbleness of the mind, and hopelesslyso.
In each of these instances from Homeric epic and Homeric hymn, -
(n-is) qualifies one who does not know who lacks know-how and skill in
the performance of some art, some craft some (tekhn). The areas of
expertise are several: the making of war, the competing in games, the performance of song, the foreseeing of the purpose of the gods the skills of the
warrior and the (dmioergoi).
But the use of - (n-is) is of course not limited to early Greek literature
only. An interesting example is provided by a Hellenistic author whom we
continue to encounter. In Aetia fr. 1.12, Callimachus writes:
] ,

oid hot]i moi Telchines epitruzdousin aoid(i),
ndes hoi Mouss ouk egenonto philoi
210
I know tha]t the Telchines murmur against mysong,

they who know nothing of the Muse and are not her friend.88
As we saw earlier, in this poem Callimachus refers to his literary detractors as
Telchines. In Greek tradition the Telchines were legendary craftsmen, chthonic
figures who belonged to the pre-Olympian past, the world of Titans and Giants.
Associated with the islands, particularly prehistoric Rhodes and Ceos, their
demise was linked imaginatively to earthquake damage that destroyed earlier
habitations in these regions.89 Strabo (14.2.7) identifies them as the first workers of iron and bronze, who (tekhnais diapherontes) excel
in skills/craftsmanship. Diodorus Siculus (5.55.2) states that they were the discoverers of certain skills ( [technn tinn]) and that they were the
first to produce [agalmata then] images of the gods: indeed,
certain divine images bear their name, such as Apollo Telchinius among the
Lindians and Hera Telchinia among both the Ialysians and Kamirans all
peoples of the eastern Aegean island of Rhodes. The Telchines were said to be
wizards ( [gotes]) as well, having power over the weather, and to whom
Diodorus (5.55.3) compares the Persian Magi. Strabo (14.2.7) tells us that it
was the Telchines who crafted the scythe with which Cronus emasculated his
father Uranus; Callimachus himself elsewhere (Hymns 4.3031) writes of the
Telchines as the craftsmen who made the trident for Poseidon.
In characterizing the Telchines as in Aetia fr. 1.12, Callimachus must
certainly not be asserting that they are unskilled or ignorant, knowing nothing, in some absolute way. Such ones are famed for their expert skill in metalwork. The assertion is surely that they are (ns) in a particular area, which
the poet specifies, the realm of the Muses, the realm of poetic composition.
In the Samian tradition preserved by Pausanias, the crafting of Heras
image is attributed to one who is not - (n-is). It is attributed to one who
has know-how in the use of the (smil). He is - (Smil-is), one
knowing the (smil), a legendary chisel wielder and image maker from
the mythic past like Daedalus, like the Telchines.

Other elements of archaic Samos are apropos to the present investigation: a
specific feature of Samian cult; and the general, robust intercultural phenomenon of which that feature is a single expression. Regarding the former matter
of Samian cult, Kyrieleis summarizes:
211
The bones found in the excavation show not only the customary practice of
sacrificing animals in Greek sanctuaries but also another, extremely peculiar, aspect of the Samian Hera sanctuary. There was an entirely unexpected,
unique find: skull fragments from a 5m long Egyptian crocodile (Crocodylus
niloticus) as well as skull and horn fragments from two African antelopes
(Alcelaphus uselaphus). These exotic animals could hardly have been living on
Samos, not even in captivity. It is reasonable to assume that these animals, or
perhaps only the heads, were brought by travelers from Egypt, as trophies of
the chase for instance and as such dedicated to the Samian goddess. These
bones of exotic animals, taken together with artworks imported from Egypt
and other countries, reflect the far-flung trade connections that Samos had in
the Early Archaic period.90
Kyrieleis also notes that the Heraion in Naukratis is a branch sanctuary of

the Samian Hera in Egypt, founded by Samian merchants in this trading centre in the Nile delta and maintained by several East Greek towns.91 Ancient
Naucratis, located within the Nile delta, on the Canopic branch of that great
river, about fifty miles southeast of Alexandria, was the first Greek settlement
in Egypt, seemingly founded during the reign of Psammetichus I, in the second half of the seventh century bc . According to Strabo (17.1.18), the town
was founded by Milesians. Naucratis grew to be an important and thriving
cosmopolitan Greek trading center, a port of trade: Aeolians, Ionians, and
Dorians here lived side by side, writes Braun, who continues: There was
nothing like it in the Greek world until the Panhellenic foundations, Thurii
and Amphipolis, of Periclean Athens.92 According to Herodotus (2.179),
Naucratis had occupied a privileged position as the site through which all
imports had to be funneled.93 In addition to the several Greek sanctuaries
in Naucratis individually built by Greeks from various locales (Samians,
Milesians, Aeginetans) was the Hellenion. This temple was owned and funded
corporately by diverse eastern Greek peoples and cities: Lesbian Aeolians from
Mytilene; Ionians from Chios, Teos, Phocaea, and Klazomenai; and Dorians
from Rhodes, Knidos, Halicarnassus (which used both Ionic script and language), and Phaselis (Herodotus 2.178).
These musings on Naucratis and a Samian presence there lead to the more
general phenomenon mentioned previously, a SamianNear Eastern cultural
interaction, for which the evidence is impressive. Again, Kyrieleis observes:
The high percentage of imported items from foreign countries, especially from
Egypt and the Near East, among the early finds from the Heraion at Samos is
212
a striking phenomenon. No other archaeological excavation in Greece has, in

fact, produced anything like the number and variety of imports found in the
Archaic sanctuary of Hera in Samos.94
Perhaps best known of these items is a piece of equine armor a bronze forehead protector decorated with nude female images beneath a winged solar
disk, bearing an Aramaic inscription written in Phoenician style (i.e. nintheighth century) Aramaic letters: That which Hadad gave our lord Hazael
from Umqi in the year that our lord crossed the river.95 The plate is of a
well-known North Syrian type.96 The recipient of the weather god Hadads
gift, the Syrian king Hazael, ruled Damascus during the second half of the
ninth century bc (At present the inscription is the earliest example of a west
Semitic script in Greece).97 What must be companion pieces, two bronze
blinders, have been found in excavations of the Eretrian (Euboia) sanctuary
of Apollo, one of which bears the same inscription.98 An intricate carving
in ivory of a springing lion, ornamentation for a piece of furniture, is also
included among the finds from the Samian Heraion, identified as a product
of Ramessid Egypt:
Since the lion was brought to Samos about six hundred years after it was made,
one could hardly assume that it reached Samos via a direct route from the
court of the pharaohs in Egypt. Perhaps a Greek in Egypt acquired the lion as
an antiquity so to speak. A comparison with the present day art market is not
altogether irrelevant because we know from ancient Egyptian legal documents
that as early as the second millennium bc grave robbers systematically plundered pharaohs tombs stuffed full of possessions which they naturally will
have sold.99
These particular items and other artifacts of Near Eastern provenience were
recovered from the excavation of a set of wells found at the Heraion. Much of
the more recent work at the site has in fact involved digging at levels beneath
the present-day water table.100 The wet conditions of the site, however, have
resulted in the preservation of typically perishable objects, notably a large
number of wooden artifacts: one finds, for example, various wooden statues,
small carved boxes with horse figures at either end, and a large number of
curved boards representing, it seems, highly stylized boat hulls.101 The quantity
of wooden votive objects preserved in the remains of the Heraion of Samos
is of course not a necessary indication that a greater number of such items
were housed here than at any other given archaic Greek sanctuary only that
213
they were better preserved. One cannot help but recall Clements report that
the (agalma) of Hera in her Samian sanctuary was at first (
[proteron]) a (sanis), and associated remarks regarding other votive
items.
6.2.9 a lexical matrix of cult and writing as

(cult) performance
What I believe it is important to take stock of at this point is the conspicuous
co-occurrence of lexical forms in (1) the copper-plaque line of text that we
have been examining in this chapter, line 16 of MS 2-2; and in (2) the Samian
tradition about Hera and her sanctuary at Samos as preserved by Pausanias.
In both we find a form of (smil): (1) its variant (ml), denoting
an engraving tool in the copper plaque; and (2) a proper name (Smil-is),
a compound built from (smil), denoting one skilled in the use of the
tool in the Samian tradition. And in both we find (3) a form of (lugos), a
term denoting withe, descended from a Proto-Indo-European root meaning
to bend: the nominal form, (lugos), occurs in the Samian tradition,
denoting a sacred tree and plaited tethers and garlands made from such a tree;
in the line 16 of the copper plaque MS 2-2, we find what I suggested is a verbal
form of (lugos). It is to the morphology and meaning of that verb that
we still must turn, but before doing so, some attention needs to be paid to the
writing of the alphabet as cult performance in the archaic Greek world not
on the geographic periphery, but in the geographic center.
6.2.9.1 t he ta l k i n g oaks o f d o d o na . In Pausaniass discussion of the
most ancient living trees, as we saw, he sets in the second tier, after the primeval (lugos) of Samian Hera, the oak ( [drus]) of Dodona. It is, writes
Herodotus (2.52), judged to be the most ancient divinatory site in Greece.
More than that, the founding of the site is in Greek tradition associated with
the pre-Greek people called the Pelasgians (see Ephorus [fourth century bc ],
FGrH 70 F 142; Strabo 7.7.10; cf. Herodotus 2.5253):102 it is to the Pelasgian
Zeus of Dodona ( [Zdeu Ddnaie Pelasgike])
that Achilles prays before Patroclus goes into battle (Il. XVI 233235).103 A
disguised Odysseus weaves the tale for Eumaeus and then Penelope of how
Odysseus had visited Dodona in order to learn how to make his way back
to Ithaka that he might hear the will of Zeus from the towering leafy oak
of the god (Homer, Od. xiv 327330; xix 296299; cf. Aeschylus, PV 829832
214
on [hai prosgoroi drues] the talking oaks of Dodona).

Herodotus (2.5455) also reports that priests of Zeus in Egyptian Thebes
had told him that Phoenicians had abducted two priestesses from Thebes,
one of whom ended up in Libya, the other in Greece each founding the
earliest divinatory sites in those respective places. However, he had it from
the prophetesses ( [hai promanties]) of Dodona themselves
(Promeneia, Timarete, and Nicandra) that it was two black doves that had left
from Thebes, one flying to Libya and the other to Dodona. Uttering sounds of
human speech, the former told the Libyans to found an oracle of Ammon; the
latter came to rest in the sacred oak of Epirus and announced that the place
was to be an oracular site ofZeus.
6.2.9.2 phaedrus a n d d o d o na . The Socrates of Platos Phaedrus, in his
repudiation of the written word of (grammata) letters alludes
to the chronological primacy of Dodona among oracles and to the prophecies
that came through the rustling of the leaves of Zeuss oak, as he dialogues with
Phaedrus about the primacy of the spoken word over the written (275BC):
, ,
. ,
, ,
Hoi de g, phile, en t(i) tou Dios tou Ddnaiou hier(i) druos logous ephsan
mantikous prtous genesthai. tois men oun tote, hate ouk ousi sophois hsper
humeis hoi neoi, apekhr druos kai petras akouein hup eutheias, ei monon
alth legoien;
And, my friend, theyve said that the oracular words of the oak in the sanctuary of Dodonian Zeus were the first such. Those who lived then lacked the
sophisticated wisdom (i.e., were not endowed with [sophos]) of the sort
that you moderns have; it was simply enough for them to hear an oak or a
stone, if truth alone it would speak.
The site of Dodona is certainly one of great antiquity; Mitchell observes:

Mycenaean contacts with Dodona date back to the fifteenth century. However,
Zeus was not the only, or even the original, occupant of the sanctuary. There
is also evidence of an earlier chthonic cult, probably connected with the Great
Goddess. Evidence for this earlier cult is both literary and archeological. The
archaeological record seems to indicate two distinct strata, the earliest of which
has been associated with the goddess worship.104
215
But though the antiquity of the site be great, as Platos Socrates avows, the
written word, (grammata) beginning at some moment in the history of the oracle did indeed play a role in divination at Dodona. Since at
least circa the last quarter of the sixth century bc inquirers had inscribed
questions on lead plaques and submitted them to oracular attendants in order
to obtain a divinatory response from the god of the oak. Some of the surviving
plaques likewise record the divine answer.105 There is no shyness at Dodona
about requesting revelation in writing and writing can deliver revelation at
Dodona.
Surely Plato must have been aware of the revelatory use of writing at
Dodona. Are we not to see here an implied contrast between an oracular
site at a time when it operated without means of the alphabet and a present
moment when the alphabet plays a central role in the imparting of divinatory leading? Between a time when aural perception was enough
(druos kai petras akouein) to hear rock and stone was sufficient so long as truth was revealed? A contrast between such a time and a
time when revelation and request for revelation is not linguistically immediate but depends upon a derivative symbolic system for delivery of revelation. Between a past of orality and a present of literacy.
Socrates continues (Phaedrus 275CD), drawing attention to the previously
encountered notion of (tekhn), a skill, an art, such as that practiced by
Smilis, he who possesses know-how of the (smil) / (ml); here
the alphabetic (tekhn) is explicitly invoked:
,
,
,
.
Oukoun ho tekhnn oiomenos en grammasi katalipein, kai au ho paradekhomenos hs ti saphes kai bebaion ek grammatn esomenon, polls an eutheias
gemoi kai t(i) onti tn Ammnos manteian agnooi, pleon ti oiomenos
einai logous gegrammenous tou ton eidota hupomnsai peri hn an (i) ta
gegrammena.
Then the one who thinks that in letters he is leaving behind a tekhn, and the
one who receives it as being something sure and steadfast, from the letters,
would be full of a lot of foolishness and ignorant of the prophecy of Ammon
thinking written words to be something more than a reminder for the one who
knows the things about which they have been written.
216
This prophecy of Ammon, that Egyptian god whose oracle in Libya is etiologically linked to Zeuss oracle at Dodona, was expounded by Plato some
lines earlier (274 D275 B) in the myth of Thamus (i.e., Ammon) and Theuth
(the Ibis-headed Thoth), an ancient god of Naucratis, writes Plato, and the
inventor of (ta grammata) letters. Thamus foretells that this
(tekhn) will not be the great boon to the Egyptian intellect that Theuth,
the (patr grammatn) father of letters, envisions, but will
instead lead to a loss of the powers of memory and to a pseudo-wisdom.
Phaedrus agrees regarding the foolishness of those who would embrace the
alphabetic (tekhn), and Socrates continues (Phaedrus 275DE):
, , , .
, ,
. ,
,
. ,
, ,

.
Deinon gar pou, Phaidre, tout ekhei graph, kai hs alths homoion
zdgraphia(i). kai gar ta ekeins ekgona hestke men hs zdnta, ean d aner(i)
ti, semns panu siga(i). tauton de kai hoi logoi; doksais men an hs ti phronountas autous legein, ean de ti er(i) tn legomenn boulomenos mathein, hen ti
smainei monon tauton aei. hotan de hapaks graph(i), kulindeitai men pantakhou pas logos homois para tois epaousin, hs d hauts par hois ouden
proskei, kai ouk epistatai legein hois dei ge kai m; plmmeloumenos de kai ouk
en dik(i) loidortheis tou patros aei deitai bothou; autos gar out amunasthai
oute bothsai dunatos haut(i).
For, Phaedrus, writing has this peculiarity its really like painting, the creations of which are like living beings; but if you should ask them something,
they remain entirely solemnly silent. Its the same with [written] words. You
might suppose them to speak, as if they have understanding; but if you should
ask them something, wanting to learn about what they are saying, they will
always signify ( [smainei]) only one and the same thing. And when
once it has been written, each word always gets passed around, equally among
those who have understanding and those who are irrelevant to the matter;
217
and [the word] does not know to whom it ought to speak, and to whom it
shouldnt. But when it is wronged (lit. badly played; < [plmmele])
and abused it always needs help from its father: for it is able neither to defend
nor to help itself.
Writing, like painting, is a symbolic system. It may record and resemble (in
some sense) the phenomena it symbolizes, but it is not the equivalent of those
phenomena. Phaedrus again agrees, and Socrates asks (Phaedrus 276A):
; , ,
;
Ti d? allon hormen logon toutou adelphon gnsion, t(i) trop(i) te gignetai,
kai hos(i) ameinn kai dunatteros toutou phuetai?
What then? Do we find a word that is a legitimate brother to this one, both
in the way it comes into being and in how much better and more powerful it
naturallyis?
Is there a comparable word legitimated in its organic composition and its

existential quality? Phaedrus responds to Socrates question with a question:
What is this word and how does it come into being? Socrates replies that it
is theword
,
, .
Hos met epistms graphetai en t(i) tou manthanontos psukh(i), dunatos men
amunai heaut(i), epistmn de legein te kai sigan pros housdei.
Which is written ( [graphetai]) with skill in the soul of the learner,
able to defend itself, and knowing before whom it ought to speak and before
whom to be silent.
Are Platos remarks that set the living oral word against its illegitimate
brother the word of letters foregrounded against a backdrop of a popular notion of writing that is, orthography as performance? An active
interpretative phenomenon dating to the time of the Greek acquisition of the
alphabet but not a forgotten memory in Platos own day finding immediate expression in the Phaedrus in the oral oracular performances of the cult
specialists of Dodona in a bygone day in opposition to the assumed written
revelations of the present moment?
218
Plato then places on Phaedruss lips a summary observation, astute in its

insight, anticipating Ferdinand de Saussures signified / signifier (signifi / signifiant) dichotomy by more than two millennia:
,
.
Ton tou eidotos logon legeis zdnta kai empsukhon, hou ho gegrammenos eidlon
an ti legoito dikais.
Youre talking about the living and animate word of the one who knows, of which
the written form could rightly be said to be an image ( [eidlon]).
Archaic Greek alphabetic practice and cultic practice were not intrinsically mutually exclusive phenomena: they could share the conceptual space
of a lexical set (involving notions of revelation). As discussed in Chapter2,
graphic symbols, (smata), can be offered to Zeus Semios on Mount
Hymettos; and the one presenting the graphemic sacrifice can offer a graphic
expression of the act being performed (and see the further discussion in the
next chapter). At Dodona, request for revelation and revelation can be mediated by the alphabet. But, as Robert Parker has observed, texts had no direct
place in the conduct of the vast majority of Greek rituals.106 After the Greek
acquisition of the alphabet, production of graphic symbols, second-order
encoding of the inherently and iteratively creative first-order phenomenon
of human language, looks to have been selected against, in the end, in most
expressions of legitimate cult ritual.
6.2.10
( lu g i z d ):
a secondary formation
In the preceding discussions, attention was focused on (lugos), the

Greek nominal reflex of the Proto-Indo-European root *lug- to bend. From
Greek (lugos) is formed a denominative verb (lugizd) to bend
(like a withe), to twist, to turn. The building of verbs with the formant -
(-izd) is a highly productive process: together with the related formant -
(-azd), the two fashion denominatives from stems of all kinds to the number of several thousands, writes Buck.107 Chantraine concurs: Le grec a constitu environ 1.000 verbes en - [-azd] et 2.000 verbes en - [-izd].108
Examples include: (hubrizd) to outrage from (hubris) an outrage; (nomizd) to hold as customary from (nomos) custom;
(teikhizd) to build a wall from (teikhos) wall.109
219
The earliest attested examples of the verb (lugizd) appear to be

those provided by Sophocles one being of perhaps particular interest for
the present investigation. In a fragment of Sophocles Ichneutae (Trackers),
a work relating the tale of how Silenus and a chorus of satyrs went in search
of the stolen cattle of Apollo, the chorus verbally accosts the nymph Cyllene,
infant Hermes caregiver, at the cave of Hermes mother Maia. Cyllene berates
the satyrs and rejects their claim that the child whom she nurses is a cattle
thief: he is no less than a child of Zeus by clear reckoning (
[saphei log(i)]; line 358). The chorus of satyrs replies (lines 371372):110
,
.
Strephou lugizdou te muthois, hopoian theleis
baksin heurisk apopskton.
Plait and withe those myths (muthoi), devise whatever
slick revelation youwish.
The two conjoined imperatives with which line 371 begins, (strephou)
and (lugizdou), are verbal counterparts of the nominal forms we saw to
co-occur in Odyssey ix 420429 and Odyssey x 167168: (eustrephs)
well-plaited and (lugos) withe. The verb (streph) occurs frequently in Homer, though typically with the sense of to turn as in to turn a
horse, or, in the medio-passive, to turn oneself . The denotation of to plait or
to weave together can, however, be seen plainly in the occurrence of
(streph) in the Homeric Hymn to Hermes 409411 within that same passage
mentioned earlier in conjunction with the occurrence of the botanym
(agnos), which, as we have seen, would be eventually identified with
(lugos). It here occurs in two forms: (1) within the complex verb
(peri-streph; line 409); and (2) as the simplex verb (streph), in the
form of the perfect participle (estrammenai; line 411). The infant
Hermes has led Apollo to his stolen cattle; Apollo then begins to plait (
[peri-streph]) withes to make a binding, intended for Hermes it would
seem,111 but the withes take root, intertwine, and entangle the cattle instead:

410
,

220
kai khersi peristrephe karteradesma

agnou; tai d hupo possi kata khthonos aipsa phuonto 410
autothen emboladn estrammenai alll(i)sin,
rheia te kai pas(i)sin ep agrauloisi boessin
Herme boul(i)si klepsiphronos
and with his hands he was plaiting strongbonds
of agnos; but straightaway they grew into the ground beneath their feet 410
where weaving together with eachother
and deftly across all the field-dwelling cattle
by the schemes of dissembling Hermes.
A second Sophoclean occurrence of (lugizd) is found in Women
of Trachis 779, where it is used to denote the turning of the ankle at its
joint. Additional fifth- and fourth-century bc usages are these. A fragment
(Kock 1.339) of the fifth-century comic playwright Eupolis is reminiscent of
the quoted lines from the Ichneutae in its coupling of (lugizd) and
(streph): (lugizetai kai sustrephei
aukhena) he wreathes and twists his neck. Similarly, in the Republic (405c)
Plato sequences (streph) and (lugizd) in writing, metaphorically, of twisting and weaving ones way to skirt justice.
These two verbs, (lugizd) and (streph), display a fundamental morphological difference. The latter, (streph), represents
a primary derivation, formed by adding the thematic suffix -o- directly to
streph-, the verb root: this is a process of primitive Indo-European origin.112 We
have already noted that (lugizd) is built with a formant - (-izd)
used to derive verbs from noun stems a highly productive morphological
process of Greek that represents, in effect, a secondary derivation.113
6.2.11
( lu z d ):
a primary formation
The form (luzd) in line 16 of copper plaque MS 2-2 is, I contend, like
(streph), a primary derivation, one formed from the root *lug- by the
attachment of a suffix of primitive Indo-European origin. In the derivation
of (streph), the thematic suffix -o- is used; in the derivation of our
verb , the suffix that is employed is the thematic suffix -yo-, that highly
productive stem-forming thematic suffix of ancient Indo-European languages
mentioned in the preceding footnote. The attachment of the -yo- suffix to
verb roots of Greek commonly results in the final consonant of the root being
221
modified in some way: thus, for example, we find alternations such as the
following:114
(18) (phulass) to guard from *phulak-yo- beside (phulaks)
guard
(pess) to cook from *peqw-yo- beside aorist tense
(e-pep-sa)
(harpazd) to snatch away from *harpag-yo- beside
(harpag) seizure
(elpizd) to hope for from *elpid-yo- beside genitive noun case
(elpid-os) ofhope
And in the same way, beside the nominal (lugos) withe, we have a primitive primary verb formation (luzd) from *lug-yo- (where *g + *y zd
by regular sound change),115 with the meaning to withe, to plait, and so, given
the demonstrated semantic relations, to (inter)weave.
6.3
(se)
a n d
(abgd)
The two remaining arguments in our structure are the object of the verb
(luzd) and the object of address (a vocative object). The former is transparently the second-person pronoun (se) you. The latter, (abgd), must
be the denotation for in other words, the way of naming this script that
is being written by the scribes of the copper plaques: an identical concatenation of letters (abcd) is, after all, the source, via Latin (Late Latin abecedrius),
of our English word abecedary or abecedarium (cf. [alphabtos],
Anecdota Graeca 181, Sch. D.T. p.320H). Hence, the argument structure
(Ml se luzd abgd) can be understood to say: O abecedarium
(), may the stylus () interweave () you ().
The subjunctive verb (luzd) lacks the familiar iota ( or ) of classical Greek. That iota is a graphic expression of the so-called primary verb endings (like those of the present indicative): a long-vowel subjunctive such as
(lu(i)) may (s)he release would reflect, in terms of a possible synchronic
analysis, a constituent structure /lu-e-ei/. The absence of iota in the CP spelling of the thematic subjunctive (luzd) provides additional confirmation
of the deep antiquity of this inscribed Greek sentence. How is thisso?
Comparative linguistic evidence reveals an archaic Indo-European form
of the subjunctive that was built using secondary endings (like those of the
imperfect indicative); Indo-Iranian preserves examples, as do various Greek
222
dialects.116 One can be seen, for instance, in Boiotian (pi), whoever

drinks, in a metrical inscription on a kantharos from Thespiai (ca. the midfifth century bc ).117 The same form, though (further) inflected with the primary athematic ending - (-si), occurs in the inscription on Nestors cup, as
we saw earlier:
h <> : []
hos d a<n> tde pisi : potri[]
But he who drinks from thiscup
The verb form (pisi) is preserved on one of the more recent fragments
to be joined to the pot, the linguistic significance of which has been clearly
articulated by Watkins:
It removes the asterisk from a whole class of subjunctives presupposed by
Jakob Wackernagel for Homeric grammar in 1897.118 Subjunctives
[ethel(i)si], [pher(i)si], * [*pi(i)si] in the Homeric textual tradition reflect * [*ethelsi], * [*phersi], [pisi] with long
vowel and primary ending - [-si] < *- [*-ti] = (or ) Indo-Iranian
-ti.119
Invoking the previously mentioned Boiotian form (pi), Watkins further

remarks:
The corresponding long-vowel subjunctive with the secondary ending,
IE *-(-t?) (similarly remade in Hom. 328)is found in the early Boeotian
metrical vase inscription IG VII 34674 . The linguistic reality of the longvowel thematic subjunctive endings, as in , , can no longer be questioned: how they are to be historically accounted for is a separate issue.120 The
status of as a linguistic archaism would appear to exclude a dating of the
graffito itself to as late as the sixth century.121
The presence of the long-vowel thematic subjunctive with secondary ending,

(luzd), in the copper plaques likewise points to an early archaic date for
the execution of the inscription.
The sentiment expressed by the scribe of the copper plaque
(ml se luzd abgd) O abecedarium (), may the stylus ()
interweave () you () reminds us of the words that Euripides In-Law
addresses to his plaques, urging them to submit to the strokes of his carving
instrument (Aristophanes, Thesm. 778779):
223
Come on! tablets of smooth-polished plaques,

Accept the etchings of the stylus,
In the case of the copper plaques, the scribe calls on the alphabet to behave
as it should; in the case of Aristophanes, the scribe calls upon the medium to
allow the alphabet to behave as it should. The two are varying expressions of
the same sought-after phenomenon.
6.4
o n sam o s : pa rt 4
In concluding this chapter I would like to return to Samos and to underscore the similarities between the etiological myth attached to the cult of
Hera and the line of Greek language that appears on the copper plaque MS
2-2. The Samian etiological tradition echoes the language of the linguistic
utterance of the copper plaques. A common lexical matrix shared by that
line of encoded language the scribes enunciation and by the Samian tradition an account of the founding of the festival of the Tonaia situates the
copper-plaque utterance within a context of religion and cult. This is not to
suggest that the plaques are themselves in any way affiliated with the celebration of the ritual, but that the enunciation of the copper plaques and the
etiological tale told of the origin of the Samian ritual appear to be common
participants in a lexical nexus. If so, this relatedness must be grounded in
some actual anthropological phenomenon not simply dual reflexes of some
metalinguistictext.
In the copper-plaque utterance we find the nominal (ml), denoting
the engraving tool that was used to inscribe the plaques with repeating alphabetic series. The scribe calls upon the alphabet to allow the (ml) to
carve the alphabetic symbols into the surfaces of the plaques.
In the etiological convention of Samos, the name assigned to the carver of
Heras (agalma) is (Smil-is), a compound, I argue, built from
(smil), the variant form of (ml), denoting one who is skilled
and knowledgeable in the use of the carving tool used to craft that
(agalma).
In the copper-plaque enunciation, the scribe invokes the alphabet to permit the (ml) to weave a graphic fabric, as it were. The verbal notion is
encoded in an archaic subjunctive formed from the root seen in -
(lug-os) withe, descended from a widely attested Proto-Indo-European verb
root meaning to bend.
224
In Samian etiological memory, the nominal (lugos) plays a seminal

role. The term denotes a sacred tree (beneath which Hera was born) a tree
characterized by flexible branches, withes, with which Heras (agalma)
was entwined and woven securely to abase.
In the mythic past of Samos, it was the craftsman Smilis, a contemporary
of Daedalus, who applied his (tekhn) and his (smil) to produce a
chiseled (agalma) for the goddess, either the (sanis), the plank,
that her sanctuary had held at some sufficiently early moment, or, as the
attested tradition holds, its replacement revealing an image of the deity. It
is the unknown inscriber of the copper plaques, wielder of a (ml) who
converted a sheet of copper an object that could be appropriately termed a
(sanis) into a revelation of the alphabetic sequence.
an addendum on ms 22, lines10/11

In Chapter4, I noted that in addition to MS 2-2, line 16, a second stretch
of graphemes, lying on lines 10 and 11, also deviates appreciably from the
periodic order of the alphabet, and that, like line 16, it is characterized by
a lambda + digamma ( F) sequence and by the apparent inclusion of the
discrete periodic subsequence alpha, beta, gamma, delta: discrete must
of course be qualified, because these abecedarium-initial members are here
disrupted by the anomalous deletion of alpha and by the insertion of (apparently) san, which looks to have been moved forward from the latter portion
of the line, from a position just following the large range of omitted letters,
marked by // in Chapter4 and here in (19). The relevant expanse from lines
10 and 11 is repeated here; just prior to the position of the deleted alpha, (A),
the preceding abecedarium had terminated; after the tau, T, the ensuing abe
cedarium begins:
(19)
(A) // F 11 T
10
The playfulness of the scribe is here more elaborate; the interweaving of letters more dense; the recovery of language from the stretch less certain than
in line16.
One might propose that the lambda + digamma ( F) sequence is to be
parsed together with the two characters that immediately follow, that is, sigma
+ xi ( ), with xi again standing in for eta, and that the scribe has thus woven
into the line another form of the verb (luzd). The sequence F ,
that is (lus), could be interpreted as the aorist (rather than present)
225
subjunctive may it interweave. The s-aorist of present-stem verb forms in -(-zdo-) in which the sequence -zdo- is derived from earlier *-g + *yo- are typically, and expectedly, formed in (ks): in other words, this aorist-tense marker
derives from *-g + *s-, with *gs becoming ks (written ) by regressive voicing
assimilation. But competing with these are aorist stems in (s) formed under
analogical pressure.122 For example, (harpazd [from *harpag-yo-] to
snatch away) produces both a phonologically expected s-aorist subjunctive
formant (harpaks-) as well as a remade s-aorist subjunctive formant
(harpas-). An aorist subjunctive (lus) would be a formation of
the latter type. But then what sense does one make of the theta () or, for
that matter, the sequence theta-tau ( T) that immediately follows F ,
prior to the onset of the next abecedarium?
If, on the other hand, one abstracts from the sequence of symbols that
immediately follows the extended range of omitted characters (again, marked
by //) all of those letters that are out of alphabetic order, a more attractive solution offers itself. The letters on which this operation is to be performed are
here underlined:
(20) // F T
By abstracting the underlined letters, that is, those out of periodic order, the
sequence can be resolved into two interwoven substrings, (21) and(22):
(21) // T
(22) F
The letters of (21) constitute the terminating sequence of the alphabet; the letters of (22) encode language, another form of the verb (luzd). One could
read that form as (luks). In other words, the scribe would here be using
the symbol xi () to represent its characteristic value of /k + s/, rather than
employing it as a playful alternative to eta, representing . On the other hand,
the scribe would be assigning to theta () not its canonical value of th, but
instead in keeping with the copper-plaque practices that we have seen is
using it in lieu of omicron () to represent a mid back vowel. Before the creation of the vowel symbol omega, representing the -vowel, omicron was used
in local Greek alphabets to spell both short and long vowels, that is, o and .
Without any doubt, the omicron symbol of the copper-plaque alphabet would
have been entrusted with that same dual value; and its theta alternative, by this
interpretation, would here be used to represent the long member of this set,
the -vowel; hence (luks). Such a form is morphologically ambiguous:
226
the verb could be either the expected s-aorist subjunctive formed in (ks;
rather than analogically in [s].) or the future indicative (also marked by an
s-morpheme; the same phonological processes apply) of the first-person
singular in either case. The scribe would thus be announcing the sentiment
May I interweave or I shall interweave, as he is, in fact, performing the very
deed, via his stylus (the [ml)]), that he proclaims to be performing.
7
The Warp and Weft of Writing
Rude rymes, the which a rustick Muse didweave

In salvage soyle, far from Parnasso mount,
And roughly wrought in an unlearnedLoome
Edmund Spenser (lines in a dedicatory verse to Lord Grey
of Wilton, from the front matter of the Faerie Queene)
7. 0
i n t r o du c t io n
Much of the preceding chapter was devoted to an interpretation of line 16 of

copper plaque MS2-2:
(1)I(?) A H Z F E HM
The line consists of fourteen arbitrarily ordered graphemes (to include the
denotation of the abecedarium) followed by, as it is preceded by (at the end of
line 15), an unusual iota-like symbol possibly a mark of punctuation. I have
read the line as (ml se luzd abgd) O abecedarium
( [i.e., abcd]), may the stylus () interweave () you (): a sentiment closely matched by the In-Laws address to his plaque in Aristophanes
Women at the Thesmophoria 778779.
But what exactly is it that the copper-plaque scribe is invoking his alphabet to do in its proper behavior? To answer that question, we must first
remind ourselves of two notions that we encountered in earlier chapters: (1)
a common condition of literacy; and (2) a metaphoric expression of poetic
composition.
227
228
7. 1
c o n f u sio n o f la n g uage a n d s c r i p t
First language is fundamentally a biological phenomenon available to all

human beings, a seemingly innate primitive expressed in each member of
the species, with the exception of a few who are cognitively, socially, or physically deprived of the ability. A writing system on the other hand is a derived
phenomenon: it is an arbitrary means of symbolically recording language. All
people have language; not all people have a writing system. There are those
who have language but who cannot write or read; no people can write and
read but lack language.
Yet these two distinct systems language and orthography are commonly
equated among literate peoples. This assimilation is well known and has been
documented many times over. As we saw, for example, Saussure called attention explicitly to linguistic-graphemic confusion in his Cours:
But the written word is so intimately connected with the spoken word it represents that it manages to usurp the principal role. As much or even more
importance is given to this representation of the vocal sign as to the vocal sign
itself. It is rather as if people believed that in order to find out what a person
looks like it is better to study his photograph than his face.1
7. 2
p o e t ic weav i n g
And second in the preceding chapter, following the discussion of Scheid

and Svenbro,2 we took note of the metaphoric concept of literary composition
as poetic weaving, expressed as such in the works of the lyric poets Pindar
and Bacchylides.
7.2.1 pindar and bacchylides

Beyond the examples compared in that earlier discussion Pindar, Olympian
Odes 6.8687, and Bacchylides, Victory Odes 5.814 Scheid and Svenbro note
yet other usages of the same metaphor in the works of these two poets. Thus,
in Nemean Odes 4.4446, Pindarsays:
, , , ,

, .
45
229
ksuphaine, glukeia, kai tod autika, phormigks,

Ludia(i) sun harmonia(i) melos pephilmenon
Oinna(i) te kai Kupr(i), entha Teukros aparkhei.
45
Sweet lyre, quicklyweave

this song in Lydian mode, beloved
45
of Oinona and Cyprus too, where Teucer reigns so far from home.
The poet addressing the lyre, calling on it to weave its song, is again reminiscent of the copper-plaque scribe calling on the alphabet, urging it to be woven
by his stylus: the weaver and the woven, instrument and product, are alternately implored. Scheid and Svenbro remind their readers of Snyders intriguing insight regarding the imagery of the invoked lyre:3 Thus the lyre is transformed into a loom, whose vertical warp corresponds to the vertical strings.4
Scheid and Svenbro5 also call attention to Nemean Odes 8.15 in which Pindar
refers to his poem as a Lydian headband ( [mitra]) skillfully worked
( [pepoikilmena]) with resonance, and to a Pindaric fragment
(179): I weave ( [huphain]) for the sons of Amythaon an elaborate
( [poikilon]) headband ( [andma]), about which a scholiast
writes: He likens his poem to weaving.6 And, finally, Scheid and Svenbro
take note of Bacchylides, Dithyrambs 19.811:
In a poem addressed to the Athenians, Bacchylides uses the same metaphor,
calling on the care for perfection (merimna) characteristic of the poets from
Ceos (his uncle Simonides and himself): Weave [huphaine] something new in
the rich beloved Athens, O famous perfectionism of Ceos!7
To this set of texts demonstrating these lyric poets conceptualization of

poetic composition as weaving, one could add Pindars closing lines (110115)
of the third Pythianode:
,
.
, ,
,
,
.
Ei de moi plouton theos habron oreksai,
elpid ekh kleos heuresthai ken hupslon pros.
110
115
110
230
Nestora kai Lukion Sarpdon, anthrpn phatis,

eks epen keladennn, tektones hoia sophoi
harmosan, ginskomen: ha d areta kleinais aoidais
khronia telethei: paurois de praksasth eumares.
115
And if a god should give me splendorous wealth, 110

my hopes to find a lofty future fame.
Nestor and the Lycian Sarpedon, the talk ofmen,
we know, from ringing words that clever craftsmen
joined: distinction long endures through songs offame:
but for few tis easy to achieve.
115
With line 113 compare Pindars Nemean Odes 3.45, where he writes of
| (meligarun tektones | kmn) craftsmen of
sweet-voiced celebrations,8 referring to the chorus members who are awaiting the Muse, (neaniai sethen opa maiomenoi)
young men desiring your voice.
(tektones) craftsmen (singular [tektn]) is a word of primitive Indo-European extraction, synchronically sharing a root with
(tekhn) art, craft, skill, several spheres of which we encountered in the preceding discussion of (ns). Together with its numerous cognates,
(tektones) points to an Indo-European etymon *teks - meaning to weave; to fabricate. Among descendant forms are Latin tex to weave, to plait, to embroider; to fabricate; textor weaver; textum woven fabric; interlaced timbers; textus style of weaving; woven fabric; the product of joining words (to produce a
text); Sanskrit tkati to hew; to fabricate; Avestan taaiti to frame; to cut with
a knife or ax; Old High German dehsa and Old Norse exla mattock; Middle
High German dehsen to beat flax and dehse spindle.9 The common origin of
terms denoting both acts of (1) weaving and plaiting on the one hand and (2)
fabricating on the other may lie in the Neolithic practice of constructing the
walls of houses with wicker and wattle.10 Pindars (tektones) are poetic
craftsmen who by their skillfully joined words bring enduring fame to heroes
such as Nestor and Sarpedon. His syntagmatic clustering of (tektones)
and (epea) in Pythian Odes 3.113 (from ringing words that clever craftsmen joined) is itself of considerable import, as we shall soonsee.
7.2.2 archaic greece

The metaphoric notion of the warp and weft of language must have been a
fundamental one in archaic Greece, both before and after the acquisition of
231
alphabetic literacy. Verbal weaving is known to Homer, depicted, for example, in Iliad III 212, in the poets description of how, before an assemblage of
Trojans, Odysseus and Menelaus

muthous kai mdea huphainon
Wove words and counsels
In her study of weaving imagery in archaic poetry, Snyder notes that Homer
extends the metaphor of weaving (lexically encoded in [huphainein])
beyond the realm of verbal composition to additional cognitive activities (in
her words, as a description of an intellectual process):
Odysseus, the suitors, Nestor, and others weave stratagems and wiles; Athena,
the only female figure in Homer for whom her weaving is not a literal occupation, helps Odysseus weave wiles. Penelope, though she cannot achieve the
status of the androgynous Athena, is nevertheless capable, through her literal
weaving, of enjoying the masculine ability to weave stratagems.11
Snyder notes too that, in addition to various metaphorical allusions to weaving, Homer draws weaving and singing together directly in his descriptions
of Calypsos and Circes weaving activities at Odyssey v 5962; x 220223, 226
228, and 254255:
Thus, while Homer himself never actually describes poetic activity as analogous to weaving at the loom, his frequent references to metaphorical and literal weaving, as well as his juxtaposition of actual weaving and singing, lay the
foundation for the lyric poets descriptions of their own webs of song.12
With regard to her latter point (lay the foundation for the lyric poets descriptions of their own webs of song), however, we should bear in mind that lyric
is a genre no less archaic than epic; as Gregory Nagy has reminded us: Lyric
did not start in the archaic period. It is just as old as epic, which clearly predates the archaic period. And the traditions of lyric, like those of epic, were
rooted in oral poetry, which is a matter of performance as well as composition
(Lord 1995:2268, Oral Traditional Lyric Poetry).13
Snyder offers an interesting observation in light of the close association
of women and weaving in Mycenaean and epic culture: It is not surprising that a woman seems to have been the first among extant writers to apply
the Homeric metaphor explicitly to her own art, the creation of song.14 The
woman is of course Sappho. Snyder references Sappho fragments 1.2 L-P, in
which Aphrodite is addressed as [pai Dios doloploke]
232
wile-weaving daughter of Zeus (she here merely echoes Homer, writes

Snyder) and 188 L-P, preserving the epithet (muthoplokos) muthos-weaver forEros.
Scheid and Svenbro take exception to certain observations that Snyder
offers in her 1981 study. They view it significant that Homer, in contrast to
the lyric poets Pindar and Bacchylides, never makes allusion to poetic weaving self-referentially he does not explicitly refer to his own verse-crafting
as poetic weaving,15 does not define song as fabric, although he was familiar
with the metaphor of language weaving,16 as scenes such as that surrounding
Iliad III 212, already described, clearly reveal. Concerning Snyders proposal,
to which allusion was made, that Homers frequent references to metaphorical and literal weaving, as well as his juxtaposition of actual weaving and
singing, lay the foundation for the lyric poets descriptions of their own webs
of song,17 Scheid and Svenbro raise the following objection:
While it may have had a certain importance for the poets who metaphorically represented their own song as fabric, the simple juxtaposition of weaving
and song in Homer is not enough to explain this development (in any case
the same juxtaposition existed in the daily practice of weavers, for song has
always accompanied work).18
Scheid and Svenbro envision that the invention of poetic weaving in the
Greek language is due to choral poets probably to Simonides, a pioneer in
this domain.19
7.2.3 common indo-european tradition

Perhaps it could be the case that Homer does not envision the epic bards
composition in performance as a process of word weaving and deems the
metaphor appropriate only for nonbardic, nonpoetic episodes of speaking:
that, however, would seem quite improbable. The concept of poetic composition as weaving long precedes Homer and long precedes the equally old
lyric. Comparative Indo-European evidence reveals that metaphors of the
weaving or crafting of poetic language are widespread among early IndoEuropean peoples. Indo-Iranian and Greek usages point to a reconstructable
common ancestral denotation for poetic production: *wekwos teks-, weaving or crafting words;20 the poet is the *wekwm teksn, weaver or crafter
of words. This matches morpheme for morpheme Pindars own
233
(epen tektones) in Pythian Odes 3.113 that we just encountered;21

if the hierarchical syntactic structure of Pindars line and that of the reconstructed phrase differ, the linear sequencing and the fundamental meaning
they impart are equivalent. Compare Pindars line and the reconstructed
Indo-European phrase with Homers previously cited
(muthous kai mdea huphainon) They wove words and counsel (Il. III 212): the lyric poets are no less heirs to Indo-European tradition
than is Homer. Compare too the Old English poet Cynewulf s metaphoric
phrasing from the closing lines of his Elene: Thus I, wise and willing, /
Wordcraft wove (wordcrft wf) and wondrously gathered (12361237).22 In
Welsh the bards are called seiri gwawd or seiri cerdd, carpenters of song, as
Williams points out,and
claimed as their own all the tools and technical terms of the craftsmen in
word, e.g. the axe, knife, square. When a rival imitated their themes or methods they told him bluntly to take his axe to the forest and cut down his own
timber.23
Old Irish preserves the phrase fig ferb fithir the master wove the word.24
7. 3
weav i n g o f a w r i t t e n t e x t
Is there then anything that does appear to be new within the conceptual realm
of poetic weaving in the work of the lyric poets noted in the preceding section?
In searching for an affirmative answer, one might point to a use of the metaphor of language weaving to characterize the composition of a written text,
as opposed to an oral composition in performance. Perhaps this is implicit in
Pindars lines in Pythian Odes 3, but it is made quite explicit by Bacchylides in
Victory Odes 5; Scheid and Svenbro draw their readers attention to this latter
point, if expressing it less absolutely than my question might seem to frame it
(and the emphasis indicated is myown):
Bacchylides considers that he wove his hymn. Not that this metaphorical
usage is dependent upon the written nature of the poem; on the contrary. It
is interesting to note, however, that it is a poem that the poet sends (pempei) to his recipient. The material and tangible nature of this epistolary ode
in fact adds a new dimension to language weaving as we have studied it until
now.25
234
The relevant text occupies lines 916 of Bacchylides Odes 5, written to celebrate the chariot victory of Hiero of Syracuse at the Olympian games of 476
b c (to whom Pindars Olympian Odes 1 was also written, occasioned by the
same victory):

,

.
sun Kharitessi bathuzdnois huphanas
humnon apo zdatheas
nasou ksenos humeteran
es klutan pempei polin,
khrusampukos Ouranias
kleinos therapn; etheleide
garun ek stthen khen
ainein Hierna.
With the deep-girded Graces ahymn
has your xenos woven,
and from the sacredisle
he sends it to your city of renown,
he, the famed servantof
golden-filleted Urania; hewants
to pour out speech from his heart
in praise of Hiero.
10
15
10
15
10
15
The woven hymn, which equates to praise of poured-out speech, is being sent
in written form from the Ionic island of Keos to Hieros Sicily. Bacchylides
poetic weaving produces an orthographic fabric: but the choral poets, I suggest, did not inaugurate the weaving of the writtenword:

ml se luzdabgd)
O abecedary (), may the stylus () interweave () you()
235
7.3.1 weaving of alphabetic letters

What we see in the copper-plaque abecedaria is an interweaving of letters.
That weaving occurs along both a horizontal dimension and a vertical dimension realized at the iterating intersections of an associative (or paradigmatic)
structure and a syntagmatic structure in Saussurian terms producing a fabric of combination and selection in Jakobsonian terms. The horizontal dimension the weft or woof of the alphabetic fabric is the continuous stringing
together of the letters in their periodic, that is, alphabetic, order: from alpha
to tau, from alpha to tau, from alpha to tau, and on and on. The vertical
dimension the warp of the alphabetic fabric is realized by the constant
interchanging of the morphological variants of the various letters at the individual letter positions within the alphabet the substitution of one letter form
for another. This process of alphabetic weaving results in the highly variegated
fabric that we see on the six faces of the copper plaques, presented row by row
and column by column warp and weft in the transcriptions of Chapter4.
7.3.2 latin alphabetic interweaving

Aside from the production of this alphabetic fabric in the copper plaques,
there is, however, still another sense in which the alphabet is woven or
plaited. Much of the previously known evidence comes from a time long
after the period in which the copper plaques were produced. As we saw in
Chapter2, the Greek alphabet would spread by way of Greek settlers in the
south of Italy to the Etruscans, who would then pass it to the Romans and to
speakers of other Italic languages (the conventionally imagined route). In the
ruins of Pompeii, the Italian city on the Bay of Naples destroyed by an eruption of Mount Vesuvius in a d 79, there are found instances of graffiti displaying the twenty-one-letter Latin abecedarium written in a peculiar way: the
abecedarium of CIL IV 5472, for example, appearsas
A X BVCT
DSERFQGPH
OINKML
That of CIL IV 9272 shows a similarly arranged alphabetic series, accompanied by a second sequence, comprised of three lettersonly:
AXBVCTDSERFQGP
HO I NKML
AXB
236
Likewise, CIL IV 5499, following the coda of an abecedarium, RSTVX, shows

an alphabetic series with the same order:
RSTVX
AXBVCTDSERFQGPHOINKML
Abecedaria having the same letter sequence also occur in CIL IV 6905, 6907,
as well as in the incomplete abecedarium of CIL IV9268.
What is the alphabetic pattern displayed in these several abecedaria? The
first half of the alphabet is written from left to right, in the normal fashion
(2), but then the alphabet turns back on itself and the remainder of the abecedarium is written right to left (3), being interspersed with the letters of the
firsthalf:
(2)
(3)
A
X
C
V
D
T
F
R
G
Q
H
P
I
O
K
N
L
M
In other words, the following sequential order is generated, where the initial,
left-to-right, portion is indicated with plain text, the remaining, right-to-left,
portion with underlining:
(4)
A X B V C T D S E R F Q G P H O I NKML
In an article treating, in part, the origin of the Latin term elementum letter of
the alphabet, Coogan draws attention in a footnote to Roman pedagogical practice vis--vis these intertwined plaited alphabets from the Bay of Naples,
remarking: These graffiti reflect a pedagogical practice described in Quintilian,
Inst. Orat. I.1.25 and Jerome, In Jerem. 25 v. 26.26 The former passage is of particular relevance to the present investigation. Quintilian writes that he disapproves
of the practice of teaching children the names and the order of the letters (litterarum nomina et contextum [on the latter term, see the subsequent discussion])
of the alphabet before the children have learned their graphic shapes it makes
it harder for them later to recognize visually the letters:
Quae causa est praecipientibus ut, etiam cum satis adfixisse eas pueris recto illo
quo primum scribi solent contextu videntur, retro agant rursus et varia permutatione turbent, donec litteras qui instituuntur facie norint, non ordine: quapropter optime sicut hominum pariter et habitus et nomina edocebuntur.
It is for this reason that instructors, even when they believe that they have sufficiently fixed the letters within the childrens minds in that linear sequencing
237
in which they are conventionally first written, then reverse the direction and
disarrange the order by various substitutions, until the children master the letters by their appearance, not by their order: and so it will be best for them to
be instructed thoroughly in both the form and name of the letters side by side
(just as with people).
Does Quintilian here have in mind the sort of plaiting of letters that is preserved
in the graffiti from Pompeii? Possibly, but he has cast his net wide: his remarks
on this pedagogical process provide a necessary but not sufficient description
of the weaving in and out of letters as attested in the south of Italy.
On the other hand, in light of those Latin abecedaria, lexical choices in
Quintilians text demand our attention. To denote the sequence in which
children conventionally first write the letters that is, the periodic order of the
symbols of the Latin alphabet Quintilian uses the word contextus, a nominal
derivative of the verb contex to make or join by weaving, itself a compound
form of tex to weave, to plait and so a member of that set of Indo-European
weaving and crafting terms to which (tektn) craftsman and so forth
belong (as we have discussed). The nominal contextus denotes most fundamentally the act of weaving; the act of constructing, and also, among other
senses, fabric; structure. In Quintilians quoted lines, his use of contextus for
the sequence of symbols suggestively denotes that fabric of letters which is the
alphabet. And the letters of that alphabet are further characterized as conforming to a rectus contextus, which I translated as linear sequencing. The
adjective rectus, in a straight line, however, also has its own affiliation with
weaving, and a seemingly quite archaicone.
Pliny (HN 8.194), citing Varro as his source, records that wool on the distaff
(colus) and spindle (fsus) of Tanaquil, wife of Tarquinius Priscus, first of the
Etruscan monarchs to rule Rome, could be seen housed in the temple of Semo
Sancus.27 Furthermore, a toga (toga regia undulata wavy [billowing?] royal
toga) that Tanaquil had made and that had been worn by Servius Tullius,
Priscuss successor to the throne, was on display at the sanctuary of Fortuna.
Pliny continues:
Inde factum ut nubentes virgines comitaretur colus compta et fusus cum stamine. Ea prima texuit rectam tunicam, qualis cum toga pura tirones induuntur
novaeque nuptae.
And so it came about that a decorated distaff and a spindle with thread accompanied young women in their wedding ceremonies. She [Tanaquil] was the
238
first to weave a tunica recta, the sort that young men who have come of age
and brides wear with a plain whitetoga.
According to Festus (p.277M), the tunica recta is referred to in this way (i.e., as
recta) because a stantibus et in altitudinem texuntur they are woven vertically
by standing [weavers]; Festus seems to equate the tunica recta with the regilla
tunica: regillis tunicis textis susum versum a stantibus woven upwards by
standing [weavers] (p.286M).28 The reference is apparently to weaving on the
warp-weighted looms common in classical antiquity.29
In the repeated reference to the order of the alphabet in the passage cited
(Inst. Orat. I.1.25), Quintilian uses not contextus a second time but ord to
denote the conventional sequence of letters (until the children master the
letters by their appearance, not by their order). But this lexeme also invokes
the metaphor of weaving: the nominal ord a line of items, a row is related
to the verb ordior, meaning to lay the warp of (a web),30 and then secondarily
to begin and to begin to speak or write. The fundamental notion of weaving recurs in related forms: exordior to lay out the warp; to lay out strands
for plaiting; exordium the warp laid out on a loom prior to interweaving the
weft both of which terms also carry notions of beginning; and redordior
to unweave, unravel. The ord of the alphabet is the sequence of letters
the alphabetic fabric that one produces weaves setting out at the beginning and passing straight on to the end, and is thus equivalent to the rectus
contextus.
Distinct from this is a variegated weave of the alphabet that is accomplished by reversal and disarranging the order. Quintilian denotes the latter
action with the verb turb (retro agant rursus et varia permutatione turbent
[then reverse the direction and disarrange the order by various substitutions]), ultimately traced to Proto-Indo-European *(s)twer- (with s-mobile),
meaning to turn, to whirl.31 The related Latin noun turb, turbinis designates
whirling implements: in addition to spinning top and a spinning object
used in the practice of magic, another of its senses is spindle whorl. The
verb turb may not in itself be a dedicated member of the lexicon of spinning and weaving, but might its usage here be intentionally (metaphorically)
suggestive of that realm of activity? Regardless, what we find preserved in
Quintilians description of a particular Roman pedagogical technique, I suggest, is the traditional Latin vocabulary of a much older practice of the weaving of the abecedarium.
239
7. 4
d io n ysi u s o f hal ic a r nassu s : l i t e r a ry,

l i n g u i st ic , a n d alphabe t ic weav i n g
Preceding the Spaniard Quintilian by a couple of generations was the Greek

historian of Rome and literary critic Dionysius of Halicarnassus, who arrived
in the eternal city circa 30 b c , bringing with him traditions of philosophy,
rhetoric, and grammatical analysis from the East. In a remarkable document
entitled (Peri suntheses onomatn), or De compositione verborum, On the Composition of Words, Dionysius reveals himself
not only to be familiar with that metaphor of language weaving as applied
to the composition of written literature that we encountered in Bacchylides
Victory Odes 5 but to be thoroughly steeped init.32
Influenced by Aristotle and, particularly, Theophrastus,33 among others,
Dionysius of Halicarnassus identifies in this work two fundamental aspects
of literary production that anticipate the dual linguistic planes of Saussure
and Jakobson. On the one hand, there is (eklog), the selection of
words out of the set of possible words that could be used in a given context:
this is notionally comparable to Saussures vertical linguistic dimension
the paradigmatic or associative axis and Jakobsons mode of selection.
On the other, Dionysius identifies (sunthesis), the composition,
combination of sentences and other linear structural elements: this compares precisely to Saussures horizontal that is syntagmatic dimension
and Jakobsons mode of combination.34 Much of Dionysiuss discussion in
this treatise is given to demonstrating what is in his view the primacy of
(sunthesis) combination over (eklog) selection: though literary (sunthesis) combination is naturally second order, he writes,
(eklog) being a necessary first step (i.e., words must be selected
before they can be combined), the former has far greater impact in the art of
language ( [en tois logois]) than does the latter; and in this way
it is consistent with other (tekhnai) arts that involve the building of
structure (Comp.2):
,
,
,
,
, .
240
enthumoumenos hoti kai epi tn alln tekhnn, hosai diaphorous hulas lambanousai sumphorton ek toutn poiousi to telos, hs oikodomik te kai tektonik kai poikiltik kai hosai tautais eisin homoiogeneis, hai sunthetikai dunameis
t(i) men taksei deuterai tn eklektikn eisi, t(i) de dunamei proterai; hst ei
kai t(i) log(i) to auto sumbebken, ouk atopon hgteon.
Consider that as with other arts that utilize varying raw materials and make
from them an end product such as architectural construction and carpentry
and embroidery and others of a similar sort the constructional capabilities occur second in order after the selectional ones, but are first in terms of
productive power: thus if the same relationship should hold with the art of
language, one must not think itodd.
As a first example of this relationship of selection and combination, Dionysius

offers Homers verses of Odyssey xvi 116, narrating Telemachuss arrival at the
hut of Eumaeus, where he is welcomed warmly by the swineherd (and will
see, and be seen by, his father, Odysseus, disguised as a beggar). Dionysius
judges that these verses are unsurpassed among poetic lines. But why are they
so? The answer comes (Comp.3):
;
,
,

.

.
;
poteron dia tn eklogn tn onomatn dia tn sunthesin? oudeis an eipoi
dia tn eklogn, hs eg peithomai; dia gar tn eutelestatn kai tapeinotatn
onomatn peplektai pasa h leksis, hois an kai gergos kai thalattourgos kai
kheirotekhns kai pas ho mdemian ran tou legein eu poioumenos eks hetoimou
labn ekhrsato. luthentos goun tou metrou phaula phansetai ta auta tauta kai
azdla;
.
ti oun leipetai m oukhi tn sunthesin tou kallous ts hermneias aitiasthai?
Is it because of selection of words or because of combination? No one
would say that it is because of selection I am persuaded of that: for with
241
commonplace and lowly words is the entire text plaited/woven words that a
farmer and a sailor and an artisan and all who show no concern for speaking
well would take up and use readily. For were the meter broken up, the very
same text would seem paltry and unenviable;
.
What else remains then but to credit combination with the beauty of
expression?
To express the metaphor of poetic word weaving, Dionysius here (line

58) uses the verb (plek) that we encountered in Chapter 6:
(plek) most fundamentally means to plait but, as Nagy brought to our attention, is sometimes used instead of (huphain) for to weave35 just as
Dionysius is clearly using it here. We also met with the observation of Scheid
and Svenbro, that plekein is used instead of huphainein when one wishes to
emphasize the aspect of interweaving.36 Dionysius is describing the Homeric
text as the end product of a process of interweaving: the two strands that
are being interwoven are those of (eklog) selection and
(sunthesis) combination the vertical and horizontal dimensions of language production the warp and the weft of the poetic fabric. The element of
(eklog) selection takes temporal priority in all those (tekhnai)
arts that involve the production of an end product, writes Dionysius, including the language art one form of which is the weaving of the Homeric text.
Just so, in the production of fabric on a loom, it is the warp that is first laid
out the vertical strand (on the warp-weighted loom): in poetic weaving,
(eklog) selection is the warp. Second in order, but of chief priority in
these (tekhnai) arts, is (sunthesis) combination. In the production of fabric on a loom, the strand of the weft the horizontal strand is
secondarily interwoven into the previously placed warp:37 in poetic weaving,
(sunthesis) combination is the weft. In the weaving of fabric on a
loom, the warp is a pre-arranged and more-or-less fixed set:38 just so, in the
weaving of a poetic fabric (eklog) selection involves the more or less
fixed (finite) set of a languages lexemes. The weft, in contrast to the warp,
is characterized by its extreme length and flexibility:39 analogously, in the
weaving of poetic fabric, (sunthesis) combination involves the syntactic component of language, capable of generating infinitely long sentence
structures. These matters are made plain by Dionysius of Halicarnassus.
The metaphor of literary weaving recurs throughout De compositione verborum. Dionysius advises the literary artist
(egkataplekein te kai sunuphainein) to intertwine and weave together
242
unpleasant and graceful words or the latter type only (Comp. 12; see also
Comp. 16. with regard to Homeric practice). He observes that the most accomplished composers of both poetry and prose,
(poitn te kai suggraphen), arrange their words
(sumplekontes epitdeis alllois) purposively weaving them together
with one another (Comp. 15). The extension of the metaphor of word weaving
from oral composition to written is here made explicit perhaps even more
strikingly so in De compositione verborum 25, where Dionysius, mixing his
metaphors and drawing the ancient figure of poetic weaving into the metaphoric realm of hair care, writesthat

,

.
ho de Platn tous heautou dialogous ktenizdn kai bostrukhizdn kai panta
tropon anaplekn ou dieleipen ogdokonta gegons et; pasi gar dpou tois
philologois gnrima ta peri ts philoponias tandros historoumena ta te alla kai
d kai ta peri tn delton, hn teleutsantos autou legousin heurethnai poikils
metakeimenn tn arkhn ts Politeias ekhousan tnde Katebn khthes eis
Peiraia meta Glauknos tou Aristnos.
And Plato, did not stop combing and curling his own dialogues and weaving/plaiting them in every way, even up to eighty years of age; for I suppose
that the stories concerning the mans love of work are well known to every
scholar, especially that one about the writing tablet that they say was discovered after he had died, containing the opening words of the Republic transposed in varied woven patterns: I went down yesterday to the Piraeus with
Glaucon the son of Ariston.
The adverb (poikils), which I have translated as in varied woven

patterns, also holds membership in the Greek lexicon of weaving. It is a
member of the family of Greek words at the head of which stands the nominal (poikilos) wrought in various colors [of woven or embroidered
stuffs];40 it occurs already in the Mycenaean documents, with the spelling poki-ro-nu-ka (Knossos tablets Ld 579, 598 etc.), describing a type of o-nu-ka, a
243
part of a garment. The Greek word-set is of primitive Indo-European origin:

linguistic relatives include Sanskrit pea- ornament, peala- artfully fashioned; Avestan pas- to color, decorate; Old Persian nipistanaiy to write;
Lithuanian pias freckle, piti to write, draw; Old Church Slavic pi
to write; Old Norse f to paint, f rnar to cut runes; Tocharian pik- to
draw, write; Latin ping to adorn with colors, to paint.41 Plato himself of
course knows and utilizes members of the lexical set, such as
(poikilmata) in Euthyphro 6c, describing the appearance and manufacture of
the peplos presented to Athena at the Panathenaia, and tying its weaving to
things told by the poets. Regarding this passage and the lexical form, Nagy
observes:
I draw attention to the metaphor of poikilia pattern-weaving, which
establishes a parallelism between poetry and fabric-work as prime media
of mythmaking. Similarly in Plato Republic II 378c, the expression muthologteon to be mythologized is made parallel to poikilteon to be patternwoven, and the subject of mythologization / pattern-weaving is none other
than the battles of gods and giants, that is, the gigantomakhiai of the Great
Panathenaia.42
The relevance, and importance, of this observation for the present investigation will become clear later in this chapter.
Dionysius of Halicarnassus does not limit his application of the metaphor
of literary weaving to the domain of words alone. For example, descriptions of
the weaving of (kla) clauses (Comp. 19 and 25)43 and of the interweaving of rhythms (Comp. 18 and 25) both occur. The domain over which the
metaphor holds can also be internal to the word: he writes of syllable weaving:
(h tn sullabn plok
pantodaps skhmatizdomen) the weaving of syllables, being arranged in all
kinds of ways (Comp.12).
Most fascinating of all, Dionysius makes use of the metaphor of weaving
letters, as in the following example here in conjunction with the weaving of
syllables. Addressing the matter of how to make the arrangement (
[harmonia]) of phrasing ( [leksis]) that is, of the combining of words
beautiful, he states (Comp.13):

,
244
aitia de kantautha h te tn grammatn phusis kai h tn sullabn dunamis,

eks hn pleketai ta onomata
it is here too a matter of the nature of the letters
and the quality of the syllables, fromwhich
the words arewoven
In Greek and Latin grammatical tradition, the unit of the syllable is treated
as a kind of orthographic/phonological primitive; this strikes us as peculiar,
especially given that the Greeks and Romans, who typically equate orthographic units with phonological units, were writing alphabetically.44 Diodorus
operates within this tradition, though he views and treats the syllable as a unit
having component parts:45 mentioning (grammata) letters of the
alphabet, he remarks that syllables are (dia toutn
plekomenas) woven out of them46 (Comp. 15; and hence, he states, syllables
display the properties of their constituent letters).
Dionysius finds a parallelism in the weaving of letters, the weaving of
syllables, and the weaving of words (Comp.16):
;
,
,

Ti d to kephalaion esti moi toutou tou logou? hoti para men tas tn grammatn
sumplokas h tn sullabn ginetai dunamis poikil, para de tn tn sullabn
sunthesin h tn onomatn phusis pantodap, para de tas tn onomatn harmonias polumorphos ho logos;...
What is the main point of my assertion? That it is by the interweaving of letters that the variegated effect of syllables comes about, and by the combination of syllables arises the varied nature of words, and by the arrangement of
words comes manifold discourse.
In these remarkable lines Dionysius reveals to his readers two interpretative mechanisms. First, the twin axes of selection ( [eklog]) and
combination ( [sunthesis]) the warp and weft of linguistic and
literary fabric are operative at the level of letters ( [grammata];
Dionysius conflates phonology and orthography, as would be fully anticipated), at the level of lexemes ( [onomata]), and at the level of
245
discourse ( [logos]). Second, the weaving mechanisms operate iteratively to produce intricate variegated ( [poikilos]) patterns: the
interweaving ( [sumplok]) of letters through selection and combination produces the pattern of syllables; the output of this operation produces syllabic units that are selected and combined to weave word patterns;
the output of this operation, in turn, produces lexical units that are selected
and combined to produce linguistic utterances, either oral or written, as the
following chart illustrates.
LETTERS
selection
combination
SYLLABLES
selection
combination
WORDS
selection
combination
LINGUISTIC
UTTERANCE
We can see here a logical regression at work as a nascent tradition of linguistic analysis is applied to the deeply ancient Indo-European metaphor of
poetic word weaving. As woven poetic words lexemes were subjected to
analytic scrutiny, their component syllables were in a parallel fashion viewed
as participating in a process of syllable weaving and a syllables component
sounds, understood and described as letters through phonic-graphic syncretism, were likewise cast as strands providing the raw material for letter
weaving. This regression is only natural given that the dual axis of warp and
weft is a linguistic primitive. The result is the realization of a woven pattern
of greater variegated intricacy; in effect, what began as metaphor of poetic
composition evolves into a metaphoric expression of a fundamental linguistic reality.
246
7. 5
st. je r ome a n d alphabe t ic i n t e rweav i n g
In the twenty-fifth chapter of the biblical book that bears his name, the seventh/sixth-century b c Hebrew prophet Jeremiah proclaims the words of the
Lord the God of Israel that came to him (verses 8b9a, 15b27):47
Because you have not listened to my words, 9aI will summon all the tribes of
the north, says the Lord: I will send for my servant Nebuchadrezzar king of
Babylon. I will bring them against this land and all its inhabitants and all these
nations roundit;
8b
Take from my hand this cup of fiery wine and make all the nations to whom
I send you drink it. 16When they have drunk it they will vomit and go mad;
such is the sword that I am sending among them. 17Then I took the cup from
the Lords hand, gave it to all the nations to whom he sent me and made them
drink it: 18to Jerusalem, the cities of Judah, its kings and officers, making them
a scandal, a thing of horror and derision and an object of ridicule, as they still
are: 19to Pharaoh king of Egypt, his courtiers, his officers, all his people, 20and
all his rabble of followers, all the kings of the land of Uz, all the kings of the
Philistines: to Ashkelon, Gaza, Ekron, and the remnant of Ashdod: 21also to
Edom, Moab, and the Ammonites, all the kings of Tyre, all the kings of Sidon,
and the kings of the coasts and islands: 23to Dedan, Tema, Buz, and all who
roam the fringes of the desert, 24all the kings of Arabia living in the wilderness,
25
all the kings of Zamri, all the kings of Elam, and all the kings of the Medes,
26
all the kings of the north, neighbors or far apart, and all the kingdoms on the
face of the earth. Last of all the king of Sheshak shall drink. 27You shall say to
them, These are the words of the Lord of Hosts the God of Israel: Drink this,
get drunk and be sick; fall to rise no more, before the sword that I am sending
amongyou.
15b
The lines of the dire prophecy cited here make mention twice of Babylon,
ruled by Yahwehs servant Nebuchadrezzar, presented at the outset of this
pericope (verse 9)as an avenging agent and at the end (verse 26)as no less
a victim than the other kings. In the case of the second mention, however,
Babylon is identified by the name Sheshakh ( ;cf. Jeremiah 51:41). It is at
this point that Coogans previously quoted reference to Jerome vis--vis the
alphabetic pedagogical practice described by Quintilian becomes pertinent.
In his commentary on Jeremiah 25:26, St. Jerome, the fourth/fifth century
Croatian-born cleric, sets out an accounting of the Sheshakh denotation of
Babylon:
247
Apud nos Graecum alfabetum usque ad novissimam litteram per ordinem legitur,
hoc est, alfa, beta et cetera usque ad o, rursumque propter memoriam parvulorum solemus lectionis ordinem vertere et primis extrema miscere, ut dicamus alfa o, beta psi, sic et apud Hebraeos primum est aleph, secundum beth,
tertium gimel usque ad vicesimam secundam et extremam litteram thau, cui
paenultima est sin. Legimus itaque aleph thau, beth sin, cumque venerimus ad
medium, lamed litterae occurrit chaph; et ut, si recte legatur, legimus Babel,
ita ordine commutato legimus Sesach.
Among us, the Greek alphabet is recited all the way to the last letter in a
straight sequence, that is alpha, beta and so on, all the way to omega; also, for
the sake of children memorizing [the alphabet], we make it a practice to turn
the straight sequence of recitation backwards and to intertwine the final elements with the initial, so that we say alpha omega, beta psi. In a corresponding
way, among the Hebrews, the first letter is aleph, the second beth, the third
gimel, all the way to the twenty-second and last letter, taw, before which is shin.
Thus we recite aleph taw, beth shin, and when we make the turn in the middle,
lamed comes face to face with the letter kap: thus if it [i.e., the alphabet] is
read straight, we read Babel [i.e., Babylon], while with the alphabetic sequence
rearranged, we read Sheshach.
The lexemes and concepts are familiar from the earlier discussion of Quintilian
Institutio Oratoria 1.1.25: unmodified ord denotes the conventional linear
sequence of letters; Jerome writes of reading straight, recte, as Quintilian
refers to rectus contextus, linear sequencing. The aforementioned are linked
to the realm of weaving, as we have seen. Jeromes selection of miscere to
denote the intertwining of letters represents the appropriation of a term that
shares semantic space with the Greek verb (lugizd), the denominative
formed from (lugos) withe, discussed in Chapter6, comparable, I propose, to the copper-plaque verb (luzd). Certainly Jerome must also be
drawing on the standard vocabulary of alphabetic activity, a Latin vocabulary
that has its roots in a far older Greek tradition of performative writing and
alphabetic scholarship.
7. 6
wes t sem i t ic alphabe t ic i n t e rweav i n g
If Quintilians description of his envisioned pedagogical process allows for the

possibility that it is one that may somehow differ in specifics from the process
we see displayed in Pompeian graffiti, Jeromes does not, mutatis mutandis.
248
Jerome describes the interweaving of the letters in terms of Greek alphabetic

tradition, rather than Latin, and notes that a comparable phenomenon is practiced apud Hebraeos among the Hebrews. The intertwining of the twenty-two
letter Hebrew script that he describes takes the followingform:
(5)
(6)
Jerome writes that when we make the turn in the middle [i.e., after the eleventh letter kap ( ;)see (5)], lamed [ ]comes face to face with the letter kap
[]. What he means by this can be seen clearly when we examine the backand-forth segments in continuous (left-to-right) intertwined sequence (where
the first half of the abecedarium is indicated by plain text, the second half by
underlining):
(7)
The only two letters that retain their contiguous placement are the middle
letters kap (eleven of twenty-two) and lamed (twelve of twenty-two) still situated face-to-face but located at the coda of the intertwined abecedarium,
rather than at the middle of the straight letter sequence. In the intertwined
abecedarium, bet ( )and shin ( )occur face-to-face, as, of course, do kap ()
and lamed():
(8)
Substituting for the consonants of Babel ([ )i.e., Babylon]), the respective

face-to-face pairings in the intertwined abecedarium (shin [ ]for bet [ ]and
kap [ ]for lamed [ )]produces Jeremiahs Sheshakh ().48
This substitution process is an otherwise known practice of Hebrew
cipher spelling called atbash, named after the pairing of alep ( )with taw ()
(hence at-) and bet ( )with shin (( )hence -bash). Earlier scholars had been
dubious about Jeromes claim that Sheshakh is an atbash representation of
Babylon, but the traditional interpretation is generally accepted today, not
only because all other suggestions have proved to be wrong, but also because
cryptographic writing of personal names is attested in Mesopotamia as
early as the seventh century b. c . e . points out Steiner, quoting Tigay 1983.49
Steiner continues, in this tightly argued study of the recensions of Jeremiah:
The atbash cipher can no longer be dismissed as a fanciful practice [a reference to a characterization by George Rawlinson in the mid-nineteenth
249
century] and there is no good reason to doubt that it existed in Jeremiahs

time.50
Other Hebrew cipher spellings are also attested, such as the so-called
albam system, involving an interlacing of letters by plaiting each half of the
abecedarium from its starting point, rather than a weaving back of the second
half; in other words:
(9)
(10)
Thus, alep ( )pairs with lamed (), bet ( )with mem () providing the denotation albam and soon.51
There is some evidence that the atbash and albam systems of abecedaric
interlacing were in use in Syria-Palestine as early as circa 1200 b c . Among
the numerous letters appearing in the first four lines inscribed on the Izbet
arah ostracon, Demsky has called attention to pairs that follow the linear
(i.e., alep through taw) letter order and its reverse, others that appear to be
pairs in atbash order, others pairs in albam order and its reverse.52 Line 5 consists of the entire abecedary written left to right, though with certain peculiarities, as noted in Chapter5 (see note64).
7. 7
alphabe t ic i n t e rweav i n g a n d d i v i sio n
Coogan has drawn attention to other evidence from Syria-Palestine, Greece,

and southern Italy that suggests a common practice of dividing the alphabet
at its midpoint, the basis for the interweaving of letters seen at Pompeii and
implicit in Hebrew atbash pairings. Noting the attestation of three abecedaria
at the later second-millennium b c Syrian city of Ugarit, Coogan observes that
two of the three (PRU II, nos. 184 and 188B, lines 45) are divided at l: they are
written on three lines, and on both the first line contains the signs from a
to l:53 the division is made after the fourteenth of twenty-seven letters.54 The
third Ugaritic abecedarium seems to show a division required by the shape
and size of the tablet on which it is inscribed. An ostracon from Qumran,
circa 30 b c , apparently a students practice text, bears four lines of letters, the
middle two inscribing an abecedarium: the initial line of the abecedarium
(line 3)ends in et (), the ensuing line (line 2)begins with lamed (), thus the
intervening yod ( )and kap ( )are missing; Coogan remarks: Although there
is sufficient space for y and k either at the end of line 3 or at the beginning
of line 2, they have been omitted, presumably because of a memory lapse.
250
But the student apparently remembered that the second half of the alphabet
began with l, m, n, and so made a fresh start, as it were, with line 2.55 In other
words, regardless of whether the omission of y and k (yod and kap) were
accidental or not, this students exercise stands as an alternative reflection of
the practice described by Jerome as the turn in the middle of the interwoven
abecedarium.
Coogan mentions three instances of divided abecedaria found in Greece
and Magna Graeca. One is an abecedarium painted on a stamnos from the
Akhaian colony of Metapontion in the south of Italy, circa the first half of the
fifth century b c .56 Regarding its script, Jeffery writes:
Important evidence for the duration of the local alphabet at Metapontion
should be given by the abecedarium which was painted round the shoulder of
a stamnos of local type, found in a grave near the city. Gamma, delta, iota,
san are shown in their local forms; vau [digamma] and qoppa are still in place,
but the unused sign sigma is not represented. Its place in the line is taken by
san. The complementary [i.e., supplementary] letters run: upsilon, phi, red
chi, and red xi written twice. The repetition of this sign at the end has been
variously explained, as a means of filling the vacant space, or as an indication
that the Metapontines were aware that a with the value of chi existed as well
as the =.57
On one side of the rim twelve letters appear, running from alpha to lambda;
on the other side are painted the remaining thirteen letters of this alphabet,
with the pots handles separating the two alphabetic segments. That the dipinto preserves an expression of an archaic concept of the alphabet existing
in two halves is in this instance unclear: one might suspect that the painters
decision to divide the alphabet in the way that he does was simply dictated by
the (approximately) symmetrical geometry of the available space, interrupted
by the two opposing handles. On the other hand, the repetition of the final
symbol (i.e., red xi) could be viewed as an indicator of the scribes unwillingness to divide the alphabet at some place other than between lambda and
mu and thus as a reflection of a particular (local) form of the two-alphabets
doctrine: had he made the division between kappa and lambda, as, mutatis mutandis, seen at Pompeii and in the abecedarium from Qumran, and as
perhaps suggested by letter pairings on the Izbet arah ostracon, the result
would have been a symmetrical arrangement of twelve letters on one side
and twelve on the other without having to fill the latter space with a repeated
redxi.
251
More interesting is a second Greek example that Coogan mentions a

double abecedarium painted on a cup from Boiotia, of uncertain provenance,
dated to the second half of the fifth century b c .58 The abecedaria again run
around the rim of the pot, and the field in which they are painted is again
bisected by opposing handles. In this case, however, a complete abecedarium
appears on each side of the cup, divided into two horizontal rows showing an
albam-sort of arrangement. On one side, the abecedarium consists of twentythree letters, alpha through upsilon without san or qoppa plus red xi (with
a xi-grapheme not occurring in its internal, that is Phoenician, position), phi,
and chi; division is made after lambda, the twelfth letter, hence producing an
initial line of twelve characters and an ensuing line of eleven. On the other
side, the abecedarium consists of twenty-five characters: the sequence of letters is the same but with two additional characters appended, which Jeffery
describes as an attempt at the Ionic forms [psi] and [omega] at the
end of the stoichos.59 In this case, the division occurs after mu, the thirteenth
letter, giving thirteen characters in line one and twelve in line two. The point
of division thus in one case matches that of the dipinto from Metapontion
(between lambda and mu); in the case of the other, however, the case of the
abecedarium with extra supplementals, division is forwarded to the next possible position between mu and nu. In so doing, the scribe has given the two
abecedaria a parallel form in that the first line in each instance contains one
more letter than the second. Again, it is unclear if an archaic concept of two
alphabets, rather than considerations of decorative symmetry, may be principally responsible for the divisions observed.
Most important and crucial of Coogans examples and much the earliest for the archaic notion of two alphabets, I believe, is the Etruscan abe
cedarium on the brown impasto goblet from Narce, dated circa the second
quarter of the seventh century b c , preserving the Euboian alphabet, which
has been mentioned several times previously.60 It is a partial abecedarium
extending from alpha to kappa and attests the xi-symbol in the alphabetic
position of eta, exactly as found in abecedaria of the copper plaques. The partial abecedarium is situated neatly on the face of the goblet, and it is clear that
its half length is not the consequence of damage to the piece.
The practice of plaiting the two alphabetic strands in atbash fashion as
attested at Pompeii and by Jerome and, likely, by Quintilian may also be bound
up with the etymological origin of the Latin word for letters, elementa. This
is a central point in Coogans insightful article. Arguments have long been
offered in favor of deriving the term from the first three letters of the second
252
half of the Latin alphabet l, m, and n: A great deal of ingenuity has been
expended in the effort to find a derivation for elementum. It seems to me,
however, that the old one from el em en, which has been rejected as too simple and natural for science, is really the only one which ought to be thought
of, wrote J. B. Greenough in the last decade of the nineteenth century.61 He
continues: In the first place, throughout Latin literature, from first to last
the word means A B Cs, literally. Greenough argues chiefly from the earliest
attestation of the word, in Lucretius 1.196 (and in ensuing lines), and Ciceros
usage of the term in Academica 1.7.26, where he indicates that elementa is a
translation from Greek (ut e Graeco vertam), and he undoubtedly has in mind
letters; components.62 Greenough concludes:
Inasmuch, then, as elementa is distinctly a translation of , it seems
almost certain that it had the meaning of A B Cs, and not any other more
abstruse meaning. As to the form, it must, it seems to me, be for el-em-ena, a
plural like A B Cs. As the tendency in the language increased to substitute the
longer forms in mentum for those in men (as in momen, momentum), this
word also went with the rest, and became elementa in the same meaning.63
And Coogans modest conclusion on this point is that in view of the conservative nature of alphabets the examples may reflect a Semitic pedagogical practice continued in the West which resulted in the second half of the
alphabet being called elementum. The entire alphabet then could be called the
elementa.64
7. 8
g r eek alphabe t ic i n t e rweav i n g a n d t he

c o ppe r plaq u es
The alphabetic traditions evidenced in the copper plaques show unmistakable

traces of also being subject to this particular form of interweaving, or plaiting,
if the atbash sequence itself finds no full expression in the plaque abecedaria.
7.8.1 a geometric subset within the

interwoven alphabet
Like the Hebrew script, the alphabets of the copper plaques consist of twentytwo characters, matching the Hebrew system letter for letter both being of
Phoenician origin. A sketch of how a similar weaving of the twenty-two-letter
copper-plaque alphabet would look is shown in (11) and(12):
253
(11)
(12)
The back-and-forth steps of the two alphabetic strands are precisely those
that are illustrated for Hebrew in (5) and (6), as described by Jerome and
well known from Hebrew atbash usage, repeated here as (13) and (14) for
comparison:
(13)
(14)
The form that a continuous interwoven copper-plaque sequence would take
is shown in (15), where the first half of the abecedarium is again indicated by
plain text and the second half by underlining:
(15)
Its, mutatis mutandis, identical Hebrew atbash counterpart is repeated as(16):

(16)
The interlaced Greek sequence of (15) reveals, I propose, the source of certain morphological characteristics of archaic Greek alphabets that we have
encountered in the preceding chapters. Most notably, we can see that the
plaited sequence omicron, eta, xi, theta, the underlined sequence in (17),
forms a mirror-image, symmetrical geometric subset { h } within the
plaited alphabet:
(17)
The first pair ( h) being distinguished from the second ( ) by the presence
of crossing strokes in the second essentially an unmarked : marked binary
pairing. Stated slightly differently: h and are mirror-image atbash pairings of an unmarked and marked pair. It is this subset arrangement, an accidental secondary consequence of the plaiting of the alphabet that gives rise, I
would argue, to the playful scribal interchange of the two square symbols, eta
(h) and xi (), and of the two round symbols, theta () and omicron (), that
allows them to function as allographs with which the scribe can weave his variegated alphabetic fabric. These interchanges are, as we saw, well attested beyond
the confines of the CP abecedaria: the interlacing of the halves of the alphabet
must date to an early period in the history of archaic Greek literacy; this atbash
254
practice was most likely taken over as a part of the process of the Greek adaptation of the Phoenician consonantal script, perhaps used for pedagogical purposes (or as a cipher?) from the outset of the Greek alphabets creation.
Let us further consider the plaited (i.e., atbash) subset { h }. These
four symbols are actually related by two sets of binary oppositions. These
oppositions can be identified using the features [ round] and [ crossed].
In the case of [ round], the + value denotes a rounded periphery and the
value a rectangular periphery. In the case of [ crossed], the + value denotes
the presence of crossing strokes dividing the interior of the symbol into quadrants or quarter circles that are absent from the graphemes valued as; the
internal geometry would be the same were the free variant given preference
in the analysis. On the basis of these binary oppositions, the four graphemes
are characterized as follows:
(18)
omicron
+ round
crossed
eta
h
round
crossed
xi
round
+ crossed
theta
+ round
+ crossed
Jakobson reminded us earlier of Honor de Balzacs dictum: Tout est bilatral dans le domaine de la pense. Les ides sont binaires. The accidental
co-occurrence of four such graphemes in interwoven sequence must certainly
have provided an ineluctable cognitive invitation to automatic binary analysis;
and the responsive scribal playfulness preserved in the copper-plaque abecedaria thus took the form of modifying the value of the extra stroke feature
while keeping the round feature constant giving theta the morphology of
omicron and omicron that of theta, eta the morphology of xi and xi that ofeta.
There is more to this matter, however. The binary opposition displayed in
the interwoven subset { h } is itself an innovation, whether the consequence of intentionality or of a cognitive predilection for local sameness.
As we saw in Chapter2, the archaic xi-symbol represents a modification
of its Phoenician precursor samek (s) and is of limited, though interesting,
distribution. Before the emergence of this symbol (), the interwoven subset would have been of the form { h x }. It seems clear that the inherited
xi-symbol x was assimilated to the peripheral shape of its atbash neighbor eta
(h) and thereby acquired a symbol-internal quadrantal geometry like that of
its atbash neighbor theta () both assimilations being accomplished by the
addition of vertical lateral strokes, producing : in this way the two rectilinear
255
symbols were brought into structural parallelism with the two curvilinear
symbols that they abut; and, hence, the binary opposition of [ round] and
[ crossed] was realized.
We also observed in Chapter 2 that the symbol occurs in the local alphabet
of Naxos with the value /k + s/. The advent of this xi-symbol, I would argue,
must like the origin of the xi-symbol lie in an atbash plaiting of the alphabet, but one with an ultimately different expression of assimilation. Beginning
with the subset { h x }, the most likely most natural initial step would be
the assimilatory shift to the subset { h }, producing the binary opposition
of [ round] and [ crossed]. An additional assimilatory change must have been
driven by the asymmetry of the graphic relationship that obtains betweenthe
two [+ round] members of the subset vis--vis that which holds between the
two [ round] members: omicron and theta ({ }) are distinguished by the
presence or absence of an internal division the former symbol produced with
no internal linear strokes, the latter with a crossed internal linear strokes; eta
and xi ({h }), in contrast, both show internal division, but distinguished by
the presence or absence of vertical bifurcation: the former symbol is produced
with a single horizontal internal linear stroke, the latter with crossed internal
linear strokes. Assimilation of the latter graphic opposition (that of {h }) to
the former graphic opposition (that of { }) produces a further change in the
form of xi, giving the symbol , which like omicron () lacks any internal bifurcation. The resulting subset, { h }, is thus one in which the four members
{A B C D} stand in an analogical relationship A : D :: C : B; stated differently,
xi is in this process graphically assimilated to a symbol (omicron) with which
it is contiguous in the periodic order of the alphabet, whereas in the former
process xi is assimilated to a symbol (theta) with which it is contiguous in the
plaited atbash order. This relationship imposes a binary opposition on the subset involving the feature [ round], as with the earlier subset; but in this instance
the second feature is one that entails simple graphic internal division of the
symbol its presence or absence: we could call the feature [ divided]. On the
basis of these binary oppositions, the four graphemes constituting the interwoven atbash subset { h } are thus characterized in thisway:
(19)
omicron
+ round
divided
eta
h
round
+ divided
xi
round
divided
theta
+ round
+ divided
256
The two assimilated subsets, { h } and { h }, thus have slightly different feature matrices, {[+] [] [ +] [+ +]} and {[+] [ +] [] [+ +]},
respectively. The graphic simplification of to effectively reverses the
feature matrices of the two middle (i.e., rectangular) members of the subset,
creating two mirror-image contrastive sub-subsets {[+] [ +]} and {[] [+
+]}, and thus internalizing a binary opposition within the subset relationship.
This outcome very likely provides additional support for the interpretation
of as a further graphic modification of , whether or not it suggests intentionality, or an autonomic cognitive process.
The same sort of scribal weaving play that gave rise to the interchange of
the two square symbols, eta (h) and xi (), and of the two round symbols,
theta () and omicron (), in copper-plaque abecedaria, and in various local
alphabets as described in Chapter2, must also be at work in the case of the
further modified xi-symbol . Its position in the subset { h } of the
plaited (i.e., atbash) abecedarium, situated adjacent to the structurally similar
eta-symbol h, led to the use of itself as an eta-symbol, representing a long
mid-vowel or the glottal fricative /h/ in alphabets of Aegean Naxos, Knidos,
Kyme, and Sicilian Naxos.
7.8.2 the n u - io ta- m u subset within the woven

alphabet
The interweaving of the alphabet in atbash fashion, I would argue, also underlies at least three other phenomena that we have encountered in the preceding
discussions one that is attested within the abecedaria of the copper plaques,
and two that are not. In Chapter2, I drew attention to what I characterized as
an odd state of affairs: some of the instances of Iota-2 in the copper-plaque
abecedaria are morphologically very close to forms of archaic Greek nu, as
used, for example, on the Dipylon oinochoe or the Mantiklos statuette, and
so provide close matches to the corresponding forms of Phoenician nun. Even
more curious is that some instances of Iota-1 are similar to somewhat different forms of Phoenician nun, such as the nun of the eighth-century Cypriot
jug of nt and of the late-ninth century Kilamuwa inscription from Zenjirli.
This state of affairs is made all the more odd by the fact that forms of iota in
the copper-plaque abecedaria are distinctly different from the forms of nu
that occur on the plaques. In that earlier discussion, I suggested that CP Iota-3
may continue the form of earliest Greek iota, a bivalent symbol created to
257
spell both phonemic /i/ and nonphonemic [y] and arbitrarily assigned a morphology distinct from that of its Phoenician protoform yod and in this way
the character would provide a functional and structural parallel to the case of
digamma. I went on to speculate that Iota-1 and Iota-2 perhaps evolved as the
alphabet continued to take shape in a milieu of Greek-Phoenician interaction.
There was a seeming fly in the ointment of this scenario, however: the evolving
morphology of Iota-1/2 bears only the grossest of similarity to Phoenician yod
(spelling /y/) and is remarkably closer, as we have just reminded ourselves, to
forms of Phoenician nun (spelling /n/). Is there any sense to be made ofthis?
There may in fact be a sensible solution, and this solution lies in the interwoven Greek alphabet of atbash form. As can be seen in (20), iota is plaited
so as to fall between the mu and nu symbols of the strand that is the second
half of the alphabet:
(20)
The several morphologies of CP mu and nu parallel one another, as discussed

in Chapter2. The nu of the copper plaques is, however, almost unique among
forms of Greek nu in being produced with four (rather than three) strokes
and, in this regard, has no Phoenician counterpart. The nu-like morphology
of CP Iota-1/2, I would suggest, arose by graphic assimilation of the character
to its neighboring graphemes in atbash sequence its neighbors, mu and nu,
being, again, morphologically parallel to one another, distinguished by the
presence or absence of a fifth stroke. As iota assimilatorily encroached upon
the graphic form of what must have been originally a three-stroke nu, given
its Phoenician model and near ubiquity among Greek alphabets, the scribal
response was to dissimilate that nu by the addition of a fourth stroke. This is
of course a dissimilatory response within that alphabetic tradition that is preserved in the copper plaques conceivably elsewhere, but not a general one:
four-stroke nu appears to be otherwise limited to the Euboian-based Etruscan
alphabet seen on the bucchero bottle from Caere (and not morphologically
identical to CP nu). But, in contrast, the morphological melding of the earliest Geek iota to the two nasal threads with which it is interwoven in atbash
sequence (mu and nu) is a broad one if I am right in seeing CP Iota-1/2 as parent of the well-attested crooked iota (as discussed in Chapter2).
Scribal play is certainly at work in this process. Such playfulness reveals
itself in the set of assimilatory and dissimilatory graphic changes occasioned
by the interweaving of alphabetic strands. But beyond this, the formal similarity of at least some instances of Iota-1 to particular forms of Phoenician nun
258
from Cyprus and Zenjirli, distinct from the typical Greek nu (that symbol
that has been characterized as having no marked local variations),65 suggests scribes were reaching outside of the Greek alphabet for a nasal graphemic model for their remadeiota.
7. 9
g r eek alphabe t ic i n t e rweav i n g beyo n d

t he c o ppe r plaq u es
The second of these aforementioned phenomena that I would hold to be products of the atbash plaiting of the alphabet, is, as noted earlier, not one attested
within the abecedaria of the copper plaques. It is one, however, that we have
encountered in our examination of these letters.
7.9.1 dotted om ic r o n
In the discussion of theta and omicron homography in Chapter2, I drew attention to dotted theta and dotted omicron and their co-occurrence in Argive,
Kyrenaean, and Theran inscriptions. Dotted theta occurs already on the early
seventh-century b c Boiotian Mantiklos statuette. When a cutting-compass
has been used, it is possible to explain an early example of dotted theta as
due only to the masons forgetting to add the cross; but obviously this cannot
always be the reason. The dotted theta was probably first evolved by those
writing rapid script with a brush, writes Jeffery.66 Concerning the view that
dotted theta arose as a reduced version, Guarducci concurs, as does McCarter:
a very early simplification of the crossed-diameter types.67
Dotted omicron is a different matter, however. As noted in Chapter 2,
Phoenician ayin is the ancestor of Greek omicron, and a dotted ayin (preserving the pupil of its ancestral, iconographic Canaanite eye-symbol) is found in
Phoenician scripts before the tenth century b c , but not later. Thus the early
(eighth and seventh century b c ) attestation of dotted omicron at Thera68 and
in Etruria69 has translated into a point of contention in the scholarly debate
over the date of the Greek acquisition of the Phoenician script, creating an
inelegance for what would seem to be an otherwise reasonable dating of the
Greek adaptation to the ninth or eighth century b c . McCarter summarizes
the awkwardness:
The dotted omicron is indeed one of the great surprises of the Greek scripts.
it is an unmistakable archaism, reminiscent of the old Canaanite pictorial
259
representation of an eye. The most numerous Semitic examples are in fact to

be found in the period preceding the development of the national scripts, that
is, in the twelfth and early eleventh centuries. Our reexamination of the early
Phoenician scripts has shown that dotted ayin survived no later than the very
early tenth century. The gap between the dates of these inscriptions and
the year 800 is considerable. It is difficult to suppose that dotted ayin survived for two centuries more without making a single appearance in surviving
inscriptions.70
In 1979, subsequent to the penning of these words, this span of silence

was somewhat diminished by the discovery in northern Syria of the Tell
Fakhariyeh inscription, a bilingual Akkadian-Aramaic document dated to the
mid-ninth century b c .71 The script of the Aramaic inscription is highly idiosyncratic, judges Kaufman.72 The occurrence of dotted ayin in the Aramaic
inscription places that symbol within striking distance of the time of origin of
the Greek alphabet. The responses of those, such as Naveh and Bernal, advocating a second-millennium b c Greek acquisition of the Phoenician script,
have countered by proposing that the Tell Fakhariyeh script is an archaizing
one.73 Even if that were the case, it would not obviate the fact that there was
a living awareness of dotted ayin in the ninth century b c though in an
Aramaic context. Those who would place the acquisition in the second millennium b c would presumably respond by stating that although an awareness
of the symbol existed in the ninth century, the distribution and frequency
of occurrence of dotted ayin was certainly limited, given the uniqueness of
this ninth-century attestation, and so less likely to have provided a model for
Greek adapters in the ninth or eighth century. The retort would then likely be
that there is a general dearth of Phoenician inscriptions from this period we
possess what must be only a tiny fraction of the inscriptional output and any
single example of Phoenician or Phoenician-derived writing must then weigh
heavily. And soon.
This debate may well be misplaced. The occurrence of dotted omicron
in archaic Greek spelling quite likely has nothing to do with Phoenician or
Canaanite dotted ayin. Given what we have seen in the foregoing discussion
of the plaiting of the alphabet in atbash fashion, the crafting of an omicron
with a dotted center is likely the product of scribal weaving play. Just as the
xi-symbols and were shaped as a consequence of alphabetic interweaving and came ultimately to be used for eta-symbols in some local alphabetic
systems, undoubtedly via playful substitution of the sort evidenced in the
260
copper plaques, so dotted omicron was shaped and entered the spelling system of certain epichoric alphabets. An interwoven archaic alphabet in which
the theta symbol is dotted rather than crossed, as already in that evidenced on
the Boiotian Mantiklos statue, perhaps dating as early as late eighth century,
would have the following idealizedform:
(21)
h x
The crucial interwoven subset {omicron eta xi theta} is now of the form
{ h x }. Scribal manipulation of the lateral [+ round] members would,
as one permutation of the play options, result in the utilization of a single
[+ round] form for both theta and omicron. A fixed expression of this play is
to be seen preserved in the mid-fifth-century Argive inscription recording a
treaty between Knossos and Tylissos, discussed in Chapter2, in which dotted theta and dotted omicron routinely co-occur. The same co-occurrence of
with the value of both the theta-consonant and the omicron-vowel is evidenced, though less regularly, in the Cretan version of the treaty, as we saw, as
well as in the archaic Theran script and in the alphabetic tradition of the Theran
colony at Kyrene. This co-occurrence is paralleled, as noted in Chapter2, by
the use of the theta-symbol to spell both theta and omicron (i.e., a crossed
omicron paralleling dotted omicron) in the archaic rock graffiti from Thera,
as well as in the early fifth-century inscription from the Argive Heraion in
which (iaromnamones) is spelled as . Recall
that this inscription appears to be the work of that same engraver who produced the Tanagra inscription from Argos that attests the use paralleling the
practice of the copper plaques of both and h for eta (spelling/h/).
In sum, Greek dotted omicron is almost certainly the product of scribal
interweaving. Its origins lie in those same scribal phenomena that produce
other attested theta / omicron and eta / xi interchanges. The presence in the
Greek alphabet of a dotted omicron very likely provides no evidence to support a second-millennium b c Greek acquisition of Semitic script.74
7.9.2 square t he ta and om ic r o n

The third phenomenon involves the square theta () and square omicron
() symbols that were examined in Chapter2. The discussion there focused
on the occurrence of these symbols in the inscription found on the rim of a
bronze lebes (ca. 600550 b c ) from Delphi and on their appearance in the
recently discovered Barako abecedarium from Attica (ca. 550 b c ); still other
261
examples were cited, including that of a seventh-century graffito from Mount

Hymettos and a sixth-century Corinthian bronze helmet.
As I remarked in Chapter2, square theta and square omicron are additional
participants in a playful scribal homography. More generic interpretations of
geometric reconfiguring (i.e., of the squaring of curvilinear characters under
the influence of their rectilinear neighbors) could be formulated, given the
interwoven atbash sequence of round and square characters that is crucially
involved (omicron, eta, xi, theta), but let us again begin with the archaic subset
{ h } that we have already seen to be at work. The archaic symbol is
pressed into service for representing not only xi but each of its atbash neighbors, eta (h) on its left as we already saw and theta () on its right: with
the former, the xi-symbol shares its peripheral geometry ([ round]); with
the latter, the xi-symbol shares its internal geometry ([+ crossed]). The use of
the symbol as a square theta, attested as early as the seventh century b c ,
is a conventionalized synchronic expression of alphabetic play anchored in
diachrony a synchronic instantiation of scribal play unfolding through the
diachronic dimension.
The same state of affairs holds fundamentally for square omicron. If I am
right in seeing the symbol as primally a xi-symbol (i.e., occupying the xiposition in the archaic subset { h }), its conscription for use as an expression of omicron differs structurally, however, from that of the metaphoric
application of the xi-grapheme , which substitutes for its two contiguous
neighbors eta and theta in atbash sequence, as described in the preceding
paragraph. In the instance of , the symbol is pressed into use for the curvilinear and the rectilinear characters that precede it in atbash sequence that
is, omicron andeta.
One might, however, imagine a stepwise process in which was reinterpreted synchronically as fundamentally an eta-symbol and its use as a xisymbol was then correspondingly viewed as secondary: the extension of
to omicron as well (as to xi) would then structurally recapitulate the use of
for eta and theta. It is this parallel metaphoric relationship that arises from
the interwoven alphabet that I anticipated when in Chapter2 I wrote that the
xi-grapheme a second-half-of-the-alphabet symbol can serve not only
in the eta-position, but in the theta-position as well two contiguous letterpositions in the first half of the alphabet. Conversely the eta-grapheme
a first-half-of-the-alphabet symbol can serve not only in the xi-position
but in the omicron-position as well two contiguous letter-positions in the
second half of the alphabet.
262
7. 1 0
at t he j u n c t u r e o f t he alphabe t ic
su bs t r i n gs
It was noted in preceding sections that the broadly attested process of alphabetic interweaving subsumes a notion of internal alphabetic division and
that Coogan cites evidence for such periodic subcategorization from Ugarit,
Qumran, Boiotia, and Magna Graeca: the evidence suggests that there is one
alphabetic substring that begins, of course, at the beginning, that is, with the
A-symbol, and that there is a second that typically begins with the L-symbol.
Coogan also suggests, as others had before him, that Latin elementum finds
its etymological origin in the first three letters of the second of the periodic
series within the alphabet: l, m, andn.
In Chapter 2, I called attention to the fact that within the abecedaria of
the copper plaques, it is kappa which can be viewed as the last letter in the
first alphabetic periodic series that is the most frequently omitted letter, left
out of eleven of the abecedaria, and that no other letter comes close to this
frequency of omission except for the ensuing letter, lambda the first letter in
the last (i.e., second) alphabetic periodic series which is omitted ten times.
Coogans observation of the Qumran abecedarium that omits the sequential
letters yod ( )and kap ( )compares interestingly, as do the abecedaria from
Metapontion and Boiotia that he mentions. As pointed out, the latter might,
one could speculate, preserve variant local traditions in which the first periodic series terminated with lambda rather than kappa.
I also noted that various other anomalies in the copper-plaque abecedaria
center around what we could now identify as the kappa-lambda juncture,
the meeting point of the two periodic substrings. Specifically, the sequence
kappa-lambda is absent five times (on MS 2), while kappa and lambda are
inverted once (MS 1-1), and lambda and mu are inverted once in conjunction with an omission of kappa (also MS 1-2). Two of the five omissions of
the kappa-lambda sequence occur within a larger anomaly: the omission of
the string kappa-lambda-mu-nu (both on MS 2-1); and the sequence kappalambda-mu is also omitted once (on MS 2-1 as well).75
These conspicuous omissions at and around the kappa-lambda juncture
by the scribes of the copper plaques must be of some significance, perhaps a
particular synchronic instantiation of a diachronically perpetuated alphabetic
phenomenon that likewise shows itself in the student exercise from Qumran,
in which case there is an absence of yod and kap despite the fact that there
is ample writing space for containing the characters. Coogan imagines that
263
the omission is here the consequence of a memory lapse on the part of the
writer, and perhaps such a processing error is indeed the consequence of the
nonexpression of the two symbols with which the initial alphabetic substring
terminates. But there must be something about the juncture of the two substrings of the alphabet the turning point of the abecedarium and the elementum that makes it conspicuous (processually or otherwise) that makes it
conspicuously available for scribal playfulness.
7. 1 1
t he wov e n alphabe t ic t e x t
The alphabetic crafting of the scribes of the copper plaques reveals a

metaphoric interpretation of producing alphabetic text the textus woven
fabric; the product of joining words. One of these scribes invokes his engraving tool ( [ml]) to weave ( [luzd]) the alphabet ( [i.e., abcd]).
The weaving of the alphabet through the production of the abecedaria of the
copper plaques manifests itself in two different ways. Most conspicuously,
the repeating lines of alphabetic text constitute the warp and weft of a highly
variegated graphemic fabric: the weft, the horizontal weaving dimension
cognitively, the syntagmatic axis of this multidimensional process of alphabetic fabrication, presents itself as the continuous sequencing of letters in
their periodic order, running repeatedly side to side across the loom of the
plaques; the warp, the vertical dimension cognitively, the associative or the
paradigmatic axis presents itself as the ongoing substitution of allographic
variations at the individual letter positions of the repeating abecedaria. But
there is also represented in the fabric of the copper plaques the depth dimension of the weaving process that of the passing in and out of threads as they
are plaited over and under one another. This depth dimension of the fabrication of the alphabetic text is provided metaphorically by the atbash practice
of interweaving the two halves of the alphabet that is implicitly assumed by
the scribes as is revealed by their explicit playful substitution for one another
of members of the geometrically salient atbash subset {omicron, eta, xi, theta}
and by the assimilatory and dissimilatory relationships exhibited by the members of the {nu, iota, mu} subset, as well as by the omission of one or more
characters at the turning point.
This last-named scribal process of a Greek plaiting of the dual alphabetic
strands a first half and a second half almost surely had its origin in Semitic
tradition and so would likely have been acquired by the Greek adapters of the
Phoenician script as an accompanying graphemic accoutrement. Beyond this,
264
the Semitic practice may have served as a metaphorical catalyst, but atbash
intertwining is only one dimension of the multidimensional process of alphabetic weaving which leads one to ask: Whence came the broader notion of
weaving an alphabetic warp and weft?
The metaphoric notion of the warp and weft of language was certainly a
fundamental one in archaic Greece, both before and after the acquisition of
alphabetic literacy. Verbal weaving is undeniably known to Homer, regardless
of what sense one makes of the absence of an explicit self-reference to poetic
weaving in the form of the Homeric epics in which we have them. The notion
of poetic weaving is well attested in lyric, equally as old as epic, and preceded
by a more ancient, broadly attested, ancestral Indo-European tradition of the
weaving of poetic words.
This interweaving of alphabetic warp and weft is a metaphorical extension of that weaving of poetic speech which is oral poetic composition and performance. It is an appropriation of the Homeric metaphor of weaving as a
description of an intellectual process76 to co-opt Snyders words rehearsed
at the beginning of this chapter. But also, it is a cognitive extension of the
structure of language itself. Language has both a horizontal and a vertical
dimension. Its horizontal dimension is what, as we have examined in some
detail, Saussure called its syntagmatic structure: in production, language
unfolds according to a permissible linear sequencing of linguistic elements.
The vertical dimension of language, in Saussurian terms, is its associative, or
paradigmatic, structure: in production, variety of linguistic expression arises
as linguistic elements are able to substitute for like linguistic elements the
phoneme /b/ substitutes for /t/ creating a distinction between bar and tar; dog
substitutes for cat; runs substitutes for jumps; and soon.
7. 1 2
t he pe r f o r ma n c e o f t he alphabe t
The notion of the weaving of the alphabet must have arisen as Greek speakers
inevitably and almost immediately equated their language with their newly
introduced alphabetic writing system, following the dictates of the human
cognitive condition, as we have examined. The weaving of language was a
phenomenon of composition in performance the production of poetic
speech. The weaving of the alphabet was a reflected image of the phenomenon of orality, of composition in performance the production of written
speech even written poetic speech.
265
We have caught glimpses of the weaving of the alphabet and of the metaphorically associated phenomenon of writing as performance throughout
this study, if they have passed mostly without remark. The graffiti from Mount
Hymettos, ranging from circa 700 to early sixth century b c , offer a principal example of alphabetic performance. As noted in Chapter2, a number of
the inscribed sherds bear dedications to Zeus, or indicate that they belong
to Zeus77 the Zeus of Mount Hymettos, who is distinguished by the epithet
Semios, derived from (sma) sign; several of the offered vases bear abecedaria;78 others are etched with a self-referential claim to the act of inscribing.79 In their present state, sherds of the latter group attest, in the simplest
case, only the single verb (egrapse) (s)he wrote or (egrapsa)
I wrote. Langdons inscription 29 is restored from three fragments to read [
] . [- -] [] ([to Di]os eimi. [- -]as de m egrap[se]n)
I belong to Zeus. X wrote me.80 Another reads h (hosper
egrapsen) as he wrote,81 and yet another - <>[ - ]
(-ai tad autos eg<r>ap[se - ]) X wrote this himself .82 Langdon surmises:
Writing must have been still so new that its accomplishment was being
stressed. He emphasizes the aspect of novelty with regard to the inscribing
of abecedaria as well: Another manner in which a votary could display his
knowledge of writing was by scratching the alphabet. But only at a time
when writing itself was new would abecedaria have been considered appropriate dedications for a deity.83
In the absence of comparable practice in later periods, and at other locales,
there would clearly be some sense in which the relevant nascent moment in
the history of the Attic alphabet is tied to the alphabets cult use on Mount
Hymettos.84 But simply because the alphabet is a new thing, does that make
it a compelling offering for the deity? It is not the novelty of the alphabet in
and of itself that makes it suitable for presentation to Zeus Semios.
Langdon notes that some of the inscriptions from Hymettos were produced
after the pots on which they were etched had already been broken, calling
particular attention to the second inscription in his catalog (H 232), scratched
on an unglazed interior surface that would have been inaccessible while the
pot was intact: (Smioi Di), to Zeus Semios. Now a single, broken
potsherd would seem to be a quite unlikely gift to dedicate to a deity: an apt
observation. Thus Langdon surmises (emphases in italics are myown):
The worshipper would believe that a specimen of this new skill which allowed
him to express in visible and permanent form that which not so long before he
266
could only think or say verbally was a most worthy gift. In the case of [inscription] 2 the fact that the dedication happened to be on a mere sherd mattered
little: it was the writing itself that was the gift.85
Langdon must certainly be correct. What is crucially important here is that

writing is deemed to be the performance of a sacred act; as Henrichs observes,
there can be no doubt that in the shrine of Zeus on Mt. Hymettos the idiosyncratic emphasis on autography as a performance and the worship of the
gods through gift giving were closely connected.86 What is novel about the
seventh-century offerings of Mount Hymettos is not the alphabet per se, but
that the act of producing alphabetic symbols is viewed to be a performance.
7. 1 3
ze u s o f t hesig n
The Zeus to whom alphabetic performance is offered is Zeus Semios. The epithet (Smios) is clearly derived from the noun (sma) sign. The
nuances of adjectival derivatives in (-io-) are manifold,87 so that, on the
basis of linguistic considerations alone, little can be said of the semantics of
(Smios) other than that we are here dealing with Zeus of the sign. The
noun that supplies the epithet is the term used on the sole occasion on which
Homer makes reference to written symbols, the story of Bellerophon and the
fatal letter that he delivered to the king of Lycia, inscribed within a
(pinaks), from the Argive king Proetus (Il. VI 168169):
, ,
,.
Pempe de min Lukinde, poren d ho ge smata lugra,
grapsas en pinaki ptukt(i) thumophthora polla,.
He sent him to Lycia, and gave him baneful signs,
Written in a folding tablet many and life-destroying,.
Beyond , much of the vocabulary here is familiar to the reader: The

verb that describes Proetuss production of the (smata) is
(graph) writes, etches: the same verb used by worshipers of Zeus Semios on
Mount Hymettos in announcing their personal production of an alphabetic
offering; the same verb used by Euripides In-Law in Aristophanes Women
at the Thesmophoria as he uses his (smil) stylus to carve symbols into
(pinakn ksestn deltoi) tablets of smooth-polished
267
plaques. The noun there used for plaque (pinaks) is here employed by
Homer for the tablet (pinaks ptuktos) folding tablet88 in
which Proetuss signs are inscribed and concealed.
Langdon construes the sense of Zeuss Hymettos epithet Semios, of the sign,
as a reference to the gods weather (smata).89 For evidence he points,
for example, to Mount Parnes in the north of Attica where there stood, as
reported by Pausanias (1.32.2), both an altar of Zeus (Smaleos) and
an altar on which sacrifices were made to Zeus (Ombrios), that is,
Zeus of the rains; Zeus Ombrios likewise had an altar on Mount Hymettos,
notes Pausanias. Regarding the site called Harma on Mount Parnes, Langdon
notes that augurs looked to this place from Athens for the flash of lightning
which was the sign for sending an offering to Delphi. Also, weather signs
were read from clouds there.90 Concerning Mount Hymettos, he writes:
Throughout most of antiquity, and even in modern times, Hymettos was
regarded as a natural weather indicator, especially for approaching rain.91
a n e xc u r ses o n ( s ma ) ,
( s male o s ) , a n d ( s ma n t r )
As intimated in the previous paragraph, Zeuss epithet (Smaleos) is,

paralleling Semios, derived from (sma) sign. In this instance, the derivational formant is - (-aleos), a peculiarly Greek conglutinate, write
Buck and Petersen: regarding the origin of the morphology, it can only be
said that it must have arisen in prehistoric times by the addition of -- [-eo-]
to -- [-alo-], but everything else is obscure.92 The formant is well attested
in Homer:93
(22) (argaleos) painful, troublesome; compare (algos) pain,
sufferings
(harpaleos) alluring, and in adverbial form (harpales) greedily; compare (harpazd) to snatch away
(austaleos) dried up; compare (au) to kindle a fire
(tharsaleos) daring; compare (tharsos) boldness
(iskhaleos) dried; compare later attested (iskhnos)
dry
(karphaleos) dried; compare (karph) to make dry
and wrinkled
268
(karkhaleos) parched (of a dryness from thirst)

(kerdaleos) cunning; compare (kerdos) gain, advantage; cunning arts
(leptaleos) delicate (of the voice [ (phn)] of a young
male singer); compare (leptos) having the husk removed; fine,
delicate and (lep) to peel off
(leugaleos) experiencing wretchedness; sorry; bringing woe;
compare (lugros) wretched, sorry; bringing woe
(mudaleos) dripping; compare (mudos) decay in
Nicander; Sanskrit mudira- cloud; Middle Irish muad fog; Lithuanian
mudyti to bathe; English smut94
(optaleos) roasted, broiled; compare (optos) roasted,
broiled; (opta) to roast, to broil
(rhgaleos) broken; torn, ragged; compare (rhks)
denoting conduits within the palace of Odysseus;95 (rhgnumi)
to break, to shatter
(smerdaleos) terrifying to see or hear; compare Old High
German smerzan, Old English smeortan to hurt (Modern English
smart)
Other early examples of the formant can be found outside of Homericepic:
(23) (aualeos) dried, parched, Hesiod; compare Homeric
(auos) dry, dried up
(graleos) aged, old, Anacreon; compare Homeric
(gras) old age
(oidaleos) swollen, Archilochus; compare Hippocratic
(oidos) swelling
(tarbaleos) fearful, Homeric Hymn to Hermes; compare
Homeric (tarbos) fear
The semantics of the derived adjectives in - (-aleos) are again diverse,
though among the archaic forms cited here, many are conspicuous in their lexical association with the notions expressing the outcome of a process of transformation: desiccation and its by-form cooking, dampening typical of putrefaction, aging, swelling, husking, fracturing. Most of the remainder of the set
has an affiliation with cognitive and emotive phenomena: intellectual puzzling
and the perception and experience of fear, misery, pain, and desire.96
269
The Greek noun (sma) sign, from which the adjectives

(Smios) and (Smaleos) are derived, is of Indo-European heritage
and of common origin with Sanskrit dhyti to think, dhyna- planning,
reflection, dh- thought, conception, religious reflection; Avestan d(y)to see, da- discerning, intelligent, daman- eye. Compare also Albanian
dtur, dtme wisdom, learning and Gothic filu-deisi cunning (translating
Greek [panourgia] at 2 Corinthians 11.3 and Ephesians 4.14).97 An
exact formal equivalent of the nominal (sma) is provided by Sanskrit
dhyman- thought.
In an important examination of Greek (sma), Gregory Nagy,98 building upon earlier work by Douglas Frame,99 has shown that (sma) sign,
like its Indo-Iranian (and other) lexical relatives (Sanskrit dhyman- thought
etc.), is fundamentally bound to the notion of cognitive activity. He does so
by examining the semantics of (sma) in conjunction with those of
(noos) mind, sense, perception (and the derived verb to notice),
finding (sma) to be the key to a specific aspect of cognition, namely,
recognition.100 Regarding (sma) in Homer, he observes: In particular,
Homeric diction deploys sma as the conventional word for the signs that
lead to the recognition of Odysseus by his phloi, those who are near and
dear to him. An appropriate word for the recognition of this sma is the
verb anagignsk [].101 (anagignsk) is used in
Odyssey xxiv 329 of Laertes recognition of that (sma) which is the scar
that Odysseus carries from a boar-tusk laceration, and in Odyssey xxiii 206 of
Penelopes recognition of the (smata) signs provided by the clothes
that she had once given to Odysseus. Inherent in this notion of recognition is
interpretation, an act often, though certainly not exclusively, linked with the
prophetic function of the (mantis) seer.
In the poetry of Homer and Hesiod, Zeus is commonly linked with the
notion of the sign, sending (smata) to be recognized and interpreted.
Nagy points to the portent of a snake in Iliad II 308319, whose swallowing of
eight hatchlings and the mother bird is interpreted by the seer Calchas as sign
of the coming destruction of Troy; and to Zeuss frequent sending of lightning
as a (sma) one might say as a code bearing distinct messages that are
to be interpreted in context by both the witnesses and the narrative itself.102
The signs, (smata), those sent by Zeus and other signs as well,
are seemingly arbitrary, Nagy points out. In order for one to recognize and
interpret the signs properly, one must be able to recognize the position of
the sign within an internally coherent system of signals:103 the nexus of bird
type and flight pattern provides to the augur one interpretation as opposed
270
to all others; the position of the Dog Star within an astral array determines
the recognition of its meteorological signal (see Iliad XX 3033); the
(smata) that Proetus inscribed in the (pinaks ptuktos) folding
tablet that Bellerophon carried to the king of Lycia can be read (i.e., recognized and interpreted) by the tablets recipient as graphic symbols that receive
a phonetic value within a system of graphic symbols.
Nagy draws attention to Hesiod, Works and Days 267269 (the translation
is Nagys):104

,
.
Panta idn Dios ophthalmos kai panta nosas
kai nu tad ai k ethel(i)s epiderketai, oude he lthei
hoin d kai tnde dikn polis entos eergei.
The eye of Zeus sees everything and recognizes [verb noe] everything.
If it so pleases him, he casts his glance downward upon these things aswell,
and it does not escape hismind
what kind of justice [dik] is this that the city keeps withinit.
The phrase (oude he lthei) and it does not escape his mind in
line 268 is one that occurs in Homeric epic on several occasions in conjunction with (sma) sign with regard to the giving and the recognizing of
(smata) signs as in Odyssey xi 126 (the translation is again Nagys),
spoken by the shade of the seer Teiresias whom Odysseus encounters during
his visit to Hades realm:
,
Sma de toi ere mal ariphrades, oude se lsei
I will tell you a sma, a very distinct one, and it will not escape your
mind
Of these two epic pericopes, Hesiodic and Homeric, Nagy observes, given
other uses of the phrase (oude he lthei) and it does not escape
his mind, that it is to be expected, in the first passage, that the cognition
of Zeus is linked with the sma; and, in the second passage, that getting the
sign is linked with its recognition (noun noos or verb noe). Zeus sends his
(smata) signs meteorological and otherwise; the requisite mortal
271
response is to read the deitys signs: What humans must do is to decode the
various signs encoded by Zeus.105
What we find in the offerings to Zeus Semios on Mount Hymettos, I submit,
appears to be the very opposite of this normal state of affairs. The respective
cognitive roles of mortals and deity are reversed. (smata) signs are
presented to Zeus (Smios) of the signs, graphic symbols encoding either
(1) linguistic messages that self-reference the performative act of encoding such
signs, or (2) the abecedarium, with its (smata) in their periodic order.
The (smata) belong to a closed system and the decoding of these cognitive offerings by Zeus Semios requires on the part of the deity a recognition
and interpretation of the significance of each (sma) within that system.
This is of course the structural dynamic elaborated by Saussure, examined in
Chapter3, and the same dynamic that Zeus utilizes when he encodes messages
in the structured systems of bird flight, meteorological phenomena, and so
on. If the epithet of Hymettian Zeus, (Smios), antedated the advent of
Attic literacy, it undoubtedly reflected this encoding cognitive activity of Zeus,
as presumably does that of Zeus (Smaleos) of Mount Parnes (with
its peculiar morphology associated with notions of cognitive transformations).
But with the monumental introduction of the structured system of graphic
(smata), the (grammata) letters of the alphabet, those signals that comprise Bellerophons (smata lugra), Zeus Semios,
the encoder of messages, could be naturally enough linked to this symbolic
system and become decoding recipient of its offerings.
The verbal notion expressing the cognitive response of Zeus Semios to such
scriptic sacrifices is not attested on Mount Hymettos. As noted earlier and
discussed by Gregory Nagy, the verb (anagignsk) is appropriated for similar cognitive acts of recognition of (smata) depicted
in Homeric epic. This archaic verb (anagignsk) is in fact one
of the principal verbs to which the once nonliterate Greeks would, upon the
acquisition of the alphabet, assign the meaning to read. Its earliest occurrence in this sense is found in Pindars Olympian Odes 10.13:

,

Ton Olumpionikan anagntemoi
Arkhestratou paida, pothi phrenos
emas gegraptai.
272
Read to me the name of the Olympic victor,

the son of Archestratos, where it has been written
in mymind.
The presentation of encoded (smata) signs to Zeus Semios is the

offering of a verbal performance. The offering is either the weaving of a linguistic phrase or the weaving of an alphabetic strand its all the same; in
either case the material of the weaving (smata) signs is deemed
to encode an expression of language (a graphic symbolic expression of orality
[or oral performance]). These are (smata) that are presented to Zeus
Semios for the decoding, just as the epic poetic language of an oral performance must be decoded by its viewers, mortal or immortal just as the later
written poetic performance of the choral poet would have to be decoded
re-cognized by its readers.
Zeus is not only (Smios) and (Smaleos); he is also
(Smantr). The nominal (smantr) is deverbative,
derived from (smain) to show by a sign, to give a sign or signal
and so, to give orders; the verb is itself transparently derived from
(sma) sign. The morphology of (Smantr) is that of an
agent noun, a word formation of primitive Indo-European origin hence,
literally, one who gives a sign/signal, or, bearing in mind the cognitive process implicit in (sma), one who encodes his thoughts/will in a sign/
signal.106 Homer uses the term (smantr) to denote those who
control and herd animals: at Iliad VIII 127, (smantr) designates
a chariot driver, at Iliad XV 325 a cattle herdsman or shepherd; with the latter, compare the related adjective with negative prefix, (asmantos), used of unshepherded flocks at Iliad X 485. The epic nominal
(smantr) may likewise identify one who exercises authority over encodes
his will in signals to humans: Homer employs this agent noun for denoting
commanders of troops at Iliad IV 431 and an agricultural foreman at Odyssey
xvii 21, as of a household master at Odyssey xix 314, that is, persons who give
the signal, agents of signaling.
In the epic poem The Shield of Heracles, attributed to Hesiod,107 Zeus is
called (then smantr pantn) smantr of all the
gods. The reference comes at line 56, within verses identifying the two fathers
of Alcmenas twin sons Heracles and Iphicles: the latter was fathered by a mortal, (dorussoos) spear-brandishing Amphitryon; in contrast, Zeus
fathered Heracles. In these lines the poet makes parallel, conjoined references
273
to Zeus as (Kronin), son of Cronus: in the first instance the god is

denoted as (kelainephs Kronin) black-clouded son of
Cronus (line 53); in the second he is (Kronin), the smantr of all
the gods:
,

, , 55
, .
Ton men hupodmtheisa kelainephei Kronini,
autar Iphikla dorusso(i) Amphitruni;
kekrimenn genen, ton men brot(i) andri migeisa, 55
ton de Dii Kronini, then smantori pantn.
Him [i.e., Heracles] [she birthed] having been subdued by the
black-clouded son of Cronus,
but Iphicles [she birthed subdued] by spear-brandishing Amphitryon;
offspring who differed the one from her mingling with mortal man, 55
the other with Zeus the son of Cronus, smantr of all thegods.
The phrase (kelainephs Kronin) black-clouded son

of Cronus constitutes a well-known Homeric formula. Regarding Homers
use of the adjective (kelainephs) and this formula, Heubeck and
Hoekstra observe: Usually the adj. is found in formulae such as
(e.g. Il. i 397)or when Zeus is addressed in his full majesty of (IndoEuropean) Sky-god, cf. e.g. , , , ,
[Zdeu, kudiste, megiste, kelainephes, aitheri nain] (Il. ii 412)108 that is, O
Zeus, most glorious, greatest, black-clouded, dwelling in the sky, the opening
words of a prayer offered by Agamemnon for victory in battle.
The parallel epithets of lines 53 and 56 of The Shield,
(24)
Cronus
(kelainephs Kronin) black-clouded son of
, (Kronin then smantr pantn)

son of Cronus, smantr of all thegods
suggest that for its poet or, more carefully, in the poetic tradition from which
the lines are drawn the semantics of the agent noun (smantr)
remained palpably within the cognitive dimension of the code of the sovereign
274
sky god Zeus, giver of signs (encoder of signs), especially that sign which is
the lightning of the black storm cloud. Like the adjectives modifying Zeus on
Hymettos and Parnes, (Smios) and (Smaleos), the agent
noun (smantr) is fundamentally linked with the cognitive process of encoding messages in signs.
But this is only to state the obvious. The verb from which the agent noun
is derived, (smain), clearly preserves this sense. For example,
Aeschylus places on Orestes lips the redundant (and redundantly translated) proverbial sentiment that a man will speak frankly to another man,
(kasmnen emphanes tekmar) and signify his
sign openly (Choe. 667). Of Apollo, and his Delphic oracle, Heraclitus offers
the puzzle that (oute legei oute kruptei
alla smainei) he neither proclaims nor hides but instead he gives a sign (fr.
93).109 Similarly, Xenophon writes that he consulted the gods to determine if
he should command (An. 6.1.31):

.
Kai moi hoi theoi houts en tois hierois esmnan hs kai iditn an gnnai hoti
ts monarkhias apekhesthai medei.
And the gods gave signs to me in the sacrifices such that even a nonspecialist would know that it is requisite for me to abstain from taking supreme
command.
And it will come as no surprise, given the practice of the worshipers of Zeus
Semios on Hymettos, that (smain) is also used to denote the signaling action of written symbols. We have seen this already, when, in Chapter6,
we observed that Plato has Socrates avow that writing, (graph), is like
painting, (zdgraphia), a symbolic system (Phaedrus 275DE):
, , , .
, ,
.
, ,
.
Deinon gar pou, Phaidre, tout ekhei graph, kai hs alths homoion
zdgraphia(i). kai gar ta ekeins ekgona hestke men hs zdnta, ean d aner(i)
ti, semns panu siga(i). tauton de kai hoi logoi; doksais men an hs ti phronountas
275
autous legein, ean de ti er(i) tn legomenn boulomenos mathein, hen ti smainei monon tautonaei.
For, Phaedrus, writing has this peculiarity its really like painting, the creations of which are like living beings; but if you should ask them something,
they remain entirely solemnly silent. Its the same with [written] words. You
might suppose them to speak, as if they have understanding; but if you should
ask them something, wanting to learn about what they are saying, they will
always signify only one and the same thing.
And so we have returned to the realm of (sma) as graphic signal but

more must be said regarding (smantr) one who gives a signal.
Beyond line 56 of The Shield of Heracles, (smantr) is seen again
as Jovian epithet, both within and without the Hesiodic corpus. In Hesiodic
fragment (MW) 5, in which Zeus is said to have fathered Graecus, eponymous ancestor of the Greeks, by a maiden Pandora,110
the poet refers to the sovereign god as
(Zeus patr then smantr pantn) Zeus the father, smantr of all the
gods.111 Similarly, in the Homeric Hymn to Hermes, when Apollo and Hermes
tell their conflicting tales of Apollos stolen cattle, Hermes, with the address
(Zdeu pater) O father Zeus, speaks (allos muthos)
another muthos, pleading his case (es
Kronina then smantora pantn) to the son of Cronus, smantr of all the
gods (line367).
With these archaic references to the sovereign sky god Zeus as smantr
of all the gods, compare the epithet - (polu-smantr) assigned
to Hades, sovereign of the Netherworld, in the Homeric Hymn to Demeter,
line31:

, .
30
Tn d aekazdomenn gen Dios ennesi(i)si

patrokasigntos polusmantr poludegmn
hippois athanatoisi Kronou polunumos huios.
30
And against her will, at Zeuss suggestion, he carried her off,

her fathers brother, polusmantr, receivingmany,
with his deathless horses, Cronuss many-namedson.
30
276
The abductee here referred to is Persephone, daughter of Zeus and Demeter;

her abductor is of course Hades, her uncle, another son of Cronus. In addition
to the compound agent noun (polusmantr), literally and
componentially one who gives many (- [polu-]) signs, Hades is in this
hymn also described by the adjectives (poludegmn; line 31)and
(polunumos; line32).
The latter, (polunumos) having many names, is used attributively of various deities (at times assigned the sense famed [i.e., named
many times]): for example, of Styx (Hes., Theog. 785); of Apollo (Hymn. Hom.
Ap. 81); of Theia (Pind., Isthm. 5.1); of Dionysus (Soph., Ant. 1115); of Nike
(Bacchyl., Epigr. 1.1); of Artemis (Ar., Thesm. 320). The fifth-century b c comic
poet Strattis (fr. 220.104) similarly uses it of the chthonic deity, coupling the
adjective with the gods alter-name Pluto.
The former adjective, (poludegmn), like
(poludekts) both constructed with the root of the verb (dekhomai)
to receive is used as a euphemistic name for Hades, Receiver of Many.
(Poludegmn) occurs four times in the Homeric Hymn to Demeter
in this way: lines 17, 31, 404 (restored), and 430; the referent of the adjective
appears to be the same in Orphic Hymns 18.1112 and other Orphic texts.112
The Stoic philosopher L. Annaeus Cornutus (first century a d ) lists it among
the names of Hades (Theol. Graec. 74.15). For the Hellenistic poet Lycophron,
(poludegmn) is the name of a lofty mountain of Italy from which
all rivers of the country flow (Alexandra 700), though a scholiast attempts to
link the mountain to Hades (Schol. Lycoph. 700.67).
As we have understood with the Zeus Semios of Hymettos, in the epithets
of lines 3132 of the Homeric Hymn to Demeter, Hades is likewise depicted as
participator in reciprocal action specifically that of signaling.113 In line 31,
the invoking of the gods attribute (polusmantr) is followed
immediately by the invocation (poludegmn): he is Hades giving
many signs, receiving many. That the rehearsed reciprocity can be primitively
construed as a reciprocity of signifying is suggestively reinforced by the juxtaposition in line 32 of the epithet (polunumos) many-named
denoting one who is assigned many designations, many significations. Whatever
array of nuances is eventually attached to Plutonian (poludegmn),
receiving many, at some appropriately early moment there must be a core sense
that reflects in its application to the chthonic sovereign reciprocal notions of
cognition such as those that attach themselves to Zeus: the giving of many signals and the receiving of many signals. The signal giving (communicativeness)
277
of Zeus itself finds expression in conjunction with these epithets as it is in line

30 an (ennesi) suggestion114 of Zeus that is responsible for putting
into Hades head the idea of seizing Persephone (as an [ennesi] of Gaea
caused Cronus to regurgitate the swallowed siblings of Zeus [Hesiod Theog.
494]; cf. Iliad V 894). Cognition and re-cognition are present.
The agent noun (polusmantr) occurs twice more in
archaic usage again in the Homeric Hymn to Demeter. Once (line 84), as
Helios reveals to Demeter the fate that has befallen her daughter Persephone
(lines 77b81, 83b87):

,

.
.

.
oude tisallos
aitios athanatn ei m nephelgereta Zdeus,
hos min edk Ad(i) thalern keklsthai akoitin
autokasignt(i); ho d hupo zdophon eroenta
harpaksas hippoisin agen megala iakhousan.
.
ou toi aeiks
gambros en athanatois polusmantr Adneus
autokasigntos kai homosporos; amphi de timn
ellakhen hs ta prta diatrikha dasmos etukhth;
tois metanaietaei tn ellakhe koiranos einai.
80
85
80
85
theres noother
of immortals to blame but Zeus, cloud-gatherer,
whos given her to Hades to be known as his youthful mate
his own brother; and down to the hazy gloom Hades took her, 80
snatching her screaming her head off with his steeds.
.
278
hes not a shameful

in-law among immortals, Adoneus [i.e., Hades], polusmantr
your own brother, shared sibling; and regarding honors,
85
he got his part by lot when at the start the threefold division wasmade;
he dwells with them whose ruler he was fated tobe.
The mention of a meteorological attribute of Zeus (line 78) is echoed by

a reference to Hades (line 80): Zeus manipulates the clouds, while Hades
inhabits the mist-shrouded darkness of the nether realm. A further comparison of the two sovereign deities is invited by similarities between the two
sets of verses cited here structural framing provided by overt reference
to the membership of each deity in the set of deathless beings (
[athanatn], line 78; [en athanatois], line 84) and emphasis
on sibling relationships ( [autokasigntos], lines 80 and 85),
linking Zeus and Hades both directly and indirectly through their common
sister Demeter. Set beside (nephelgereta Zdeus) Zeus
cloud-gatherer is (polusmantr Adneus) Hades
polusmantr.
And regarding honors (line 85), Homer preserves the tale of how the
three sons of Cronus and Rhea Zeus, Hades, and Poseidon obtained their
respective realms of sky, netherworld, and sea by a casting of lots, and does so
with language that is mirrored by the preceding lines from the Homeric Hymn
to Demeter. Asserting his equality with Zeus, an irate Poseidon tells Zeuss
messenger Iris that (Il. XV 190194)

, ,

.
190
toi egn elakhon polin hala naiemen aiei

pallomenn, Ads d elakhe zdophon eroenta,
Zdeus d elakh ouranon eurun en aitheri kai nephel(i)si;
gaia d eti ksun pantn kai makros Olumpos.
190
And I indeed took the gray sea to live in forever

190
when the lots were cast, and Hades won the hazy gloom,
and Zeus took wide heaven in the air and clouds;
but earth is common still to all, and towering Olympustoo.
279
As others have noted,115 this theme of a three-way division of the cosmos by

a trio of deities who cast lots for their portions is a Mesopotamian motif,
preserved in various Akkadian versions of the Atrahasis tradition, the earliest attested of which is the Old Babylonian version of the seventeenth
century b c .116
The third occurrence of (polusmantr) in the Homeric
Hymn to Demeter is found at line 376. As Hades prepares to return Persephone
from his nether realm to Demeter, the poet sings:

.
375
Hippous de proparoithen hupo khruseoisin okhesphin

entuen athanatous polusmantr Adneus.
375
And his deathless steeds he hitched in front, beneath

the golden chariot, the polusmantr Adoneus.
375
The agent noun (polusmantr), in each of these passages,

is commonly rendered by some term denoting leader: for example, West117
translates as Major General, and both Evelyn-White118 and Cashford119 as
Ruler of Many. The notion of - (polu-) many (combining form of the
adjective [polus], much, many) seems to be inherent to the character
of Hades: we noted other nominals so compounded that are applied to the
god of the netherworld: (polunumos) having many names; and
(poludegmn) and (poludekts) receiver of many.
Hades can also be called (poluarkhos), ruler of many (Cornutus,
Theol. Graec. 74.15); and that other name by which he is commonly invoked,
(Ploutn), is semantically linked to the fundamental sense of (polu-), if not etymologically related: the common noun (ploutos),
from which the divine name is derived, denotes riches, abundance.120 For
Hesiod (Op. 465), it is a chthonic god to whom the farmer should pray at the
outset of the plowing season in order to enjoy an abundant grain harvest:
(eukhesthai de Dii khthoni(i)
Dmteri th hagn(i)) pray to chthonic Zeus and holy Demeter; elsewhere
(Theog. 767), Hesiod identifies Hades ( [iphthimos Aids] stalwart Hades) as the (theos khthonios) chthonic god, coupling his
mention with reference to (epain Persephoneia) dreadful Persephone.
280
References to chthonic Zeus occur beyond Hesiods poetry, and in

instances the appellative clearly identifies Hades.121 Thus, Homer sings the
tale of Phoenix and how he had slept with his fathers concubine, at his
neglected mothers behest, and how the father then sought his revenge on the
son, calling on the Erinys, that Phoenix would be made impotent: the curse
( [epara]) was fulfilled and vengeance meted out by the gods
(Zeus te katakhthonios kai epain
Persephoneia) Zeus beneath-the-earth and dreadful Persephone (Il. IX 457).
Compare with this the poets words some hundred lines later when he tells of
the anger of Althaea, Meleagers mother, following Meleagers slaying of her
brother; she prays to the gods (Il. IX 569572):
,
, ,

, .
kiklskous Adn kai epainn Persephoneian,
prokhnu kathezdomen, deuonto de dakrusi kolpoi,
paidi domen thanaton; ts d erophoitis Erinus
ekluen eks Erebesphin, ameilikhon tor ekhousa.
570
570
summoning Hades and dreadful Persephone

sitting splayed, and drenching her bosom with tears 570
to give death to her son; and the Erinys that walks in darkness
heard her out of Erebus, that one whose heart is unbending.
And so Meleager will die, and the death of his uncle be avenged. The two
lines (Il. IX 457 and 569), and the passages of which they are a part, provide
a parallel. The alternation of chthonic Zeus and Hades within a formulaic reference to the nether god and his dread queen utilized within accounts
that invoke the Erinys and the avenging of crimes against family members
strongly suggests an equation of the two deities so named.
From a much later period (fifth/sixth century a d ), compare Nonnos,
Dionysiaca 44.258259:

,.
Hai de Dios khthonioio dusante neumati korss
Eumenides Penthos epestratonto melathr(i),.
281
And at the dreadful sign of the head of chthonicZeus

the Eumenides stormed the hall of Pentheus,.
Elsewhere (Dion. 27.7677), Nonnos explicitly equates Hades and the chthonic
Zeus. The (neuma) sign that chthonic Zeus gives that releases the
Eumenides (i.e., the Erinyes) is presumably a nod ( [kors] head is an
emendation for manuscript [Rheis] of Rhea), a fundamental denotation of this word (from [neu] to nod).
Other examples can be adduced from Greek tragedy. In Aeschyluss
Suppliant Maidens, the chorus of Danads says that the maidens will supplicate , (ton gaon, ton
poluksentaton Zna tn kekmkotn) Zeus of the earth, the much-welcoming receiver of those who have died (lines 156158); the sovereign sky god
Zeus is at times identified as (Zdeus Ksenios) Zeus of the guesthost relationship:
protector of suppliants and xenoi (guest strangers), the god who walks by
the side of the esteemed xenoi (Odyssey 9.270271). Homer has the swineherd Eumaeus tell Odysseus disguised as a beggar and received kindly into
the herdsmans humble hovel that all xenoi and beggars are from Zeus;
my gift [hospitality] is small but philos (Odyssey 14.5759) again, formulaic phrasing, uttered similarly by Nausicaa upon finding Odysseus (Odyssey
6.207208).
It is the gods, and Zeus most particularly, who set the standards for the treatment of the guest stranger for the proper social response one of philos
towards the xenos.122 And the social relationship of xenos (guest stranger) and
xenodokos (host) is one that has particular affiliations with kings and no less
so with the king of gods.123
Aeschyluss chthonic Zeus is likewise affiliated with hospitality to the

wanderer the wandering dead: in keeping with aforementioned notions of
abundance associated with the nether god, he is - (poluksentatos Zdeus) much-welcoming / welcoming-many Zeus. A bit further
along in the play, Danas, father of the young women, affirms that there exist
those atrocities from which there is no escape even beyond the grave, in
Hades domain (lines 230231):
, ,
.
230
282
Kakei dikazdei tamplakmath hs logos,

Zdeus allos en kamousin hustatos dikas.
230
There, so its said, among the dead

230
124
another Zeus judges a final judgment of sins.
Among other references to chthonic Zeus, a particularly interesting one is

provided by Sophocles in Oedipus at Colonus. Just prior to the aged Oedipuss
death, Oedipus and his daughters recognize a signal of that death
(phthoggos eksaiphns pikros) a sudden bitter sound. It is
chthonic Zeus who gives this signal (16061607):
,

ktupse men Zdeus khthonios, hai de parthenoi
rhigsan hs kousan
chthonic Zeus thundered, and the maidens
shivered when they heardit
The signal is typical of those meteorological signs that Zeus the sovereign
deity of the sky gives, as we saw earlier; for Sophocles, chthonic Zeus can likewise be a (smantr) in the literal sense of theterm.
We observed earlier that the agent noun (smantr) one who
gives a sign/signal is derived from the verb (smain) to show by
a sign, to give a sign or signal, which is in turn derived from (sma)
sign. There is a conspicuous sense of the noun (sma) with which the
netherworld deity is intuitively connected: (sma) commonly denotes
grave, tomb. The referent may be a tumulus or cairn raised above a grave,
as in the case of the mound described in Iliad II 811815, a great barrow situated on the plain before Troy:
(athanatoi de te sma poluskarthmoio Murins) and the gods [call it] the
grave mound of Myrine (line 814). Or (sma) may denote some other
funerary marker such as the wooden post with two stones inclined against
it that serves as the turning post for the chariot race at the funeral games for
Patroclus; Nestor describes its appearance at Iliad XXIII 327330, and then
surmises (lines 331332):
,
.
283
teu sma brotoio palai katatethntos,

to ge nussa tetukto epi protern anthrpn.
Its either a tomb marker of a man who died longago,
or its a turning post wrought in times of earlier men.
Whatever the form of the marker, such a (sma) is a sign, signal of the
presence of a grave. Though the distinction may be subtle, by a slight semantic extension, (sma) can denote the tomb itself, rather than, most
immediately, a funerary monument serving as signal. This sense seems to be
predominant, for example, in Platos remarks on body as tomb at Gorgias
492E493A (Socrates is speaking):
. ,
,
, ,
;

, ,.
Alla men d kai hs ge su legeis deinos ho bios. ou gar toi thaumazdoim an, ei
Euripids alth en toisde legei,legn
tis d oiden, ei to zdn men esti katthanein,
to katthanein dezdn?
kai hmeis t(i) onti iss tethnamen; d gar tou egge kai kousa tn sophn,
hs nun hmeis tethnamen kai to men sma estin hmin sma,.
But then, as you yourself tell it, life is strange. For I wouldnt be astonished if
Euripides speaks the truth when hesays,
Who knows if to live is to bedead,
and to be dead is tolive?
And we may very well be dead; for I have actually heard a wise man say that
we are ourselves now dead and that the body is our tomb,.
Compare Phaedrus 250C: here Plato writes of pure souls not being entombed
within the body, locked up like an oyster, where the term expressing the
notion not entombed is (asmantos). This adjective,
(asmantos), is derived from the previously discussed denominative verb
(sman) and commonly means unmarked, as in Herodotus 2.38,
where the historian describes an Egyptian cultic practice: a priest must
inspect a bull that is to be sacrificed for purity, and, if the bull is found to be
284
pure, the priest signals ( [sman]) the finding by wrapping papyrus

around the bulls horns and stamping the papyrus with a seal; a bull that is
not marked ( [asmantos]) in this way as pure must not be sacrificed. The adjective can also carry the sense without a leader that is, without a (smantr) in the secondary sense of that agent noun; we
have of course seen this already: it was pointed out in the previous discussion
of (smantr) that the adjective (asmantos) is used of
unshepherded flocks at Iliad X485.
But for Plato this (sma) tomb which the (sma) body constitutes, and with which it forms a phonological minimal pair (sma / sma), is,
to judge by Socrates etymologizing at Cratylus 400BC, not semantically far
removed from the words fundamental notion of sign/signal. Regarding the
origin of the word (sma), Socrates conjectures:
,
.
Kai gar sma tines phasin auto einai ts psukhs, hs tethammens en t(i) nun
paronti.
For some say that it [i.e., the body ( [sma])] is the tomb ( [sma])
of the spirit, as [the spirit] has been buried within our present being.
He continues with an added consideration:

,
.
Kai dioti au tout(i) smainei ha an smain(i) h psukh, kai taut(i) sma
orths kaleisthai.
And further because by this [body/tomb], the spirit signals whatever it would
signal, and thus it is rightly calledsma.
In light of what we have observed in the preceding pages, an analogical equation relating the celestial sovereign and the nether sovereign (the
chthonic Zeus) suggests itself:
(25)Zeus : (smantr) :: Hades : - (polusmantr).
In those ways in which sky god Zeus can be conceptualized as
(smantr) his netherworld equivalent can be as well, with the added, and
contextually expected, and so (nearly) redundant, cognitive element of
285
- (polu-), in keeping with notions of plenty and multiplicity that are

intrinsic to Plutos / Hades character. For Sophocles the chthonic god can
even give that signal, that encoded message, with which Zeus the sign giver is
conspicuously associated thunder: a message that is all too easily decoded
by Oedipus and his daughters as a signal of his impending death. Giving and
the reciprocal receiving of signals is an act fundamental to the sovereign deity
in both his celestial and chthonic forms. It is appropriate that the new signaling system the alphabet should be perceived as votive material appropriate
to the signaling deity.
As sovereign god of the dead, Hades must also be readily linked to the conception of (sma) as funerary signal, a signal that encodes the presence
of a burial or encodes the notion of the space within which the remains are
entombed. Hades - (polu-smantr) could be easily construed
as the agentive force behind many such encoded messages. Celestial Zeus
(Smios), (Smaleos), (smantr) is agent of the
(sma) sign and as such is deemed to be right recipient of the presentation of signs on Mount Hymettos; chthonic Zeus / Hades / Pluto is likewise
agent of the (sma) sign and, one might reasonably expect, would be
deemed to be appropriately associated with the presentation of written symbols in a funerary context within the physical space of the (sma)
which is the grave.
Such a funerary presentation, I would suggest, is to be found in the letters inscribed on the base of a Protocorinthian conical oinochoe from Kyme,
dated circa 700 b c .125 The piece has already been mentioned several times in
the present work: the first time in the discussion of CP gamma in Chapter2,
at which point I noted that perhaps the earliest example of a gamma like that
of the copper plaques, though reversed in stance, is to be found in a partial abecedarium on this wine jug from the Euboian colony of Kyme. It was
also noted that the jug attests the unusual morphology of the CP epsilon (also
common to graffiti of Mount Hymettos) and the CP digamma. The oinochoe
is inscribed on its base with two mirror-image partial abecedaria; as we saw in
the discussion of the morphology of CP eta and xi in Chapter2, both of these
abecedaria preserve an eta-symbol of the shape , a symbol that is also used
as a xi-character on Aegean Naxos.
The two partial abecedaria are of interest both for their physical orientation
relative to one another and for the variation in letter shape that they show. One
of the abecedaria126 is markedly curvilinear, sweeping counterclockwise in a
half circle from its starting point to a roughly straight scored line that forms an
axis dividing the two abecedaria. Its eight characters are inscribed from left to
286
right: the sequence of letters is alpha, beta, gamma, delta, epsilon, digamma, eta,
zeta; in other words, the periodic order of the last two characters is reversed.127
The direction of the letters is consistent with the dextroverse orientation of the
line with the exception of gamma, the stance of which is reversed relative to the
remainder of the line (i.e., it shows a sinistroverse orientation).
The second abecedarium is likewise dextroverse in terms of overall periodic
order, but its six letters show a sinistroverse orientation. It is also curvilinear,
though less so than the first, but also separated farther from the dividing axis
at its starting point than at its end point. Both alpha and epsilon are missing
and eta and zeta are again metathesized, giving the line a letter sequence beta,
gamma, delta, eta, zeta. The first three letters of this abecedarium are distinctly
different in shape from the corresponding letters in the other abecedarium:
beta has the unique morphology of Corinthian beta though rotated leftward, gamma has a pi-like curved crossbar, and delta has a decapitated apex,
giving it a form similar to that of the box-shaped eta which occurs in both of
the abecedaria.
The general physical impression that the two abecedaria give is thus that of
a continuous curving line of letters running from alpha through zeta up to the
dividing axis and continuing from that axis in reverse order from zeta to beta.
The predominant alternating dextroverse and sinistroverse stance of the symbols of the two lines (not an alternating directionality of the two lines) could be
viewed as comparable to a sort of meandering hybrid boustrophedon, looping
at the point at which the two zetas meet; following is a schematic representation of the field of the abecedaria with conventionalized letter shapes:
(26)
LETTER STANCE
(27)
LETTER STANCE
287
The arrangement fundamentally produces an interlacing of symbols such

as those that we saw in the interwoven abecedaria considered earlier in this
chapter. In this instance, however, rather than making a turn at the middle
of the alphabet and continuing the periodic order of the letter signs in the
opposite direction, the inscriber makes the turn after (the reordered) zeta
and repeats (most of) the symbols of the partial abecedarium with which he
or she began, in such a way that both lines follow the same direction but display their symbols in reverse orientation. The result is a woven pattern of a
different sort, with each individual grapheme paired, approximately, with an
identical grapheme, though the pairs show mirror-image stances in the case
of asymmetric characters and, in some instances (beta and gamma), variant
morphology. The pattern is made more variegated still by the departure from
regular periodic order and by character omissions.
There is more, however, to this funerary artifact. Another hand has inscribed
a line in Euboian script, possibly metrical as Watkins has proposed,128 around
a portion of the rim of the base of the lekythos. The line reads h
(hisamenetinnuna). Cassio,129 following upon Ribezzo,130 interprets the line as
follows:131
h ()
hisa menei tinnuna(i)
fatale pagare le stessecose.
Watkins offers the English rendering It remains to pay equal retribution,

invoking Alcmans line (fr. 1.36) (esti tis sin tisis) there
exists some vengeance of the gods, and characterizing the sentiment as a variant of the golden rule or the judgment of Rhadamanthus:132 wholly appropriate, he continues, for an object destined, as this was, for the grave of a
child or adolescent; as Cassio notes, adults were cremated.133
The alphabetic fabric woven on the base of the pot one suspects it to be
the handiwork of the deceased child or youth consists of a string of plaited
(smata). It seems a fitting grave good in light of the accompanying
line. One is reminded again of the previously rehearsed tale of Phoenix in Iliad
IX134 of how his father, Amyntor, (stugeras d
epekeklet Erinus; line 454)invoked the abominable Erinys after Phoenix had
slept with Amyntors concubine to make Phoenix impotent and of how Zeus
beneath-the-earth ( [katakhthonios]) and dreadful Persephone
(eteleion eparas; lines 456457) brought the curse to fulfillment. Amyntor has his equal retribution, as the author of the line on this
288
oinochoe looks to have his or hers. The weaving of (smata) makes for
an appropriate performance in the realm of chthonic Zeus, sign giver, in late
eighth-century Kyme, no less so than such performances are properly offered
to Zeus Semios on late eighth-century Mount Hymettos in Attica and one
may well suspect that the envisioned retribution would no less entail a
(sma) tomb.
7. 1 4
home r sba n e
Over the years various scholars have called attention to the fact that a high
number of the earliest known Greek inscriptions are written in verse, in a
poetic meter. A few have even imagined that this signals that the Greek alphabet was created for the express purpose of writing down Homeric poetry. I
believe that we can with some confidence now say that such a view is precisely
opposite the actual state of affairs. Aside from the inherent unlikelihood of
such a high-minded motivation for the creation of a system of writing, the
copper plaques with their recording of the alphabetic strand, again and again,
have brought to our attention that after the introduction of the alphabet to
Greece and the advent of Greek literacy, letter weaving was placed on a par
with word weaving. The outcome was the democratization of poetic performance. Not everyone could be an oral poet not everyone could extemporaneously compose and perform poetic epic; but almost anyone who acquired
the use of the alphabet could turn its use to linguistic composition of another
sort: to creations inspired by an individuals own private Muse to the production of recorded speech, which, under the proper circumstances, would
acquire a certain permanence and thus notoriety, if only modest notoriety,
akin to that of the memorable compositions in performance of the bards.
And it is for this reason, I would argue, that many of the earliest surviving
examples of Greek writing are verse compositions typically brief, but linguistic handiwork available to some audience, larger or smaller a kind of
everymans hexameter.
This performance response to the introduction of a new element of technology to the Greeks fully parallels responses in recent years to new elements
of technology introduced to society globally. The bard disappeared from our
ancestral cultures many centuries ago; in contemporary culture other performance phenomena hold sway notable examples being the visual media
of film and television. The performance activities of film and television are
limited to a few. Many among us may possess certain basic skills and talents
289
that we share, to a lesser or greater degree, with those few who are sanctioned for participation in the performance phenomena of film and television. The many are not sanctioned to engage in visual performance by the
studios of Hollywood, New York, London, and Bollywood, only the few; yet
introduction of new technologies has in recent years greatly expanded the
set of individuals capable of participating in visual performance phenomena before potentially vast audiences. The production of ones own video
performance and the posting of such performances on the World Wide
Web viral video has, in some sense, become, or is readily becoming, a
mainstream alternative to studio film and television. Lacking the grandeur,
cinematography, elaboration, and so forth of the latter, viral video compares
chiefly as an individualistic expression of talent or exhibitionism or existence or something providing the creator with a means to participate in
a performance phenomenon performance that is offered to anyone who
happens to be in a recipient position and is willing to watch anyone with
the hardware and software required to permit the prospect of constituting
the audience.
In the same way, the alphabet arrived in Greece in a time in which extemporaneous oral poetic performance was the performance phenomenon par
excellence when it is the bard who holds a certain celebrity status in the
entertainment world of archaic Greece. The bard is sanctioned to fulfill this
highly exclusionary linguistic role within the archaic Greek community. There
are many others who undoubtedly share with the bard a certain poetic creativity, and all share with the bard a common access to language, the medium
of the oral poets performance. With the arrival of what must be viewed as
a breakthrough technological advance for any people a writing system
and in this case an especially learnable writing system, the alphabet Greek
individuals found themselves in possession of the functional equivalent of a
video recorder: a device for capturing the essence of performance in this
case linguistic performance. Into the rich oral poetic environment of archaic
Greece an alternative for weaving language was thus introduced. The output of the recording device (the alphabet) is an image of language but the
image of language quickly immediately perhaps was awarded the status
of language, so that the activity of graphically recording individual, unique
expressions of language was equated with the activity of language weaving
the action of the poet. With this new recording device individuals could produce inscribed poetic expressions brief, certainly in comparison with oral
poetic composition as we know it, and perhaps often lacking in what we could
290
call real artistic merit but the individuals own expression of creativity, the
individuals own weaving of words, realized through a weaving of letters, even
simply through the weaving of the abecedarium.
Once viral hexameter was born, once the individual expression of language
via the alphabet grew and spread among the Greeks, there would be over time
a resultative loss a gradual disappearance of the phenomenon of oral composition of poetry in performance from Greek society. This was not a necessary
outcome, one could argue, but it was the outcome. The alphabet, the woven
(smata) not the boys of Ios would prove to be Homersbane.
Notes
1 Background
1 Heubeck 1986:18.
2 The preliminary findings of this technical physical examination of the plaques appear in
Scott etal. 2005. And see now the 2007 article by Brixhe, which examines the alphabetic
forms of the Schyen plaques only on the basis of a visual examination of the corroded
plaque surfaces.
3 Atiya1991.
4 See http://ccdl.libraries.claremont.edu/collection.php?alias=/cce and http://ccdl.libraries.
claremont.edu/cdm4/browse.php?CISOROOT=/cce.
5 See Woodard 1997; 2010;2012.
6 Personal communication and http://www.schoyencollection.com/firstalpha2.html.
7 Atiya 1985a:2737, 3941; Atiya Archive Bx23Fd9.
8 On this chronology, see Atiya 1985a:1, 4, 1520, 2324, 3942, 4648; and http://www.lib.
utah.edu/collections/middle-east/atiya.php.
9 All letters referenced in this paragraph are to be found in Atiya Archive Bx35Fd2.
10 Correspondence (Atiya Archive Bx18 Fd4) dated December 1961 and January 1962 refers to
Atiya returning from Egypt with ten tons of books more than ten thousand titles.
11 Regarding the number of forays to Egypt in one of his interview sessions with Atiya
(Atiya 1985b:2425), Cooley says: You were given the twenty thousand dollars initially to
make your journey to Egypt for collecting. How many times was this repeated? I know in
my period here at the library you made several journeys to Egypt to collect. Atiya replies:
Always got something, always got something. Nothing of the monumentality of the first
trip.
12 Kraus 1978:185.
13 Atiya 1967:75.
14 Atiya Archive Bx5Fd5.
15 Kraus 1978:370372.
16 Atiya 1985b:20.
17 http://www.celebratingresearch.org/libraries/utah/atiya.shtml.
18 Atiya 1985b:21.
291
292
Notes to Pages 618
19
20
21
22
23
24
25
26
Atiya 1985b:20.
Atiya 1985b:25.
Atiya 1985a:2930.
Kraus 1978:356.
http://ccdl.libraries.claremont.edu/cdm/singleitem/collection/cce/id/1569/rec/13.
Valuable Collection of Ancient Scrolls Presented to Yale.
Emmel 1980:53n1.
The relevant papyri are P.CtYBR inv. 17391878, 22482268, 28962943, 2956, 29782981,
29873033, 30373056, 3548. Other items in the collection that were acquired from H. P.
Kraus are these: in 1965, P.CtYBR inv. 2125, 23292342, 25872752, 29652968; in 1966,
P.CtYBR inv. 20802124, 23432353, 2982, 2983, 30583075.
27 Kraus 1978:276.
2 The Associative Structure of the Copper

Plaques
1 See, inter alia, LSAG2 pl. 1, 1; EG 1:135136, with fig.28ab. The inscription of the Dipylon
oinochoe will be examined more closely later in this chapter.
2 As described by Immerwahr 1990:7 unless, of course, the text of the copper plaques
should prove to be earlier.
3 See EG 1:225, with fig.87; Heubeck 1979:123, with Abb.48.
4 McCarter (1975b:141) has suggested that Guarducci has read the graffito upside down and
that the letters are not Greek at all but Phoenician.
5 See Bile 1988:29; Heubeck 1979:125; EG 2:330332, with fig. 111; LSAG2 468469;
Masson1976.
6 See LSAG223.
7 The reference is to Levi1969.
8 LSAG2469.
9 See Cordano 1984:285286, 297, 308, no.26; LSAG2 461; pl. 77,8.
10 On Greek colonization of Sicily generally, with bibliography, see Holloway 1991; for Naxos,
see especially pp.4549.
11 See Guarducci 1985:711, with fig.1; LSAG2 455; pl. 76,5.
12 See LSAG2 83; pl.5,6.
13 See Cordano 1984:286, 298, 309, no.31; Guarducci 1970; LSAG2 464; pl. 78,2.
14 See LSAG266.
15 Immerwahr 1990:131.
16 See Langdon 1976:1920, with fig. 8, 28. Two examples also occur on H 500 (Langdon
1976:3435, with fig.15, 127), where the direction of writing is ambiguous. A probable example with the same stance as CP Alpha-1 occurs in the fragmentary inscription of H 387; see
Langdon 1976:1920, with fig.8, 31 and pl. 5. Compare too H 527; Langdon 1976:2829, with
fig.12, 83 and pl.9.
17 P 25922; see Lang 1976:31, with pl. 11,F20.
18 See LSAG2109.
19 See LSAG2133.
20 See LSAG2 314; pl. 60,29a.
Notes to Pages 1823
293
21 See LSAG299.
22 See LSAG2 pl. 12, 1. Jeffery (p.101) expressed uncertainty over the script being Phokian;
Johnston, in his revision of LSAG (p.438), affirms that it can be so considered.
23 See LSAG2 471; pl. 79, 7; EG 1:265266, with fig.119. The backstroke of the Samian letter
may continue slightly beyond the point at which it intersects with the crossbar and in this
way would differ from the typical CP Alpha-1.
24 See LSAG2 328 and pl. 63, 4. Jeffery points out the graphemic variation displayed in the inscription: This is an excellent example of the cursive Ionic script on stone; we may note the hasty,
indifferent use of 12 and 23[references to symbols in Jefferys notes on letter forms on
pp.325326]. The inscription is written up the side of the veil; for it was evidently a characteristic
Ionic practice to cut the inscription on the statue itself rather than on the base a sharp contrast
with the Attic habit, in which it is rare indeed to find the dedication anywhere but on the base.
25 See LSAG2 329 and pl. 63,5.
26 See LSAG2 pl. 65, 41c,d.
27 See LSAG2 340; pl. 66, 56. On this inscription and others left by Greek mercenaries in
Egypt, see Macfarquhar1966.
28 See EG 1:261262 with fig.115.
29 From later periods an alpha of the type of CP Alpha-1 is attested among western Ionic
alphabets, as in the fifth-century bc alphabets of the Euboian city of Styra, along with the
same type in reversed stance (see LSAG2 86; pl. 6, 26), and the Euboian colony of Zankle in
Sicily (see LSAG2 243; pl. 49,5).
30 See LSAG2 348; pl. 67,6.
31 See the examples provided by LSAG2 348; pl. 67, 4a and9.
32 LSAG2 133,314.
33 See LSAG2 434; pl. 73,4.
34 See McCarter 1975a:139.
35 See Langdon 1976:1618, with fig.7, 20 and pl.4.
36 Langdon 1976:17.
37 See Immerwahr 1990:xxii.
38 See Immerwahr 1990:133n5, and also p.61 on the last-named piece.
39 See Pandolfini and Prosdocimi 1990:3637.
40 See LSAG266.
41 See LSAG2 pl. 65,41A.
42 See Immerwahr 1990:xxii.
43 Immerwahr 1990:132. On the Little Master cups, see Immerwahr 1990:4555.
44 See LSAG2 479480 and pl. 80,2.
45 The total does not include the following ambiguous forms: -1/3 (W-1, line 4); -1/3 (W-1,
line 14); -1/3 (MS 1-1, line 5); -2/3 (MS 1-1, line 11); -2/3 (MS 2-2, line 7); -1/3 (MS 2-2,
line 14); -2/4 (MS 2-2, line 17). Included in the total is the singularly unusual alpha of MS
2-1, line 12: the intended morphology appears to be -2; in the narrow transcription of this
plaque that appears in Chapter4, the symbol is accordingly marked -2(?).
46 McCarter 1975a:78.
47 Sol-Sol 1966, with Tavv. 12; Gibson 1982:6466 (bibliography on p.65), with pl. III, 1;
Cross 1971, especially the figure on p.190; McCarter 1975a:47, 132133; Moscati 1999:146, for
photographic image.
48 LSAG223.
49 See Teixidor 1976:67, no.25, with bibliography.
294
Notes to Pages 2328
51 He writes (1975a:78): In both the Dipylon and Lacco Ameno examples of sidelong
alpha, the nose of the horizontal v points backward away from the direction of writing.
The cross-bar may slant in either direction. When the form is rotated and set on its feet,
the cross-bar continues to slope down forward in the direction of writing or back away
fromit.
52 Cross 1980:3.
53 See LSAG2 66 and pl. 61, 1a(i).
54 See LSAG2 pl. 1, 3c; Blegen 1934:15 and 17, fig.5. Jeffery (p.66) notes another occurrence on
a seventh-century graffito from the agora.
55 This Theran beta evolves into a distinct form in the sixth century; see LSAG2308.
56 See McCarter 1975a:78: The early development of beta is also quite problematic. The
form used on Thera is probably the most primitive type since it alone resembles early
Phoenician bet, albeit inverted, of any period. Note too that, as Jeffery states, beta has
more local variations than any other letter (LSAG223).
58 See Langdon 1976:1618 with fig.7, 20 and pl.4.
59 Kenzelmann-Pfyffer, Theurillat, and Verdan 2005:70,44.
60 See Immerwahr 1990:8, 134 and pl. 1, fig.2; Lang 1976:7 and pl. 1,A1.
61 See LSAG2 9091 and pl. 7, 1; EG 1:145146; Heubeck 1979:120, with Abb.44.
62 See, inter alia, LSAG2 236237 and pl. 48, 18; EG 1:228229, with fig.89; Buonamici 1932:
Tav. I, fig.1; Pandolfini and Prosdocimi 1990:1921 and Tavv.III.
63 See LSAG2 471 and pl. 79, 7; EG 1:265266, with fig.119.
65 McCarter 1975a:7879.
66 See Pandolfini and Prosdocimi 1990:3233, with Tav.XI.
67 The total does not include the following ambiguous, or otherwise uncertain, forms: -2/
(W-1, line 3); -2/ (W-1, line 4); -2 (?) (MS 1-1, line 18); -2/ (MS 2-2, line3).
68 LSAG223.
69 On the form of the name of the letter in Phoenician, see McCarter 1975a:93n71, with references to work on the problem.
70 Some instances of CP gamma have a crossbar that joins the spine of the letter at an angle
that is somewhat greater or less than ninety degrees, owing to the crossbar being slightly
elevated or slightly depressed and/or the spine deviating from vertical. Comparison with
the morphology of neighboring letters clearly reveals, I would judge, that these are not
attempts to produce the variant forms of gamma noted in the preceding discussion.
71 See LSAG2 pl. 48, 18; EG 1:228229; Pandolfini and Prosdocimi 1990:1921 and Tavv. III. The
raised crossbar of the gamma of the Marsiliana abecedarium may terminate in a hook (in
pi-like fashion); see, for example, the photograph of Pandolfini and Prosdocimi 1990, Tav.I.
72 EG1:90.
73 Immerwahr 1990:135. See his discussion for a possible seventh-century example in a
dipinto.
74 See, inter alia, Cordano 1984:282283, 291, 304, no.7.
75 See LSAG2 9091 and pl. 7, 1; EG 1:145146; Heubeck 1979:120, with Abb.44.
76 See LSAG2 471; pl. 79, 7; EG 1:265266 and fig.119.
77 Jeffery observes (LSAG2 289) that among the scripts of the Ionic islands of the Aegean
1[i.e., ] is the more common form (e.g. Naxos, Paros, Delos), but 2[i.e., ] occurs in
early inscriptions of Amorgos , Samothrace , and in Tenos ( 5thc.).
Notes to Pages 2831
295
78 See LSAG2 293 and pl. 56, 15; EG 1:157, with fig.40.
79 See LSAG2 293; pl. 56,17.
80 Compare the name inscribed in marble in Ionic letters from Samothrace
(IG XII viii 226; mid-sixth century); in which case, however, the top stroke of gamma is
angled upward, like that of the Marsiliana writing tablet noted earlier; see LSAG2 299; pl.
57,56.
81 See LSAG2 pl. 1, 3a; Immerwahr 1990:136; Langdon 1976:42.
82 See LSAG2 291 and pl. 55, 3; EG 3:554, fig.230b.
83 See LSAG2 293 and pl. 56, 15; EG 1:157.
84 See LSAG2 300301 and pl. 58, 61; EG 1:162164.
86 See LSAG2 120121, 441, and pl. 18, 1; Heubeck 1979:121, with Abb. 46; Stillwell and Benson
1984:4041 and pl. 122, (1)143ab.
87 While others would place them in the seventh century (or later); on their date, see Stillwell
and Benson 1984:56.
88 See EG 3:330332, with fig.111; LSAG2 468(8a)469; Masson1976.
89 See LSAG2 235236 and pl. 47, 1; EG 1:226227 with fig.88ab; Heubeck 1979:109116, with
Abb. 41; Arena 1994:1819 and Tav. I 2; Dubois 1995:2228.
90 See LSAG2 pl. 1, 1; EG 1:135136, with fig.28ab.
91 See LSAG2 pl. 1,3b.
92 See LSAG2 291 and pl. 55, 2; EG 1:154156, with fig.38ac.
93 See, for example, LSAG2 pl. 61, 1a (i); IG XII iii537.
94 See LSAG2 311 pl. 59, 1a; EG 1:187188, with fig.59ab; Whitley 1997:637, fig.1.
95 Jeffery (LSAG2 311)judges that the script of this seventh-century Cretan fragment generally shows a marked resemblance to that on Nikandras statue at Delos.
96 LSAG2 328; see pl. 63,2.
97 LSAG224.
98 But note the presence of a right-triangle delta in the Nestors cup inscription (mentioned earlier) from Pithekoussai, a colony of Euboian Eretria and Chalkis. Sixth-century
inscriptions from Eretria show (approximately) equilateral-triangle deltas. Jeffery (LSAG2
79)conjectures: This may be due to the influence of Attica or the islands. Guarducci (EG
1:217) concurs: Per probabile influsso dellAttica e delle Cicladi,
99 KAI 12. See, inter alia, Dussaud 1924:135142; Gibson 1982:1216.
101 See Bessios, Tzifopoulos, and Kotsonas 2012:360361.
103 See LSAG2 189190; pl. 35, 7. Jeffery (p.189) notes: The fragments were said to come from
Lebadea in Boiotia, but hydriai of this particular type have been found on many different
Greek sites, and the expert view holds that they were manufactured in Lakonia.
104 See LSAG2 446; pl. 75,2.
105 For both inscriptions, see LSAG2 210 and215.
106 IG V ii 555; see, inter alia, Brown 1977:59.
107 On the distinction of two fundamental types of archaic epsilon tailed and tail-less varieties see McCarter 1975a:81. The bifurcation mirrors the mix of Phoenician precursors
(McCarter 1975a:93). A tailed variety of he is seen in Cypriot Phoenician scripts as early as
the ninth century bc ; a variety with a short extension of the spine both above and below
the crossbars occurs earlier (see examples in McCarter 1975a:128133).
108 LSAG266.
296
Notes to Pages 3137
109 See LSAG2 238 and pl. 47, 3; Heubeck 1979:124; Arena 1994:29 and Tavv. VIIIIX; Dubois
1995:4142.
112 McCarter 1975a:9394. He further notes regarding the form of digamma: It may have
arisen as a deliberate modification of nearby he [i.e., Greek epsilon], when the letter-form
of waw was commandeered for the representation of the Greek vowel u (psilon). But this
is only speculative and raises further problems.
113 Immerwahr (1990:8, 140)sees a right-angular digamma in the partial abecedarium on a
loom weight from the Athenian agora (MC 907), dated circa late eighth century bc (see
his pl. 1, 3). While the letters are worn, the crossbars appear to descend slightly. See also
Lang 1976:7 and pl. 1, A 1; and Brann 1961:146 (with fig.1, R 22)and pl. 23,R22.
115 See Bessios, Tzifopoulos, and Kotsonas 2012:353354.
116 See LSAG2 91, 94; pl. 7,2b.
117 LSAG2 25, where Jeffery writes vau, rather than [w], in the second line of the citation.
118 LSAG2148.
119 SEG XI 290; XIV 314. See LSAG2 148, pl. 24, 5. See also Lejeune 1982:177178; Buck
1955:48.
120 See LSAG2 144; Jeffery tentatively suggests that this usage might possibly reflect a form of
script received without the agency of Corinth , which would imply that the alphabet of
this part of the Peloponnese came by another entry, independently of Corinth. On necessary emendations of Jefferys discussion of the alphabets of Phleious, Kleonai, Nemea,
and Tiryns, see Johnstons comments on pp.443444 of his revision ofLSAG.
121 And note that Johnston in his revision (see preceding note) remarks (p.444) that the
script of Tiryns is now attested as similar to Argive.
122 Compare the much later (second-century) Cretan spelling (genitive) for the
related month name . The spelling may reflect a phonetic change; see Bile
1988:118, 182; also Lejeune 1982:55,178.
123 See Lejeune 1982:177178 with 177n3.
124 On this somewhat complex matter, see especially Brixhe 1976:4657.
125 See Bile 1988:7576; Brixhe 1976:56; Johnstone1978.
126 Brixhe 2004a:283. I wish to express my appreciation to Professor Brixhe for kindly bringing this work to my attention and providing me with a copy of thesame.
127 Heubeck 1986:14.
128 LSAG228.
129 Coote 1975:48; see the accompanying figure. For bibliography of the work that preceded
Cootes study, see his p.50. For a photographic image, see Moscati 1999:145.
130 See McCarter 1975a:55: This short form of zayin is well known from the Aramaic inscriptions of the ninth century, where the three strokes show a tendency to break down into
a z produced in one stroke with three segments. That the same development had taken
place in contemporary cursive Phoenician is suggested by the occurrence in the new
Kition Bowl of a z-form zayin. This form was previously unattested in Phoenician inscriptions older than the early sixth century.
131 McCarter 1975a:94n75.
132 On Phoenician et and its value, see Woodard 1997:136, with n. 10; Hackett
2004:367370.
133 See LSAG2 2829.
Notes to Pages 3740
297
134 See McCarter 1975a:95: The various early Greek scripts show a diversity of forms of early
eta, including two, three, and four-barred examples, which corresponds to the potential
for variety that existed in Phoenician et throughout the history of the early Phoenician
scripts.
135 See, inter alia, Naveh 1980:2425; Cross 1974; Albright 1941:20, with fig.3 (inverted).
136 The letter will be identified by the name eta throughout this work, both when it is used
to represent a vowel and when it is used to represent the consonant /h/. The letter name
heta, which is sometimes used of the symbol with regard to the latter value, will not be
utilized herein.
137 On the Aiolic graffiti, see LSAG2359.
138 See, inter alia, LSAG2 291 and pl. 55, 2; EG 1:154156, with fig.38ac; Buck 1955:189190;
Heubeck 1979:124125, with Abb.52.
139 Spelled in red alphabets. For discussion of these alphabetic classifications, see
Woodard 1997:140141.
140 On the epithet, see Heubeck and Hoekstra 1989:261.
141 LSAG2 289. See also the summary of Guarducci at EG 1:152153. See also Buck 1955:1819,
189192; Smyth 1898:167, 327328; Thumb and Scherer 1959:251252.
142 See Johnstons remarks in his revision of LSAG, p.467. See also Matthaiou 1984:171172.
143 Inscription of the builder Ason; c. 550? IG XII v 252 (LSAG2 305; see also pl. 56,28).
144 Fragment of base for a dedication by Mikkiades from Paros; c. 550530? IG XII v 147
(LSAG2305).
145 EG 1:157. On Amorgos, see also LSAG2 293. The alphabet of Amorgos shows influence by
the Naxian and Samian scripts.
146 See EG 1:157, fig.40; LSAG2 293 and pl. 56,15.
147 Buck 1955:190; see also, LSAG2 298; Daux1949.
148 For Theran orthographic practice, see LSAG2 309, 317; EG 1:349. In the latter, Guarducci
notes that eccezionalmente, pu rappresentare le breve. On the phonetic value and
history of the eta-vowel in Theran dialect, see Buck 1955:29,168.
149 See LSAG2 345; on the phonetic value and history of the eta-vowel in the Rhodian dialect,
see Buck 1955:29,166.
150 See EG 1:327. Note the spelling
for (he dedicated) in the inscription appearing on a small bronze wheel, circa mid-sixth century, from Kamiros (EG 1:330,
fig.152; LSAG2 pl. 67, 13): Jeffery (LSAG2 349)refers to it as the badly spelt dedication.
151 See LSAG2 345346; EG 1:167.
152 So it is commonly identified, though see Johnstons remarks in A. Johnston 1998:252.
153 See Cassio 19911993 [1994]; Watkins 1995a:4245; LSAG2 116117, 237, and pl. 18, 2;
Cordano 1984:282283, 291 (for bibliography), 304, no.7.
154 See Cassio 19911993 [1994].
155 See LSAG2351.
156 See Guarducci 1985:711, with fig.1, and 1921.
157 For a different reading and interpretation of the inscription, see Arena 1994:7980.
158 The etymology of the goddesss name is unknown; Chantraine (1984:352) suggests that it
is probably pre-Hellenic. On the spelling of word-internal /h/ in h [], see Guarduccis
treatment of the problem (1985:14); one might wonder if the symbol could possibly
represent a mid-vowel in this word, as the identical symbol does in the orthographic system of Knidos.
159 Guarducci 1985:22.
160 See Guarducci 1985:19.
298
Notes to Pages 4044
161 See LSAG2 291 and pl. 55, 3; EG 3:554, fig.230b.

162 Guarducci 1985:20: Il si presenta ora associato al sigma ora, come nellepigrafe di Eny,
in valore di semplice spirito aspro.
163 Guarducci also draws attention to another Naxian inscription from Delos, perhaps late
seventh century bc , which Jeffery (LSAG2 292)characterizes as the puzzling fragments
of four lines on an unidentified piece of marble sculpture also from Delos, containing
apparently part of a dedication and a sculptors metrical signature. Guarducci (1985:20)
infers the use of in this inscription to represent /h/: il , posto davanti a vocale, ha
evidentemente il medesimo valore di semplice spirito aspro.
164 On the matter of blue and red alphabets, see, inter alia, Woodard 1997, especially 140141
and 208216.
Tav. I, fig.1; Pandolfini and Prosdocimi 1990:1921 and Tavv.III.
166 See Buonamici 1932: Tav. I; Bonfante and Bonfante 1983:108109; Pandolfini and
Prosdocimi 1990:2223 and Tav.IV.
167 See Pandolfini and Prosdocimi 1990:2426 and Tavv.VVI.
168 See, inter alia, Pandolfini and Prosdocimi 1990:2629, with Tavv. VIIVIII.
169 See LSAG2 236237 and pl. 48, 19; Buonamici 1932: Tav . II; Pandolfini and Prosdocimi
1990:2932 and Tavv.IXX.
170 See Pandolfini and Prosdocimi 1990:3233, with Tav. XI. The letters of the dipinti depart
from regular alphabetic order, so that it is impossible to be certain of the intended phonetic value of in this case; Pandolfini and Prosdocimi (1990:33) write: Siamo verisimilmente in presenza di uno pseudo-alfabeto che assolve funzione decorativa.
171 See Buonamici 1932: Tav. IV; Pandolfini and Prosdocimi 1990:3436.
172 See LSAG2 237 and pl. 48, 21; Pandolfini and Prosdocimi 1990:2122 and Tav.III.
173 LSAG2236.
174 For examples, see LSAG2 pl. 48, 1823.
175 Kenzelmann-Pfyffer, Theurillat, and Verdan 2005:60,3.
176 In an article that accompanies Kenzelmann-Pfyffer, Theurillat, and Verdan 2005, Wachter
(2005) treats this instance of the letter as necessarily occurring in an abecedarium,
labeling it a dead letter (ein toter Buchstabe). Contra the view that has no attested
use outside of an abecedarium, see the ensuing evidence provided by Argive epigraphic
tradition, revealing the symbol to be very much alive into the fifth centurybc .
177 Kenzelmann-Pfyffer, Theurillat, and Verdan 2005:7778.
178 For a second possible example of the symbol in a graffito from Eretria (dated to the
first half of the eighth century bc ), see Kenzelmann-Pfyffer, Theurillat, and Verdan
2005:75,64.
179 See Meritt 1945 and 1952:351355. Also see LSAG2 164 and pl. 29, 30. For the text of the
inscription, see especially Meritt 1945:144145 and Meiggs and Lewis 1980:35.
180 See Meritt 1945:136n8 regarding the use of the two different characters to spell/h/.
181 See Buck 1955:190.
182 See LSAG2 349 and 356 (no.15).
183 See LSAG2 348, 356, and pl. 67,4b.
184 Buck 1955:75, 190, 298. Rhodes uses a red alphabet; thus the spelling, to be more graphically precise, is , where is not dark blue chi /kh/ but red xi /k + s/. Rhodes uses the red
chi and .
185 For the inscription on this early fifth-century vessel, the kylix of Philto, see LSAG2 349
350, 357 (no.27).
Notes to Pages 4451
299
186 LSAG2 pl. 67, 1; EG 1:328, with fig.163. The final letter is only partially visible but the pleonastic spelling is probable in the light of parallels; see LSAG2346.
187 LSAG2 347348.
188 Dialects that Buck singles out, elsewhere in the same work (Buck 1955:75), as showing the
same redundant spelling of /k + s/ clusters.
189 For examples, see Buck 1955:75.
190 LSAG229.
191 McCarter 1975a:9596.
192 EG 1:94: Nel corso del VI secolo appare per non di rado un nuovo tipo. For sixthcentury occurrences in Attica and Eretria, see LSAG2 66,79.
193 See EG 1:94; LSAG2 89; Heubeck 1979:120, with Abb.44.
194 See, inter alia, LSAG2 pl. 48, 1819, 2122; Pandolfini and Prosdocimi 1990:20 (Tavv. III);
22 (Tav. III); 23 (Tav. IV); 28 (Tavv. VIIVIII); 31 (Tavv. IXX); and soon.
195 See LSAG2 91, 94, pl. 7,2b.
196 See LSAG2 322 and pl. 62,26.
198 McCarter 1975a:60. See subsequent discussion on each of these Cypriot Phoenician
inscriptions.
199 There is also co-occurrence of marked order here to the extent that eta has been deleted,
beyond which xi and theta have been inserted between sigma and tau in the preceding
abecedarium (which terminates in line 11); see the addendum to ChapterSix.
200 LSAG2 164165; see Richardson and Wheeler 1902:197202 for thetext.
201 See, for example, LSAG2 29; McCarter 1975a:82.
202 See EG 1:94; LSAG2 89; Heubeck 1979:120, with Abb.44.
203 See McCarter 1975a:60, 99,116.
204 For earliest-known examples of Canaanite ayin, see Darnell etal. 2005:84.
205 See Bessios, Tzifopoulos, and Kotsanas 2012:347349.
206 See, for example, IG XII iii 536, 540. On Thera, dotted omicron can be used to mark a long
(as opposed to short o), but the practice appears not to be consistent. See, for example, the remarks of Guarducci (EG 1:349). The segmental length of the sound encoded
by the grapheme has no crucial bearing on the problem of the origin of Greek dotted
omicron.
207 See, for example, LSAG2 pl. 48, 1920; Pandolfini and Prosdocimi 1990:2425 (Tavv.
VVI), 28 (Tavv. VIIVIII), 31 (Tavv. IXX).
208 See, inter alia, Naveh 1973; 1987:175186; 1991 for claims for an early date of adoption.
On the dotted ayin of the Aramaic Tell Fakhariyeh inscription, see the comments in
Woodard 1997:243n98, with references (and read the final sentence of the note as The
Tell Fakhariyeh inscription preserves a ninth-century dotted ayin).
209 SEG XI 316. For the text of the inscription, see Vollgraff 1910; 1913; 1948; see also LSAG2
165. Vollgraff 1948 offers extensive commentary on the Argive document as well as on
several fragments from Tylissos that appear to preserve a Cretan text of the treaty. On the
treaty, see, inter alia, Meiggs and Lewis 1980:99105.
210 For the text, see Vollgraff 1948:89 and plates III. Vollgraff discovered the inscription
in two parts, labeled A and B in his text; within the forty-eight lines of text is included a
lacuna of two unrecovered lines situated between parts A andB.
211 Les lignes 44 ( partir du mot ) 48 ont t ajoutes aprs coup sur la pierre par
un autre graveur: Vollgraff 1948:88. There is also variation in the forms of upsilon and
phi (see LSAG2 165n1), which may mean, as Jefferys surmises regarding the rider, either
300
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
Notes to Pages 5155
that it was an addition made by a different mason at a later date [per Vollgraff], or since
sidelong xi [i.e., rotated ninety degrees] is retained that the same mason added it [i.e.,
the rider] later, having meanwhile adopted some newer styles (LSAG2165).
Vollgraff 1910:335.
For a broad orthographic transcription of the text, see Volgraff 1948:78. For a photograph of the fragment, see IC 1:307.
LSAG2 165; she enumerates: I [i.e., ], 1[i.e., ] or 4[i.e., ], 1[i.e., ], 45[i.e., and
, respectively], 3[i.e., ], 3[though Guarduccis photographic image reveals , Jefferys
2], 3[ ].
See LSAG2 320; pl. 62, 20. The formal equivalence of the symbols is no less so if DobiasLalous (1970) suggestion that the central dot of the omicron was made by a tracing compass while that of the theta was not should prove accurate. The other two (undotted)
omicrons in the inscription appear to give no evidence of a point created by a compass (or
otherwise). Some scholars have claimed that the use of a drawing compass lies at the origin of dotted omicron; as McCarter (1975a:8586) and others have pointed out, a compass
can clearly not in all cases be the source of the central dot, as in the very early rock graffiti
from Thera and other inscriptions in which a central dot is present but the irregularity of
the circumference of the letter eliminates the possibility that a compass wasused.
Chamoux 1953:264n1.
Dobias-Lalou 1970:249.
See Boardman 1980:153159. For a recent treatment of the founding of the Greek colony
at Kyrene, see Calame2003.
See Johnstons comments in LSAG2 470: The picture of the Kyrenaic alphabet that we
now have is not particularly tidy: the Theran alphabet is well represented, but from 550
non-Theran forms also appear, probably under Rhodian influence ; this script gradually dominates.
See IG XII iii 541 and Supplement, p.308.
On the sense of the adverb, see Lerat 1944:7n1; Lejeune 1939:56n1a, 102n1.
Compare the shape of the previously noted second possible example of the symbol in
a graffito from the sanctuary of Apollo Daphnephoros at Eretria (first half of the eighth
century bc ); see Kenzelmann-Pfyffer, Theurillat, and Verdan 2005:75,64.
Lerat 1944:9.
See LSAG2 190n4.
See Jacquemin 1999:59n172.
See LSAG2 91, 94; pl. 7,2b.
See, inter alia, Pandolfini and Prosdocimi 1990:20 and Tavv.III.
Though Jeffery is inclined to identify the inscription as Boiotian rather than Euboian. See
the remarks in my discussion of CPPi-1.
LSAG2237.
See Gibson 1982:6668 and 182, fig.6; for bibliography, see p.67.
See Carpenter 1933:1214.
Sass 2005:134.
McCarter 1975a:56.
Lerat 1944:1112.
See his remarks on p.12 regarding the deliberate reproduction of the shape of these symbols at a later period.
See LSAG2 179182 and pl. 34,12.
See LSAG2 140, 143, and pl. 23, 2; EG 1:335336, with fig.169a.
Notes to Pages 5559
301
238 See LSAG2 141, 143, 442, and pl. 23, 8; EG 1:337338, with fig.170.
239 LSAG2 138. But Jeffery does not divorce the introduction of the symbols from a utilitarian
interpretation: of the advent of the diamond-shaped theta she writes probably because
it was easier to cut into stone or bronze. Again that may be so, but ease of production
may simply have been a benefit of the change, not the motivation for it. She notes regarding the diamond-shaped symbols: Sporadic examples of this practice appear in other
places, but only in Sikyon could it be called consistent. Presumably it would have been
equally easier to cut diamond-shaped theta and omicron in other locales.
240 Langdon (2005:176) observes: Its square shape is an uncommon Attic letter form. He
describes the head of Barako qoppa as a squarish circle.
241 Langdon 2005:176.
242 See LSAG2 202 and pl. 39, 64. Pi also has perceptible curvilinearity.
243 See LSAG2191.
244 See LSAG2 314, 316, and pl. 60, 27; EG 1:190191 with fig.61.
245 Lang 1982:78.
246 On sigma morphology in the Attic script, see particularly Immerwahr 1990:157160, with
references.
247 Lang 1982:83.
248 However, I believe that her claim that deviant forms of sigma were used to record forms of
the fricative /s/ that somehow differed (phonetically? / phonologically?) from the /s/ represented by forward-facing three-stroke sigma will not hold up to linguistic scrutiny. The
use of double sigma that she invokes in her discussion is certainly a reflection of speaker
intuitions about the placement of syllable boundaries.
249 Lerat 1944:9.
250 See Langdon 1976:43. An early sixth-century example of square omicron and perhaps of
square theta as well can be found on the fragments (soft poros covered with stucco)
from the Athenian Agora; see DAA 2. Interesting examples are provided by the Pericles
ostracon (P16755), mid-fifth century; see Shear 1940:23, with fig.2. Note too the occurrence of the symbol on the fragmentary inscription of DAA 31, where it is found alongside rounded (and dotted) omicron and rounded theta, circa early fifth centurybc .
252 See Chapter 7 for more detailed consideration of this matter.
253 Noted by Brixhe 2007:29. The image of CP iota that Brixhe provides as an illustration
(fig.7) is an example of that type that I have labeled as Iota-1 (see immediately following),
and one having an upward stretching stroke that is angled slightly toward the direction of
writing.
254 See LSAG2 320 and pl. 62, 23; EG 1:321324, with fig.159ab.
255 See LSAG2 308; EG 1:322324.
256 See Teixidor 1976:67, no.25.
257 See Gibson 1982:3941, and 182, with fig.5; for bibliography, see p.40.
258 The total does not include the following ambiguous forms: -1(?) (W-1, line 12); -2/3 (W-2,
line 13); -2/3 (MS 1-1, line 15). Note too the unusual shape of the iota in line 2 of MS 2-1,
labeled as Iota-2: it lacks a distinct vertical extension at the distal end of the diagonal cross
stroke, showing instead a shallow-inclining extension.
259 See, inter alia, LSAG2 pl. 1, 1; EG 1:135136, with fig.28ab.
260 See Immerwahr 1990:146147.
1984:4041 and pl. 122, (1)143ab.
302
Notes to Pages 5960
262 As, for example, in the rock graffiti of IG XII iii 536540 (ca. eighth century bc and later)
and the amphora graffito (ca. first half of the seventh century) of LSAG2 pl. 61, 1a(ii).
263 See, for example, the law code from Dreros; LSAG2 pl. 59, 1a; EG 1:187188, with
fig.59ab.
264 LSAG2 29. Jeffery (pp.18, 29)suggests that the particular form of Phoenician yod that provided iota was a cursive form like the yod found on Samaritan sherds.
265 McCarter (1975a:96) analyzes the S-shaped iota as having its origin in the archaic, deeply
rounded form of tenth- and ninth-century yod [which] found its last examples in the
script of the Kition bowl (from Cyprus; ca. 800bc ).
266 LSAG2 2930; EG 1:95; McCarter 1975a:8283. On the selection of sigma and san by the
various archaic alphabets, see Woodard 1997:184188. The wording of Jefferys claim
requires examining; what she actually says is essentially that the presence of crooked iota
implies the presence of san: one may guess that the alteration from crooked [iota] to
straight was made by one of the first receivers [of the alphabet], whose dialect required
the use of (unvoiced) sigma (e.g. Rhodes, Euboia, Ionia ?), and straight iota was thus
passed on to the numerous places which ultimately derived their script by this channel,
and which, according to this view, never had the chance of using crooked iota (LSAG2
2930). By this reckoning it would seem that the selection of san or sigma went hand-inhand with the selection of crooked or straight iota that is, the same dialects also never
had a chance of using san; the nature and intricacies of the implicational hierarchy that
the proposal entails are in need of clarification and closer scrutiny.
267 LSAG2427.
268 See LSAG2 453 and pl. 76,2.
269 A. Johnston 1983:64.
270 Johnston refers to the inscription in a more recent work, where he writes: What is important is that we have here, at a date which is as secure as any and seems to be before 775, an
alphabetic text, perhaps in Greek language, employing at least one vowel, inscribed on an
imported pot at a Latin site. The next alphabetic texts of the Mediterranean world are a
generation or so later, and we may note that one of them, the famous Dipylon oinochoe
, employs a set of letter shapes perhaps closer to the Phoenician model, esp. the iota, if
we read straight iota on the Osteria sherd (A. Johnston 2003:263).
271 On the possibility of a ninth-century dating based on recalibration of carbon 14 datings,
see the summary remarks in Sass 2005:156, with references. For Sass a ninth-century dating would abrogate a Greek reading of the inscription. Elsewhere (see the discussions
of Woodard 1997:205245), I have proposed that a ninth-century date for the earliest
phases of the Greek alphabet may be more realistic than an eighth-century date (not an
idea original to me); and I would not view a ninth-century dating of the Osteria dellOssa
inscription to be necessarily problematic for the identification of the language of the
inscription as Greek.
272 Bietti Sestieri, De Santis, and La Regina1990.
273 Ridgeway 2004:42.
274 Watkins 1995a:38.
275 See the diagrams and accompanying descriptions in Holloway 1994:110112.
276 See Holloway 1994:108.
277 On the flask and its inscription, in addition to Bietti Sestieri, De Santis, and La Regina
1990; Ridgeway 2004; Watkins 1995a; and Holloway 1994, see Bietti Sestieri and De Santis
2000:53 and Bietti Sestieri 1992:184185.
278 McCarter 1975a:9394.
Notes to Pages 6268
303
279 On the limited preservation of PIE *y in Mycenaean Greek, see, inter alia, Lejeune
1982:162, 165, 167169,171.
280 En grec alphabtique, la notation de ces consonnes de transition nest rgulire quen
pamphylien; Lejeune 1982:163164. For the forms cited, see Brixhe 1976:168172.
281 See Buck 1955:52; see also Thumb and Kieckers 1932:113114.
282 On this inscription on a gold ring, probably from the Heraion, see Tracy1986.
283 Buck 1955:52; LSAG2 140, 143, and pl. 23, 2; EG 1:335336, with fig.169a.
284 Buck 1955:49. As with the use of iota for writing phonetic [y], the use of digamma to spell
phonetic [w] is regular only in Pamphylian (among Greek users of the alphabet); see
Brixhe 1976:5253; Lejeune 1982:163164. Lejeune further notes (p.164) that la letter
survit mme jusquau vie sicle en ionien des les, en ionien dEube et Athnes, dans cet
emploi
285 See Lejeune 1982:163.
286 Brixhe 2004a:283.
287 On the pre-alphabetic status of the Greek syllabary of Cyprus, see, inter alia, Woodard
1997:56. On an orthographic continuity connecting the two Greek syllabic scripts with
the Greek alphabet, see Woodard1997.
288 See LSAG2 34; EG1:98.
289 This, at least, is the conventional interpretation of the value of qoppa. But see Woodard
2012, where I have argued for evidence that qoppa had been utilized by the Greek adapters of the Phoenician consonantal writing system to encode a Cypriot dialect-specific
labiovelar sound that disappeared not long after the creation of the alphabet; the letter
qoppa was then stranded without distinctive phonemic value and was reinterpreted as an
allograph of kappa, as indeed it must have been so utilized automatically by a majority of
the first non-Cypriot-speaking users of the alphabet.
290 On the Greek adaptation of Phoenician consonantal characters for spelling vowels, see
Woodard 1997:135136, 148, 159, 250251.
291 Brixhe (2007:29) draws attention to the form of iota seen in an inscription on a
bronze tablet from the Akhaian colony at Francavilla Marittima (on the east coast of
Calabria, north of Sybaris; on the Akhaian materials from the south of Italy, see recently
Papadopoulos 2001). He would identify this crooked, three-stroke iota-morphology as
comparable to that of the Schyen copper plaques. While I would be hesitant to make
that equation, Brixhe is certainly accurate regarding an evolutionary trajectory when,
taking note of the serpentine-like iota visible in the same document, he observes: La
coprsence du ita serpentin habituel montre quon pouvait passer facilement de lun
lautre. For this Akhaian inscription, see, inter alia, EG 1:110111, with fig.14.
292 And for that matter, one might conversely suggest that CP Iota-2 could easily have been
modified to produce the special Pamphylian digamma .
293 And an evolution of Greek Iota-3 into Iota-1/2 in a Phoenician milieu would not necessarily entail an intentionality of assigning separate values toeach.
294 Brixhe 2004a:282284.
295 On the Phrygian alphabets, see Brixhe 2004b:781.
296 In Brixhe 2004a this seems to be a specifically Phrygian process; but he develops his
argument in conjunction with Greek practice, writing that the Phrygian pair (/i/) and
(/y/) voque le double trac, rectiligne et serpentin, du iota grec, et . In a more recent
article (2007:1617), Brixhe explicitly assigns the bifurcation to the Greeks, drawing in
evidence provided by the Lemnian stele and the language of Zn-Samothrace; on the
latter materials, see Brixhe2006.
304
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
Notes to Pages 6971
LSAG230.
EG1:95.
McCarter 1975a:83. See the earlier discussion of the Nikandra statue.
See Peruzzi 1973: Tav. IV; Heubeck 1979:123, with Abb. 49; A. Johnston 1983:64, with fig.3;
LSAG2 453; Dubois 1995:2930.
LSAG2 453; see also A. Johnston 1983:64.
See Peruzzi 1973:2526 and Tav.IV.
Coldstream 1977:300.
See Heubeck 1979:119, with Abb. 43; LSAG 6970 and pl.1,2.
See, inter alia, Buonamici 1932: Tav. I, fig.1; Pandolfini and Prosdocimi 1990:1921 and
Tav. I Tav.II.
See Pandolfini and Prosdocimi 1990:2629 and Tavv. VIIVIII.
See EG 1:341342, with fig.172; Arena 1998:116 and Tav. XVI40.
See LSAG2 471 and pl. 79, 7; EG 1:265266, with fig.119.
See LSAG2 291 and pl. 55, 2; EG 1:154156, with fig.38ac.
Langdon 1976:43. See p.15 with fig.6, 9 and pl. 3 for H 130; and fig.7, 11 and pl. 3 for
H217.
For this particular fragment, see Arena 1994:17 (1.c) with Tav.I1.
See LSAG2 235236 and pl. 47, 1; EG 1:226227, with fig.88ab; Heubeck 1979:109116, with
See LSAG2 pl. 1, 1; EG 1:135136, with fig.28ab.
See LSAG2 120121, 441, and pl. 18, 1; Heubeck 1979:121, with Abb. 46; Stillwell and Benson
1984:4041 and pl. 122, (1)143ab.
M 2248 and 2219 respectively: see Bessios, Tzifopoulos, and Kotsanas 2012:339343,
368369.
See EG 1:94; LSAG2 89; Heubeck 1979:120, with Abb.44.
See LSAG2 91, 94, and pl. 7,2ab.
See LSAG2 231, 234, and pl. 45, 2; EG 1:275276, with fig.126.
See LSAG2 238 and pl. 47, 3; Heubeck 1979:124; Arena 1994:29 and Tavv. VIIIIX; Dubois
1995:4142.
See LSAG2 237 and pl. 48, 21; Pandolfini and Prosdocimi 1990:2122 and Tav.III.
See Pandolfini and Prosdocimi 1990:2223 and Tav.IV.
See LSAG2 236237 and pl. 48, 19; Pandolfini and Prosdocimi 1990:2932 and
Tavv.IXX.
Langdon 1976:43. In one of these cases, however, H 224, at least a small gap seems to
occur between the points at which the arms touch the spine, though the photograph of
the sherd lacks clarity; see Langdon 1976:36, with fig.15, 136 and pl.12.
See LSAG2 311 and pl. 59, 1a; EG 1:187188 with fig.59ab; Whitley 1997:637, fig.1.
See LSAG2 319. To this she adds the practice of making a kind of ligature with mu or nu
and a vowel.
As in IG XII iii 552, 573, 582 and in the Supplement, 1440 and1463.
Though kappa with curled ends can be seen, as in IG XII iii 552, where one of the kappas
has legs that curl at the ends, but the second has legs thatbow.
See EG 1:184, with fig.56; LSAG2 313, 315, and pl. 59,2.
The bowing of the upper arm of the kappa of line 9 is subtle.
See McCarter 1975a:83, where he also notes: By contrast, many early examples of kappa
show a vertical shaft that is very little or not at all longer than the widest extent of the
horizontal v. Of course, this type gradually became standard everywhere.
Notes to Pages 7276
305
331 For these two inscriptions, see, respectively, Langdon 1976:27 with fig.11, 66 and pl. 8; and
p.36 with fig.15, 141 and pl.13.
332 Not included in these counts is the morphologically ambiguous Kappa-1/2 of MS 1-2,
line11.
333 These numbers do not include the omission of the better part of one entire alphabetic
sequence (epsilon thorough san), in line 10 of MS2-2.
334 For exemplars, see McCarter 1975a:132133. And angular Phoenician lamed is also known
from much earlier periods; see McCarter 1975a:37 and 58, on which latter page he notes
that although the short and rounded lamed of the Karatepe inscriptions is unquestionably the ancient form and recalls second-millennium examples, the long and angular
form was already present in the script of the Airam sarcophagus; and since both types
survive into the eighth century, neither length nor degree of roundness may be regarded
as a dependable measure of antiquity.
335 See LSAG2 pl. 47, 1; EG 1:226227, with fig.88ab.
336 See, for example, LSAG2 pl. 61, 1a (ii); IG XII iii, 536, 537, 538, 539,etc.
337 See LSAG2 pl. 67, 1; EG 1:328.
1984:4041 and pl. 122, (1)143ab.
339 See LSAG2 322 and pl. 62,26.
341 LSAG2 3031.
342 See EG 1:341342, with fig.172; Arena 1998:116, with Tav. XVI40.
343 See Immerwahr 1990:xxii,149.
345 See Langdon 1976:25, with fig.10, 52 and pl. 6. The inscription on the fragments of the cup
H 69 (/H 70)probably also preserves a similar lambda, though in this case a letter more
of the type of CP Lambda-1; see Langdon 1976:18, 43; and Blegen 1934:1819, with fig.6,
nos. 13 and14.
346 See Guarducci 1985:711, with fig.1. Compare Arena who interprets the letter as a gamma:
Arena 1994:7980.
348 EG 1:95. Similarly, Jeffery (LSAG2 30)summarizes: Those which use the form closest to
the Semitic were the group Euboia, Attica, Boiotia, and Opountian Lokris, and certain
scattered places in Crete Dreros, Knossos, Eltynia, and Praisos.
350 The total does not include the ambiguous form of MS 2-2, line 1, classified as Lambda-2/3;
nor does it include the lambda-like form (Iota-2?) of line20.
351 See Gibson 1982:2528 (bibliography on page 26), with pl. II, 2; Cross 1972, especially
fig.1; McCarter 1975a:4243, 130131; Carpenter 1958:4748, with pl. 5, figs. 2 and3.
352 See Gibson 1982:6871 (bibliography on p.69), with fig.7; McCarter 1975a:4849, 132133;
Carpenter 1958:4748, with pl. 5, fig.4.
353 See Sol-Sol 1966, with Tavv. 12; Gibson 1982:6466 (bibliography on p.65), with pl. III,
1; Cross 1971, especially the figure on p.190; McCarter 1975a:47, 132133; Moscati 1999:146,
for photographic image.
354 See LSAG2 8182; McCarter 1975a:84; EG1:95.
355 See, inter alia, Pandolfini and Prosdocimi 1990: Tavv. III.
356 See Coote 1975 (especially the figure on p.48); McCarter 1975a:44, 132133. For a photographic image, see Moscati 1999:145.
306
Notes to Pages 7679
357 See Dubois 1995:2122; Arena 1994:17 and Tav. I 1; A. Johnston 1983:6364, with fig. 1;
LSAG2 453454 and pl. 76,2.
358 See LSAG2 321, 324, and pl. 62,25b.
359 See the examples at EG 1:184, fig.56; and LSAG2 pl. 59,2.
360 On the date of the Gortyn legal materials and general discussion with bibliography, see
Davies 2005:306.
1995:4142.
362 Jeffery writes that the form is regular in Eretria and Kyme until the start of the 5th c.
(LSAG279).
363 Again, Jeffery states that m-1[ ] -2[ are] normal for Crete, Melos and Sikinos (LSAG2
309). Examples of this morphology of mu can also be seen, alongside that type having an
elongated spine leaning away from the direction of writing, in use at the temple of Apollo
Pythios at Gortyn (see EG 1:184).
364 See Dubois 1995:2122; Arena 1994:17 and Tav. I 1; A. Johnston 1983:6364, with fig. 1;
LSAG2 453454 and pl. 76,2.
365 LSAG2 454. Johnston makes explicit reference to the occurrence of both archaic fivestroke and archaic four-stroke forms of mu and to four-stroke and three-stroke forms of
sigma.
366 See LSAG2 235236 and pl. 47, 1; EG 1:226227, with fig.88ab; Heubeck 1979:109116, with
367 See Bessios, Tzifopoulos, and Kotsonas 2012:339343; cf. the highly stylized mu of
2249 (also late eighth or early seventh century) on pp.337339.
368 See Pandolfini and Prosdocimi 1990:2426 and Tavv. V andVI.
369 See Buonamici 1932: Tav. IV; Pandolfini and Prosdocimi 1990:3436.
370 See LSAG2 8485 and pl. 5, 9 A12.
371 See LSAG2 311 and pl. 59, 1a; EG 1:187188, with fig.59ab; Whitley 1997:637, fig.1.
372 See LSAG2 322 and pl. 62, 27. The form also occurs in the previously mentioned inscription from the temple of Apollo Pythios in Gortyn, and so co-occurs with mu having a
spine sloping away from the direction of writing; see LSAG2 pl. 59, 2. A similar coexistence of four-stroke mu-forms one type with vertical spine, another with spine sloping
outward is noted by Immerwahr (1990:149) in his study of the Attic alphabet. He writes
that the slanting and vertical forms of high mu are equally early, and continues,
observing, both shapes are standard on vases until the middle of the sixth century and
still occur in the third quarter.
373 See Gibson 1982:6668 and 182, fig.6; for bibliography, see p.67.
374 The mem of lines 1 and 5. See Gibson 1982:7276 and pl. VIII, 1; for bibliography, see
pp.7374.
376 See LSAG2 236237 and pl. 48, 19; Pandolfini and Prosdocimi 1990:2932 and
Tavv.IXX.
377 For a photographic image of the syllabary, see Buonamici 1932: Tav. II; and Pandolfini
and Prosdocimi 1990: Tavv. IXX. For a drawing of the syllabary (and the accompanying
alphabet), see Pandolfini and Prosdocimi 1990:31. For brief discussion of the bottles syllabary vis--vis its abecedarium, see Einarson 1967a:18 and his corrections of that work in
Einarson 1967b:262.
378 LSAG231.
Notes to Pages 7985
307

380 Jeffery (LSAG2 237)mentions the four-stroke nu (and six-stroke mu): The bucchero bottle from the Regolini-Galassi tomb may be dated tentatively in the second half of the seventh century. Here, as a further aid to the learner, a syllabary was inscribed spirally round
the bottle (, , , &c.); mu and nu have elaborated forms, with an extra stroke at the
end. Einarson must certainly be mistaken when, in his Notes on the Development of
the Greek Alphabet (1967a:18), he transcribes the four-stroke nu symbols of the syllabary
as further occurrences of mu (though he makes no overt correction of it in his later published Corrigenda [1967b]) that is, as a second set of mu + vowel syllables, following
the set of six-stroke mu + vowel syllables. In the Corrigenda (p.263) he correctly notes
the occurrence of four-stroke nu in the abecedarium that winds around thebase.
383 See Gibson 1982:2528, with pl. II, 2; Cross 1972; McCarter 1975a:4243, 130131.
384 See LSAG2 pl. 1, 1; EG 1:135136, with fig.28ab; Heubeck 1979:116118, with Abb.42.
385 See Immerwahr 1990:155.
386 See the high-resolution image of Whitley 1997:637, fig.1.
387 See EG 1:187188 and fig.59ab. The drawing of LSAG2 pl. 59, 1a, mistakenly reproduces
an epsilon in both instances. In her volume on the Cretan dialect, Bile (1988:241n337)
states that she considers the variation of and to reveal un example de faute
graphique, in light of common orthographic practice in other archaic Cretan documents (see her transcription [after Guarducci] of the Dreros text on p. 29). She notes
(p.240n334) that the same orthographic variation occurs, inter alia, in the inscription
on a bronze mitra from Lyttos (ca. 500 bc ; see Jeffery and Morpurgo Davies 1970; Bile
1988:3740), itself an important document for the social history of Greek alphabetic
writing (see Thomas 1992:6971), where (Bile no. 28, l. 6) occurs beside
(l. 5)to serve as remembrancer/recorder. Intentional scribal variation in
these particular documents cannot, however, be a priori eliminated.
388 See Langdon 1976:32, with fig.13 (and pl. 11), no.106.
389 See EG 1:137 and fig.29; LSAG2 69, 76, and pl. 1, 4; Lang 1976:30.
390 See LSAG2 pl. 10, 22. Jeffery (LSAG2 85), however, groups this inscription from Eretria
(Euboia) with finds from Boiotia on the basis, in part, of two of its letters: a four-bar eta
symbol (spelling [h]) that is attested in a Boiotian inscription from Thebes (see LSAG2
190n4 and the earlier discussion of a Laconian inscribed lebes preserving the same letter) and this very pi symbol, which she appears to liken to the Boiotian pi that matches
CP Pi-3, a rectilinear form of CP Pi-1. However, the four-bar eta and a pi of the CP Pi-1
type also occur in Euboian abecedaria from Etruria, the former in the Marsiliana writing
tablet and the latter in the syllabic inscription from Caere and in an oracular inscription
from Kyme (see the next note).
391 See EG 1:229230, with fig. 90; LSAG2 238, 240; Arena 1994:33 and Tav. X 4; Dubois
1995:4446 (the facsimile of the inscription on p.44 of Dubois does not accurately represent the symbol, incorporating what appear to be surface scratches into its outline).
392 Thus Guarducci (EG 1:230) describes it as being of the archaic curvilinear variety:
Lantichit dellepigrafe confermata dal pi a uncino( ).
393 See the examples at EG 1:184, fig.56; and LSAG2 pl. 59,2.
394 See LSAG2 313 and pl. 59, 12. For additional, helpful images, see IC 1:297298.
395 See, for example, the rock graffito of LSAG2 pl. 61, 1a (i) (IG XII iii 357)and the gravestone
inscriptions of pl. 61, 5 (IG XII iii 762; ca. late seventh century bc ) and 12 (iii) (IG XII iii
308
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
Notes to Pages 8587
779; ca. second half of the sixth century); for further examples, see, inter alia, the rock
graffiti of IG XII iii 540 and537.
See LSAG2 466 and pl. 78, 10. The pi of early inscriptions written in the alphabets of Naxos
and Samos (on the relationship, see LSAG2 293)appears, however, typically as the common right-angular type, like that of CP Pi-4, discussed later; see examples at LSAG2 pl. 55
(Naxos) and pl. 63 (Samos).
For the inscription from the sanctuary of Apollo, see LSAG2 92 and pl. 7, 4. Regarding these
two types of pi (i.e., types equivalent to CP Pi-2 and Pi-3) in archaic Boiotian inscriptions,
Jeffery (LSAG2 89)makes the summary observation that the curved or crooked forms
are normal in the 7th and 6th c.; thereafter 3[= CP Pi-4] becomes regularly used. Note,
however, that her 3[= CP Pi-4] is attested already in seventh-century Boiotian writing.
See Pandolfini and Prosdocimi 1990:1921 and Tavv.III.
See Buonamici 1932: Tav. I; Bonfante and Bonfante 1983:108109; Pandolfini and
See, inter alia, Buonamici 1932: Tav. II; Pandolfini and Prosdocimi 1990:2932 and
Tavv.IXX.
Though not the form of rho found in the abecedarium inscribed on the bottle. Such a rho
does occur, for example, on the bird-shaped bucchero bottle mentioned earlier; see, inter
alia, Pandolfini and Prosdocimi 1990:2223 and Tav.IV.
The letter appears somewhat more rounded in the excellent facsimile of Pandolfini and
Prosdocimi 1990:31 than on the vase itself.
Again, the character in the facsimile of Pandolfini and Prosdocimi 1990:31 is drawn overly
round, showing a curved serif at the distal end of the diagonal crossbar; in contrast the
symbol shows a vertical tick with an angular join to the diagonal in photographic images.
The peculiar example of pi on MS 2-1, line 14, is fundamentally a Pi-2, but the scribe has
attached an angular tick to the distal end of the letters loopinghead.
See Bessios, Tzifopoulos, and Kotsanas 2012:343344.
See LSAG289.
See LSAG2 91, 94, and pl. 7,2b.
See LSAG291.
See LSAG2 183 and190.
See Lerat 1944. Recall that Lerat (pp.89) leaves open the possibility that the inscription might not be Lakonian, largely on the basis of the particular form of eta (for /h/),
the same form having four horizontal bars that one finds in the Euboian script of the
Marsiliana writing tablet, but not known in Lakonian.
On the dating of these objects, see also Johnstons remarks at LSAG2448.
Adding, which is a further disconcerting feature for those seeking to date the early
Lakonian alphabet; LSAG2190.
And, indeed, gamma appears to have been used in lieu of pi on plaque W-1, line 15. A
painted inscription on a sherd from Aigina, found at the temple of Apollo and dated circa
late eighth century bc , shows this type of pi (i.e., the type of CP Pi-4), but without the
small descending stroke again, looking very much like an archaic gamma; McCarter
(1975a:86) judges the letter to be of dubious connections. For an image of the sherd, see
LSAG2 pl. 16, 1. The letter is wrongly drawn in the illustration of EG 1:196, as Guarducci
pointsout.
On copper plaque MS 1-2, line 10, there is a Pi-4 having a tick that extends both above
and below the horizontal arm. An identically shaped pi appears in the drawing of the
Notes to Pages 8790
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
309
grave stele of Polynoe from Korkyra, circa 650600 bc (see LSAG2 232233 and pl. 46, 8),
though the whereabouts of the object are apparently now unknown.
Again, to gain an impressionistic sense of how widely dispersed the graph is among the
archaic alphabets, one need only glance at the Table of Letters at the end of LSAG2.
See LSAG2 235236 and pl. 47, 1; EG 1:226227, with fig.88ab; Heubeck 1979:109116, with
On the meter of the first line especially and of the entire inscription, see Watkins 1976
and 1995a:4142. Some, such as Powell (1991:165), have embraced the opposing and, in
my view, unlikely interpretation that the line is prose; in Powells case, the interpretation
seems to be required by the social scenario in which he imagines the inscription to have
been etched onto thecup.
See the detail of the inscription in Metzger 1965, pl.17.
See Metzger 1965, pl. 17 for a photographic image.
Kenzelmann-Pfyffer, Theurillat, and Verdan 2005:70,43.
See Langdon 1976:18, 20 with fig.8 (and pl. 4), no.27.
See LSAG2 69 and pl. 1,3b.
See Langdon 1976:22, with fig.9 (and pl. 5), no.42.
See LSAG2 91 and pl. 7,3bc.
Recall that we noted that in the second commemorative inscription on that lebes there
also occurs a pi like that of CPPi-4.
See LSAG2 156 and pl. 26, 3; EG 1:127128 and fig.24b.
See LSAG2 291 and pl. 55, 3; EG 3:554 and fig.230b.
See LSAG2 300301 and pl. 58, 61; EG 1:162163 and fig.45.
See Cordano 1984:281282, 290, and303.
See Cordano 1984:282, 291, and 303; LSAG2 125n3.
The total does not include the following ambiguous forms: -1/3 (MS 1-1, line 2); -1/3
(MS 1-1, line 6); -1/3 (MS 1-1, line17).
LSAG233.
McCarter 1975a:99.
McCarter 1975a:86.
McCarter continues regarding this asymmetric form (1975a:86n64): The best examples
are found in old Etrurian abecedaries on bucchero ware objects. He makes particular
note of one of the examples in Jefferys LSAG, in which the back leg of san is omitted entirely: Jeffery, pl. 48 (19). There seems here, however, to be a mistake in Jefferys
illustration. The object under consideration is the bottle from Caere, mentioned several
times above, on which an abecedarium is found encircling the base of bottle with syllabic combinations engraved above it. Examination of photographs of the bottle (see, e.g.,
Buonamici 1932: Tav. II; Pandolfini and Prosdocimi 1990: Tavv. IXX [and see also the
drawing on p.31]) reveal the distinct presence of the back leg that is missing in Jefferys
rendering. Note also that the transcription of san in the abecedarium of Jefferys plate 48
(22) appears to be inaccurate as well; this is the inscription on the bird-shaped bucchero
bottle from Viterbo, also mentioned several times already. In Jefferys illustration, the
initial stroke (the back leg) appears to be longer than the three remaining strokes, each
of which is about equal in length. The morphology of the letter would thus be fundamentally like that of the CP sans (as detailed later), at least with respect to relative stoke
lengths. Photographs of the inscription (e.g., those appearing in Bonfante and Bonfante
1983:108109; and Pandolfini and Prosdocimi 1990: Tav. IV) reveal, however, a letter having an initial stroke and a final stroke (back leg and front leg respectively) that are of
310
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
Notes to Pages 9093
about equal length (the first perhaps being slightly longer), and two inner strokes producing a v-form that is shorter (i.e., a letter approximating the shape of the idealized
symbol).
McCarter 1975a:86.
Three fundamental morphologies are identified: San-1, San-2, and San-3. Yet finer classificatory distinctions could perhaps be made, but will not be. The limited nature of
comparative evidence from other archaic alphabets mediates against this, and establishing additional morphological classes would entail identifying sets with few members.
Particular features of individual examples will sometimes be noted, however.
In one of its occurrences, W-1, line 18, the two inner strokes of San-1 (those producing a
v-form) are shorter than the two outer strokes, as in the typical shaped archaic san,
though, in contrast to the later, the back leg of the CP character remains longer than the
front leg. The San-1 of W-1, line 17, and W-2, line 14, have marginally shorter internal
strokes.
In a few instances, the head of San-1 is noticeably raised, like that of CP Nu-2: see MS
1-1, line 17; MS 1-2, line12.
See LSAG2 103 and pl. 12,5.
The names written with san also use qoppa, not otherwise attested in the Phokian alphabet. Jeffery writes (LSAG2 101): Are these the names of Phokian and non-Phokian masons,
imported to erect this building? Or are they alien donors and Phokian masons? Or are
they all Phokians i.e. was san used in some parts of Phokis, and not in others? I hesitate
to accept the last hypothesis, if the appearance of san in Phocis should prove in fact to be
confined to Delphi, where examples of non-Phokian scripts naturally abound.
For discussion of the evidence, see LSAG2 100101.
LSAG2438.
For examples, see the spellings of and in LSAG2 pl. 12, 5; notice that in
the san of , all four strokes are approximately the same length.
See LSAG2 100101; compare EG 1:143 and244.
See LSAG2 pl. 45, 1; EG 1:275, fig.125ae. For discussion of the inscription (with references
to earlier work), see LSAG2 230; EG 1:274275; and Wachter 2001:168169.
On the conjunction of philos, xenos, and hetairos in Greek epic and its ancestry, see
Woodard 2007b.
See LSAG2 pl. 45, 1 (b); EG 1:275, fig.125b.
See LSAG2 pl. 45, 1 (c); EG 1:275, fig.125c.
See LSAG2 pl. 45, 1 (c); EG 1:275, fig.125a. Some investigators have read h as h. The
nominative h must, however, be the correct reading, occurring at the beginning of the
inscription; compare the relative clauses in the inscriptions of the Dipylon oinochoe and
Nestors cup, and see the evidenced amassed by Wachter (2001:168).
LSAG2 230; EG 1:275.
In a few instances, the head of San-2 is detectably raised, like that of CP Nu-1: see MS
2-1, line 20; MS 2-2, line 13. Note also the sixth from last letter of MS 2-2, line 10, which
could be either San-2 or Nu-1, its occurrence out of regular alphabetic order making its
intended identity difficult to confirm.
Note the san-symbols of, for example, W-2, lines 4 and 17, designated San-3, with what I
judge to be a very subtly outward leaning long stroke.
A case might admittedly be made for reclassifying some examples as San-1, or for identifying some as transitional between San-1 and San-3.
Notes to Pages 9397
311
456 The san of W-1, line 5, is designated San-3: although the long stroke bends rightward,
the bottommost point of the stroke is hardly any farther to the right than the topmost
point; however, the symbol appears to be an attempt to execute a letter like that of the
more distinct San-3 found three lines above it (W-1, line 2). Compare also the similar
san of MS 1-2, line 2, designated San-3, with very little, if any, rightward extension ultimately. The San-3 of MS 1-2, line 18, has a similar morphology, but with a pronounced
upturnedhead.
457 This tendency is prominently displayed on plaque MS 1-1. Consider, for example, the
occurrences of San-1 in lines 3, 5, 9, 12, and 18, each of which has a long stroke with a curving rightward progression but matched by the vertical spines of surrounding letters.
458 See EG 1:127, fig.24b; LSAG2 156 and pl. 26,3.
459 For discussion of the statues and of the history of scholarship treating them, see LSAG2
154156. For an image of the sculptors inscription, see LSAG2 pl. 26, 4B. For recent discussion of these well-known statues, see Hall 1995:594596.
460 See Pandolfini and Prosdocimi 1990:2426 and Tavv.VVI.
461 For Guarducci (EG 1:242): Nel santuario di Era argiva, ma proveniente da Kleonai o da
Tirinto. And Jeffery (LSAG2 149)allows: It might also be Tirynthian; but it cannot be
Argive . See, however, Johnstons remarks in the revision of LSAG regarding Jefferys
discussion of the alphabets of Phleious, Kleonai, Nemea, and Tiryns: The discovery of
[leges sacrae at Tiryns] written in Argive script, requires considerable alterations to statements throughout this section, as does the transfer of [a stele inscribed with a lex sacra
attributed to Tiryns] to Kleonai; (p.443) and The script of Tiryns is now attested as
similar to Argive (p.444).
462 See EG 1:242243 and fig.102ab. See also LSAG2 149, with n. 1, and pl. 25,11.
463 The spine of epsilon bows slightly but lacks the pronounced sweep ofsan.
464 Note also the Kleonaian inscription from Nemea on which san appears five times (see
LSAG2 148 and pl. 24, 5). In four of its occurrences san has a typical symmetrical -shape,
but once it is homographic with the four-stroke mu of the inscription, having a back leg
that is longer than its front leg. This instance of san thus shares the trait of a lengthened
vertical back leg in common with CP San-1, though again differs from the typical CP san
in having a front leg that is longer than the two inner strokes of the letter. It is, however,
virtually identical with the anomalous San-1 of copper plaque W-1, line 18, having an
atypically long frontleg.
465 On the morphology of eighth-century Phoenician ade, see McCarter 1975a:6061.
466 See Gibson 1982:6871 (bibliography on p.69), with fig.7; McCarter 1975a:4849, 132133;
Carpenter 1958:4748, with pl. 5, fig.4.
467 Not included in these counts are the two instances in which discrimination between San-1
and San-3 was withheld (marked -1/3), one in W-2, one in MS 1-1. Also not counted is
the symbol of MS 2-2, line 10, which occurs out of alphabetic position: the letter may
be either San-2 or Nu-1 (see the discussion immediately following); as we shall see in
Chapter4, there is perhaps some evidence suggesting that the letter is in fact san, though
the guidance provided by that evidence could be abrogated by scribal playfulness.
468 See the precedingnote.
469 Jeffery (LSAG2 34) observes that this type seems to be established everywhere by the
middle of the sixth century, perhaps earlier.
470 McCarter 1975a:87, 100. Jeffery (LSAG2 333334) and Guarducci (EG 1:98) seem to have
overlooked that a Phoenician prototype exists for each of the early Greek types and
312
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
Notes to Pages 9798
implicitly (Jeffery) or explicitly (Guarducci) identify the form with an undivided head as
being secondary.
Note the pronounced diamond shape of the head in some instances: see MS 2-1, lines 7
and9.
McCarter writes (1975a:87): The Rhodian cup (4) and the fragmented Samian abecedary
(20) share a form of qoppa in which the vertical shaft bisects the round head. Elsewhere
the shaft does not break the circle. The Rhodian cup (no. 4 in his survey) is a subgeometric cup bearing a graffito that Jeffery (LSAG2 347)characterized in 1961 as by its
appearance, as early as any inscription which we have, except the Dipylon oinochoe;
see LSAG2 pl. 67, 1; EG 1:328. The Samian abecedary (no.20 in McCarters survey) is
the aforementioned graffito on a cup from the Heraion of Samos. To this list could be
added the qoppa of eighth-century Theran rock graffiti; see, for example, IG XIII iii 538,
540,543.
In contrast to McCarters (1975a:87) observation that in most cases the length of the shaft
of early qoppa approximates the diameter of the circular head.
Note particularly the lunate quality of the spines of qoppa on MS 2-1, lines 3 and 6. The
curving aspect of the spine of qoppa could perhaps be of some morphological consequence, though will not be herein adopted as a parameter for more subtly typifying the
character, as comparative evidence does not at present seem to motivate such a decision.
See Lang 1976:17 and pl. 7 D 4; very similar is the qoppa from a late seventh or early
sixth-century graffito, illustrated in D 5. Compare too the two occurrences of qoppa in
the graffiti from Mount Hymettos: H 534 (Langdon 1976:25), with fig.10 (and pl. 7);and
H 552 (1976:38), with fig.15 (and pl.13).
See LSAG2 291 and pl. 55, 2; EG 1:154156, with fig.38ac.
See Pandolfini and Prosdocimi 1990:25 and Tavv. V andVI.
See Buonamici 1932: Tav. II; Pandolfini and Prosdocimi 1990:31. Qoppa is missing from
the abecedarium. The qoppa of the Marsiliana tablet shows the spine partially penetrating
the head (see, inter alia, LSAG2 pl. 48, 18)as does that in an incomplete abecedarium on
an amphora from Veii, late seventh century (see Pandolfini and Prosdocimi 1990:2629
and Tav.VII).
See Buonamici 1932: Tav. I; Bonfante and Bonfante 1983:108109; Pandolfini and
Prosdocimi 1990:2223.
See LSAG2 70 and pl. 2, 10a andc.
LSAG234.
McCarter 1975a:100. On the form of the Phoenician name of this Semitic character (r),
see McCarters p.100, n.87.
See LSAG2 pl. 1, 1; EG 1:135136, with fig.28ab.
See Langdon 1976:44, with a listing of occurrences of eachform.
See, inter alia, Pandolfini and Prosdocimi 1990:31 and Tavv. IX and X. An angular type
of rho is common among the Euboian abecedaria of Etruria, as can be easily observed
by perusing the images in, for example, Pandolfini and Prosdocimi 1990, though here,
as elsewhere, the angular varieties often do not display a sharply triangular shape, but
have themselves a rounded quality, being almost a blend of the pronounced angular and
rounded varieties of the letter.
Notes to Pages 98101
313
489 See LSAG2 235236 and pl. 47, 1; EG 1:226227, with fig.88ab; Heubeck 1979:109116, with
490 See EG 1:184, with fig.56.
491 As with certain other characters, the distinction between the two types of rho is quite
pronounced in many instances, but more subtle insome.
492 And, again, this could conceivably be of some typological consequence, though will not
be used herein as a parameter for more subtle classification ofrho.
493 See LSAG2 353354 and pl. 69, 43; Segre 1952 Tab.CXXV.
494 Segre, in his 1952 study of the sherd (p. 217), speculated that the graffito is of Carian
production. Only in 1981 did clarity begin to emerge in the study of Carian. In recent
years much progress has been made, though much about the language remains uncertain.
For an overview of the current understanding of the Carian language, with references to
additional work, see Melchert2004.
495 LSAG2 354. Also see p. 154.
496 See Blegen 1934:20 and 22, with fig.8, no.17; see also Langdon 1976:17 and44.
497 And overall the Hymettos symbol bears great similarity to CPPi-2.
498 See Blegen 1934:18 and fig. 16, no. 14. Compare also the rho of the erased form on
Blegens inscription no.1, pp.1012, with fig.1 (and pl.1).
499 See the exemplars in McCarter 1975a:128132.
500 These numbers do not include the rho of plaque MS 2-1, line 10, classified as Rho 1/2,
owing to the difficulty in seeing the character.
501 LSAG234.
502 And in his revision of LSAG, Johnston (p.428) remarks, The phrase, the use of one form
or the other is never invariable, while substantially correct, may mislead; in some areas
exceptions are extremely rare.
503 On the name of the Semitic character, see McCarter 1975a:100101n88; see also Woodard
1997:175186.
504 See Peruzzi 1973: Tav. IV; Heubeck 1979:123, with Abb. 49; A. Johnston 1983:64; LSAG2
453; Arena 1994:21 and Tav. IV 2; Dubois 1995:2930. Compare the sigma on a skyphos
from Methone ( 2247), dated to the late eighth or early seventh century bc (Bessios,
Tzifopoulos, and Kotsonas 2012:369370); the letter, however, appears to have a fifth
stroke added to its upper extremity. On five-stroke sigma, see subsequent discussion.
505 See LSAG2 110 and pl. 16, 1; EG 1:196197.
506 See Langdon 1976:1213.
507 For a listing of the occurrences of each type, see Langdon 1976:44.
508 See Langdon 1976:13, 15, with fig.6 (and pl. 2), no.2. Compare H 213 (Langdon 1976:15).
509 Langdon 1976:15.
510 Langdon (1976:7980) observes: The worship of Zeus on Hymettos was an acknowledgement of one of his basic aspects, that of weather god. We now clearly see Zeus
acknowledged as the supreme god who ruled from the mountain heights and controlled
the weather. It was to him that farmers paid homage. Hesiod provides us with the earliest literary evidence of Zeus importance for agriculture when he exhorts farmers to
pray to Zeus Chthonios and Demeter before beginning the seasons first plowing (Erga,
465467). Regarding the epithet, Henrichs (2003:50) observes: As far as I know, the
Zeus Semios of the Hymettos graffiti is the earliest nonliterary instance of the interplay
of divine names and cult epithets that ranks as one of the defining features of Greek
polytheism.
314
511 Jeffery 1978:203.

512 Jeffery (1978:203) summarizes Langdons four categories, based on content, as (i) deity
and (i.e., votive; Zeus is already specified on some of the earliest, c. 700): (ii)
abecedaria and (i.e., direct demonstrations of literacy): (iii) names and other
words (including parts of sentences; i.e., personal names and comments): (iv) miscellaneous (i.e., too fragmentary for any likely interpretation).
513 Thomas (1992:60) echoes Jefferys observations: The rest of the seventh-century finds at
the Mt. Hymettos sanctuary throw another curious sidelight on early attitudes to writing. A mass of graffiti was found here many of which were clearly inscribed on what was
already a broken sherd rather than a whole pot. Several consist merely of a personal name
and the word , i.e. so-and-so wrote this, and abecedaria were dedicated later in
the sixth century. It seems that the very dedication of bits of writing was thought appropriate for this shrine. The sanctuary was dedicated to Zeus Semios, that is, Zeus of the
weather signs, and signs must therefore have been interpreted to include written signs.
Hence one dedication reads, I am the property of Zeus. So-and-so wrote me [Langdon
1976:20]. It is even possible that people brought along to the sanctuary any piece of writing they could find.
514 Found in the name ; regarding the fifth stroke, which is clearly visible,
Guarducci (EG 1:146) muses: ma forse senza una precisa intenzione.
515 LSAG2 34 and 183, where Jeffery states that there is no constant version of sigma in
Lakonian. The characteristic type is usually of five strokes, but on occasion even of
eight. See also pp.186187.
516 See LSAG2 34, 341, 473474, and pl. 79, 8; EG 1:270272, with fig.123.
517 That one which in 1990 Johnston suggested to bear possibly our earliest Greek texts; see
LSAG2 454 and pl. 76, 2; see also Johnston 1983:6364.
519 See LSAG2 89, 91, and pl. 7,2b.
521 LSAG2290.
522 See, inter alia, Pandolfini and Prosdocimi 1990:2426, with Tavv.VVI.
523 See LSAG2 236237 and pl. 48, 19; Buonamici 1932: Tav. II; Pandolfini and Prosdocimi
1990:2932, with Tavv.IXX.
524 See LSAG2 325, where Jeffery describes the form as untidy and fairly common in the
7th6th c. Compare the sigma of copper plaque MS 2-1, line 6, where the upper half of
the character is marked by such roundedness; see also the similar sigma of line 10, though
with more angularity.
525 LSAG2 3435.
Tav. I, fig.1; Pandolfini and Prosdocimi 1990:1921, with Tavv. III n.I 1.
Prosdocimi 1990:2223, with Tav.IV.
528 See Pandolfini and Prosdocimi 1990:2426, with Tavv.VVI.
529 See Pandolfini and Prosdocimi 1990:2629, with Tavv. VIIVIII.
530 See LSAG2 236237 and pl. 48, 19; Buonamici 1932: Tav. II; Pandolfini and Prosdocimi
1990:2932, with Tavv.IXX.
531 See Arena 1994:22 and Dubois 1995:32.
532 For a close-up view, see Metzger 1965: pl.XVII.
315
1995:4142.
534 See LSAG2 238, 240, and pl. 47, 4; Arena 1994:13 and Tav. VI 2; Dubois 1995:49.
535 See LSAG2 238, 240, and pl. 47, 7; Arena 1994:27 and Tav. VI 1; on the interpretation of the
form (Lenos), see Dubois 1995:5052.
536 See Heubeck 1979:119, with Abb. 43; LSAG2 6970 and pl. 1, 2. With ][,
Heubeck compares Homeric both.
537 See EG 1:274275, with fig.125c. The crossbar of the tau of ] may descend
slightly away from the direction of writing; see EG fig.125b.
538 See Lang 1976:17 and pl. 7D1.
539 There are also instances in which the spine leans or bends back as it descends from the
head. In a subset of these there is a concomitant forward lean of the crossbar, and, hence,
the (approximate) right-angularity of the crossbar and spine is preserved (see MS 1-1, line
6; MS 2-1, line 22; MS 2-2, lines 1 and 9); in the case of the other examples, the crossbar
remains approximately horizontal (see W-1, line 18; W-2, lines 1 and 14; MS 1-1, lines 5, 9,
12, and 14; MS 2-1, line 9; MS 2-2, line12).
540 McCarter 1975a:101.
541 Jeffery (LSAG2 34)linked early Greek tau explicitly with protoforms from Cyprus, noting
that the Greek letter has its nearest equivalent in Semitic in the form on the bronze
bowls from Cyprus.
542 See Teixidor 1976:67, no.25.
543 See Gibson 1982:6668 and 182, fig.6.
544 See Gibson 1982:7276 and pl. VIII,1.
545 EG 1:99100.
546 See Pandolfini and Prosdocimi 1990:2932 and Tavv.IXX.
547 See LSAG2 238, 240, and pl. 47, 4; Arena 1994:13 and Tav. VI 2; Dubois 1995:49.
548 See Metzger 1965: pl.XVII.
549 Immerwahr 1990:160.
550 The spine may also pass through the crossbar, perhaps significantly so, in the case of the
tau of MS 2-2, line 19, though the X-ray image of the letter is quite faint.
3 Physical and Chemical Examination

1
2
3
4
5
6
7
8
9
10
11
See http://www.doernerinstitut.de/en/aufgaben/aufgaben_15.html.
Scott2002.
Tylecote1976.
Liritzis 1996:199.
Liritzis 1996:198.
Liritzis 1996:199.
Further details of the principles of metallographic examination can be found in Scott 1991
and 2011, to which the interested reader is referred.
Scott2002.
Scott and Dodd2003.
Gettens1963.
Gillard etal.1994.
316
4 The Syntagmatic Structure

1 That is to say the second half of that most primitive form of the Greek alphabet, that one
of twenty-two letters, as preserved in the copper plaques, running from alpha through tau,
with no supplemental characters (i.e., coterminous with the Phoenician source script). The
first half of the alphabet consists of the eleven-member sequence F ;
the second half of .
2 Immediately above this tau (at the start of line 11), a large portion of the top crossbar of zeta
is absent, and the lambda of the same line (11) appears to have an unusually short spine. Both
may, however, be the consequence of damage inflicted on the plaque face by attempts to rid
the plaque of its patina at sometime in the past. To the right and slightly above the deformed
tau of line 12 (i.e., in the right margin of the plaque) is what appears to be a tiny rho, written dextroverse (opposite to the direction of writing otherwise used on the plaque). Is this
actually an engraved mark or simply an artifact of surface corrosion? The right margin of the
middle portion of this plaque face, and to a lesser degree the left, is notably pocked, and in
several of these areas one is tempted to impose a graphic shape onto certain of these marks.
3 On the formal similarity of gamma and lambda and the potentiality of ambiguity, see, inter
alia, LSAG2 23; EG 1:90; McCarter 1975a:80.
4 And one familiar with the orthographic processes of both the Mycenaean Linear B script
and the Cypriot Greek syllabary cannot help but sit up and take notice at the use of deletion
and insertion processes in the inscribing of these lines. On the Mycenaean and Cypriot processes, see Woodard 1997, especially chaps. 15 and8.
5 Langue et criture
1 The foundational work on basic linguistic concepts produced by this father of modern linguistics has never been supplanted, only reexamined and refocused in some areas, blurred
in others. Regarding issues pertinent to the present study, Saussures work is, I believe,
unsurpassed and provides an unmarked point of reference and departure. As Robert
Burchfield (2006) observes in the opening pages of his work on the history of the English
language:
The work of Ludwig Wittgenstein (in philosophy) and of Ferdinand de Saussure (in linguistics) has been of fundamental importance. Other scholars, for example, M. Bral,
Leonard Bloomfield and A. Martinet, have made signal contributions at a more practical
level. But since 1945 linguistics as a subject has been riven and dismembered by disastrous civil wars between eminent scholars, most of them still unresolved, and the theoretical outlook is gloomy.
2 Ainsi lide de sur nest lie par aucun rapport intrieur avec la suite de sons s r
qui lui sert de signifiant; il pourrait tre aussi bien reprsent par nimporte quelle autre:
preuve les diffrences entre les langues et lexistence mme de langues diffrentes: le signifi
buf a pour signifiant b f dun ct de la frontire, et o k s (Ochs) de lautre
(Saussure 1995:100). Here, as elsewhere in this chapter, the English translation presented is
that of Roy Harris (see Saussure 1986:6768).
3 Situe la fois dans la masse sociale et dans le temps, personne ne peut rien y changer, et,
dautre part, larbitraire de ses signes entrane thoriquement la libert dtablir nimporte
5
6
7
8
317
quel rapport entre la matire phonique et les ides. Il en rsulte que ces deux lments unis
dans les signes gardent chacun leur vie propre dans une proportion inconnue ailleurs, et
que la langue saltre, ou plutt volue, sous linfluence de tous les agents qui peuvent atteindre soit les sons soit les sens. Cette volution est fatale (Saussure 1995:110111). For Harriss
translation, see Saussure 1986:76.
I am for the moment skirting the issue of what Saussure termed relative arbitrariness
(larbitraire relatif). Ainsi vingt est immotiv, mais dix-neuf ne lest pas au mme degr, parce
quil voque les termes dont il se compose et dautres qui lui sont associs, par exemple dix,
neuf, vingt-neuf, dix-huit, soixante-dix, etc.; pris sparment, dix et neuf sont sur le mme
pied que vingt, mais dix-neuf prsente un cas de motivation relative (Saussure 1995:181).
Harris (see Saussure 1986:130) translates: The French word vingt (twenty) is unmotivated,
whereas dix-neuf (nineteen) is not unmotivated to the same extent. For dix-neuf evokes the
words of which it is composed, dix (ten) and neuf (nine), and those of the same numerical
series: dix (ten), neuf (nine), vingt-neuf (twenty-nine), dix-huit (eighteen), soixante-dix
(seventy), etc. Taken individually, dix and neuf are on the same footing as vingt, but dixneuf is an example of relative motivation. In the Greek phrase h
, relative arbitrariness can be identified. For example, [klephsei], consists of
three structural units: the root, which here takes the form - [kleph-]; the suffix- [-s-],
marker of either future (indicative) or aorist (subjunctive); and the inflectional ending -
[-ei]. These three phonic components recur with their affiliated conceptual components
throughout the ancient Greek linguistic system and to that extent can be said to be relatively
motivated. In his discussion of relative arbitrariness, Saussure himself draws attention to
the Greek marker of future tense: En grec ds je donnerai exprime lide de futur par
un signe qui veille lassociation de ls, sts, tps, etc. (Saussure 1995:181). Harris (see
Saussure 1985:130) translates: In Greek, ds (I will give) expresses the idea of futurity by
a sign which links it associatively with other future tense forms like ls, sts, tps,etc.
Dubois 1995:41.
The author expresses his appreciation to Professor Maurizio Bettini for providing a nativespeakers Italian equivalent.
Rix 1985:207.
La masse sociale nest point consulte, et le signifiant choisi par la langue, ne pourrait pas
tre remplac par un autre. Ce fait, qui semble envelopper une contradiction, pourrait tre
appel familirement la carte force. On dit la langue: Choisissez! mais on ajoute: Ce
sera ce signe et non un autre. Non seulement un individu serait incapable, sil le voulait, de
modifier en quoi que ce soit le choix qui a t fait, mais la masse elle-mme ne peut exercer
sa souverainet sur un seul mot; elle est lie la langue telle quelle est (Saussure 1995:104).
For Harriss translation, see Saussure 1986:71.
Comme on constate un tat de choses identique dans cet autre systme de signes quest
lcriture, nous le prendrons comme terme de comparaison pour clairer toute cette question. Enfait:
1. les signes de lcriture sont arbitraires; aucun rapport, par exemple, entre la lettre t et le
son quelle dsigne;
2. la valeur des lettres est purement ngative et diffrentielle; ainsi une mme personne
peut crire la lettre t avec des variantes. La seule chose essentielle est que ce signe ne
se confonde pas sous la plume avec celui de l, de d, etc.;
3. les valeurs de lcriture nagissent que par leur opposition rciproque au sein dun systme dfini, compos dun nombre dtermin de lettres. Ce caractre, sans tre identique au second, est troitement li avec lui, parce que tous deux dpendent du premier.
318
10
11
12
13
14
15
16
17
18
19
20
21
22
La signe graphique tant arbitraire, sa forme importe peu, ou plutt na dimportance

que dans les limites imposes par le systme;
4. le moyen de production du signe est totalement indiffrent, car il nintresse pas le
systme. (Saussure 1995:165166)
For Harriss translation, see Saussure 1986:117118).
Thus, in the general discussion of writing systems in the Cours, we read Langue et criture
sont deux systmes de signes distincts; lunique raison dtre du second est de reprsenter
le premier; lobjet linguistique nest pas dfini par la combinaison du mot crit et du mot
parl; ce dernier constitue lui seul cet objet (Saussure 1995:45). Harris (see Saussure
1986:2425) translates: A language and its written form constitute two separate systems of
signs. The sole reason for the existence of the latter is to represent the former. The object
of study in linguistics is not a combination of the written word and the spoken word. The
spoken word alone constitutes that object.
Consider here, too, the remarks of Benveniste 1966:28: En effet la facult symbolique chez
lhomme atteint sa ralisation suprme dans le langage, qui est lexpression symbolique par
excellence; tous les autres systmes de communications, graphiques, gestuels, visuels, etc.
en sont drivs et le supposent.
Compare Woodard 2003:160: Spoken language and written language are not to be confused: it is probable that this book will fall into the hands of people who through familiarity
with its Latin script are able to give its letter symbols their approximate sound values, but
who may not recognize what language they record. Almost everyone with certain basic
cognitive abilities knows one language, but many are unable to read or write. Human languages may have existed for one hundred millennia, but writing systems have developed in
only the last five.
Jakobson 1990:455456.
With the possible exception of the Korean Han Gul alphabetic script.
The contrast occurs less frequently in Attic-Ionic than in most dialects, consequent to the
historical change of inherited // to the vowel represented by eta in Attic and, especially,
Ionic.
The example, and the one that follows, are provided by W. Sidney Allen in his 1974 discussion of Greek vowel quantity (p.86). On contrastive vowel length in ancient Greek, see,
inter alia, Allen 1974:8485; Lejeune 1982:190191, 218; Meier-Brgger 1992:114116; Palmer
1980:213214; Rix 1976:4546, 4951.
On the Linear B strategies for spelling consonant strings, see Woodard 1997, especially
pp.6278 and 112132.
Such at any rate is the conventional view, though the conventional view may not be sufficiently nuanced: see Woodard2012.
If the motivation should turn out to be not purely phonetic, the fact remains that more
than a single symbol is used to encode a single arbitrary phoneme.
For discussion of Kirchhoff s (1887) classification of blue alphabets (both dark blue and
light blue), red alphabets, and green alphabets, see Woodard 1997:140147. On the
concept of light red alphabets, see Woodard 1997:215216.
The symbols xi and psi do not occur in all local Greek alphabets: some lack both; some lack
only psi. On the distribution of the symbols, the asymmetric nature of their distribution,
and the associated implicational hierarchy, see Woodard 1997:140, 147161.
In the light blue alphabets, which function without either or , the /ks/ sound sequence
is spelled componentially as and the /ps/ as . On this system and an interpretation of
its origin, see Woodard 1997:140141, 208209, and 214216.
319
23 On lack of linguistic distinctiveness in ancient, medieval, and modern writing, see Sirat
1994:407425.
24 Sirat (1994:425), after examining a graphemically imprecise handwritten page produced
by one of her colleagues (and subsequently typed and printed without problem [p.119]),
observes of earlier and contemporary writing practice: What about the readers of all these
[i.e., ancient and medieval] books, these documents, these letters which seem so difficult
to read graphically, which are so far away from the modern conception of the laws of communication? The readers read them as easily as contemporary pharmacists read physicians
prescriptions, which are notoriously illegible: they pose no challenge to the pharmacists
and mistakes are very, veryfew.
25 Saussure 1995:5152.
26 Thus Benveniste (1966:24) remarks: Lalphabet latin, lalphabet armnien sont des exemples admirables de notation quon appellerait phonmatique. Un analyste moderne naurait
presque rien y changer: les distinctions relles sont reconnues, chaque lettre correspond
toujours et seulement un phonme, et chaque phonme est reproduit par une lettre toujours le mme.
27 Jakobson 1990:236.
28 I say effectively because of the so-called matres lectionis (mothers of reading): the term
refers to the limited use of consonant symbols to signal the presence of a vowel sound,
principally in word-final position, in Aramaic and Hebrew. See, inter alia, Cross 1989:86;
Gelb 1963:166168; McCarter 1975a:125n1.
29 On Mycenaean spelling practice, see Ventris and Chadwick 1973:4248; on the graphic
representation of consonantal sequences, see Woodard 1997:6266 and 112132.
30 See Aravantinos, Godart, and Sacconi 2001 and the various contributions in Deger-Jalkotzy
and Panagl 2006. See also Palaima 20002001 and2003.
31 For a description of syllabic Cypriot spelling practices, see Masson 1983:5157 and 6878.
On the spelling of consonantal sequences, see Woodard 1997:6675 and 112132.
32 See Woodard 1997, especially chaps. 7 and8.
33 Saussure himself mentions both the Cypriot syllabary and the Semitic consonantal script
in his discussion of speech sounds in the Cours; the flavor of the remarks is recognizably
that of the early twentieth century (see Saussure 1995:6465; the translation that follows is
that of Harris (see Saussure 1986:3940), informed, writes Harris (p.39), by shorthand
notes taken at three lectures Saussure gave in 1897 on The Theory of the Syllable):
La chane acoustique ne se divise pas en temps gaux, mais en temps homognes, caractriss par lunit dimpression, et cest l le point de dpart naturel pour ltude phonologique.
cet gard lalphabet primitif mrite notre admiration. Chaque son simple y est reprsent
par un seul signe graphique, et rciproquement chaque signe correspond un son simple,
toujours le mme. Cest une dcouverte de gnie, dont les Latins ont hrit. Dans la notation
du mot brbaros barbare, , chaque lettre correspond un temps
| | | | | | | |
homogne; dans le figure ci-dessus la ligne horizontale reprsente la chane phonique, les
petites barres verticales les passages dun son un autre. Dans lalphabet grec primitif,
on ne trouve pas de graphies complexes comme notre ch pour , ni de reprsentations
doubles dun son unique comme c et s pour s, pas non plus de signe simple pour un
son double, comme x pour ks [see my remarks within the English translation below].
Ce principe, ncessaire et suffisant pour une bonne criture phonologique, les Grecs lont
ralis presque intgralement.
320
Les autres peuples nont pas aperu ce principe, et leurs alphabets nanalysent pas la chane
parle en ses phases acoustiques homognes. Les Cypriotes, par exemple, se sont arrts
des units plus complexes, du type pa, ti, ko, etc.; on appelle cette notation syllabique; dsignation quelque peu inexacte, puisquune syllabe peut tre forme sur dautres types encore,
par exemple pak, tra, etc. Les Smites, eux, nont marqu que les consonnes; un mot comme
brbaros aurait t not par eux BRBRS.
Harris (Saussure 1986:3940) translates:
The sequence of sounds we hear is not divided into segments of equal duration, but into
segments identifiable as auditory units. This fact provides us with a natural starting point
for the study of speech sounds.
From this point of view, one cannot fail to admire the Greek alphabet in its most primitive
form. Each sound unit is represented by one symbol, and conversely each symbol invariably
corresponds to a single sound. It was a system of brilliant simplicity, later taken over by the
Romans. In the spelling of the word brbaros (barbarian), each of the letters
| | | | | | | |
stands for a single segment. In the diagram given here, the horizontal line represents the
sequence of sounds, while the vertical strokes indicate the transitions between each sound
and the next. In the primitive Greek alphabet, there are no combinations like our modern
French ch for . Nor are there variable representations of a single sound, like our c and s for
the sound s. Nor are there single characters representing a combination of sounds, like our
x for ks [a notable inaccuracy; Saussure (in the form of his students) includes a footnote
further qualifying some of these observations, but offers no corrective remarks concerning
and its value of ks, and with no mention of , spelling ps (RDW)]. This principle, which
is both a necessary and a sufficient condition for good transcription, was adopted almost
without exception by the Greeks.
The principle was not grasped by other nations, and consequently their alphabets do not
analyse sound sequences into constituent auditory units. The Cypriot system, for example,
went no further than complex segments of the type pa, ti, ko, etc. This is usually referred to
as syllabic notation, although the term is not very accurate inasmuch as there are other
syllabic patterns, e.g. pak, tra, etc. The Semitic system simply indicated consonants; so a
word like brbaros would have been written BRBRS.
34 The eta-position can, of course, be also a consonantal position in the case of those alphabets using the eta-symbol for /h/, or for both /h/ and the long vowel . I will, however,
present evidence for the eta of the copper-plaque abecedaria being (at least) vocalic in
itsuse.
35 In addition, homography involving the -symbol is attested cross-alphabetically over a
broad geographic range: it spells (intra-alphabetically) both the single consonant /h/ and
the consonantal sequence /ks/ in the Aegean Naxian script (as just noted), the single consonant /h/ in the alphabets of Kyme and Sicilian Naxos, and the long vowel in the alphabet of Knidos.
36 In the Cours, Saussure (1995:180181) acknowledges regarding language, Le principe fondamental de larbitraire du signe nempche pas de distinguer dans chaque langue ce qui
ne lest radicalement arbitraire, cest--dire immotiv, de ce qui ne lest que relativement.
Une partie seulement des signes est absolument arbitraire; chez dautres intervient un phnomne qui permet de reconnatre des degrs dans larbitraire sans le supprimer: le signe
peut tre relativement motiv. Harris translates (see Saussure 1986:130): The fundamental
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
321
principle of the arbitrary nature of the linguistic sign does not prevent us from distinguishing in any language between what is intrinsically arbitrary that is, unmotivated and
what is only relatively arbitrary. Not all signs are absolutely arbitrary. In some cases, there
are factors which allow us to recognize different degrees of arbitrariness, although never
to discard the notion entirely. The sign may be motivated to a certain extent. For more on
relative arbitrariness, see my note at the beginning of this chapter.
On pharyngeal consonants, see Woodard 1997:136 and 188n9.
Though the Cypriot dialect of Phoenician appears to have developed a consonant + consonant reflex of the sound that was spelled by the letter zayin. Phoenician zayin is the precursor of the Greek zeta, which is also used to spell a sequence of two consonants, namely
[z+d]. See Woodard 1997:161175.
See Woodard 1997:115123.
The translation is that of Jenny Oates, appearing in Jean 1992:195.
Ray 2007:100, who continues: He is the Byron of scholarship, and he is also an outsider
like Keats. His intellectual talent, combined with the years of detailed research which led
to his breakthrough, holds us in awe, and his early death movesus.
Davies 1990:96. See also Sol, Valbelle, and Davies 2006:208.
Ray 2007:11; see also Davies 1990:82; Sol, Valbelle, and Davies 2006:193.
Davies 1990:96.
See the map of the Qena Bend in Darnell etal. 2005:74. Darnell etal. 2005 provides the
editio princeps of two inscriptions written with this script.
See Darnell etal. 2005:8690.
See Darnell etal. 2005:9091.
See Darnell etal. 2005:73 with bibliography of published work on the inscriptions.
On the inscriptions from Serabit el-Khadim, see, inter alia, Sass1988.
Darnell etal. 2005:100n130. They point out (p.90) that a late Middle Kingdom date for
the time of the carving of the Wadi el-Hl inscriptions is consistent with Gardiners
original dating of the Serabit texts. The reference is to Sir Alan Gardiner, who in 1916
deciphered the crucial word blt, lady, identified the script as West Semitic, and proposed
an Egyptian origin (Pardee 1997a:354); see Gardiner1916.
To logograms and phonograms we could add determinatives: signs used for semantic disambiguation. For general discussion of the Egyptian writing system and the types of symbols
employed, see, inter alia, Loprieno 2004:163167 and 192217 (with references to additional
bibliography). See also Davies 1990:102112; Sol, Valbelle, and Davies 2006:214234;
Ritner1996.
Loprieno 2004:163.
Symbol O4 in the Gardiner sign list; see the sign lists in Allen 2000:423448; also in
Loprieno 2004:192217. The Egyptians also use the symbol as a phonogram spellingh.
This is one of the more important forms in these inscriptions, because at last it provides
us with the obvious precursor of the West Semitic bt; Darnell etal. 2005:77.
On Proto-Canaanite consonantal script, see, inter alia, Sass 1988; and Pardee 1997b and
2004a.
http://penn.museum/cgi/hieroglyphsreal.php Write Like an Egyptian: Your name in
hieroglyphs, the way an Egyptian scribe might have writtenit.
See Loprieno 2004:164.
See Loprieno 2004:164 and 206; Allen 2000:437.
See Loprieno 2004:164 and 214; Allen 2000:445. The Penn Museum Web site has more
recently modified its algorithm to transcribe both c and k with this symbol.
322
60 BDB 1063. As in Ezekiel 9:4, 6: Go through the city, through Jerusalem, and put a mark
(tw) on the foreheads of those who sigh and groan over all the abominations that are
committed in it. Cut down old men, young men and young women, little children and
women, but touch no one who has the mark (NRSV). See Driver 1976:170.
61 See Darnell etal. 2005:8385.
62 The symbol, Z11 in Gardiners list, is used also with the biconsonantal value jm; see Allen
2000:447; Loprieno 2004:216. On Proto-Sinaitic taw, see, inter alia, Sass 1988:133, who
writes, First identified in the Proto-Sinaitic inscriptions by Macalister (1906) on the basis
of its resemblance to the Phoenician letter. If the inventor of the alphabet [i.e., of the West
Semitic script that took shape in early second-millennium Egypt] could not do without a
hieroglyphic prototype, he could have found it in sign Z9 [sic] and the like. The letter
requires hardly any comment.
63 The sort of arbitrariness envisaged by Saussures point 1 is perhaps more acutely realized
in the case of other nonalphabetic scripts. For example, while the occurrence of a CV
symbol ka in the Linear B inventory of graphemes is also derivatively arbitrary derived
from (1) the arbitrary occurrence of the consonant sounds /k/, /kh/, and /g/; and (2) the
arbitrary vowel sounds // and /a/; and (3) their arbitrary concatenation the choice of
the graphic shape to encode the phonic strings /k/, /k/, /kh/, /kh/, /g/, and /g/ may
be fundamentally unmotivated, at least not motivated in the sense that the selection of
graphic shapes for Proto-Sinaitic symbols was motivated by West Semitic linguistic signs.
However, the formative history of the Cretan linear scripts is cloaked in uncertainty; one
must bear in mind the highly iconic nature of the symbols of Cretan Hieroglyphic (and of
the contentious script of the Phaistos disk).
64 See, inter alia, Demsky 1977; Cross 1980:815. The order of letters in this abecedarium,
presumably the work of a student, cannot be said to be fully canonical. Regarding the
order, Cross writes (p.13): Some, if not all, of its peculiarities must be ascribed to error:
the replacement of waw with mem and omission of waw, the interchange of et and zayin,
the repetition of qop, and displacement of re. There remains the sequence of pe-ayin. This
may be an error. On the other hand, there is evidence, as Demsky [1977] has observed, of
this existence of an alternate order pe-ayin in biblical acrostic poetry and in the abecedaries of Kuntillet Arjd.
65 The characterization consonantal must be qualified: the script includes three syllabic
symbols encoding the sequence of glottal stop (//) plus the vowels /a/, /i/, and /u/. On the
syllabic symbols and the Ugaritic writing system generally, see Pardee 2004b:288291.
66 There is also a Ugaritic short consonantal script, not well attested, and a system preserving the South Arabian letter order; see Pardee 2004b:290.
67 See, inter alia, Pardee 1997c, with annotated references to earlier work. Pardee writes
(pp.7778): Comparison between the Ugaritic abecedaries and the order of letters in the
later Northwest Semitic languages shows that a long alphabet was in use in the fourteenth
century bce . The basic consonantal inventory of Ugaritic was represented by twentyseven signs, whereas the basic southern Canaanite alphabet consisted of twenty-two signs.
The five extra Ugaritic signs are interspersed among the signs known from the southern
Canaanite alphabet. If the inventor of the Ugaritic cuneiform system had been imitating
a short alphabet, he would in all likelihood have tacked the extra signs he needed on at the
end of the alphabet.
68 Though perhaps not one identical to that of Ugaritic; Pardee (1997c:77) again: Because
the Ugaritic phonological system does not perfectly match the writing system (e.g., is used
inconsistently), it is likely that the alphabet was originally borrowed from a West Semitic
69
70
71
72
73
74
75
76
77
78
79
323
language with a slightly different consonantal inventory and/or that the use of the long linear alphabetalready had a history at Ugarit.
Pardee 1997c:77.
Watt 1987 and1989.
Watt here includes the following footnote: The best known cases in which one people
has adopted anothers alphabet but at or after the point of adoption reordered the letter
list are probably those of the Arabic and runic abecedaria (cf. respectively Naveh 1982:162
and Diringer 1968:404; Page 1987:89, 20)and the Ethiopic syllabarium (Driver 1976:138;
Naveh 1982:51), to which list the Irish ogham abecedarium may be added, with its fantastic reordering dependent on tree names and Old Irish wood-poaching fines (Pedersen
1931:233). Regarding the variant letter order of Arabian consonantal scripts, see Macdonald
2004:494495 and Healey 1990:219220. On the Ethiopic (Geez) syllabary and its development from the South Semitic consonantal script, see Gragg 2004:431433.
Watt 1989:6465.
The order and form of supplementals in the dark blue alphabets.
Dune part, dans le discours, les mots contractent entre eux, en vertu de leur enchanement, des rapports fonds sur le caractre linaire de la langue, qui exclut la possibilit de
prononcer deux lments la fois. Ceux-ci se rangent les uns la suite des autres sur la
chane de la parole. Ces combinaisons qui ont pour support ltendue peuvent tre appeles
syntagmes. Le syntagme se compose donc toujours de deux ou plusieurs units conscutives (par exemple: re-lire; contre tous; la vie humaine; Dieu est bon; sil fait beau temps,
nous sortirons, etc.). Plac dans un syntagme, un terme nacquiert sa valeur que parce quil
est oppos ce qui prcde ou ce qui suit, ou tous les deux (Saussure 1995:170171). For
Harriss translation see Saussure 1986:121.
On the Arabian consonantal scripts, see Macdonald 2004:494495 and Healey 1990:219
220; see also Nebes and Stein 2004:455465 with table15.1.
Faarlund (2004:909) writes: The runes were organized in a specific order, like an alphabet;
such a runic alphabet is called a futhark, from the values of the first six runes. Although
there was some individual variation, the futhark was remarkably uniform throughout the
area and through the four centuries of use. On the origin of the runic futhark, see Faarlund
pp.910911.
And the further Saussurian specification in its place in a syntagma, any unit acquires
its value simply in opposition to what precedes, or to what follows, or to both does this
hold with regard to the alphabetic syntagma? There is, of course, a sense in which it does
in which each letter is negatively valued, dependent upon the values of all other letters for
its own. This, however, lies properly within the realm of what was earlier considered: the
distinctiveness of graphemic shapes secondary to the distinctiveness of the sounds that
they encode. Saussure again points to a one-versus-all contrast, but herein of course lies
the difference: the syntagmatic structure of the alphabet consists of graphemes that reflect
the sound values of the phonemic inventory of a language; the syntagmatic structure of a
language consists of meaning-bearing morphological units, themselves composed of phonological units.
la syntaxe, cest--dire, selon la dfinition la plus courante, la thorie des groupements
de mots, rentre dans la syntagmatique, puisque ces groupements supposent toujours au
moins deux units distribues dans lespace. Tous les faits de syntagmatique ne se classent pas dans la syntaxe, mais tous les faits de syntaxe appartiennent la syntagmatique
(Saussure 1995:188); for Harriss translation, see Saussure 1986:135.
See Saussure 1995:468n250.
324
80 These are offered as evidence for the syntagma belonging, at least in part, to langue (as
opposed to parole).
81 On rencontre dabord un grand nombre dexpressions ; ce sont les locutions toutes
faites, auxquelles lusage interdit de rien changer, mme si lon peut y distinguer, la
rflexion, des parties significatives (cf. quoi bon? allons donc! etc.). Il en est de mme,
bien qu un moindre degr, dexpressions telles que prendre la mouche, forcer la main
quelquun, rompre une lance, ou encore avoir mal (la tte, etc.), force de (soins, etc.),
que vous en semble?, pas nest besoin de , etc., dont le caractre usuel ressort des particularits de leur signification ou de leur syntaxe. Ces tours ne peuvent pas tre improviss,
ils sont fournis par la tradition (Saussure 1995:172). For Harriss translation, see Saussure
1986:122123.
82 See Watt 1987 and1989.
83 See Watt 1987:910.
84 Watt 1987:13; he goes on to specify the particular changes that occurred in the transformation of the Ras Shamra Matrix into the Byblos Matrix (see pp.1314).
85 Watt 1989:6970.
86 See Watt 1989:7374; In sum, then, we have this picture: of the Matrixs eight Columns,
two are too meager to be subject to the principle of maximal separation; the other six must
obey or disobey. All six obey. We can hardly avoid concluding that the principle of maximal separation organizes the contents of the Columns at least as rigorously as it organizes
the Columns themselves (p.74).
87 See Watt 1989:62, fig.1, for a full depiction of the matrix structure.
88 Watt 1989:70. On pp.7273 he restates the pedagogical interpretation, emphasizing that
he considers it to be conjectural: Columns in the Matrix had the consequence that similar sounds were separated in the abecedarial recitation, perhaps making it a little easier to
learn, and certainly making it easier to monitor. Of course we do not know (and presumably will never know) whether the rationale just presented has any truth to it; but since
nothing in the present argument really depends on it we can view it as a simple fantasy
with no harm done. What matters is the apparent ordering of the Columns and the principle (separation of similar sounds) that can be hypothesized to underlie that order.
89 See references in Maddieson 1984:136.
90 Maddieson (1984:126) writes: Some languages have been analyzed in the linguistic literature as having fewer than three phonemic vowels. The best known are Kabardian
and Abaza (Allen 1965; Anderson 1978). However, more conservative analyses of these
languages can be defended in which less of the contrast between syllables is attributed to
the consonants. This approach results in an analysis in which they have only three vowels
it is clear that these are languages with a small number of language contrasts under any
analysis.
91 See, inter alia, Jakobson 1990:283285. This article, coauthored with Linda Waugh, is
excerpted from Quest for the Ultimate Constituents and The Network of Distinctive
Features appearing in Jackobson and Waugh 1979:80176.
92 Regarding vowel inventories and configurations, Maddieson (1984:136) observes: The
most prevalent patterns seem to be the so-called triangular systems, particularly those
of average size, and notably the five-vowel systems. For example, over a quarter of the 209
languages in the Stanford Phonology Archive have a triangular vowel system consisting of
/i, , a, , u/, while less than 5% have any of the other 5-vowel configurations; the square
4-vowel and 6-vowel systems combined total less than 10% (Crothers 1978).
325
93 Maddieson (1984:153154), after examining the vowel systems of 317 of the worlds languages, concludes: The great majority of vowel systems in our sample assume configurations which are predictable from a theory of vowel dispersion, considered in the light of
some basic facts about the overall number of vowels, their degree of peripherality, and
the like. He continues (p.154): Thus the number of obvious exceptions to a vowel dispersion hypothesis in the whole of the UPSID [UCLA Phonological Segment Inventory
Database] data is extremely small. About 86% of the languages have vowel systems that
are built on a basic framework of evenly dispersed peripheral vowels. About another 10%
approach this specification. This strongly indicates that a vowel dispersion theory correctly captures a principle governing the distribution of vowels in natural languages.
94 The features compact-diffuse are defined acoustically as displaying a strong concentration of energy in the mid-frequency region of the spectrum [i.e., compact] versus spread
of energy over a wider frequency region [i.e., diffuse]; Jakobson 1990:259. For more
detailed discussion of the acoustic properties of the features, see pp.273285.
95 The features grave-acute are defined acoustically as displaying a concentration of energy
in the lower [i.e., grave] versus upper frequencies [i.e., acute] of the spectrum; Jakobson
1990:260. For more detailed discussion of the acoustic properties of the features, see
pp.268271.
96 Jakobson 1990:286.
97 Jakobson 1990:284.
98 Jakobson 1990:284.
99 Nor is it necessarily to claim that the structure of the matrix could not be sensitive to
binary features of the axes grave-acute and compact-diffuse. Watt (1989:8687) opts for
an articulatory basis for the matrix structure, rather than one defined in terms of acoustic
distinctive features; he writes:
First of all, as we have already noticed, the phonological categories of the Matrix
are conspicuously articulatory in character. Not only do the eight Columns consist
of sounds whose articulatory classifications make sense, but they consist of sets for
which categorization in terms of conventional distinctive features, so long as their
basis is fundamentally acoustic, would be scattered and awkward. It should be an
attribute of any attempt to interpret the Columns of the Matrix that those Columns
sound classes are classes indeed, each a natural set exhibiting enough homogeneity for its defining phonological attributes to hold for all of its members. Interpreted
articulatorily, the Columns do this quite well, as witness the appropriateness of the
labels identifying their places of articulation within the vocal tract (a lingual feature
such as Coronal, which is highly correlated with location in the vocal tract, would
also be serviceable). In contrast the more purely acoustic a phonological feature is
(or the subtler the articulatory distinction at issue), the less satisfactorily that feature
labels the Matrixs columns. This is not to say that the Columns of the Matrix could
not be described in terms of distinctive features (that would be absurd), but it is to say
that any such features must apparently have the somewhat antique air of characterizing sounds by almost exclusive reference to approximate place of articulation, with
subsidiary reference (in the subcolumns a and b) to articulatory manner.
100 Jakobson 1990:287288.
101 Honor de Balzac, Illusions perdues: cited by Jakobson and Waugh 1979:80.
102 In the notes to his critical edition of the Cours, Tullio de Mauro writes (in Saussure
1995:468n248): Lusage a consacr le terme paradigmatique, absent chez Saussure mais
326
103
104
105
106
107
108
109
110
111
112
113
114
115
116
suggr par des passages dans lesquels les paradigmes flexionnels sont cits comme
exemples typiques de rapports associatifs.
Saussure 1995:171 and173.
More literally, acoustic imagery: sur la simple communaut des images acoustiques (par
exemple enseignement et justement); Saussure 1995:174. In the critical edition of the
Cours, there is an editorial footnote added regarding this matter (Saussure 1995:174n1),
translated by Harris (Saussure 1986:124n1) in this way: This case is rare and may be
treated as abnormal. For the mind naturally discards all associations likely to impede
understanding and discourse. Nonetheless, the existence of such associative groups is
proved by the category of feeble puns based upon the ridiculous confusions which may
result from homonymy pure and simple. E.g. Les musiciens produisent les sons et les
grainetiers les vendent (Musicians produce [sounds/bran], which seedsmen sell son
meaning both sound and also bran). Such cases must be distinguished from those in
which word association, although fortuitous, is backed by a certain connexion of ideas:
e.g. French ergot (spur, spike) and ergoter (to quibble), or German blau (blue) and
durchbluen (to beat, to thrash). What is involved here is a new interpretation of one or
other of the terms. These are cases of popular etymology. Although of interest in the
study of semantic change, from a synchronic viewpoint they merely fall into the category
of enseigner, enseignement, etc. mentioned above.
Lesprit saisit la nature des rapports qui les relient dans chaque cas et cre par l autant
de sries associatives quil y a de rapports divers. Ainsi dans enseignement, enseigner,
enseignons, etc., il y a un lment commun tous les termes, le radical; mais le mot enseignement peut se trouver impliqu dans une srie base sur un autre lment commun, le
suffixe (cf. enseignement, armement, changement, etc.); lassociation peut reposer aussi sur
la seule analogie des signifis (enseignement, instruction, apprentissage, ducation, etc.), ou
au contraire, sur la simple communaut des images acoustiques (par exemple enseignement et justement). Donc il y a tantt communaut double du sens et de la forme, tantt
communaut de forme ou de sens seulement. Un mot quelconque peut toujours voquer
tout ce qui est susceptible de lui tre associ dune manire ou dune autre (Saussure
1995:173174). For Harriss translation, see Saussure 1986:123124.
Jakobson 1990:118. This essay, Two Aspects of Language and Two Types of Aphasic
Disturbances, first appeared in Jakobson and Halle1956.
Jakobson 1990:129.
Jakobson 1990:124.
For discussion of this similarity disorder, see Jakobson 1990:120125.
Jakobson 1990:126.
The patient confined to the substitution set (once contexture is deficient) deals with
similarities, and his approximate identifications are of metaphoric nature, contrary to the
metonymic ones familiar to the opposite type of aphasics (Jakobson 1990:126). For discussion of the contiguity disorder, see pp.125128.
Jakobson 1990:130.
Jakobson 1990:130.
Jakobson 1990:131. For his discussion of the mentioned expressions of the dichotomy, see
pp.130133.
See Jakobson 1990:131.
Havelock 1982:50. Havelocks observation is accompanied by the presentation of very little evidence and, surveying the essays in this volume, one wonders if this may be the only
noncontroversial claim made by the author regarding the alphabet, and, hence, if many of
117
118
119
120
121
122
123
327
his readers are aware of its noncontroversial nature: a lonely fact set adrift within a flotilla
of unfortunate fictions. The observation is itself abutted by a particularistic discussion of
Chinese writing. Regarding the latter, one might contrast remarks made by Sirat vis--vis
claims made about the superiority of the alphabet. Bound up with this claim are judgments of the greater ease with which an alphabet is acquired and its efficiency for printing
(Sirat 1994:449450):
This is true. However, the printing press with mobile types was not invented in Europe
in 1430 but in China between 1041 and 1049 (Drge, Ishigami-Iagolnitzer, and Cohen,
1986). Chinese printing of books can be cumbersome, but in 1983, 35,700 titles were
published (5,804millions of copies), 3,415 journals (1,769millions of copies) and 773
daily papers (16,366millions of copies) (Lin Guojun and Zheng Rusi, 1989:130).
This proves that systems where the number of symbols is not small and easily learnt
nevertheless function quite well. It does not disprove the scientific superiority of the
Occident. It proves however that the Latin alphabet is not the cause or even one of the
causes of modern science: it is merely the writing modern science uses, and is associated with. During the Middle Ages, the superiority of Arabic science (mathematics,
astronomy, medicine) over the Latin and Byzantine ones was not caused by the Arabic
alphabet either. The Latin alphabet is a part of the culture and tradition where occidental science arose. Modern science can be written in Chinese characters and it is very
probable that Japanese and Chinese scientists will be, in the next century, more numerous than occidental ones.
While I am uncomfortable with Sirats use of the terms prove and disprove, the
appropriateness and accuracy of her point is, I believe, self-evident.
See Whitney 1867 and Harriss remarks in Harris 1987:83.
Pour bien faire sentir que la langue est une institution pure, Whitney a fort justement
insist sur le caractre arbitraire des signes; et par l, il a plac la linguistique sur son axe
vritable. Mais il nest pas all jusquau bout et na pas vu que ce caractre arbitraire spare
radicalement la langue de toutes les autres institutions. On le voit bien par la manire
dont elle volue; rien de plus complexe: situe la fois dans la masse sociale et dans le
temps, personne ne peut rien y changer (Saussure 1995:110). For Harriss translation, see
Saussure 1986:76.
One could include in this list events that might most appropriately be termed evolutionary, but which still clearly entail personal agency. For example, beginning about the turn
of the millennium, some persons adapted the form of the script that had evolved for
spelling the Phoenician variety of Canaanite language in order to record the Hebrew and
Aramaic languages, the latter developing over time numerous individual forms (see, inter
alia, Creason 2004:393395, with references). Personal agency is particularly evident in
the case of the latter: in the seventh or eighth century bc , some person or persons engineered the use of four consonantal graphemes for signaling the presence of vowels under
certain conditions (the so-called matres lectionis).
Watt 1989:64.
Dubois 1995:41.
The former Cyrillic-using speakers of Balkan Romance would perhaps have recognized
a somewhat greater graphemic affinity with ancient Greek inscriptions, though the core
elements of their language remained Latinate.
Langue et criture sont deux systmes de signes distincts; lunique raison dtre du second est de reprsenter le premier; lobjet linguistique nest pas dfini par la combinaison
du mot crit et du mot parl; ce dernier constitue lui seul cet objet. Mais le mot crit se
328
124
125
126
127
128
129
mle si intimement au mot parl dont il est limage, quil finit par usurper le rle principal; on en vient donner autant et plus dimportance la reprsentation du signe vocal
qu ce signe lui-mme. Cest comme si lon croyait que, pour connatre quelquun, il vaut
mieux regarder sa photographie que son visage (Saussure 1995:45). For Harriss translation, see Saussure 1986:2425.
Saussure 1995:5152.
Sherratt 2003:232.
Though this is not to suggest that conventional English orthography has no bearing
on phonological realities; see the classic treatment of Chomsky and Halle 1968 in this
regard.
Les premiers linguistes sy sont tromps, comme avant eux les humanistes. Bopp luimme ne fait pas de distinction nette entre la lettre et le son; le lire, on croirait quune
langue est insparable de son alphabet. Ses successeurs immdiats sont tombs dans le
mme pige; la graphie th de la fricative a fait croire Grimm, non seulement que ce son
est double, mais encore que cest une occlusive aspire; de l la place quil lui assigne dans sa
loi de mutation consonantique ou Lautverscheibung. Aujourdhui encore des hommes
clairs confondent la langue avec son orthographie; Gaston Deschamps ne disait-il pas de
Berthelot quil avait prserv le franais de la ruine parce quil stait oppos la rforme
orthographique? (Saussure 1995:46). For Harriss translation, see Saussure 1986:25.
See Meier-Brgger 2003:12.
The critical comments of Tullio de Mauro in Saussure 1995:430. On the remarks, de
Mauro references Brunot and Bruneau 1956:xxxiii.
6 Of Styluses and Withes

1 See, inter alia, Khn 2.574.10; 2.581.10; 2.712.17.
2 See, inter alia, Khn 2.686.67; 2.711.14; 2.712.15. For an example of the spatulate probe,
see Caton 1914: pl. XI.29.
3 For a photograph of such a probe, see Caton (1914) pl. XI. 25. Caton notes (p.117) that
No. 25 is exactly like a probe of to-day; it has the two olivary thickenings at the ends.
4 And he says that if the anatomist needs a smaller probe than this type, then a
(mltis) can be used (Khn 2.581.10[De anat. admin.]).
5 It is interesting to note, if only as an aside for the moment, that Galen goes on to say these
several (elasmata) are made of hard wood: Galen uses boxwood ( [puksos]) for his own, as it never breaks (Khn 2.574.1012[De anat. admin.]).
6 Khn 10.889.15; 18a.376.8; and 18a.478.16.
7 As suggested by inscriptional spelling practices; other evidence, however, points to a
much later preservation of the distinction: see Allen 1974:71, who notes that, for example, Gothic and Old Armenian transcriptions of Greek register a phonetic distinction
between the eta-vowel and the iota-vowel (as opposed to showing confusion between the
sounds of long iota and the long vowel represented by the digraph (the so-called spurious diphthong).
8 At the coarsest level of analysis, one might be prone to characterize - (spathomil) as having arisen by a process of misspelling, or, perhaps somewhat less coarsely, a
form of folk etymology, given the variation - (amphi-mlon) versus -
(amphi-smil). LSJs identification of (amphismil) and (amphismilon) as probable errors for (amphimlon) fails to take account of the
evidence.
329
9 See also, inter alia, Callim., Epigr. 55; Herod. 4.19; and Aesch., Supp. 463, where the daughters of Danas liken themselves to the (pinakes) hanging from the images of gods
(and see my comments in Scott etal. 2005).
10 Votive objects linked with the notion of (kharis) granted in return by a
recipient deity; Day (1994:57) writes regarding the similarity between the inscription of
the Mantiklos statuette and Mentors prayer at Odyssey iii 5859: The Odyssean parallel
for Mantiklos prayer, though unique in literature, illustrates the kind of context in which
words typically appear. For example, only certain things exhibit in poetry
metal objects, cattle, textiles, what Homer calls or , the latter, of course,
the word for a dedication in epigrams. These objects function primarily as top-rank gifts
given by aristocrats, either to one another as marriage gifts, guest gifts, athletic prizes, and
grave goods, or to the gods as sacrifices or dedications.
11 See, inter alia, the discussion in Sommerstein 1994:198203; Tzanetou 2002:341344.
12 Thus, concerning lines 693695 in which In-Law threatens to sacrifice the baby on the
altar, Sommerstein writes (1994:198): The language is strongly tragic, and the lines may
well be quoted or adapted from Telephus or some other Euripidean play, since the reference
to thigh-bones does not fit the current dramatic situation.
13 Sommerstein 1994:204.
14 Sommerstein 1994:204.
15 (Chilias) means thousand, and the letter phi symbolizes 500 in alphabetic numeral
notation.
16 (Phidis): that is, phi twice, as itwere.
17 On the use of reed pens for writing, see Pliny, HN 16.157158: Egyptian reeds are especially
good, but those from Knidos and those growing in Asia circa Anaeticum lacum are considered even more desirable.
18 Regarding the ritual, Vernant (2006:186) summarizes: This ceremony was celebrated on
the fifth, seventh, or tenth day after birth and sometimes coincided with the naming of the
infant. Its specific function was to consecrate the official recognition of the newborn child
by its father. The ritual is obviously aimed at enrolling the child in the space of the oikos,
attaching it to the hearth of which it is the issue. According to the existing evidence, the
ritual consisted of two parts between which a distinction should be made: on the one hand,
the ring run around the hearth by one or more naked people holding the baby in their
arms; on the other the laying of the child at a given moment (probably before the running)
directly on the ground. For references to the ritual procedures, see p.452n125.
19 Acosta-Hughes (2002:121) calls attention to a formative influence of the Homeric Hymn to
Hermes on Callimachuss Iambus 12. But see, too, the followingnote.
20 At the center of the poem is a divine assembly gathered to celebrate Hebes birth, a birthday celebration that mirrors the mortal one, which in turn has occasioned this poem. The
gods gather at Heras invitation to engage in gift giving. The description of an assembly of
the gods on a festive occasion has a long tradition in earlier Greek poetry and myth; the
weddings of Peleus and Thetis and of Cadmus and Harmony are part of this tradition. At
these two occasions Apollo is also the singer, or the singer together with the Muses, writes
Acosta-Hughes (2002:124), along with his notation, Cf. Pindar Fr. 32 S.-M., Theogn. 1516
(Cadmus and Harmony); Il. 24.5563, Menander Rhetor 2[6] The Epithalamium (Peleus
and Thetis).
21 Acosta-Hughes (2002:130), for example, writes: Barbers conjecture at line 27 [ is, as
Gould noted, especially attractive, as Athenas gifts would then be done in relief. The reference appears to be to Barber 1955. For support, Acosta-Hughes cites Herodas Mimiambus
330
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
4.5759, where both the verb [glupha] and the image of Athena as chiseler
appear. See also his p. 130n25.
Acosta-Hughes 2002:136.
On which, see, inter alia, Martin 1989; 1996:119133.
On this possible interpretation, see Kerkhecker 1999:226.
On Callimachuss opposition to epic and the tradition of his personal conflicts
with Apollonius of Rhodes, see, inter alia, Lefkowitz 1980; Traill 1988; Clauss 1993;
DeForest1994.
Though compare remarks recorded in the Geoponica, the tenth-century encyclopedic
work on agricultural technique, in which the prescribed treatment for a vine that does not
bear involves cutting open the trunk of the vine with a (smil) or a (teretron; a boring tool), or better, (sphni drun(i)) with a wedge made of oak
(Geoponica 5.35).
Or a root *(s)meh1- (i.e., with the number-one laryngeal) if the verb to probe
were actually native to the (Sicilian) Doric dialect of the mimographer Sophron of Syracuse
(fifth century bc ), in whose fr. 146b of the edition of Olivieri it appears, and if the fragment is rightly ascribed to Sophron. Regardless of the matter of proper attribution, if the
word should belong to a Sicilian linguistic milieu, it is likely taken over in its Ionic form (as
opposed to an expected Doric evolutionary outcome of * -) from the technical lexicon
of medicine.
Walde and Pokorny reconstruct the root as the set of variants smi- : smi- : sm- and note:
Eine s-lose Wfz. ist wohl mai- (mi-) hauen, abhauen. See the next footnote.
Walde and Pokorny reconstruct the root as mai- and gloss: hauen, abhauen, mit einem
scharfen Werkzeug bearbeiten, wohl eigentlich mi- und s-lose Form neben smi- : smi- :
sm- schnitzen, mit einem scharfen Werkzeug arbeiten.
See Melchert 1994:49; Watkins 1998:43; Cowgill and Mayrhofer 1986:175; Jasanoff1978.
On Indo-European ablaut, see, inter alia, Watkins 1998:5153; Meier-Brgger 2003:144152;
Hoenigswald, Woodard, and Clackson 2004:540541.
See Melchert 1994:170.
Compare the discussions in Chapter2 of the alternations between consonantal w and y and
vocalic u and i respectively.
See Melchert 1984:99; 1994:58, 170; LIV545.
See Watkins 1998:43.
Recall that this metathesis is an active synchronic phenomenon in Proto-Indo-European.
From a cognitive processing perspective, this example reveals (i.e., requires) that the
metathesis of */Ch2wr/ to */Cwh2r/ precedes the syllabification process (or rule) whereby
the sonorants (nasals, liquids, glides) are assigned a vocalic value when they occur between
two other consonants or between a consonant and a word boundary (on the context of
syllabification, see, inter alia, Watkins 1998:44; Meier-Brgger 2003:9899; Hoenigswald,
Woodard, and Clackson 2004:538539). The word-final */-r/ would still become syllabic
in the case of metathesized */puh2r/, but with elimination of the laryngeal consonants in
descendant languages, desyllabification must potentially accompany. Compare Rix 1970
for the Greek treatment of word-initial sequences of laryngeal followed by syllabic liquid.
The centum languages are those of the western aspect of the ancient Indo-European speech
area (Hellenic, Anatolian, Italic, Celtic, Germanic), plus Tocharian in the FarEast.
See Walde-Pokorny 2:413414; LIV 416. In the latter, the author of the entry expresses
misgivings about including Lithuanian lgnas flexible, owing to a perceived violation of
Winters Law that this would entail. But Winters Law is itself a matter of considerable
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
331
uncertainty: see, inter alia, Szemernyi 1996:153154, where that author writes, The weight
of exceptions is in any case considerable, and the christening of Winters Law was perhaps
too hasty.
Compare the use of the related adjective (eustreptos), describing ox-hide line
for hoistingsail.
Nagy 2002:76.
Nagy 2002:78.
On whom, in this regard, see also the remarks of Scheid and Svenbro 1996:24, with n. 79,
and pages 122124, with n.41.
Scheid and Svenbro 1996:206n26.
Scheid and Svenbro 1996:118.
Scheid and Svenbro 1996:118. The translation of Bacchylides is that of the English edition of
Scheid and Svenbroswork.
Here Scheid and Svenbro refer to the hypothesis of an etymological link between huphainein and humnos [hymn]. See their references at p.206n28. Regarding the etymological
question, see Nagy 2002:7071 with references.
Scheid and Svenbro (1996:118119) also call attention to Bacchylides Dithyramb 19.811: In
a poem addressed to the Athenians, Bacchylides uses the same metaphor, calling on the
care of perfection (merimna) characteristic of the poets from Keos (his uncle Simonides
and himself): Weave [huphaine] something new in the rich beloved Athens, O famous
perfectionism of Keos!
See Heubeck, West, and Hainsworth 1988:138; Russo, Fernndez-Galiano, and Heubeck
1992:8081,93.
See, inter alia, Bremmer 1983:310312, with references to earlierwork.
The dating of the Homeric Hymn to Hermes has proved to be a matter of some uncertainty, or contentiousness. Allen, Halliday, and Sykes (1963:276) argue for a date not later
than seventh century. Many would assign a more recent date, ranging between the late
sixth and the early fourth century. For brief discussion with references, see, inter alia, S.
Johnston 2002:109n1. West (2003:14) writes: It is generally agreed on grounds of style
and diction that it must be the latest of the major Hymns. The subject matter was already
familiar to Alcaeus, whose lyric hymn to Hermes covered the same ground. Alcaeus very
likely knew a Homeric hymn on the subject. But to date our Hymn as early as 600 is
implausible: it contains too many words and expressions that are not paralleled before the
fifth century. The likelihood is that it is a later descendant of the hymn that Alcaeus knew.
This is possibly so; but on the other hand, the Hymn to Hermes that we have appears to
preserve primitive Indo-European motifs that surface in Indic epic, and to lack the Near
Eastern influence evidenced in other major Homeric Hymns: see Allen and Woodard 2013.
Acosta-Hughes (2002:122) calls attention to the formative influence of the Homeric Hymn
to Hermes on Callimachuss Iambus 12 (considered in note 18): a poem that finds a wide
variety of resonance in the literature of Ptolemaic Alexandria.
See, for example, Bremmer 1983:311, with n. 68. The connection is apparently made secondary to textual emendation, on which see the immediately following discussion intext.
See the discussion of Allen, Halliday and Sykes 1963:330332.
On these plant names, see Gow and Scholfield 1988, Index I (beginning on p.228).
Chantraine 1968:2:195: n. image en bois dune divinit Non homrique, p.-. dorien.
Cest un quivalent de . Le terme sapplique aux vieilles idoles de bois, hritage des
plus anciens cultes. Terme mditerranen sans tymologie,.
On the festival, see, inter alia, OBrien 1993:5462.
332
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
Frazer 1965:4:122.
Kyrieleis 1993:135.
And, in addition, see Pliny, HN 16.234240.
Theophrastus (Hist. pl. 4.13.2) mentions the olive tree on the Athenian Acropolis, the palm
on Delos (beneath which Apollo and Artemis were born), the olive tree at Olympia, the
leaves of which provided wreaths for athletic contestants, and oaks of Ilium. He also reports
the opinion that Agamemnon planted the plane tree at Kaphyai (rather than Menelaus), as
well as the same type of tree at Delphi; Pliny (HN 16.238) agrees on this point.
Megaw 19631964:24.
Catling 19881989:94.
Megaw 19631964:24.
Kyrieleis (1993:126127) writes that the first temple of Hera was constructed as early
as the eighth century bc , a long narrow structure typical of that time, 100 ft long. This
Hekatompedos was replaced by the gigantic dipteral temple built before the middle of the
sixth century bc . The so-called Rhoikos temple must have been destroyed shortly after completion. Its successor was a somewhat larger temple, shifted about 40m further west, started
in the Late Archaic period. The earliest altar predates the temple: The inconspicuous beginnings of the altar may perhaps date back to late Mycenaean times (Kyrieleis 1993:128).
Compare the remarks of Kyrieleis (1993:135) regarding the juniper tree: The tree probably
belonged to a sacred grove, mentioned in an ancient source; that grove will have been east
of the altar. At best the tree serves to indicate that the altar lay directly at the edge of the
grove.
So LSJ, citing SEG 977a 10 (Delos, ii b. c .).
On the recorded voice of Orpheus, see Detienne2003.
See Sthlin 1972:36.
Frazer 1965:4:123.
On Daedalus and the (daidala), see Morris1992.
Bechtel 1917:603.
IG XII 8 no.298.14.
For general discussion of this type of compound in Sanskrit and other ancient IndoEuropean languages, see Burrow 1955:210213.
The compound vtrahan- is of Indo-Iranian origin; see Dumzil 1970:115138.
Burrow 1955:212.
For archaic Greek compounds with a similar structure but a different semantic relationship
(bahuvrihis, or possessive compounds), see Hamp1985.
An additional example of such a compound may be provided by the name of the Celtic
priest, as in Gaulish Druid, preserved in the Latin borrowing Dru-ides: the second member
of the compound is almost certainly from Proto-Indo-European *wid- and the first from
*deru- / *dreu- to be firm, solid, source of, inter alia, the word for tree in various IndoEuropean languages, including English, and, more restrictive semantically, for oak in
some. But whether dru-wid- means knowing trees/oaks, with a presumed intended reference to ritual and/or magic practice (as envisioned by Pliny [HN 16.249250]), as opposed
to meaning something like strong seer, has long been a matter of scholarly disagreement.
Addressing such compound formations, Buck (1933:356) summarizes: The stem was originally the same as that of the uncompounded word. But there are many analogical substitutions, especially a great spread of the o-stem at the expense of others. Palmer (1980:260)
writes regarding compounds: A large proportion of first members were noun stems of the
thematic declension whose stems ended in -o-. As a consequence this vowel was regarded
as a composition suffix par excellence and became attached to other stems.
333
78 Examples are drawn from Buck 1933:356.

79 Sanskrit reveals that the primitive accent occurred on the second member of the compound (consider the preceding examples). The accent of - (Smil-is), -- (Smilid-os) is a Greek development.
80 Buck 1933:356.
81 On the (dmioergos), see, inter alia, Finley 2002:4445; Burkert 1992:23;
Woodard 2010:45.
82 Chantraine 1968:273; Buck 1955:134. Buck allows that the forms - (dmi-orgos)
and o, [dmi-orgos] could come also from *- (*dmio-wergos),
that is, with loss of the first two segments of the second member of the compound, but
advocates the elision solution as the simplest, which must certainly be correct, especially given the parallel treatment of - (dmi-ergos) in Central and Eastern
Aegean Doric, and most especially given the close proximity of Astypalaea and Nisyros
to Samos.
83 Buck 1955:133134.
84 On Calands Law and the Caland System, see, inter alia, Fortson 2004:123; MeierBrgger 2003:288291; Szemernyi 1996:193194; Rau2009.
85 Chantraine 1968:1028.
86 The presence of a negative formant as the first element of the compound is certain, though
its etymological history is not. Often analogy is invoked to account for -, as by Heubeck,
West, and Hainsworth (1988:357): --. - is doubtless by analogy with ,
, where it is the product of an early contraction. See also, Chantraine 1968:750:
Compos dont le second terme appartient certainement la racine de () savoir, le
premier terme de sens ngatif remontant la ngation de phrase *ne qui figure en composition avec contraction ancienne dune voyelle initiale dans , etc. Compare the
sense of the Germanic forms built on Proto-Indo-European *wid- (see Walde-Pokorny
1:238; Lehmann 1986:379): Gothic un-wita ignorant, fool; Old English un-wita; Old High
German un-wizzo fool; Old High German fora-wizzo foreknowing.
87 Kirk 2000:260.
88 For a different restoration of the beginning of line 1, see Acosta-Hughes and Stevens
2002:238n5, with bibliography.
89 So described by Acosta-Hughes and Stevens in their study of Callimachus and his literary
adversaries (2002:241). The authors refer their readers (241n12) to Herter 1934:197224 for
general discussion. See the same note for additional references.
90 Kyrieleis 1993:138.
92 Braun 1982:38.
93 See the discussion in Braun 1982 (4041); also see Graham 1982:134135 and, for a
more recent overview, Boardman 2004:139142. For a recent comprehensive study, see
Mller2001.
95 Mykytiuk 2004:119, referencing Ephal and Naveh 1989 and Bron and Lemaire1989.
96 Luraghi 2006:39; Dalley 1985:3839.
98 See Luraghi 2006:3940, with references. These are very likely offerings made by Greeks
who had served as mercenaries in armies of the Near East. Luraghi (pp.4041) observes
that these pieces are of a slightly larger group of objects, which includes four more blinkers and one frontlet from Samos and two frontlets and one blinker from the sanctuary of
Athena at Miletos. They all belong together in terms of chronology and style.
334
100 The water table was already shallow at the time the wells were in use. Kyrieleis (1993:136
137) writes: The wells were surrounded by limestone paving which shows, among other
things, that the ground here had already become swampy at the time these wells were
dug and had to be given a firm surface. The filling inside the wells consisted of potsherds,
whole pots and broken votive offerings. These fillings, as well as the layers below the paving of the wells, can be dated by means of the Corinthian pottery. Thus it can be shown
that the wells were constructed in the late seventh century bc and were soon afterwards
filled in by the early sixth century bc . All the finds from these contexts date to the seventh or, at the latest sixth century bc . The many animal bones, pieces of charcoal and iron
cooking spits show that the wells were filled with debris from the sanctuary. The votives
found in the well had either been damaged or had become unsightly before they were
discarded or buried.
101 See Kyrieleis 1993:141145.
102 Or, alternatively, the site was said to have been founded by Deucalion after the primeval
flood; see Nicol 1958:130 with references.
103 In his prayer Achilles refers to the oracular interpreters of Dodona, the Selloi, whose
feet are unwashed and who sleep on the ground (
[hupophtai aniptopodes khamaieunai]. Philostratus (Imag. 2.33) calls them Helloi (and
describes their individual functions), as did Pindar, according to a scholion on Il. XVI 234
(see Nicole 1966: vol.1).
104 Mitchell 2001:342, who references, inter alia, Dakaris 1993:78 and Hammond 1967;
Mitchell continues: A Mycenaean gold bezel dating from the fifteenth century has also
been discovered at Mycenae depicting what appears to be an ancient earth mother under
the sacred oak with her three priestesses (reflected in Herodotus three priestesses of Zeus:
2.55.2?), which Dakaris [1993:8] links with Dodona. Pausanias also seems to make reference to a hymn sung by the priestesses which seems to refer to a mother-goddess figure
(Paus. 10.12.10) [Hammond 1967:368369], and a fragment of the Catalogue of Women
also appears to allude to the chthonic and pastoral nature of the Dodonian cult ([Hesiod]
Fr. 240 Merkelbach-West).
105 See, inter alia, LSAG2 228,230.
106 Parker2008.
107 Buck 1933:261. He continues (p.262): In general, - [-azd] is more common from stems and neuter n-stems, - [-izd] from other stems.
108 Chantraine 1984:236. Chantraine (1973:340) notes that the formant is also used in certain
instances to form deverbative verbs. On doublets of verbs in - (-izd) and (-e),
see Chantraine 1984:241; 1973:339340.
109 Examples are drawn from Buck 1933:262.
110 The text is that of Lloyd-Jones1996.
111 Though see the remarks of Allen, Halliday, and Sykes 1963:330331.
112 See, inter alia, Chantraine 1984:213216; Rix 1976:211212; Buck 1933:260.
113 I say in effect since many of these verbs are formed by a process of analogy that has its
origin in straightforward secondary derivation. The addition of the commonly occurring
suffix -yo- to stems ending in -- (-id-)/-- (-ig-) or -- (-ad-)/-- (-ag-) produced the
respective sequences -- (-izdo-) and -- (-azdo-) by regular sound change (i.e. *d + *y
zd and *g + *y zd); see the next paragraph in the body of the text. The denominative
formants -- (-izdo-) and -- (-azdo-) that is - (-izd) and - (-azd) with the
first person singular ending attached spread analogically to a great many nouns whose
stems had ended in neither *d nor *g, as indicated previously.
335
114
115
116
117
118
119
120
See Buck 1933:261.

See note 113.
See Buck 1955:119.
IG VII 3467. See LSAG2 93 and pl. 9, 18; Buck 1955:228.
Wackernagel 1897:5051.
Watkins 1976:26.
Watkins footnotes the comment (1976:26n3): At the present I am still inclined to regard
the secondary ending as basic and to derive *, from * + , + , like
from + , as a Greek innovation.
121 Watkins 1976:2627.
122 On the formation, see, inter alia, Buck 1933:262.
7 The Warp and Weft of Writing

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Saussure 1986:25.
Snyder 1981:194195.
Scheid and Svenbro 1996:117: the translations of Pind. Nem. 8.15 and fr. 179 that follow are
myown.
Scholion on Pind. Nem.7.79.
On Pindar and the (kmos), see Morgan 1993, with references to earlierwork.
Walde-Pokorny 1:717.
See, inter alia, Mallory and Adams 1997:139. On Neolithic Pontic-Caspian structures so
fabricated, see Anthony 2007:143, 166, and285.
Snyder 1981:194. She cites (n. 4)the following passages as relevant: Metaphorical references to weaving occur in Il. 3.21113, 6.18789, and 7.32425; Od. 4.67780, 5.35657,
9.42023, 13.30307, and 13.38688. Penelopes wiles involving weaving are alluded to in
Od. 2.9395; 2.10409; 15.51217; 19.13840; and 24.12850.
Snyder 1981:194.
Nagy 2007:19.
Snyder 1981:195.
Snyder 1981:194.
Greek (epen tektones); Avestan vaas-tati-; Sanskrit vcas- tak-; see
Schmitt 1967:14, 297298.
And this is one of the passages cited by Schmitt (1967:14, 297298) in his development
of the idea of the Indo-European metaphor. Schmitt (p.14) credits James Darmesteter
for having called attention to its Indo-European status in an article that appeared
in1878.
As cited in Ferber 1999:229. See also Schmitt 1967:300. Ferber cites numerous examples of
the prolongation of the use of the metaphor in English literary tradition.
Williams 1944:7. See also Davies1995.
336
24 See Mallory and Adams 1997:436437; Campanile 1990; Schmitt 1967. Scheid and Svenbro
(1996:121) are not unaware of the evidence for prehistoric Indo-European metaphors of
poetic weaving; they write that instead of reducing the difficult history of poetic weaving in Greece to a few lexical facts quickly integrated into the global Indo-European history of language weaving, we wanted to respect the fact that poetic weaving is resolutely
absent from the Homeric poems and to try to understand why. One admires the investigatory task that these two superb scholars have set for themselves; but the argumentum ex
silentio, which applies only to the specific lexical choices in the oral epics performatively
(re)composed forms in which we have them, must certainly be tempered by the comparative
evidence provided within Greece and beyond. Their further contention that the reductive
tendency of the comparativist approach is inclined to erase the specific characteristics of a
culture or of a cultural history in favor of an abstract synthesis, high in the Indo-European
sky, where everything is the same because it must be is perhaps a right characterizing of
some work but certainly does not describe the necessary case. Interpreting what Homer
does have to say about word weaving locally within a comparative archaic Greek context,
framed by the greater, and formative, Indo-European context, mandates no compulsory
view of Homer as ignorant of the metaphor of poetic word weaving and surely suggests a
general archaic Greek awareness of the metaphor.
25 Scheid and Svenbro 1996:118.
26 Coogan 1974:63n16.
27 See Woodard 2006:188.
28 For recent linguistic discussion of recta and regilla, see Nielsen 2004:202.
29 Does the phrase woven upwards denote the practice of packing the weft upward as one
weaves from the top on a warp-weighted loom? See Barber 1991:92 for a description of that
process.
30 OLD1266.
31 See Walde-Pokorny 1:749750. For Ernout-Meillet (1959:707708) the Latin lexemes are
likely borrowed from Greek rather than being direct inheritances.
32 In his summary remarks on the work, Lesky (1996:831) aptly notes: It deals with the two
elements of style, vocabulary ( [eklog]) and the arrangement of words (
[sunthesis]), the latter with a great many examples. Many of these observations, especially
those on the combination of sounds, may make us realize to what extent the effects of
ancient literary language are no longer accessible tous.
33 See the followingnote.
34 In his essay On Isocrates (3), Dionysius attributes to Theophrastus the idea that there are
three elements that bring about (to mega kai
semnon kai peritton en leksei) greatness and majesty and eminence in style (with specific
reference to the style of the speeches of Isocrates): these are (eklog
tn onomatn) selection of words; (harmonia) fitting together of those words;
and (skhmata) figures of speech that frame these. With regard to which, particularly the last-named of the three and its distinction from the first two, Fortenbaugh
(2005:292293), in his commentary on the fragments of Theophrastuss works on rhetoric
and poetics, remarks:
This threefold division may well have determined (at least in part) the structure of
Theophrastus work On Style (666 no.17a). The selection of words will have been treated
first, their arrangement second, and figures third, perhaps in a separate section or chapter.
And if figures were picked out for independent treatment, then Theophrastus may have
given special impetus to the cataloguing of figures, which later generations practiced to
35
36
37
38
39
40
41
42
43
44
45
46
47
48
337
excess. We should, however, be clear that the Theophrastean triad was no creation ex
nihilo. It is already implicit in Aristotle, who begins his discussion of style by focusing on
diction (ordinary and strange words, metaphor, etc. 3.24), then takes up composition
(including prose rhythm and periodic structure 3.89), and along the way gives special consideration to several well-known figures: namely, antithesis, balanced clauses and
similarity in sound (3.9 1409b351410b1). Here figures come after periodic structure (i.e.,
they are discussed in the second half of Rhetoric 3.9), so that by position they are marked
off from periodic structure. But strictly speaking, figures are not assigned to a separate
section or chapter. They are introduced as a feature of periodic structure and as such are
part of Aristotles treatment of that topic. I hesitate, therefore, to assert without qualification that Theophrastus gave independent treatment to figures. He may have made
clear that they are distinct from periodic structure and nevertheless treated figures (at
least those of antithesis, balanced clauses and similarity in sound) together with periodic
structure as important enhancements.
If, then, for Theophrastus (as implicitly so for Aristotle) figures were fundamentally distinct from selection and combination, the essentially dyadic structure of selection and combination that emerges in (Peri suntheses onomatn) may have
found a well-articulated expression prior to Dionysius own work on the subject.
Nagy 2002:78.
Scheid and Svenbro 1996:206n26.
Nagy too has called attention to the correlation of the horizontal and vertical axes of language and the horizontal and vertical fibers of the loom; in his discussion of weaving vis-vis notions of beginning, he writes (Nagy 2002:79): Let us apply here the Prague School
construct of a horizontal axis of combination interacting with a vertical axis of selection.
From the standpoint of working at the loom, you cannot move horizontally from one point
to the next unless each given oncoming point has already been set for you vertically.
Barber 1991:79.
Barber 1991:79.
LSJ1430.
See Chantraine 1968:924; Walde-Pokorny 2:910; Mallory and Adams 1997:414; Gamkrelidze
and Ivanov 1995:123,543.
Nagy 2002:9293.
And compare De compositione verborum 2, where, after discussing the number of parts
of speech that were recognized by Aristotle, the Stoics, and subsequent grammarians, he
remarks that regardless of how many distinct parts are identified, the weaving / plaiting
( [plok]) of these makes clauses.
The notion is clearly bound up with instruction in alphabetic writing; I believe that the
use of syllable as an orthographic primitive is most likely an artifact of a pedagogical
tradition that has its origin in a (pre-alphabetic) time and place in which the Greeks were
writing syllabically. Iron-Age Cyprus would of course be the likely candidate. On the transmission of an orthographic tradition from Cyprus to the Greek mainland, see Woodard
2010:3945; 1997:256258.
See especially De compositione verborum15.
Where the preposition (dia) with the genitive case indicates the material out of which
a thing is made: see LSJ 389A III2.
The translation is that of the New English Bible.
This interweaving of the Hebrew abecedarium is also known to have had a pedagogical
usage, mentioned by Abot de-Rabbi Nathan, as reported by Demsky 1977:20 and Steiner
338
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
1996:82n35. Steiner translates as follows: He and his son went and sat with teachers of
children. He said to them, My master, teach me Torah. R. Akiva held the head of the tablet
and his son (held) the head of the tablet. He wrote aleph beth for him, and he learned it.
He wrote aleph tau for him, and he learnedit.
Steiner 1996:82; Tigay 1983:179.
Steiner 1996:83. A second example of atbash spelling can be seen at Jeremiah 51:1; see
Steiner, pp.81, 8384.
Regarding these two systems, and a third that he identifies as atbah, Kahn (1996:79)
writes:
These three substitutes are used here and there throughout Hebrew writing, particularly
atbash, which is the most common. Their importance consists, however, in that the use
of atbash in the Bible sensitized the monks and scribes of the Middle Ages to the idea of
letter substitution. And from them flowed the modern use of ciphers as distinct from
codes as a means of secret communication.
See Demsky 1977:1920. Regarding the atbash and albam practices, he remarks: Exercises
of this sort in learning the alphabet were probably as old as the order and letter names
themselves and were employed in the dissemination and learning of the alphabet throughout the ancient world.
Coogan 1974:61.
Not including local additions to the original alphabet, Coogan 1974:62; Demsky
(1977:20n15) writes: For an Ugaritic example in writing letters and their reverse, see PRU
II:198.
Coogan 1974:61.
Coogan 1974:62. IG XIV 2420.4; see Arena 1996:98 and Tav. XIX 12; EG 1:115117, with
fig.17ab.
LSAG2 256. See also pp.261 and 291, and pl. 50,19.
See LSAG2 9495 and pl. 10, 20. Compare Jefferys hand-drawn copy at http://poinikastas.
csad.ox.ac.uk/Papers/CG/1000/J.CG.Boi.20.p01.jpg.
LSAG294.
See LSAG2 237 and pl. 48, 21; Pandolfini and Prosdocimi 1990:2122 and Tav.III.
Greenough 1890:97.
And to these he adds Luciliuss use of the Latin borrowing stoechia (28.805811) in the sense
elements, which merely shows that stoechia was already in a manner in use in the language in its technical sense ready to be translated as soon as a more patriotic Sprachgefhl
should suggest that course.
Greenough 1890:99. He continues: The only objection that can be made is the fact that
these letters are not the first of the alphabet. But they are the most glib and easy in their
utterance; and though we have no example of the same kind in regard to letters, yet, as
has been suggested by a friend, we do have Solfeggio and sol-fa-ing, of precisely the same
nature, and chosen for precisely the same reason.
Coogan 1974:62.
LSAG231.
LSAG229.
EG 1:94; McCarter 1975a:82.
As in IG XII iii 536,540.
See, for example, LSAG2 pl. 48, 1920; Pandolfini and Prosdocimi 1990:2425 (Tavv. VVI),
28 (Tavv. VIIVIII), 31 (Tavv. IXX).
McCarter 1975a:116.
339
71 See Abou-Assaf, Bordreuil, and Millard1982.

72 Kaufman 1984:571.
73 See, inter alia, Naveh 1973; 1987:175186; 1991; see also Bernal 1987:13. See also Cross 2003,
especially chap.4.
74 The morphology of the typical archaic Greek eta-symbol could be folded into this scenario if need be, accounted for by a process of graphic assimilation. A second-millennium
bc source has also been invoked for this three-barred Greek character. A three-barred
Phoenician et is found in the inscription on the eleventh-century Nora fragment from
Sardinia (and, as pointed out in Chapter2, the squarish variety of CP eta is particularly
reminiscent of this form of the Phoenician ancestral symbol). McCarter (1975a:116) appropriately observes: It is also true, however, that some variety of et is found throughout the
period of the early Phoenician scripts and that this variety is reflected in the various forms
of Greek heta / eta. The unqualified postulation of a second millennium borrowing of the
alphabet, therefore, would not account for all the evidence.
75 And, as noted in Chapter2, these anomalies do not include that omission of the alphabetic
sequence of epsilon thorough san on MS2-2.
76 Snyder 1981:194.
77 See Langdon 1976:1317 for dedicatory inscriptions.
78 See Langdon 1976:1718 for abecedaria.
79 See Langdon 1976:1821 for these inscriptions.
80 H 223, H215, H 181. See Langdon 1976:20.
81 Langdon 1976:18; inscription 27 (H189).
82 Langdon 1976:20; inscription 30 (H156).
83 Langdon 1976:46.
84 Though a similar practice in a later period and at other locales may not be absent from the
record. In the discussion of CP iota, reference was made to the occurrence of a similar type
in a late sixth-century Melian graffito on a marble column. The inscription reveals the column to be an agalma for a child of Zeus, dedicated by Ekphantos: h
h (soi gar epeukhomenos tout etelesse grophn). If h (grophn)
is to be understood as the participle (graphn), then Ekphantos made this vow to
you and fulfilled it by writing. Compare the presence of the same form in an inscription
from Olympia (ca. last quarter of the sixth century bc ) that makes reference to Melians
(IGA 12); see LSAG2 320,324.
85 Langdon 1976:46.
86 Henrichs 2003:51. Though Henrichs characterizes Langdons words that It was writing
itself which was the gift by remarking that Langdons suggestion is attractive, but remains
speculative; then follows his observation: Still, there can be no doubt that in the shrine of
Zeus on Mt. Hymettos My view differs from Henrichss own at least as he expresses it
in his 2003 article to the extent that the conceptualization of the producing of alphabetic
strings as a performative act is not idiosyncratically limited to the cult of Zeus on Mount
Hymettos.
87 -yo- furnishes the most productive type of IE adjectives, some primary , but mostly secondary derivatives from nouns and denoting any sort of relationship to the latter (Buck
1933:316317). The suffix -yo- or -io- (-iyo-) was one of the most common of all suffixes
both in IE times and throughout the history of the Greek language, which, before the modern period, forms at least 10,000 words with it (Buck and Petersen 1949:43).
88 On folding tablets in the Near East and in Mycenaean Greece, see Kirk 2000:181182, with
references, and Shear1998.
340
89 Langdon 1976:15.
90 Langdon 1976:101. On the former point he cites Strabo 60.2.11 and on the latter
Theophrastus De signis tempestatum3.47.
91 Langdon 1976:3. He cites Theophrastus, De signis tempestatum 1.20.24;3.43.
92 Buck and Petersen 1949:35. See also Chantraine 1933:253: Le suffixe en -- sest combin
avec la finale - selon un procs que nous ne pouvons reconstituer. Cette combinaison a
d tre sentie comme expressive. Cette expressivit apparat plein dans un mot comme
[penikhraleos, needy] (hapax, Anthologie VI, 190) o - se combine
avec [penikhros, needy]. See also Benveniste 1962:4449: On en possde des
relevs complets et une description prcise dus en partie M. Fraenkel (Griech. Denom.,
p.10; KZ., XLII, p.114), tout particulirement M. Debrunner, IF, XXIII, p.143 et aussi
XXI, p.36 sq. et repris chez Chantraine, Formation des noms, p.253 sq. Mais sur lorigine
de -al- lincertitude persiste.
93 Buck and Petersen 1949:35: The suffix was well developed in Homer, but it spread out
more and more in later poets, the total number of words formed by it being about112.
94 See Walde-Pokorny 2:250251; Chantraine 1968:717718.
95 See Russo, Fernndez-Galiano, and Heubeck 1992:212213.
96 Looking at the set of adjectives in - (-aleos) from all periods, Chantraine (1933:254
255) observes them to be distributed en un certain nombre de groupes smantiques. A.
Des mots signifiant sec ou humide, chaud ou froid, etc. On peut joindre ce groupe
quelques mots exprimant lide de poussire, salet, etc. B. Mots signifiant audacieux,
terrible, craintif. Un certain nombre des adjectifs tudis ci-dessus expriment un
manque ou un dfaut physique. Quelques adjectifs exprimant lide de cuit, brl.
To these, Chantraine adds a miscellaneousset.
97 Walde-Pokorny 1:831832.
98 Nagy 1990:202222.
99 Frame1978.
100 Nagy 1990:203. Emphasis ismine.
101 Nagy 1990:203.
102 Nagy 1990:204.
103 Nagy 1990:206.
104 Nagy 1990:211.
105 Nagy 1990:211. Which response, Nagy adds (p.212), citing Hesiod, Works and Days 483
484, is a hard thing todo.
106 See Benveniste 1948:3031. Benveniste groups (smantr) with other agent
nouns in - (-tr) that carry the sense of chef: (mstr) adviser;
(hgtr) leader; (kosmtr) commander. He notes (p.30): Ce sont tous des
noms valeur participiale.
107 On the question of the Hesiodic authorship of The Shield of Heracles, see Woodard
2007b:8485, with reference to Nagy 1990:79.
108 Heubeck and Hoekstra 1989:174.
109 For discussion of the fragment, see Robinson 2003:143144.
110 On the question of the identity of this Pandora, see, inter alia, Osborne 2005:89; West
1985:52.
111 In addition Fragment (MW) 195 repeats line 56 of The Shield.
112 See Tsagalis 2008:267, with notes.
113 The discussion that follows is not intended to be an exhaustive treatment of chthonic
Zeus but addresses that divinity as signifying deity. On chthonic Zeus, see, inter alia,
Burkert 1985:200201, with references to earlierwork.
341
114 On (ennesi), see Nagy 2008. See also Risch 1937:115.

115 See Burkert 1992:8891; 2004:3438; cf. West 1997:159160.
116 See Foster 2005:227280. Another version of this story is found in Tablet XI of the
Gilgamesh epic (Foster 2005:228).
117 West2003.
118 Evelyn-White1936.
119 Cashford2003.
120 The Greek adjective (polus), much, many, combining form - (polu-), is from
the Proto-Indo-European root *pelh1- to fill; (ploutos) riches, abundance is
from *pleu- to flow. Plato rightly makes the etymological connection with
(ploutos) at Cratylus 403A, where he has Socrates affirm that the sovereign chthonic deity
is given the name Pluto (hoti ek ts gs katthen
anietai ho ploutos) because wealth is sent up from below out of the earth.
121 The discussion that follows draws extensively on the list of relevant passages presented by
West (1978:276). See also Farnell 1907:3:280288.
122 Benveniste 1969:1:342; Finley 2002:92.
123 Woodard 2007b:146.
124 See also Aeschylus Agamemnon 13851387.
125 See, inter alia, Cordano 1984:282283, 291, 304, no.7; Dubois 1995:3640.
126 The lower in the orientation given the inscriptions in (26)(27); see subsequent discussion in thetext.
127 On the reversed order of the two letters, Dubois (1995:40) notes: La succession H Z
comme le remarque Cassio, est rcemment apparue dans un dbut dabcdaire corinthien archaque grav sur une lampe attique. For the latter inscription, Dubois cites
Boegehold 1992:411.
128 See Watkins 1995a:44: A seven-syllable catalectic dimeter is not impossible, with a choriamb in the first foot.
129 Cassio 19911993 [1994].
130 Ribezzo 1920:7577.
131 Cassio compares the infinitive () (tinnuna[i]) with Homeric -, where the first
vowel scanslong.
132 Compare Sophocles Oedipus Rex 810 and Aeschylus Suppliant Women 435, noted by
Cassio and Watkins.
133 Watkins 1995a:44. See also Watkins 1995b:104.
134 Compare Iliad III 278279; XIX 259260; and especially Odyssey xiii214.
Bibliography
Abou-Assaf, Ali, Pierre Bordreuil, and Alan R. Millard. 1982. La statue de Tell Fekherye et son
inscription bilingue assyro-aramenne. Paris.
Acosta-Hughes, Benjamin. 2002. Polyeideia: The Iambi of Callimachus and the Archaic Iambic
Tradition. Berkeley and Los Angeles.
Acosta-Hughes, Benjamin, and Susan A. Stevens. 2002. Rereading Callimachus Aetia
Fragment 1. C Phil. 97:238255.
Albright, William F. 1941. New Light on the Early History of Phoenician Colonization. Bulletin
of the American Schools of Oriental Research 83:1422.
Allen, James P. 2000. Middle Egyptian. Cambridge.
Allen, N. J., and Roger D. Woodard. 2013. Hermes and Gandharvas. Nouvelle mythologie
compare.
Allen, T. W., W. R. Halliday, and E. E. Sykes. 1963. The Homeric Hymns. 2nd ed. Oxford.
Allen, W. Sidney. 1965. On One-Vowel Systems. Lingua 13:111124.
1974. Vox Graeca. 2nd ed. Cambridge.
Anderson, Stephen R. 1978. Syllables, Segments and the Northwest Caucasian Languages. In
Bell and Hooper 1978, pp.4758.
Anthony, David W. 2007. The Horse, the Wheel and Language: How Bronze-Age Riders from the
Eurasian Steppes Shaped the Modern World. Princeton.
Aravantinos, Vassilis L., Louis Godart, and Anna Sacconi, eds. 2001. Les tablettes en linaire
B de la Odos Pelopidou. dition et commentaire (Thbes fouilles de la Cadme 1). Pisa
andRome.
Arena, Renato. 1994. Iscrizioni greche arcaiche di Sicilia e Magna Grecia, III, Iscrizioni delle
colonie euboiche.Pisa.
1996. Iscrizioni greche arcaiche di Sicilia e Magna Grecia, IV, Iscrizioni delle colonie achee.
Alessandria.
1998. Iscrizioni greche arcaiche di Sicilia e Magna Grecia, V, Iscrizioni di Taranto, Locri
Epizefiri, Velia e Siracusa. Alessandria.
Atiya, Aziz S. 1967. Codex Arabicus (Sinai Arabic Ms. No. 514). In Lehmann-Haupt 1967,
pp.7585.
1985a. An Interview by Everett L. Cooley. By Everett L. Cooley. Everett L. Cooley Oral
History Project, nos. 328 and 329 (August 12, 1985).
343
344
Bibliography
1985b. An Interview by Everett L. Cooley. By Everett L. Cooley. Everett L. Cooley Oral

History Project, nos. 330 and 331 (August 13, 1985).
ed. 1991. The Coptic Encyclopedia. NewYork.
Bakker, Egbert J., ed. 2010. The Blackwell Companion to the Ancient Greek Language. Oxford.
Barber, E. A. 1955. Callimachea Varia. CR 5:241242.
Barber, E. J. W. 1991. Prehistoric Textiles: The Development of Cloth in the Neolithic and Bronze
Ages, with Special Reference to the Aegean. Princeton.
Bechtel, Friedrich. 1917. Die historischen Personennamen des Griechischen bis zur Kaiserzeit.
Halle.
Bell, Alan, and Joan B. Hooper, eds. 1978. Syllables and Segments. Amsterdam.
Benveniste, mile. 1948. Noms dagent et noms daction en indo-europen. Paris.
1962. Origines de la formation des noms en indo-europen. 2vols. Paris.
1966. Problmes de linguistique gnrale. Vol. 1. Paris.
1969. Le vocabulaire des institutions indo-europennes. 2vols. Paris.
Bernal, Martin. 1987. On the Transmission of the Alphabet to the Aegean before 1400 B.C.
BASOR 267:119.
Bessios, Matthaios, Yannis Z. Tzifopoulos, and Antonis Kotsanas. 2012.
I. Thessaloniki.
Bietti Sestieri, Anna Maria. 1992. The Iron Age Community of Osteria dellOsa. Cambridge.
Bietti Sestieri, Anna Maria, and Anna De Santis. 2000. The Protohistory of the Latin Peoples.
Milan.
Bietti Sestieri, Anna Maria, Anna De Santis, and A. La Regina. 1990. Elementi di tipo cultuale e doni personali nella necropoli laziale di Osteria dellOsa. Scienze dellAntichit
34:6588.
Bile, Monique. 1988. Le dialecte crtois ancien. Paris.
Blegen, Carl. 1934. Inscriptions on Geometric Pottery from Hymettos. AJA 38:1028.
Boardman, John. 1980. The Greeks Overseas. London.
2004. Settlement for Trade and Land in North Africa. In Tsetskhladze and De Angelis
2004, pp.137149.
Boardman, John, and N. G. L. Hammond. 1982. The Cambridge Ancient History. 2nd ed.
Vol. III, part 3, The Expansion of the Greek World, Eighth to Sixth Centuries B.C.
Cambridge.
Boegehold, A. L. 1992. Two Graffiti from Ancient Corinth. Hesperia 61:409412.
Bonfante, Giuliano, and Larissa Bonfante. 1983. The Etruscan Language: An Introduction.
Manchester.
Brann, Eva. 1961. Late Geometric Well Groups from the Athenian Agora. Hesperia 30:
93146.
Braun, T. F. R. G. 1982. The Greeks in Egypt. In Boardman and Hammond 1982, pp.3256.
Bremmer, Jan. 1983. Scapegoat Rituals in Ancient Greece. Harv. Stud. 87:299320.
Brixhe, Claude. 1976. Le dialecte grec de Pamphylie. Paris.
2004a. Nouvelle chronologie anatolienne et date dlaboration des alphabets grec et phrygien. Acadmie des inscriptions et belles-lettres, pp. 271289.
2004b. Phrygian. In Woodard 2004, pp.777788.
2006. Zn et Samothrace: lueurs sur la langue thrace et nouveau chapitre de la grammaire
compare? Acadmie des inscriptions et belles-lettres, pp. 121146.
2007. Les alphabets du Fayoum. Kadmos 46:1538.
Bron, Franois, and Andr Lemaire. 1989. Les inscriptions aramennes de Hazal. RA
83:3544.
Bibliography
345
Brown, Edwin L. 1977. The Divine Name Pan. Transactions of the American Philological
Association 107:5761.
Brown, Francis, S. R. Driver, and Charles A. Briggs. 1976. Hebrew and English Lexicon of the Old
Testament. Corrected edition. Oxford.
Brunot, Ferdinand, and Charles Bruneau. 1956. Prcis de grammaire historique de la langue
franaise. 4th ed. Paris.
Buck, Carl. 1933. Comparative Grammar of Greek and Latin. Chicago.
1955. The Greek Dialects. Chicago.
Buck, Carl, and Walter Petersen. 1949. A Reverse Index of Greek Nouns and Adjectives.
Chicago.
Buonamici, Giulio. 1932. Epigrafia Etrusca. Florence.
Burchfield, Robert. 2006. The English Language. With an introduction by Robert McCrum and
an afterword by John Simpson. London.
Burkert, Walter. 1985. Greek Religion. Cambridge, Massachusetts.
1992. The Orientalizing Revolution: Near Eastern Influence on Greek Culture in the Early
Archaic Period. Cambridge, Massachusetts.
2004. Babylon, Memphis, Persepolis: Eastern Contexts of Greek Culture. Cambridge,
Massachusetts.
Burrow, T. 1955. The Sanskrit Language. London.
Calame, Claude. 2003. Myth and History in Ancient Greece: The Symbolic Creation of a Colony.
Translated by Daniel W. Berman. Princeton.
Campanile, Enrico. 1990. La ricostruzione della cultura indoeuropea.Pisa.
Carpenter, Rhys. 1933. The Antiquity of the Greek Alphabet. AJA 37:829.
1958. Phoenicians in the West. AJA 62:3553.
Carroll, Lewis. 1866. Alices Adventures in Wonderland. London.
Cashford, Jules. 2003. The Homeric Hymns. London.
Cassio, Albio Cesare. 19911993[1994]. La pi antica iscrizione greca di Cuma e () in
Omero. Die Sprache 35:187207.
Catling, H. W. 19881989. Archaeology in Greece, 198889. AR 35:3116.
Caton, Richard. 1914. Notes on a Group of Medial and Surgical Instruments Found near
Kolophon. JHS 34:114118.
Chamoux, Franois. 1953. Cyrne. Paris.
Chantraine, Pierre. 1933. La formation des noms en grec ancien. Paris.
1968. Dictionnaire tymologique de la langue grecque: Histoire des mots. Paris.
1973. Grammaire homrique. Vol. 1. Paris.
1984. Morphologie historique du grec. 2nd ed. Paris.
Chomsky, Noam, and Morris Halle. 1968. The Sound Pattern of English. NewYork.
Clackson, James, and Birgit Anette Olsen, eds. 2004. Indo-European Word Formation.
Copenhagen.
Clauss, James Joseph. 1993. The Best of the Argonauts. Berkeley and Los Angeles.
Cobet, Carel Gabriel. 1861. Callimachea. Mnemosyne 10: 389437.
Coldstream, J. N. 1977. Geometric Greece. London.
Comrie, Bernard, Stephen Matthews, and Maria Polinsky, eds. 2003. The Atlas of Languages.
Rev. ed. London.
Coogan, Michael David. 1974. Alphabets and Elements. BASOR 216:6163.
Coote, Robert B. 1975. The Kition Bowl. BASOR 220:4750.
Cordano, Federica. 1984. Luso della scrittura in Italia meridionale e Sicilia nei secoli VIII e
VII a.C. Opus 3:281309.
346
Bibliography
Cowgill, Warren, and Manfred Mayrhofer. 1986. Indogermanische Grammatik. Heidelberg.

Creason, Stuart. 2004. Aramaic. In Woodard 2004, pp.391426.
Cross, Frank Moore. 1971. The Old Phoenician Inscription from Spain Dedicated to Hurrian
Astarte. Harvard Theological Review 64:189195.
1972. An Interpretation of the Nora Stone. BASOR 208:1319.
1974. Leaves from an Epigraphic Notebook. Catholic Bible Quarterly 36:490493.
1980. Newly Found Inscriptions in Old Canaanite and Early Phoenician Scripts. BASOR
238:120.
1989. The Invention and the Development of the Alphabet. In Senner 1989, pp.7790.
2003. Leaves from an Epigraphers Notebook: Collected Papers in Hebrew and West Semitic
Palaeography and Epigraphy. Winona Lake, Indiana.
Crothers, J. 1978. Typology and Universals of Vowel Systems. In Greenberg, Ferguson, and
Moravcsik 1978, pp.169172.
Dakaris, Sotirios. 1993. Dodona. Athens.
Dalley, Stephanie. 1985. Foreign Chariotry and Cavalry in the Armies of Tiglath-Pileser III
and Sargon II. Iraq 47:3148.
Daniels, Peter T., and William Bright, eds. 1996. The Worlds Writing Systems. Oxford.
Darmesteter, James. 1878. Iranica VI. Une mtaphore grammaticale de la langue Indoeuropenne. MSL 3:319321.
Darnell, John Coleman, F. W. Dobbs-Allsopp, Marilyn J. Lundberg, P. Kyle McCarter, Bruce
Zuckerman, and Colleen Manassa. 2005. Two Early Alphabetic Inscriptions from the Wadi
El-Hl: New Evidence for the Origin of the Alphabet from the Western Desert of Egypt.
Boston.
Daux, Georges. 1949. Un rglement cultuel dAndros. Hesperia 18:5872.
Davies, John. 2005. The Gortyn Laws. In Gagarin and Cohen 2005, pp.305327.
Davies, Morgan T. 1995. Aed ir coed i dorri cof : Dafydd ap Gwilym and the Metaphorics of
Carpentry. Cambrian Medieval Celtic Studies 30:6785.
Davies, W. V. 1990. Egyptian Hieroglyphs. In Hooker 1990, pp.74135.
Day, Joseph W. 1994. Interactive Offerings: Early Greek Dedicatory Epigrams and Ritual.
Harv. Stud. 96:3774.
DeForest, Mary Margolies. 1994. Apollonius Argonautica: A Callimachean Epic. Leiden.
Deger-Jalkotzy, Sigrid, and Oswald Panagl, eds. 2006. Die neuen Linear B-Texte aus Theben.
Vienna.
Demsky, Aaron. 1977. A Proto-Canaanite Abecedary Dating from the Period of the Judges and
Its Implications for the History of the Alphabet. Tel Aviv 4:1427.
Derrida, Jacques. 1967. De la grammatologie. Paris.
1968. Smiologie et grammatologie. Information sur les sciences sociales 7(3):135148.
Detienne, Marcel. 2003. The Writing of Orpheus: Greek Myth in Cultural Context. Translated by
Janet Lloyd. Baltimore and London.
Dietz, F. R. 1836. Severi iatrosophistae de clysteribus liber. Diss., Knigsberg.
Dimoudis, Nikolaos, and Apostolos Kyriatsoulis, eds. 1998. Die Geschichte der hellenischen
Sprache und Schrift vom 2. zum 1. Jahrtausend v.Chr.: Bruch oder Kontinuitt?. Altenburg.
Diringer, David. 1968. The Alphabet. NewYork.
Dobias-Lalou, Catherine. 1970. Pour une chronologie: Des inscriptions archaques de Cyrne.
Revue de philologie 44:228256.
Donner, H., and W. Rllig. 19661969. Kanaanische und aramische Inschriften. 2nd ed. 3vols.
Wiesbaden.
Bibliography
347
Drge, J.-P., M. Ishigami-Iagolnizter, and M. Cohen. 1986. Le livre et limprimerie en ExtrmeOrient et en Asie du Sud. Bordeaux.
Driver, G. R. 1976. Semitic Writing. Rev. ed. Oxford.
Dubois, Laurent. 1995. Inscriptions grecques dialectales de Grande Grce I Colonies eubennes.
Colonies ioniennes. Emporia. Geneva.
Dbner, Friedrich. 1969. Scholia Graeca in Aristophanem. Reprint edition. Hildesheim.
Dumzil, Georges. 1970. The Destiny of the Warrior. Translated by Alf Hiltebeitel. Chicago.
Dussaud, Ren. 1924. Les inscriptions phniciennes du tombeau dAhiram roi de Byblos.
Syria 5:135159.
Eidinow, Esther. 2007. Oracles, Curses, and Risk among the Ancient Greeks. Oxford.
Einarson, Benedict. 1967a. Notes on the Development of the Greek Alphabet. C Phil.
62:124.
1967b. Corrigenda and Addenda: Notes on the Development of the Greek Alphabet. C Phil.
62:262263.
Emmel, Stephen. 1980. A Fragment of Nag Hammadi Codex III in the Beinecke Library: Yale
Inv. 1784. BASP 17:5360.
Ephal, Israel, and Joseph Naveh. 1989. Hazaels Booty Inscriptions. Israel Exploration Journal
34:192200.
Ernout, Alfred, and Antoine Meillet. 1959. Dictionnaire tymologique de la langue latine:
Histoire des mots. 4th ed. Paris.
Evelyn-White, Hugh G. 1936. Hesiod, Homeric Hymns, Epic Cycle, Homerica. Rev. ed. Loeb
Classical Library. Cambridge, Massachusetts.
Faarlund, Jan Terje. 2004. Ancient Nordic. In Woodard 2004, pp.907921.
Farnell, Lewis Richard. 1907. The Cults of the Greek States. 5vols. Oxford.
Ferber, Michael. 1999. A Dictionary of Literary Symbols. Cambridge.
Finley, M. I. 2002. The World of Odysseus. 2nd rev. ed. with emendations. London.
Fortenbaugh, William W. 2005. Theophrastus of Eresus: Sources for His Life, Writings, Thought
and Influence. Commentary, vol. 8: Sources on Rhetoric and Poetics (Texts 666713).
Leiden.
Fortson, Benjamin W., IV. 2004. Indo-European Language and Culture. Oxford.
Foster, Benjamin R. 2005. Before the Muses: An Anthology of Akkadian Literature. Bethesda,
Maryland.
Frame, Douglas. 1978. The Myth of Return in Early Greek Epic. New Haven, Connecticut.
Frazer, J. G. 1965. Pausaniass Description of Greece. 6vols. Reprint edition. NewYork.
Gagarin, Michael, and David Cohen, eds. 2005. The Cambridge Companion to Ancient Greek
Law. Cambridge.
Gamkrelidze, Thomas V., and Vjaeslav V. Ivanov. 1995. Indo-European and the Indo-Europeans.
Translated by Johanna Nichols. Berlin.
Gardiner, Alan. 1916. The Egyptian Origin of the Semitic Alphabet. Journal of Egyptian
Archaeology 3:116.
Gelb, I. J. 1963. A Study of Writing. 2nd ed. Chicago.
Gettens, Rutherford J. 1963. The Corrosion Products of Metal Antiquities. Annual Report of
the Board of Regents of the Smithsonian Institution for 1963, pp. 547568.
Gibson, John C. L. 1982. Phoenician Inscriptions. Vol. 3 of Textbook of Syrian Semitic Inscriptions.
Oxford.
Gillard, R. D., S. M. Hardman, R. G. Thomas, and D. E. Watkinson. 1994. The Mineralization
of Fibres in Burial Environments. Studies in Conservation 39:32140.
348
Bibliography
Gow, A. S. F., and A. F. Scholfield. 1988. Nicander: The Poems and the Poetical Fragments.
Reprint edition. Salem, New Hampshire.
Gragg, Gene. 2004. Geez (Aksum). In Woodard 2004, pp.427453.
Graham, A. J. 1982. The Colonial Expansion of Greece. In Boardman and Hammond 1982,
pp.83159.
Greenberg, Joseph H., Charles A. Ferguson, and Edith A. Moravcsik, eds. 1978. Universals of
Human Language. Vol. 1: Theory and Methodology. Stanford.
Greenough, J. B. 1890. Some Latin Etymologies. Harv. Stud. 1:93105.
Guarducci, Margherita. 19351950. Inscriptiones Creticae. 4vols.Rome.
1967. Epigrafia Greca.Rome.
1970. Cibele in unepigrafe arcaica di Locri Epizefiri. Klio 52:133138.
1985. Una nuova dea a Naxos in Sicilia e gli antichi legami fra la Naxos siceliota e lomonima
isola delle Cicladi. Mlanges darchologie et dhistoire de lcole franaise de Rome
97:734.
Hackett, Jo Ann. 2004. Phoenician and Punic. In Woodard 2004, pp.365385.
Hgg, Robin, ed. 1983. The Greek Renaissance of the Eighth Century B.C.: Tradition and
Innovation. Stockholm.
Hall, Jonathan. 1995. How Argive Was the Argive Heraion? The Political and Cultic Geography
of the Argive Plain, 900400 B.C. AJA 99:577613.
Hammond, N. G. L. 1967. Epirus. Oxford.
Hamp, Eric P. 1985. Some Archaic Greek Compounds. AJP 106:222225.
Hansen, P. A. 19831989. Carmina epigraphica graeca. 2vols. Berlin.
Harris, Roy. 1987. Reading Saussure. La Salle, Illinois.
Havelock, Eric A. 1982. The Literate Revolution in Greece and Its Cultural Consequences.
Princeton.
Healey, John F. 1990. The Early Alphabet. In Hooker 1990, pp.196257.
Henrichs, Albert. 2003. Writing Religion. In Yunis 2003, pp.3858.
Herter, H. 1934. Telchinen. RE 5A:197224.
Heubeck, Alfred. 1979. Schrift. Gttingen.
1986. Die Wrzburger Alphabettafel. Wrzburger Jahrbcher fr die Altertumswissenschaft
Neue Folge 12:720.
Heubeck, Alfred, and Arie Hoekstra. 1989. A Commentary on Homers Odyssey. Vol. 2.
Oxford.
Heubeck, Alfred, Stephanie West, and J. B. Hainsworth. 1988. A Commentary on Homers
Odyssey. Vol. 1. Oxford.
Holloway, R. Ross. 1991. The Archaeology of Ancient Sicily. London.
1994. The Archeology of Early Rome and Latium. London.
Holy Bible: New Revised Standard Version. 1998. Oxford.
Honeigswald, Henry, Roger D. Woodard, and J. P. T. Clackson. 2004. Indo-European. In
Woodard 2004, pp.534550.
Hooker, J. T., ed. 1990. Reading the Past. Berkeley and Los Angeles.
Hordern, J. H. 2004. Sophrons Mimes: Text, Translation, and Commentary. Oxford: .
Hunter, Richard, ed. 2005. The Hesiodic Catalogue of Women. Cambridge.
Immerwahr, Henry. 1990. Attic Script: A Survey. Oxford.
Jacoby, Felix. 19541969. Die Fragmente der griechischen Historiker. Reprint edition. Leiden.
Jacquemin, Anne. 1999. Offrandes monumentales Delphes. Athens.
Jakobson, Roman. 1968. Child Language, Aphasia, and Phonological Universals. The Hague.
Bibliography
349
1990. On Language. Edited by Linda R. Waugh and Monique Monville-Burston. Cambridge,

Massachusetts.
Jakobson, Roman, and Morris Halle. 1956. Fundamentals of Language. The Hague.
Jakobson, Roman, and Linda Waugh. 1979. The Sound Shape of Language. Bloomington,
Indiana.
Jasanoff, Jay. 1978. A Note on Hittite tya- steal. MSS 37:9192.
Jean, Georges. 1992. Writing. Translated by Jenny Oates. NewYork.
Jeffery, Lillian. 1961. The Local Scripts of Archaic Greece. Oxford.
1978. Review of Merle K. Langdon, A Sanctuary of Zeus on Mount Hymettos (Princeton,
1976). JHS 98:202203.
1990. The Local Scripts of Archaic Greece. Rev. ed. Supplement by A. Johnston. Oxford.
Jeffery, Lillian, and Anna Morpurgo Davies. 1970. and : A New
Archaic Inscription from Crete. Kadmos 9:118154.
Johnston, Alan. 1983. The Extent and Use of Literacy; the Archaeological Evidence. In Hgg
1983, pp.6368.
1998. Epichoric Alphabets: The Rise of the Polis or a Slip of the Pen. In Dimoudis and
Kyriatsoulis 1998, pp.419431.
2003. The Alphabet. In Stampolidis and Karageorghis 2003, pp.263276.
Johnston, Sarah Iles. 2002. Myth, Festival, and Poet: The Homeric Hymn to Hermes and Its
Performative Context. C Phil. 97:109132.
Johnstone, W. 1978. Cursive Phoenician and the Archaic Greek Alphabet. Kadmos
17:151166.
Jones, D. M., and N. G. Wilson. 1969. Prolegomena de comoedia. Scholia in Acharnenses, Equites,
Nubes. Groningen.
Kahn, David. 1996. The Codebreakers: The Story of Secret Writing. NewYork.
Kaufman, Stephen A. 1984. Review of A. Abou-Assaf, P. Bordreuil, and A. Millard, La statue de
Tell Fekherye et son inscription bilingue assyro-aramenne. JAOS 104:571573.
Kenzelmann-Pfyffer, Anne, Thierry Theurillat, and Samuel Verdan. 2005. Graffiti dpoque
gomtrique provenant du sanctuaire dApollon Daphnphoros Ertrie. ZPE 151:5183.
Kerkhecker, Arnd. 1999. Callimachus Book of Iambi. Oxford.
Kirchhoff, A. 1887. Studien zur Geschichte des griechischen Alphabets. 3rd ed. Gtersloh.
Kirk, G. S. 2000. The Iliad: A Commentary. Vol. 2. Reprint edition. Cambridge.
Kock, T. 1880. Comicorum atticorum fragmenta. Vol. 1. Leipzig.
Khler, Wolfgang. 1929. Gestalt Psychology. NewYork.
Kraus, H. P. 1978. A Rare Book Saga: The Autobiography of H. P. Kraus. NewYork.
Khn, Karl Gottlob. 19641965. Medicorum graecorum opera. Hildesheim.
Kyrieleis, Helmut. 1993. The Heraion at Samos. In Marinatos and Hgg 1993, pp.125153.
Lang, Mable. 1976. The Athenian Agora. Vol. XXI: Graffiti and Dipinti. Princeton.
1982. Writing and Spelling on Ostraka. In Studies in Attic Epigraphy, History and Topography.
Presented to Eugene Vanderpool, pp.7587. Hesperia Supplement 19. Athens.
Langdon, Merle K. 1976. A Sanctuary of Zeus on Mt. Hymettos. Hesperia Supplement 16. Athens.
2005. A New Greek Abecedarium. Kadmos 44:175182.
Lefkowitz, Mary R. 1980. The Quarrel between Callimachus and Apollonius. ZPE 40:119.
Lehmann, Winfred P. 1986. A Gothic Etymological Dictionary. Leiden.
Lehmann-Haupt, Hellmut, ed. 1967. Homage to a Bookman: Essays on Manuscripts, Books and
Printing Written for Hans P. Kraus on His 60th Birthday Oct. 12, 1967. Berlin.
Lejeune, Michel. 1939. Les adverbes grecs en -. Paris.
350
Bibliography
1982. Phontique historique du mycnien et du grec ancien. Paris.

Lerat, L. 1944. Ddicace archaque de Delphes. Revue archologique 21:514.
Lesky, Albin. 1996. A History of Greek Literature. Translated by James Willis and Cornelis de
Heer. Reprint edition. London.
Levi, D. 1969. Un pithos iscritto da Fests. 21:153176.
Liddell, H. G., R. Scott, and H. S. Jones (with the assistance of R. McKenzie). 1996. A GreekEnglish Lexicon. Rev. ed. Supplement by P. G. W. Glare (with the assistance of A. A.
Thompson). Oxford.
Lin Guojun, and Zheng Rusi. 1989. Lhistoire du livre en Chine. Translated by A.-M. Harvey and
O. Pasteur. Beijing.
Liritzis, Veronica McGeehan. 1996. The Role and Development of Metallurgy in the Late Neolithic
and Early Bronze Age of Greece. Jonsered, Sweden.
Lloyd-Jones, Hugh. 1996. Sophocles: Fragments. Loeb Classical Library. Cambridge, Massachusetts.
Lobel, Edgar, and Denys Page. 1955. Poetarum lesbiorum fragmenta. Oxford.
Loprieno, Antonio. 2004. Ancient Egyptian and Coptic. In Woodard 2004, pp.160217.
Lord, Albert B. 1995. The Singer Resumes the Tale. Edited by M. L. Lord. Ithaca, NewYork.
Luraghi, Nino. 2006. Traders, Pirates, Warriors: The Proto-History of Greek Mercenary
Soldiers in the Eastern Mediterranean. Phoenix 60:2147.
Macalister, R. A. S. 1906. Review of W. M. F. Petrie, Researches in Sinai (London, 1906). Palestine
Exploration Fund Quarterly Statement 39:318321.
Macdonald, M. C. A. 2004. Ancient North Arabian. In Woodard 2004, pp.488533.
Macfarquhar, C. F. 1966. Early Greek Travellers in Egypt. Greece and Rome 13:108116.
Maddieson, Ian. 1984. Patterns of Sounds. With a chapter contributed by Sandra Ferrari Disner.
Cambridge.
Mallory, J. P., and D. Q. Adams, eds. 1997. Encyclopedia of Indo-European Culture. London and
Chicago.
Marinatos, Nanno, and Robin Hgg, eds. 1993. Greek Sanctuaries: New Approaches. London
and NewYork.
Martin, Richard. 1989. The Language of Heroes: Speech and Performance in the Iliad. Ithaca,
NewYork.
Masson, Olivier. 1976. La plus ancienne inscription crtoise. In Morpurgo Davies and Meid
1976, pp.169172.
1983. Les inscriptions chypriotes syllabiques. Paris.
Matthaiou, Angelos P. 1984. . Horos 2:167173.
McCarter, P. Kyle. 1975a. The Antiquity of the Greek Alphabet and the Early Phoenician Scripts.
Harvard Semitic Monographs 9. Missoula, Montana.
1975b. A Phoenician Graffito from Pithekoussai. AJA 79:140141.
Megaw, A. H. S. 19631964. Archaeology in Greece, 196364. AR 10:330.
Meier-Brgger, Michael. 1992. Griechische Sprachwissenschaft II: Wortschatz; Formenlehre;
Lautlehre. Berlin.
2003. Indo-European Linguistics. With contributions by Matthias Fritz and Manfred
Mayrhofer. Translated by Charles Gertmenian. Berlin.
Meiggs, Russell, and David Lewis. 1980. A Selection of Greek Historical Inscriptions to the End of
the Fifth Century B.C. Reprint edition. Oxford.
Melchert, H. Craig. 1984. Studies in Hittite Historical Phonology. Gttingen.
1994. Anatolian Historical Phonology. Amsterdam.
2004. Carian. In Woodard 2004, pp.609613.
Meritt, Benjamin D. 1945. The Argives at Tanagra. Hesperia 14:134147.
Bibliography
351
1952. Greek Inscriptions. Hesperia 21:340380.

Metzger, Henri . 1965. Sur la date du graffite de la coupe de Nestor. Revue des tudes anciennes
67:301305.
Meyers, Eric M., ed. 1997. The Oxford Encyclopedia of Archaeology in the Near East. New York
and Oxford.
Mitchell, Lynette G. 2001. Euboean Io. CQ 51:339352.
Mller, Astrid. 2001. Naukratis: Trade in Archaic Greece. Oxford.
Morgan, Kathryn A. 1993. Pindar the Professional and the Rhetoric of . C Phil.
88:115.
Morpurgo Davies, Anna, and Wolfgang Meid, eds. 1976. Studies in Greek, Italic and IndoEuropean Linguistics Offered to Leonard R. Palmer. Innsbruck.
Morris, Sarah P. 1992. Daidalos and the Origins of Greek Art. Princeton.
Moscati, Sabatino. 1999. The Phoenicians. NewYork.
Mller, Karl, and Theodorus Mller. 18411870. Fragmenta historicorum graecorum. Paris.
Mykytiuk, Lawrence. 2004. Identifying Biblical Persons in Northwest Semitic Inscriptions of
1200539 B.C.E. Atlanta.
Nagy, Gregory. 1990. Greek Mythology and Poetics. Ithaca, NewYork.
1996. Homeric Questions. Austin, Texas.
2002. Platos Rhapsody and Homers Music. Washington, D.C., and Athens.
2007. Lyric and Greek Myth. In Woodard 2007a, pp.1951.
2008. Greek: An Updating of a Survey of Recent Work. http://chs.harvard.edu/wa/pageR?tn=
ArticleWrapper&bdc=12&mn=2278#noteref_n.267.
Naveh, Joseph. 1973. Some Semitic Epigraphical Considerations on the Antiquity of the Greek
Alphabet. AJA 77:18.
1980. The Greek Alphabet: New Evidence. Biblical Archaeologist 43:2225.
1982. Early History of the Alphabet. Jerusalem.
1987. Early History of the Alphabet. 2nd rev. ed. Jerusalem.
1991. Semitic Epigraphy and the Antiquity of the Greek Alphabet. Kadmos 30:143152.
Nebes, Norbert, and Peter Stein. 2004. Ancient South Arabian. In Woodard 2004,
pp.454487.
Nicol, D. M. 1958. The Oracle of Dodona. Greece and Rome 5:128143.
Nicole, J., ed. 1966. Les scolies genevoises de lIliade. 2vols. Reprint edition. Hildesheim.
Nielsen, Benedicte. 2004. On Latin Instrument-Nouns in */-lo-/. In Clackson and Olsen
2004, pp.189213.
OBrien, Joan V. 1993. The Transformation of Hera. Lanham, Maryland.
Olivieri, Alexander. 1947. Frammenti della Commedia Greca e del Mimo nella Sicilia e nelle
Magna Grecia. Vol. 2. Naples.
Osborne, Robin. 2005. Ordering Women in Hesiods Catalogue. In Hunter 2005, pp.524.
Page, R. I. 1987. Runes. Berkeley and Los Angeles.
Palaima, Thomas G. 20002001. Review of Vassilis L. Aravantinos, Louis Godart, and Anna
Sacconi, eds., Les tablettes en linaire B de la Odos Pelopidou. dition et commentaire
(Thbes fouilles de la Cadme 1) (Pisa and Rome, 2001). Minos 3536:475486.
2003. Review of Vassilis L. Aravantinos, Louis Godart, and Anna Sacconi, eds., Les tablettes
en linaire B de la Odos Pelopidou. dition et commentaire (Thbes fouilles de la Cadme 1)
(Pisa and Rome, 2001). AJA 107:113115.
Palmer, L. R. 1980. The Greek Language. Atlantic Highlands, New Jersey.
Pandolfini, Maristella, and Aldo Prosdocimi. 1990. Alfabetari e insegnamento della scrittura in
Etruria e nellItalia antica. Florence.
352
Bibliography
Papadopoulos, John K. 2001. Magna Achaea: Akhaian Late Geometric and Archaic Pottery in
South Italy and Sicily. Hesperia 70:373460.
Pardee, Dennis. 1997a. Proto-Sinaitic. In Meyers 1997, 4:354355.
1997b. Proto-Canaanite. In Meyers 1997, 4:352354.
1997c. Alphabet. In Meyers 1997, 1:7579.
2004a. Canaanite Dialects. In Woodard 2004, pp.386390.
2004b. Ugaritic. In Woodard 2004, pp.288318.
Parker, Robert. 2008. Ignorance and Knowledge of the Divine. Paper presented at the conference on Revelation in Ancient Greek Religion, University of Illinois at Chicago.
Pedersen, Holger. 1931. The Discovery of Language: Linguistic Science in the Nineteenth Century.
Translated by John Webster Spargo. Cambridge, Massachusetts.
Perruzi, Emilio. 1973. Origini di Roma II. Bologna.
Pintaudi, Rosario, ed. 2005. Papyri graecae Schyen. Florence.
Powell, Barry. 1991. Homer and the Origin of the Alphabet. Cambridge.
Ramat, Anna Giacalone, and Paolo Ramat, eds. 1998. The Indo-European Languages. London.
Rau, Jeremy. 2009. Indo-European Nominal Morphology: The Decads and the Caland System.
Innsbruck.
Raubitschek, A. E. 1949. Dedications from the Athenian Acropolis: A Catalogue of the Inscriptions
of the Sixth and Fifth Centuries B.C. With the collaboration of Lillian H. Jeffery. Cambridge,
Massachusetts.
Ray, John. 2007. The Rosetta Stone and the Rebirth of Ancient Egypt. Cambridge,
Massachusetts.
Ribezzo, F. 1920. Le iscrizioni greco-arcaiche di Cuma. Rivista indo-greco-italica di filologia,
lingua, antichit. 3:7187.
Richardson, Rufus B., and James R. Wheeler. 1902. Inscriptions from the Argive Heraeum. In
Waldstein 1902, pp.197224.
Ridgeway, David. 2004. Phoenicians and Greeks in the West: A View from Pithekoussai. In
Tsetskhladze and De Angelis 2004, pp.3546.
Risch, C. J. 1937. Wortbildung der homerischen Sprache. Berlin.
Ritner, Robert. 1996. Egyptian Writing. In Daniels and Bright 1996, pp.7384.
Rix, Helmut. 1970. Anlautender Laryngal vor Liquida oder Nasalis sonans im Griechischen.
MSS 27:79110.
1976. Historische Grammatik des Griechischen: Laut- und Formenlehre. Darmstadt.
1985. Das letzte Wort der Duenos-Inschrift. Mnchener Studien zur Sprachwissenschaft
46:193220.
2001. Lexikon der indogermanischen Verben. Wiesbaden.
Robinson, T. M. 2003. Heraclitus. Reprint edition. Toronto.
Roehl, Hermann. 1882. Inscriptiones graecae antiquissimae. Berlin.
Russo, Joseph, Manuel Fernndez-Galiano, and Alfred Heubeck. 1992. A Commentary on
Homers Odyssey. Vol. 3. Oxford.
Sapir, Edward. 1921. Language: An Introduction to the Study of Speech. NewYork.
1949. Selected Writings. Edited by D. Mandelbaum. Berkeley.
Sass, Benjamin. 1988. The Genesis of the Alphabet and Its Development in the Second Millennium
B.C. Wiesbaden.
2005. The Alphabet at the Turn of the Millennium. TelAviv.
Saussure, Ferdinand de. 1986. Course in General Linguistics. Translated by Roy Harris. La Salle,
Illinois.
1995. Cours de linguistique gnrale. dition critique prpare par Tullio de Mauro. Paris.
Bibliography
353
Schaeffer, Claude F.-A., and Jean Nougayrol. 1955. Le Palais royal dUgarit. Paris.
Scheid, John, and Jesper Svenbro. 1996. The Craft of Zeus: Myths of Weaving and Fabric.
Translated by Carol Volk. Cambridge, Massachusetts.
Schmitt, Rdiger. 1967. Dichtung und Dichtersprache in Indogermanischer Zeit. Wiesbaden.
Schwyzer, E. 1939. Griechische Grammatik. Bd. 1. Munich.
Scott, David A. 1991. Metallography and Microstructure of Ancient and Historic Metals. Los
Angeles.
2002. Copper and Bronze in Art: Corrosion, Colorants, Conservation. Los Angeles.
Scott, David A., and Lynn Schwartz Dodd. 2003. Technical Examination of an Egyptian Gilded
Bronze Figurine of the God Osiris. Journal of Cultural Heritage 3:333345.
Scott, D. A., R. D. Woodard, P. K. McCarter, B. Zuckerman, and M. Lundberg. 2005. Greek
Alphabet (MS 108). In Pintaudi 2005, pp.149160.
Segre, Mario. 1952. Tituli Calymnii. Bergamo.
Senner, Wayne M., ed. 1989. The Origins of Writing. Lincoln, Nebraska.
Shear, Ione Mylonas. 1998. Bellerophon Tablets from the Mycenaean World? A Tale of Seven
Bronze Hinges. JHS 118:187189.
Shear, T. Leslie. 1940. The Campaign of 1940. Hesperia 10:18.
Sherratt, Susan. 2003. Questions of Script and Identity in Early Iron Age Greece and Cyprus.
Oxford Journal of Archaeology 22:225242.
Sirat, Colette. 1994. Handwriting and the Writing Hand. In Watt 1994, pp.375460.
Smyth, Herbert. 1898. The Greek Dialects: Ionic. Oxford.
Snyder, Jane McIntosh. 1981. The Web of Song: Weaving Imagery in Homer and the Lyric
Poets. CJ 76:193196.
Sol-Sol, J. M. 1966. Nueva inscripcin fenicia de Espana (Hispania 14). Rivista degli studi
orientali 41:97108.
Sol, Robert, Dominique Valbelle, and W. V. Davies. 2006. The Rosetta Stone: The Decipherment
of the Hieroglyphs. With a preface by Neil MacGregor. London.
Sommerstein, Alan H. 1994. Aristophanes: Thesmophoriazusae. Warminster.
Sthlin, Otto. 1972. Clemens Alexandrinus: Protrepticus und Paedagogus. Berlin.
Stampolidis, Nicholas, and Vassos Karageorghis, eds. 2003. Sea Routes: Interconnections in the
Mediterranean, 16th 6th c. BC. Athens.
Steiner, Richard C. 1996. The Two Sons of Neriah and the Two Editions of Jeremiah in the
Light of Two Atbash Code-Words for Babylon. Vetus Testamentum 46:7484.
Stillwell, Agnes Newhall, and J. L. Benson. 1984. The Potters Quarter: The Pottery. With contributions by Alan L. Boegehold and Cedric G. Boulter. Corinth, vol. 15, part III. Princeton.
Stumpf, Carl. 1907. Zur Einteilung der Wissenschaften. Abhandlungen der Kniglichen
Preussischen Akademie der Wissenschaften, pp.61ff. Berlin.
Suka, Roman, ed. 2012. The Sounds of Indo-European 2: Phonetics, Phonemics, and
Morphophonemics. Lincom Europa. Munich.
Szemernyi, Oswald J. L. 1996. Introduction to Indo-European Linguistics. Oxford.
Tadmor, Hayim, and Moshe Weinfeld, eds. 1983. History, Historiography and Interpretation.
Jerusalem.
Teixidor, Javier. 1976. The Phoenician Inscriptions of the Cesnola Collection. Metropolitan
Museum Journal 11:5570.
Thomas, Rosalind. 1992. Literacy and Orality in Ancient Greece. Cambridge.
Thumb, Albert, and E. Kiekers. 1932. Handbuch der griechischen Dialekte. Erster Teil. Heidelberg.
Thumb, Albert, and A. Scherer. 1959. Handbuch der griechischen Dialekte. Zweiter Teil.
Heidelberg.
354
Bibliography
Tigay, J. H. 1983. An Early Technique of Aggadic Exegesis. In Tadmor and Weinfeld 1983,
pp.169189.
Tracy, Stephen V. 1986. An Early Inscribed Gold Ring from the Argolid. JHS 106:196.
Traill, David A. 1998. Callimachus Singing Sea (Hymn 2.106). C Phil. 93:215222.
Tsagalis, Christos C. 2008. Inscribing Sorrow: Fourth-Century Attic Funerary Epigrams. Berlin.
Tsetskhladze, Gocha R., and Franco De Angelis, eds. 2004. The Archaeology of Greek
Colonisation: Essays Dedicated to Sir John Boardman. Revised edition with a new preface.
Oxford.
Tylecote, Ronald F. 1976. A History of Metallurgy. London.
Tzanetou, Angeliki. 2002. Something to Do with Demeter: Ritual and Performance in
Aristophanes Women at the Thesmophoria. AJP 123:329367.
Ventris, Michael, and John Chadwick. 1973. Documents in Mycenaean Greek. 2nd ed.
Cambridge.
Vernant, Jean-Pierre. 2006. Myth and Thought among the Greeks. Translated by Janet Lloyd
with Jeff Fort. Reprint edition. NewYork.
Vollgraff, Wilhelm. 1910. Inscription dArgos. BCH 34:33154.
1913. Inscription dArgos. BCH 37:279309.
1948. Le Dcret dArgos relatif un pacte entre Knossos et Tylissos. Amsterdam.
Wachter, Rudolf. 2001. Non-Attic Greek Vase Inscriptions. Oxford.
2005. Annex zu Fragment 3 der Graffiti von Eretria. ZPE 151:8486.
Wackernagel, Jacob. 1897. Vermischte Beitrge zur griechischen Sprachkunde. Programm zur
Rektoratsfeier der Universitt Basel, pp.362.
Walde, Alois, and Julius Pokorny. 19271930. Vergleichendes Wrterbuch der indogermanischen
Sprachen. Berlin and Leipzig.
Waldstein, Charles, ed. 1902. The Argive Heraeum. Vol. 1. Cambridge, Massachusetts.
Watkins, Calvert. 1976. Observations on the Nestors Cup Inscription. Harv. Stud. 80:2540.
1995a. Greece in Rome: Influence, Integration, Resistance. Harv. Stud. 97:3550.
1995b. How to Kill a Dragon: Aspects of Indo-European Poetics. Oxford.
1998. Proto-Indo-European: Comparison and Reconstruction. In Ramat and Ramat 1998,
pp.2573.
Watt, W. C. 1987. The Byblos Matrix. Journal of Near Eastern Studies 46:114.
1989. The Ras Shamra Matrix. Semiotica 74:61108.
ed. 1994. Writing Systems and Cognition: Perspectives from Psychology, Physiology, Linguistics,
and Semiotics. Dordrecht, Boston, and London.
West, M. L. 1978. Hesiod: Works and Days. Oxford.
1985. The Hesiodic Catalogue of Women. Oxford.
1997. The East Face of Helicon: West Asiatic Elements in Greek Poetry and Myth. Oxford.
2003. Homeric Hymns; Homeric Apocrypha; Lives of Homer. Loeb Classical Library.
Cambridge, Massachusetts.
Whitley, James. 1997. Cretan Laws and Cretan Literacy. AJA 101:635661.
Whitney, William Dwight. 1867. Language and the Study of Language. London.
Williams, Ifor. 1944. Lectures on Early Welsh Poetry. Dublin.
Williams, Ifor, and Thomas Roberts, eds. 1935. Cywyddau Dafydd ap Gwilym ai Gyfoeswyr.
Cardiff.
Woodard, Roger D. 1997. Greek Writing from Knossos to Homer. Oxford.
2003. Writing Systems. In Comrie, Matthews, and Polinsky 2003, pp.160207.
ed. 2004. The Cambridge Encyclopedia of the Worlds Ancient Languages. Cambridge.
2006. Indo-European Sacred Space. Urbana and Chicago.
Bibliography
355
ed. 2007a. The Cambridge Companion to Greek Mythology. Cambridge.

2007b. Hesiod and Greek Myth. In Woodard 2007a, pp.83165.
2010. Phoinikia Grammata: An Alphabet for the Greek Language. In Bakker 2010,
pp.2546.
2012. Labiovelar Development in Greek and an Alphabetic Repercussion. In Suka 2012,
pp.299313.
Yunis, Harvey, ed. 2003. Written Texts and the Rise of Literate Culture in Ancient Greece.
Cambridge.
Index
Abaza,324
abecedrius,221
Abou Simbel,18
Achaia, 29, 73, 250,303
Achilles, 107, 186, 192, 213,334
acrophonic principle, 157,160
adaptation, 51, 63, 64, 65, 66, 146, 151, 152, 153,
156, 157, 158, 159, 160, 163, 168, 173, 254,
258, 259, 263, 303,327
Admete,198
Aegina,211
Aelian,196
Aeneas Tacticus,186
Aeschylus, 185, 213, 274, 281, 329,341
Aethlius,202
agalma, 93, 199, 200, 203, 204, 205, 213, 223,
224, 329,339
Agamemnon, 183, 273, 332,341
agnos, 195, 196, 197, 202, 219,220
agnus castus. Seeagnos
Ahiram sarcophagus, 30,305
Aigina, 18, 19, 73, 101, 199, 205,308
Aiolic, 37, 191,297
Ajax, 207,208
Akkadian, 259,279
albam, 249, 251,338
Albanian,269
Alcaeus, xi, 184,331
Alcman,287
Alcmena,272
alep, 16, 22, 23, 24, 166, 249,338
Alexandria, 4, 186, 203, 211,331

alpha, 2, 3, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 28, 36, 40, 42, 70, 74, 135, 137, 138, 140,
144, 163, 166, 171, 224, 235, 247, 250, 251,
286, 292, 293, 294,316
alphabetic anomalies, 119, 120, 127, 128, 129,
131, 133, 135, 137, 178, 262,339
alphabetic fabric, 235, 238, 253,287
alphabetic weaving, 3, 4, 223, 234, 235, 237,
238, 243, 244, 245, 247, 249, 252, 253, 256,
259, 263, 264, 265, 272,288
Althaea,280
ambiguity, 59, 65, 66, 96, 147, 149, 150, 186,
187,316
Ammon, 214, 215,216
Ammonites,246
Amorgos, 28, 29, 33, 39, 80, 85, 102, 207,
294,297
Amphidromia,186
Amphipolis, 21,211
Amphitryon, 272,273
Amyntor,287
Anacreon, 195, 196,268
Anaphe, 47, 73,97
Anatolia,163
Anatolian,330
Ancient North Arabian,323
Andros,39
Ankylion, gravestone of, 47, 73,97
Antioch, 201,202
Antiphus,192
357
358
Antyllus, 180,181
aphasia, 170,171
Aphrodite, 40, 88,231
Apollo, 18, 25, 52, 71, 99, 101, 187, 188, 199, 201,
208, 212, 219, 274, 275, 276, 308, 329,332
Apollo Daphnephoros, 42, 45,300
Apollo Milesios,18
Apollo Ptoios,85
Apollo Pythios, 71, 77, 85, 98,306
Apollo Telchinius,210
Apollonion,52
Apollonius of Rhodes,330
Apollonius of Tyana,201
Arabia,246
Arabic, 6, 7, 323,327
Aramaic, 36, 58, 212, 259, 296, 299, 319,327
arbitrariness, 137, 138, 140, 141, 142, 143, 145,
146, 147, 148, 150, 151, 152, 153, 154, 156,
160, 161, 172, 174, 176, 177, 178, 181, 228,
269, 317, 318, 321,322
Archilochus,268
Ares, 40,186
Argive, 33, 43, 45, 50, 51, 52, 53, 62, 93, 94, 99,
150, 198, 258, 260, 266, 296, 298, 299,311
Argos, 29, 33, 37, 43, 50, 51, 88, 99, 198,260
Aristophanes, 179, 182, 183, 184, 222, 223,
227,266
Aristotle, 239,337
Arkadia, 27, 30, 101,200
Arkadian, 144,207
Armenian, 163,328
Artemis, 203, 276,332
Ashdod,246
Ashkelon,246
Asia Minor, 29, 44,73
assimilation, 52, 225, 228, 255, 257, 263,339
associative structure, 169, 170, 171, 235, 239,
263,264
Astarte, 22, 23,76
Astypalaea, 207,333
atacamite, 115,117
atbah,338
atbash, 248, 249, 251, 252, 253, 254, 255, 256,
257, 258, 259, 261, 263, 264,338
Athena, 3, 18, 187, 201, 231, 243, 329,333
Athenaeus, 196,198
Athens, 16, 17, 22, 23, 29, 43, 55, 58, 70, 85, 98,
105, 182, 196, 201, 202, 211, 229, 267,331
Athenian Acropolis, 69, 88, 104, 200, 201,332
Index
Athenian Agora, 17, 25, 97, 104, 296,301
Atiya, Aziz, 3, 4, 5, 6, 7, 8,291
Attic, 17, 18, 20, 25, 27, 51, 55, 58, 59, 63, 73, 74,
83, 84, 95, 97, 101, 191, 265, 271, 293, 295,
301, 303, 305, 306,318
Attica, 24, 29, 30, 32, 33, 37, 58, 70, 74, 75, 98,
101, 260, 267, 288,299
Aulis,40
Avestan, 230, 243, 269,335
Axos,34
ayin, 48, 50, 258, 259, 299,322
azurite, 110, 111,118
Bal Lebanon inscriptions, 48, 54, 78,105
Baalat,160
Babylon, 246, 247,248
Bacchylides, 195, 228, 229, 232, 233, 234, 239,
276,331
Baltic,198
Barako abecedarium, 55, 260,301
bards, 206, 232, 233, 288,289
Beinecke, EdwinJ.,7
Bellerophon, 15, 183, 266, 270,271
Berthelot, Pierre Eugne Marcellin,
176,328
bet, 247, 248, 249,338
beta, 24, 25, 26, 42, 55, 120, 135, 137, 138, 140,
224, 247, 286, 287,294
Beth-shemesh ostracon,24
biconsonantal symbols, 156,322
binocular bench microscopy, 2, 107,110
Bloomfield, Leonard,316
Bodmer, Martin,7
Boiotia, 29, 32, 37, 43, 54, 70, 75, 85, 98, 101,
102, 251, 262, 295, 305,307
Boiotian, 19, 25, 27, 32, 40, 44, 47, 50, 53, 85,
86, 88, 91, 222, 258, 260, 300, 307,308
Book of the Dead,7
Bopp, Franz, 175, 176,328
boustrophedon, xii, 18, 38, 69, 88,286
Bral, Michel,316
bucchero, 20, 42, 70, 77, 78, 79, 80, 85, 94, 97,
98, 102, 103, 105, 257, 307, 308,309
Buz,246
Cadmus, 119,329
Caere, 42, 70, 79, 80, 82, 85, 97, 98, 102, 103,
105, 257, 307,309
Calchas,269
Index
Callimachus, 177, 186, 187, 188, 189, 203, 204,
209, 210, 329, 330,331
Calypso,231
Canaanite, 50, 160, 162, 164, 165, 166, 168, 172,
173, 258, 259, 299, 322,327
carbonates, 113, 114, 115,116
Carian, 163, 198, 199,313
Carolingian script,173
Carthage, 73, 76,95
Celtic, 330,332
Chalkis, 17, 40, 42, 63, 75, 92,295
Chalkodamas, dedication of, 88,93
Charon,189
chaste-tree, 196,201
chi, 2, 250, 251,298
Chinese,327
Chios, 18, 20, 23,211
chisels, 111, 112, 187, 188, 190,210
chlorides, 109, 113, 114, 115,
116,118
Circe, 192,231
Clement of Alexandria, 203, 204, 205,
206,213
Clotho,189
Codex Arabicus,5
cognitive weaving, 193, 195,231
combination, 169, 170, 235, 239, 240, 241,
244,337
contextus, 237,238
contiguity disorder, 170,326
copper alloy, 108, 109, 110, 111, 113, 114,115
copper Dead Sea scroll,112
Coptic, 3, 5, 6, 7, 155,163
Corinth, 27, 28, 37, 70, 73, 296,334
Corinthian, 33, 44, 55, 58, 80, 99, 261,286
Corkyra,309
Cornutus,276
Cretan, 16, 18, 29, 33, 51, 55, 76, 78, 85, 110,
260, 295, 296, 299, 307,322
Cretan Hieroglyphic,322
Crete, 16, 28, 29, 32, 34, 37, 59, 71, 74, 75, 77,
83, 84, 90, 98, 110, 305,306
Cronus, 60, 210, 273, 275, 276, 277,278
cuneiform, 161, 172,322
cuprite, 110, 114, 115, 116,117
cutting implements, 179, 180, 181, 182, 186,
187, 189, 192, 203, 207, 210, 213, 223,233
Cybele,17
359
Cycladic islands, 18, 32, 37, 39, 45, 97,

110,295
Cyclops, 193, 194,195
Cyllene,219
Cynewulf,233
Cypriot, 58, 256, 295, 299, 303,321
Cypriot syllabary, 62, 63, 64, 65, 146, 148, 149,
153, 303, 316, 319,320
Cyprus, 23, 36, 54, 73, 76, 78, 84, 105, 115, 173,
229, 258, 302, 303, 315,337
Cyril and Methodius,163
Cyrillic,327
Daedalus, 199, 204, 205, 210, 224,332
Damascus,212
Dame(as), grave stele of,85
Danads,281
Danas, 281,329
Daphne,201
Dedan,246
Deidamas, epitaph of, 28, 29, 39, 80,85
Delian,60
Delos, 28, 29, 38, 40, 70, 73, 80, 88, 97, 102,
187, 200, 294, 295, 298,332
Delphi, 39, 44, 53, 54, 55, 62, 87, 91, 94, 260,
267, 274, 310,332
delta, 19, 21, 23, 28, 29, 38, 40, 55, 56, 59, 135,
136, 137, 138, 140, 224, 250, 286,295
Demeter, 182, 276, 277, 278, 279,313
Demophon,209
Deschamps, Gaston, 176,328
determinatives,321
Deucalion,334
dextroverse, 16, 34, 98, 286,316
digamma, 31, 32, 33, 34, 39, 51, 54, 55, 60, 61,
62, 63, 65, 66, 67, 68, 83, 136, 137, 138, 151,
159, 224, 250, 257, 285, 286, 296,303
Diodorus Siculus,210
Dionysius of Halicarnassus, 239, 240, 241,
242, 243, 244,336
Dionysus, 196,276
Dioscorides,185
Dipylon oinochoe, 15, 22, 23, 24, 29, 44, 58,
59, 70, 73, 74, 83, 84, 85, 88, 92, 98, 105,
256, 292, 302, 310,312
dissimilation, 257,263
distinctiveness, 146, 148, 149, 319,323
divination, 213, 214,215
360
Index
Dodona, 200, 201, 202, 213, 214, 215, 216, 217,

218,334
Doric, 17, 19, 39, 44, 73, 85, 206, 207, 330,333
Doric Dodecanese,99
Doric Hexapolis,98
Doric islands, 73,98
Dreros law code, 29, 71, 75, 78, 83, 84, 85, 90,
302, 305,307
criture, 140, 143, 147, 157, 158, 159, 171, 173,
174, 175, 176, 177, 317, 318, 319,327
Edom,246
Egypt, 1, 3, 4, 5, 6, 7, 8, 18, 115, 154, 155, 156, 157,
158, 159, 160, 162, 185, 211, 212, 214, 216,
246, 283, 291, 293, 321, 322,329
Middle Kingdom, 156,321
Old Kingdom,156
Ptolemaic and Roman periods, 154,156
Egyptian language,155
Egyptian script, 154, 156, 157, 158, 159, 160, 163,
172,321
demotic script, 154,155
hieratic script, 154, 155,156
hieroglyphic script, 154, 155, 156, 158,322
Eileithyia,60
Ekpropos, inscription of, 32, 47, 54, 70, 86,
88,102
Ekron,246
Elam,246
electron probe microanalysis, 107, 108,109
elementum, 236, 251, 252, 262,263
Eleusis,209
Eleutherna, 34,55
Elis, 27, 29, 101, 200,207
elision, 206, 207,333
Eltynia, 74,305
engraving, 4, 79, 112, 136, 186, 188, 192, 204,
213, 223,263
Enyo, 17, 40, 53, 74,298
Ephorus,213
Epidauros,54
Epirus, 201,214
epitaxis,116
epsilon, 29, 30, 31, 32, 33, 37, 39, 70, 85, 104,
133, 136, 137, 285, 286, 295, 296, 305, 307,
311,339
Eretria, 19, 25, 42, 45, 54, 75, 76, 78, 85, 88,
295, 298, 299, 300, 306,307
Eretrian, 30, 32, 43, 45, 63, 70, 77,212

Erinyes, 280, 281,287
Eros,232
Erotian, 179,197
eta, 33, 36, 37, 39, 40, 41, 42, 43, 44, 45, 46, 48,
49, 50, 52, 53, 54, 55, 56, 59, 83, 85, 95,
120, 128, 129, 131, 133, 135, 137, 149, 150,
178, 182, 224, 225, 251, 253, 254, 255, 256,
259, 260, 261, 263, 285, 286, 297, 299, 307,
308, 318, 320, 328,339
Etruria, 20, 25, 42, 45, 53, 70, 85, 94, 97, 98,
102, 103, 149, 258, 307, 309,312
Etruscan, 25, 26, 27, 41, 42, 47, 50, 69, 77,
78, 79, 102, 103, 105, 162, 163, 235, 237,
251,257
Euboia, 19, 29, 32, 37, 42, 45, 78, 91, 98, 182,
212, 302, 305,307
Euboian, 16, 17, 19, 25, 27, 39, 40, 41, 42, 45, 47,
50, 53, 54, 59, 69, 70, 75, 76, 77, 78, 79,
80, 82, 85, 86, 91, 92, 94, 97, 98, 101, 102,
103, 105, 142, 149, 162, 251, 257, 285, 287,
293, 295, 300, 307, 308,312
Eumaeus, 213, 240,281
Eumenides, 280,281
Eupolis,220
Euripides, 182, 183, 203, 222, 266, 283,329
Euryalus,208
Eurystheus,198
Eustathius,196
Ezekiel,322
Fayum,1,3
Festus,238
film,288
Formello, 42, 77, 94, 97, 102,103
Fortuna,237
Francavilla Marittima,303
Franois Vase,21
futhark,323
Galen, 179, 180, 181, 182, 196,328
gamma, 27, 28, 31, 42, 57, 87, 131, 133, 135, 137,
138, 140, 224, 285, 286, 287, 294, 295, 305,
308,316
Gaulish,332
Gaza,246
Geez,323
Gela,97
Index
geometric,201
Germanic, 163, 174, 190, 330,333
Getty Museum, 2,106
gimel, 27,247
glides, spelling of, 32, 33, 34, 59, 61, 62, 63, 64,
65, 66, 67, 190, 191,206
Gold Pendant inscription, 76,95
Gortyn, 71, 75, 77, 85, 98,306
Gothic, 163, 190, 192, 206, 269, 328,333
Graces, xi, 195,234
Graecus, 186,275
Greek-Phoenician interaction, 66, 67,257
Grimm, Jacob,176
groundwater, 114, 115, 118, 212,334
Hadad,212
Hades, 189, 270, 275, 276, 277, 278, 279, 280,
281, 284,285
Halicarnassus, 211,239
Han Gul,318
Harma,267
Harmonia,329
Hazael,212
he, 30, 295,296
Hebe, 187,329
Hebrew, 27, 101, 160, 246, 247, 248, 249, 252,
253, 319, 327, 337,338
Hector,207
Helen,30
Helios,277
Hellenic,330
Hephaestus, 186,187
Hera, 18, 25, 187, 198, 199, 200, 201, 202, 203,
204, 205, 210, 211, 212, 213, 223, 224,
329,332
Hera Cithaironia,203
Hera Telchinia,210
Heracles, 31, 198, 272, 273,340
Heraclitus,274
Heraia,30
Heraion, 18, 28, 50, 52, 70, 94, 199, 201, 202,
203, 211, 212, 260, 303,312
Hermes, 185, 189, 197, 219, 220, 275, 329,331
Herodas, 189,329
Herodotus, 211, 213, 283,334
Hesiod, 40, 268, 269, 270, 272, 275, 276, 277,
279, 280, 313, 334,340
Hesychius, 180,184
361
et, 37, 54, 296, 297, 322,339

Heubeck, Alfred,1,3
hexameter, xii, 38, 84, 87, 91, 92, 104, 197,
288,290
Hippocrates, 179, 180, 197, 198,268
Hittite,191
Homer, 3, 15, 40, 107, 183, 192, 193, 194, 197,
202, 203, 206, 209, 213, 219, 222, 231, 232,
240, 241, 242, 264, 266, 267, 268, 269,
270, 271, 272, 273, 278, 280, 281, 288,
290, 315, 329, 331, 336, 340,341
homography, 46, 51, 52, 56, 59, 90, 96, 149,
150, 258, 261,320
Hou,155
Hugo, Victor, 154,160
Hyacinthus,33
Ialysos,210
idioms, 164,165
Ikaros,203
Ilium,332
Imbrasos (river and plain),199
Indo-European, 62, 63, 152, 174, 176, 189, 190,
191, 192, 198, 202, 205, 206, 213, 218, 220,
221, 223, 230, 232, 237, 238, 243, 245, 264,
269, 272, 273, 330, 332, 333, 335, 336,341
Indo-Iranian, 221, 222, 232, 269,332
Indra,205
Institute for Advanced Study,4,5
interweaving, 194, 221, 222, 224, 225, 226, 227,
234, 235, 238, 241, 243, 244, 248, 250, 252,
253, 254, 255, 256, 257, 258, 260, 261, 262,
263,337
Ionia, 32, 102,302
Ionic, 18, 21, 25, 27, 28, 33, 37, 38, 39, 40, 44,
45, 62, 85, 101, 102, 178, 179, 186, 195,
198, 199, 204, 207, 211, 251, 293, 294, 295,
318,330
central Ionic, 40,207
eastern Ionic, 3, 18, 25, 28, 133, 207,303
western Ionic, 40, 293,303
Ionic Dodekapolis, 37,102
Ionic islands, 73,98
Ios,290
iota, 2, 33, 56, 57, 58, 59, 60, 61, 62, 63, 64, 66,
67, 68, 71, 82, 133, 137, 138, 144, 151, 152,
153, 159, 182, 221, 227, 250, 256, 257, 258,
263, 301, 302, 303, 305, 328,339
362
Index
Iphicles, 272,273
Israel,246
Isus,192
Italic,330
Ithaka, xi,213
candleholder from,70
oinochoe from, 91, 92,104
Izbet arah ostracon, 161, 249,250
Jakobson, Roman, 144, 147, 167, 168, 169, 170,
171, 239,254
Japanese,327
hiragana and katakana,173
Jeremiah, 246, 248,338
Jerome, 236, 246, 247, 248, 250, 251,253
Jerusalem, 246,322
Judah,246
Julian, prefect of Egypt,185
Julius-Maximilians-Universitt Wrzburg,1
juniper tree, 202,332
Kabardian,324
Kalymna,99
Kamiros, 210,297
kap, 247, 248, 249,262
Kaphyai, 200,332
kappa, 42, 69, 70, 71, 72, 73, 120, 128, 133, 145,
146, 250, 251, 262, 303, 304,305
Karatepe inscriptions,305
Keos, 38, 102, 195, 210, 229, 234,331
Kilamuwa inscription, 58,256
Kiseleff, Alexander,1
Kition, 23,36
bowl from, 36, 48, 76, 84, 296,302
Klazomenai,211
Kleobis and Biton, statues of,94
Kleonai, 33, 94, 296,311
Knidos, 39, 40, 44, 53, 211, 256, 297, 320,329
Knossos, i, 33, 51, 75, 150, 242, 260,305
Kolophon, 18,101
Kos,207
Kraus, H. P., 1, 3, 4, 5, 7, 8, 35,292
Kritoboule, inscription of, 104,105
Kuntillet Arjd,322
Kydonia, 18,19
Kyme, 27, 30, 31, 32, 39, 42, 45, 70, 75, 77, 85,
104, 105, 141, 142, 143, 173, 177, 256, 285,
288, 306, 307,320
Kyrenaean,258
Kyrene, 52, 207, 260,300
Lacco Ameno, 16,294
Laertes, xi,269
Lakonia, 29, 30,295
Lakonian, 53, 55, 87, 102, 307, 308,314
lambda, 2, 27, 72, 73, 74, 75, 79, 92, 97, 119,
129, 131, 135, 137, 138, 224, 250, 251, 262,
305,316
lamed, 73, 74, 75, 247, 248, 249,305
langue, 140, 147, 158, 159, 171, 172, 173, 174, 175,
176, 177, 317, 318, 320, 323, 324, 327,328
Larisa, 29,37
Lebadea,295
Lehmann-Haupt, Hellmut,5
Lemnos,303
Libya, 214,216
Lindos, 19,210
Linear B, 62, 64, 65, 145, 148, 149, 153, 316,
318,322
Linenthal, Richard,8
linguistic sign, 140, 141, 142, 147, 148, 156, 157,
158, 160, 169, 172, 174, 177, 178, 182, 198,
321,322
Lithuanian, 192, 243, 268,330
logograms, 156, 157,321
Lokris, 29,33
Lokroi Epizephyrioi,17
looms, 25, 195, 229, 231, 238, 241, 263, 296,
336,337
Lucian, 180,189
Lucilius,338
lugos, 192, 194, 195, 196, 197, 198, 199,
200, 201, 202, 205, 213, 218, 219, 221,
224,247
Luxor,155
Lycia, 15, 183, 266,270
Lycian, 60, 163, 188,230
Lycophron,276
Lydian, 163,229
Lyttos,307
Magna Graeca, 250,262
Magnesia,37
Maia,219
malachite, 109, 110, 111, 113, 114, 115, 116,
117,118
Index
Malta stele, 78,105
Mantiklos statuette, 25, 27, 47, 50, 58, 70, 102,
256, 258, 260,329
Marsiliana dAlbegna, writing tablet from, 25,
27, 42, 47, 54, 69, 76, 85, 90, 98, 103, 294,
295, 307, 308,312
Martinet, Andr,316
Massarosa,20
matres lectionis, 319,327
matrix, alphabetic, 165, 166, 168,
324,325
matrix, lexical, 213,223
maximal separation, principle of, 165, 166,
167, 168,324
Medes,246
Megara, 17, 18, 19, 27,37
Megara Hyblaia, 17,88
ml, 178, 179, 180, 181, 182, 186, 189, 191, 192,
213, 215, 221, 222, 223, 224, 226, 227,
234,263
Meleager,280
Melos, 32, 56, 57, 76, 77, 306,339
Melpea,30
mem, 76, 78, 249, 306,322
Menander Rhetor,329
Menelaion,30
Menelaus, 30, 200, 231,332
Menodotus,198
Mentor,329
Mesrop,163
Metaneira,209
metaphor, 166, 170, 171, 187, 188, 189, 220, 228,
229, 231, 232, 233, 238, 239, 241, 242, 243,
245, 263, 264, 265, 331, 335, 336,337
Metapontion, 250, 251,262
metathesis, 119, 120, 190, 191, 286,330
Methone, 30, 32, 50, 70, 77, 86,313
metonymy, 170,171
Middle High German,230
Middle Irish,268
Miletos, 18, 211,333
mineralization, 116,117
Moab,246
monoconsonantal symbols, 156, 157, 158, 159,
163,172
Monte Acuto, 42,97
Monteriggioni, 42,77
mottling pattern, 111
363
Mount Hymettos, 17, 19, 20, 24, 25, 28, 29, 30,
31, 56, 70, 72, 74, 85, 88, 98, 99, 100, 101,
102, 218, 261, 265, 266, 267, 271, 274, 276,
285, 288, 312, 313, 314,339
Mount Ida,192
Mount Parnes, 267,271
Mount Sinai, 4,5,7
mu, 72, 75, 76, 77, 78, 79, 80, 81, 82, 85, 89, 90,
95, 96, 97, 119, 137, 250, 251, 256, 257, 262,
263, 304, 306, 307,311
Muses, 186, 188, 210, 227, 230, 288,329
muthologoi,201
muthos, 202, 204, 205, 232,275
Mycenae,334
Mycenaeans, 62, 63, 64, 145, 147, 148, 153, 202,
214, 231, 242, 303, 319, 332, 334,339
Mytilene, 184,211
Narce, 42, 70,251
Naukratis, 18, 211,216
Nauplius,182
Nausicaa,281
Naxos, Aegean, 17, 28, 29, 32, 37, 38, 39, 40, 43, 44,
45, 53, 69, 70, 73, 80, 85, 88, 97, 101, 102, 149,
255, 256, 285, 292, 294, 297, 298, 308,320
Naxos, Sicilian, 17, 40, 45, 53, 74, 256,320
Nebuchadrezzar,246
Nemea, 33, 296,311
Nestor, 230, 231,282
Nestors cup, 29, 69, 70, 77, 83, 87, 88, 92, 98,
103, 105, 222,295
Nicander, 197, 198,268
Nikandra statue, 29, 38, 69, 70, 73, 80, 97, 102,
295,304
Nike,276
Nile River, 4, 155,211
Nisyros, 207,333
Nonnos, 119, 175, 280,281
Nora fragment, 37,339
Nora stone, 76,84
Northwest Greek,207
nt jug, 23, 58, 105,256
nu, 2, 51, 58, 67, 72, 76, 79, 80, 81, 82, 85, 87,
90, 92, 96, 120, 128, 136, 138, 242, 251,
256, 257, 258, 262, 263, 270, 304, 307,
310,311
nun, 58, 61, 67, 82, 84, 185, 256, 257, 283,284
nymphs, 201,219
364
Odysseus, 192, 193, 195, 208, 213, 231, 240,

268, 269, 270,281
Oeax, 182,184
Oedipus, 185, 282,285
ogham,323
Oinona,229
Old Babylonian,279
Old Church Slavic, 190,243
Old English, 190, 233, 268,333
Old Frisian,190
Old High German, 190, 230, 268,333
Old Irish, 190, 192, 233,323
Old Norse, 190, 230,243
Old Persian,243
Olen,60
Olsen, Fred,5
Olympia, 17, 206, 332,339
Olympichus,204
omega, 2, 52, 225, 247,251
omicron, 48, 49, 50, 51, 52, 53, 54, 55, 56, 59,
83, 87, 104, 120, 128, 129, 133, 135, 150, 225,
253, 254, 255, 256, 258, 259, 260, 261, 263,
299, 300,301
omission at alphabetic juncture, 72, 249,262
Opountian Lokris, 75,305
optical metallography, 2, 107, 108,113
ord, 238,247
Orestes, 183,274
Oribasius, 180,181
Orpheus, 203,332
orthography as performance, 213, 217, 264,
265, 266,288
Osteria dellOssa, 59, 60,302
Ozolian Lokris, 17,73
Palamedes, 182,184
Pamphylian, 34, 62,303
Pan,30
Panathenaia, 3,243
papyri, 6, 7,292
paradigmatic structure, 169, 235, 239,
263,264
paratacamite, 109, 115,117
Paros, 21, 38, 40, 101, 102, 294,297
Pasidikos, grave stele of, 78,97
patina, 1, 2, 35, 108, 109, 110, 113, 114, 115, 117,
118,316
Patroclus, 213,282
Index
Paul of Aegina,180
Pausanias, 60, 199, 200, 201, 202, 204, 205,
207, 210, 213, 267,334
pe, 84,322
pedagogy, alphabetic, 166, 236, 237, 238, 246,
247, 252, 254, 324,337
Pedanius Dioscorides, 196,197
Pelasgians,213
Peleus,329
Peloponnese, 30, 32,296
Penelope, 195, 213, 231, 269,335
periodic order of the alphabet, 137, 138, 139,
140, 161, 162, 163, 164, 166, 172, 176, 177,
224, 225, 235, 237, 238, 255, 262, 263, 271,
286,287
Persephone, 182, 276, 277, 279, 280,287
Petersen, TheodoreC.,7
Phaeacia,208
Phaedrus, 214, 215, 216, 217, 218,275
Phaistos, 16,28
Phaistos disk,322
Phanias,186
Phaselis,211
phi, 2, 184, 185, 250, 251, 299,329
Philae,155
Philippus of Thessalonica,186
Philistines,246
Philostratus, 201,334
Philto, kylix of,298
Phleious, 33, 73, 296,311
Phoenician, 2, 16, 22, 23, 24, 25, 26, 27, 30, 32,
33, 35, 36, 37, 41, 47, 48, 50, 54, 58, 59, 60,
61, 63, 64, 65, 66, 67, 68, 69, 73, 74, 75,
76, 78, 82, 83, 84, 87, 89, 95, 97, 98, 100,
101, 104, 105, 146, 147, 149, 151, 152, 153,
154, 157, 160, 161, 163, 164, 165, 173, 212,
251, 252, 254, 256, 257, 258, 259, 263, 292,
294, 295, 296, 297, 299, 302, 303, 305, 311,
312, 316, 321, 322, 327,339
Phoenix, 280,287
Phokaia,211
Phokis, 17, 18, 29, 73, 91, 293,310
phonetic complements,156
phonograms, 156,321
Photius,180
Phrygian, 19, 34, 63, 68, 163,303
pi, 27, 29, 43, 54, 83, 84, 85, 86, 87, 88, 89, 93, 131,
146, 286, 294, 300, 301, 307, 308, 309,313
Index
pinakes, 15, 119, 183,329
Pindar, 195, 202, 228, 229, 230, 232, 233, 234,
271, 276, 329, 334,335
Pithekoussai, 16, 28, 42, 59, 69, 70, 76, 77, 83,
87, 98, 101, 102, 103, 105,295
plaiting, 187, 192, 193, 194, 195, 198, 213, 219,
220, 221, 230, 235, 236, 237, 238, 241, 242,
249, 251, 252, 253, 254, 255, 256, 257, 258,
259, 263, 287,337
Plato, 189, 194, 214, 215, 216, 217, 218, 220, 242,
243, 274, 283, 284,341
playfulness, scribal, 46, 56, 136, 149, 224, 225,
253, 254, 257, 259, 260, 261, 263,311
pleonastic spelling, 44, 45,299
Pliny, 196, 197, 237, 329,332
Plutarch,40
Pluto, 276, 279, 285,341
polusmantr, 275, 276, 277, 278, 279,284
Polycrates,196
Polyphemus,193
Pompeii, 235, 237, 247, 249, 250,251
Poseidon, 201, 210,278
Posideion,22
Potters Quarter sherds, 28, 58, 70, 73,80
Prague School,337
Praisos, 74,305
Priam,192
Prinias,85
probes, 179, 180, 181, 182, 328,330
Procles,203
Proetus, 15, 266,270
Proto-Canaanite, 23, 147, 157, 161, 172,321
Protocorinthian, 31, 39, 69, 70, 88, 104,
141,285
Proto-Sinaitic, 155, 156, 157, 160, 161, 172,322
Psammetichus I,211
psi, 2, 145, 146, 247, 251,318
Qena, 155,321
qop, 64,322
qoppa, 33, 64, 93, 97, 99, 136, 137, 145, 250, 251,
301, 303, 310,312
Qbur el-Walaydah,24
Quintilian, 236, 237, 238, 239, 246, 247,251
Qumran, 249, 250,262
recrystallization,113
rectus contextus, 237, 238,247
365
reeds, 185, 186,329

repouss,112
re, 98, 100,322
Rhadamanthus,287
Rhea,278
Rhegion,101
Rhizenia,34
rho, 55, 86, 98, 99, 100, 129, 131, 136, 137, 183,
308, 312, 313,316
Rhodes, 19, 32, 39, 44, 45, 101, 149, 207, 210,
211, 298,302
Subgeometric cup from, 44, 73,312
Rhodian, 44, 52, 53, 297,300
root-noun compounds,205
runes, 163, 164,323
Russian, 171,190
ade, 89, 95,311
Salamis,208
Samaritan, 89,302
samek, 41, 153,254
Samian, 18, 26, 29, 198, 200, 201, 202, 203,
204, 205, 210, 211, 212, 213, 223, 224, 293,
297,312
Samos, 3, 18, 25, 28, 29, 70, 101, 195, 198, 199,
200, 201, 202, 203, 204, 205, 207, 210,
211, 212, 213, 223, 224, 308, 312,333
Samothrace, 28, 294, 295,303
san, 33, 42, 55, 59, 69, 83, 89, 90, 91, 92, 93, 94,
95, 96, 128, 136, 138, 146, 193, 224, 250,
251, 282, 302, 305, 309, 310, 311,339
sanis, 202, 203, 204, 213,224
Sanskrit, 176, 191, 205, 206, 230, 243, 268, 269,
332, 333,335
Sappho,231
Sardinia, 84,339
Sarpedon,230
satyrs,219
Saussure, Ferdinand de, 140, 141, 142, 143,
146, 147, 149, 150, 151, 160, 161, 163, 164,
168, 169, 172, 173, 174, 175, 218, 228, 239,
264, 271, 316, 317, 319, 320, 322, 323,
325,328
scanning electron microscopy, 2, 107,
108,116
Schyen, Martin, 1, 2, 3,4,7
selection, 169, 170, 235, 239, 240, 241, 244,
336,337
366
Selinous, 17,44
sma, 183, 218, 265, 266, 267, 269, 270, 271,
272, 275, 282, 283, 284, 285, 287, 288,290
smain, 216, 272, 274, 282,283
smantr, 272, 273, 274, 275, 282, 284, 285,340
Semitic, 32, 33, 50, 59, 64, 84, 95, 98, 147, 152,
153, 154, 155, 156, 157, 158, 159, 160, 161,
162, 172, 212, 247, 252, 259, 260, 263, 305,
312, 313, 315, 319, 320, 321, 322,323
Semo Sancus,237
Serabit el-Khadim, 155, 160,321
Servius Tullius,237
Seville statuette, 22,76
shan,59
shin, 247,248
Sicily, 17, 44, 234, 292, 293,330
Sidon,246
Siena,i,42
sigma, 2, 36, 44, 55, 59, 69, 87, 91, 101, 102, 103,
136, 137, 146, 224, 250, 298, 299, 301, 302,
306, 313,314
signifiant, 142, 156, 157, 160, 218, 316, 317,340
signifi, 156, 157, 218, 316,326
Sikinos, 76, 78, 97,306
Sikyon, 27, 54, 62, 73,301
Silenus,219
similarity disorder, 170,326
Simonides, 229, 232,331
Sinai, 155,156
sinistroverse, 15, 34, 101,286
Siphnos,38
Skelmis,204
smil, 180, 181, 182, 183, 184, 185, 186, 187, 188,
189, 190, 191, 203, 204, 205, 206, 210, 213,
215, 223, 224, 266, 328,330
Smilis, 199, 200, 202, 204, 205, 207, 213,
215,224
s-mobile, 189, 190, 191,238
Smyrna, 101,102
Socrates, 214, 215, 216, 217, 274, 283, 284,341
Sophocles, 219, 276, 282, 285,341
Sophron,330
South Arabian, 322,323
Sparta, 43, 93,200
spindle whorls, 60,238
spindles, 230,237
spinning, 60,238
Stoics,337
Strabo, 40, 210, 211, 213,340
Index
Strattis,276
stylus, 19, 79, 99, 100, 182, 183, 184, 188, 192,
221, 222, 223, 226, 227, 229, 234,266
Styra,293
Styx,276
subgeometric, 44, 73, 94,312
sulfides, 109,114
supplemental characters, 2, 33, 59, 63, 65, 146,
152, 162,316
syllable weaving, 244,245
syntactic structure, 148, 164, 165, 170,171
syntagmatic structure, 163, 164, 165, 168, 169,
170, 171, 230, 235, 239, 263, 264, 323,324
Syracuse, 70, 73, 195, 234,330
Syria-Palestine, 161, 173, 200, 201, 212,249
Syros,38
Tanagra stele, 43, 50,260
Tanaquil,237
Tarquinius Priscus,237
Tataie, aryballos of, 70, 77, 104, 141, 173,177
tau, 2, 3, 33, 70, 77, 103, 104, 105, 133, 136, 137,
141, 150, 151, 160, 163, 166, 171, 224, 225,
235, 299, 315, 316,338
taw, 2, 33, 104, 105, 151, 160, 166, 247, 248,
249,322
Teiresias,270
Telchines, 188, 209,210
Telemachus,240
Telephus, 183,329
television,288
Tell Fakhariyeh, 259,299
Tema,246
Tenos, 28,294
Teos, 195,211
Teucer,229
textile patterning,118
textus, 230,263
Thamus,216
Thasos, 21, 28, 88, 102,205
Thebes, Boiotian, 47, 86, 88, 148,307
Thebes, Egyptian, 155,214
Theia,276
Theocritus,194
Theophrastus, 201, 239, 332, 336, 337,340
Thera, 25, 28, 29, 32, 39, 45, 50, 52, 59, 71, 73, 76,
85, 91, 98, 149, 207, 258, 260, 294, 299,300
Theran, 25, 26, 52, 71, 80, 83, 91, 258, 260, 294,
297, 300,312
Index
Thesmophoria, 182, 196,266
Thespiai, 203,222
Thessaly, 37, 73,101
theta, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56,
59, 83, 87, 97, 120, 128, 133, 135, 136, 137,
150, 225, 253, 254, 255, 256, 258, 260, 261,
263, 299, 300,301
Thetis, xi,329
Theuth,216
Thomson, RoyH.,5
Thoth,216
Thracian,203
Thurii,211
Tiryns, 33, 296,311
Titans,210
Tocharian, 243,330
Tonaia, 199,223
Torah,338
triconsonantal symbols,156
Troad,110
Troy, 40, 107, 182, 200, 269,282
two alphabets, 250, 251, 262,263
Tylissos, 33, 51, 150, 260,299
Tyre,246
Ugaritic, 161, 164, 165, 166, 172, 249,
322,338
Umbrian,191
Umqi,212
upsilon, 2, 33, 51, 61, 63, 65, 67, 68, 151, 152,
250, 251,299
Uranus,210
Utah Middle East Center,6
Valerius Babrius,180
Varro,237
Vedas, 205,206
Veii, 26, 42, 69,103
viral hexameter,290
viral video,289
Viterbo, 42, 70, 78, 85, 97, 103,309
Wackernagel, Jakob,222
Wadi el-Hl, 155, 156, 157, 160,321
warp, 177, 194, 229, 230, 235, 238, 241, 244,
245, 263, 264,336
wau, 31, 32, 33, 61, 63, 64, 65, 66, 68, 151,153
waw, 32, 33, 61, 63, 64, 65, 66, 67, 151, 152,
296,322
367
weaving, 3, 60, 119, 187, 192, 193, 194, 219, 220,

221, 230, 231, 232, 237, 238, 241, 242, 243,
247, 335, 336, 337. Seealsoalphabetic
weaving; cognitive weaving; syllable
weaving; word weaving
weft, 194, 230, 235, 238, 241, 244, 245,336
Welsh,233
Whitney, William Dwight, 172,327
withes, 192, 193, 194, 195, 196, 197, 198, 213,
218, 219, 221, 223, 224,247
Wittgenstein, Ludwig,316
word weaving, 188, 195, 227, 228, 229, 231, 232,
233, 239, 241, 242, 243, 244, 245, 264,
288,289
Wulfila,163
Xenophon,274
xi, 38, 41, 42, 43, 44, 45, 46, 48, 49, 50, 52, 53,
56, 59, 77, 79, 83, 120, 128, 129, 131, 133,
135, 136, 137, 145, 146, 149, 150, 153, 178,
224, 225, 250, 251, 253, 254, 255, 256, 259,
260, 261, 263, 270, 285, 298, 299, 300,318
X-radiography, 26, 31, 107, 108, 111, 112, 116,
117, 131,315
X-ray diffraction analysis, 2, 107, 108,117
X-ray fluorescence analysis, 2, 107,
108,109
Yahweh,246
yod, 36, 59, 61, 63, 64, 66, 67, 68, 151, 249, 257,
262,302
Zamri,246
Zankle,293
zayin, 36, 296, 321,322
Zenjirli, 58, 256,258
zeta, 34, 35, 36, 39, 120, 131, 133, 137, 286, 287,
316,321
Zeus, 56, 63, 187, 213, 214, 216, 219,
232, 266, 269, 270, 272, 273, 274, 275,
276, 277, 278, 279, 280, 281, 282, 284,
314, 334,339
chthonic Zeus, 279, 280, 281, 282, 284,
287,313
Zeus Ombrios,267
Zeus Semaleos, 267, 269, 271, 272, 274,285
Zeus Semios, 101, 102, 218, 265,
266, 267, 269, 271, 272, 274, 276,
285, 288, 313, 314

Roger D. Woodard-The Textualization of The Greek Alphabet-Cambridge University Press (2014)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Roger D. Woodard-The Textualization of The Greek Alphabet-Cambridge University Press (2014)

Uploaded by

Copyright:

Available Formats

Th e Te x t u a l i z a t i o n o f t h e G re e k A l p h a b e t

The Textualization of the Greek

32 Avenue of the Americas, New York, NY 100132473, USA

For Ed Brown, teacher and friend, and to the memory

and chemical examination of the

copper plaques . ...................................................................................

syntagmatic structure of the copper plaques . ................ 119

et criture ............................................................................. 140

styluses and withes . ....................................................................

6.2.6 (Sanis) and (Smilis)

warp and weft of writing . ...................................................... 227

Isabella Vitti. Thanks too go to two anonymous manuscript referees for

American Journal of Archaeology

Harvard Studies in Classical Philology

Libation Bearers (Choephoroe)

Anthologia Palatina (Anth. Pal.; The Greek

Cornutus (L. Annaeus)

Works and Days (Opera et dies)

Homeric Hymn to Apollo (Hymn. Hom.Ap.)

Oxyrhynchus Papyri (P Oxy.)

Scholia in Aristophanem (Schol. Ar.)

Women at the Thesmophoria (Thesmophoriazusae)

Scholia in Lycophronem (Schol. Lycoph.)

The Textualization of the Greek Alphabet

The Textualization of the Greek Alphabet

The Textualization of the Greek Alphabet

Atiya was a consummate collector. In an interview with Everett Cooley,

of Congress Sinai expedition. After describing his supervisory role in the

The Textualization of the Greek Alphabet

MS 1-1. X-radiograph of obverse face of Martin Schyen plaque 1. Courtesy of

MS 1-2. X-radiograph of reverse face of Martin Schyen plaque 1. Courtesy of

MS 2-1. X-radiograph of obverse face of Martin Schyen plaque 2. Courtesy of

MS 2-2. X-radiograph of reverse face of Martin Schyen plaque 2. Courtesy of

The Textualization of the Greek Alphabet

The second alpha of is unmistakably of the CP Alpha-1 type

The Associative Structure of the Copper Plaques

The Textualization of the Greek Alphabet

copper-plaque type can be seen in a graffito ( [) on the handle of an

The Associative Structure of the Copper Plaques

The Textualization of the Greek Alphabet

The Associative Structure of the Copper Plaques

The Textualization of the Greek Alphabet

Accounting for the original morphology and morphological development

The Associative Structure of the Copper Plaques

The proposal of a ninety-degree rotation in the development of alpha is not

The Textualization of the Greek Alphabet

The Associative Structure of the Copper Plaques

The Textualization of the Greek Alphabet

The Associative Structure of the Copper Plaques

Throughout the copper plaques, gamma is morphologically quite consistent.

The Textualization of the Greek Alphabet

Throughout the copper plaques, the shape of delta is generally consistent,

The Associative Structure of the Copper Plaques

Epsilon too has a relatively consistent appearance throughout the copper

The Textualization of the Greek Alphabet

The Associative Structure of the Copper Plaques

In a few instances of CP epsilon, the tail is exceptionally short.107 Close

The digamma, or , of the copper plaques morphologically parallels epsilon,

The Textualization of the Greek Alphabet

more innovative right-angular digamma followed, or paralleled, a comparable

The Associative Structure of the Copper Plaques

The Textualization of the Greek Alphabet

The appearance of zeta in the copper plaques is an intriguing matter. Two