Textual and Other Notes on Aeschylus

6/09

These notes are mostly designed to explain some of the textual choices made in
passages from the seven surviving plays of the Aeschylean corpus in the first two
volumes of my Loeb edition (Cambridge MA, 2008). I intend subsequently to publilsh a
further article containing notes on the fragmentary plays.
Reports of the manuscripts and testimonia are based on M.L. Wests Teubner edition
(Stuttgart, 1990)1, and the sigla are those set out on pp. lxxxi-lxxxv of that edition
except that (i) some of the superscript abbreviations have been expanded, (ii) Wests
symbols for scholia
are replaced by M,
referring to the four
main classes of scholia which West describes on pp. xx-xxi, and (iii) Wests siglum ,
denoting in effect the recension of Demetrius Triclinius, is replaced by Tricl. in the
plays of the Byzantine triad and by f in Agamemnon and Eumenides (where, except in
Ag. 1-348, copies including emendations by Triclinius are our sole primary witnesses to
the text other than M where available).
The passages discussed are printed at the head of each section, normally in a form as
close as possible to the paradosis (on matters relevant to the discussion).
(1) Persians 162

162

Qsscr: v.l.

noverat

ut vid. (
)

The tenor of the speech, and of the whole scene, requires the Queen to be saying in 162
that she is not unafraid. L. Belloni in his commentary (2Milan, 1994) tries to get this
meaning out of the transmitted text by taking
to be governed by
as a genetivo di relazione and translating del tutto temendo in me
stessa2; but such a phrase would more naturally suggest the absurd meaning very
afraid of myself (cf. Pl. Rep. 386b do you think that someone who believed in the
terrors of the underworld would be
and would prefer death in battle to
defeat and slavery?) An alternative move has been to try and get rid of
. D.L.
Page (Oxford, 1972) printed Lawsons
: but for one
thing, as Belloni in effect points out, the Queen is not prophesying fear but feeling it, and
for another
is not merely a hapax in Aeschylus (Belloni) but unknown before
the Antonine age.

Wests companion volume, Studies in Aeschylus (Stuttgart, 1990), is hereafter referred to simply
as West Studies. First references consisting only of the name of a scholar with a place and date
of publication are to editions or translations of Aeschylus or of the play under discussion.
2
Syntactically parallel is the construal of W.J. Verdenius, Museum Philologicum Londiniense 7
(1986) 141, who takes the genitive as one of limitation. The parallels he cites the use of the
genitive after
and
, also Prom. 416
and Eur. Bacch. 40
suggest that by this he means that the Queen is saying she is not
unafraid so far as concerns herself, i.e. is apprehensive of what may befall her personally. While
this might seem a very reasonable state of mind for a person in the Queens position, it does not in
fact suit either the character or the context: both in this speech (168-9) and everywhere else in
the play, her anxiety is consistently not for herself but for her son.

2
West was surely right to leave the first three words of 162 unchanged; the Queens point
is, as he saw, that the statement she is in effect making that excessive wealth can lead
to ruin is not her own, not new, but a piece of ancient wisdom 3. Of the parallels he
cites, the key one is Eur. fr. 484.1
But in that case the necessary
negative for
must be found after
instead of before, and the
scholia
with their
encourage us to take this view. The superscript variant
in Q
is not worth much as evidence, since it could easily be a conjecture based on the scholia
or a mere error due to the presence of
a few words earlier or of
immediately below: it might still, of course, be a correct conjecture or a lucky error, but
it does have at least two disadvantages. It forces us to get rid of
, making the
construction rather harsh; and there is nothing in the first seven words of the sentence
in particular, no nominative adjective or adjectival phrase that would naturally be
linked to
by a coordinating conjunction like
The fact is that for the sentence to run smoothly, what we would really need is
which metre will not allow. What is more, it would be helpful if the
second half of Wests paraphrase, sententiam non meam ipsius sed a maioribus
acceptam, were spelt out in the text, as it wisely is by the Euripidean Melanippe (indeed
the other two parallels West cites, Ag. 750 and Cho. 314, show that an ancient saying
is the essential part of the expression and not my own invention is dispensable). In
other words, we need a lacuna of at least one line between
and the next
surviving word. In the Loeb I printed that word as
supposing that the -scholia had
preserved it correctly while in the direct textual tradition, after the passage had been
damaged,
had been inserted to patch up the construction and then
had been
lost by accident or in an attempt to mend the metre; but
or
(with a lost
negative preceding) are also possible.
If it is only a single line that has been lost, one might speculatively restore

But the lacuna may be longer than this.

(2) Persians 249-254
I think it is worth while to draw attention to the fact that the first six lines of the
Messengers opening seven-line speech all begin with an o-vowel. This is actually a
favourite trick of Aeschylus at the beginning of a speech expressing distress. At the end
of the Messengers long narrative, the first six lines of the Queens response (517-522)
begin with an o- or u-vowel, and her first eleven lines begin with a vowel of some sort;
similarly in her last speech in the play, after the departure of the Ghost of Darius, all
seven lines (845-851) begin with a vowel. When Eteocles in Seven against Thebes
learns that his brother is attacking the gate which he has reserved himself to defend, his

It is cited as such in Ag. 750-6. A.F. Garvie, Lexis 19 (2001) 6, do[es] not understand why
Atossa should want to emphasise so strongly that her
is not her own. The answer may be
given in words of P.E. Easterling (in R.D. Dawe et al. ed. Dionysiaca: Nine Studies Presented to
Sir Denys Page [Cambridge, 1978] 153) which Garvie himself had quoted elsewhere (on Cho. 3134
): When special attention is drawn to a sayings
antiquity or fame we should assume that the poet [and/or, I would have added, the speaker]
thinks it particularly significant.

3
first three lines (653-5) begin with ,4 after which he pulls himself together (it is not
proper to cry or lament); when Clytaemestra learns of the supposed death of Orestes,
her first three lines (Cho. 691-3) begin with an o-vowel. (In all cases, aspiration is
ignored.)
As a control, I have sampled three 50-line iambic passages5 from other parts of these
three plays, with the following results:
Lines
Pers. 353-402
Seven 568-618
Cho. 535-584

50
50
50

Initial
vowel
23
19
19

Initial
o-vowel
3
5
3

Initial
u-vowel
0
0
1

If we cautiously assume that normally 45% of all Aeschylean trimeters begin with a
vowel and 10% begin with an o- or u- vowel, then the probability of seven successive
lines beginning with a vowel by coincidence is 0.37%, of eleven such lines 0.015%, and
of six successive lines beginning with an o- or u-vowel the chance probability is 0.0001%
or one in a million.
Sophocles uses this device only once, I think, in his surviving plays: Oedipus last three
lines before he rushes into his palace to blind himself all begin with an o-vowel (OT
1183-5). Euripides may use it in a slightly different way in Trojan Women: when
Andromache learns that her son is to be thrown to his death from the walls of Troy, of
the first seventeen lines of her speech (740-758)6 nine begin with an o-vowel, and there
are never three successive lines which do not.
(3) Persians 282-3 ~ 288-9

282-3

288
289
4

e v. 281 post

transp. Page
codd. paene omnes:
, unde
Y:
Heimsoeth praeeunte Hermann
codd.:
Weil:
West
codd.: del. Heimsoeth:
Weil
fere codd.:
Boeckh

G.O. Hutchinson (Oxford, 1985) ad loc. drew attention to this, and noted the four initial s in
Pers. 249-253; but is not the only vowel that can serve this purpose, though it is evidently the
vowel that serves it most emphatically.
5
From the Seven passage I omit line 601, which like most editors I regard as spurious.
6
Omitting 742-3, which J. Diggle, Euripidis Fabulae II (Oxford, 1981), S.A. Barlow, Euripides:
Trojan Women (Warminster, 1986), and M.J. Cropp (on Eur. fr. 62i in C. Collard et al. Euripides:
Selected Fragmentary Plays II [Oxford, 2004] 87) regard as an interpolation from Euripides
Alexandros.

The strophe and antistrophe have to be considered together, as they are by West
Studies 80-82. I will here discuss only the last two lines of each, having nothing to add
to Wests treatment of the first two (except to note that I have adopted his
in
280).
At the end, Boeckhs transposition in 289 enables us to keep the excellent
in 283, and has rightly won general acceptance7. In 288, taken on its own,
nothing needs to be done except to transpose
; this gives sound metre
and blameless sense. The adjective
lacks a defining genitive, but so it does in Cho.
247 and 794, unless we make it share one uncomfortably with a word meaning son
(
); there, as here, the context shows which lost loved one is meant. Also
relevant is Soph. Trach. 563. Here
is usually understood as meaning as his wife,
a sense not otherwise attested before the last decade of the century (Eur. Or. 929; IA
397, 807). But the passage is a clear echo of the last lines of the preceding choral ode
(Trach. 529-530)8. There it was said that after helplessly watching Heracles and
Achelous fight for her, Deianeira departed from her mother, like an abandoned calf
(
and note that
can also mean orphaned: Soph. OC 1719, Pl.
Laws 927d). Here she speaks of herself,
, following
with
Heracles: as previously we saw her parted from her mother, so now we are reminded of
her being sent away with Heracles by her father. She was indeed, in effect, bereaved
of both and she can hardly even be said to have gained a husband in exchange. To
understand
here in its traditional sense powerfully reinforces the theme of
Deianeira the unprotected, unsupported 9. I conclude that in early tragedy
could
mean simply bereaved and did not need to govern a genitive specifying which loved
one had been lost; after about the middle of the fifth century the word for a time
dropped out of use entirely, to be revived by Euripides in his last years in the new sense
wife, as if it were a derivative of
Quite possibly this innovation was inspired by
Trach. 563 itself, misunderstood by Euripides as it has been by so many modern
interpreters. The scholia, incidentally, though their interpretation of the sentence as a
whole is an impossible one, do take
to mean
As for the much-criticized
in Pers. 288, it is to be understood from the Persian
point of view: the husbands and sons of Persian women have been killed in a futile
cause10. It is true, as West 82 says, that Persian men too have now lost their sons, and
that we have been reminded of this at 63 and 245; but a specific reference to women
here will direct our attention to the Persian woman on stage, who has been silent since
7

K. Sier, Hermes 133 (2005) 412-3, keeps the transmitted text in 289 while drastically emending
283 (replacing
by
) because he thinks that 284 (
) must be an immediate reaction to hearing the name of Salamis. This requires
him to accept an otherwise unmotivated metrical pause between
and
, and to
suppose that an exclamation of the form
+ exclamatory genitive, thoroughly at home in tragic
lyrics (cf. 928, Cho. 1007; Eur. Hipp. 814, Hec. 182, Supp. 847, HF 899, Hel. 211-2), got into the
text by accident. In any case, the Messenger has himself uttered the name of Salamis not very
long ago (273); indeed, if we accept a transposition (interchanging 272-3 and 278-9; first
proposed by J. Stavrids, Quelques remarques critiques sur les Perses dEschyle [Paris, 1890] 1114) which Sier himself in the same article champions on cogent grounds, the name was heard in
the last sentence he uttered before 284.
8
As was pointed out by D. Armstrong, BICS 33 (1986) 101-2 though he takes the view, wrongly
I think, that
here is to be taken as ambiguous between the older and the later sense.
9
Literally unsupported, at the time she is referring to, by Heracles; when we hear of Nessus
carrying Deianeira across the river on his shoulders (564), we are entitled to wonder why
Heracles entrusted his bride to the centaur rather than carrying her himself on those shoulders
which had once borne the weight of the sky.
10
So H.J. Rose in his commentary (Amsterdam, 1957): It means, I take it, that all this distress
and loss has been for nothing, since Xerxes is defeated.

5
the Messenger entered (cf. 290-2) and who does not yet know whether her own son is
among the dead. The male chorus, throughout the play, in the most marked contrast to
the Queen, never once grieve explicitly for their own sons, as Aeschylus could easily
have made them do, but only for the Persian nation as a whole.
Hence 288-9 is best read as
(2ia cho ith) Can we, then, restore 282-3 to correspond? If we accept that
is sound, we will need to emend
so that it
will scan x x and include the necessary (West Studies 81) mention of the
gods as subject of the clause. The first three words pose no problem, provided we take
as neuter plural (desirable in any case to provide
with an object) rather
than as a (spondaic) adverb. The favourite way to bring in the gods has been that of
Heimsoeth, to insert a monosyllabic <
in place of the augment of
; a further
insertion will then be required in 282, e.g.
(transposed from 281
where it ruins the responsion)
(Page). West loc.cit. rejects this unmentioned,
presumably because, like Hermanns emendation of 280 which he discusses, it involves a
breach of Porsons Law; in any case
is better explained as a gloss (ibid.)
The simplest solution for 282-3 is to remove the augment of
and then insert a
single word not before the verb but after it and for this word to be not
but
, which serves as its synonym in 724, 811, 1005, Seven 77, 96, 173, Supp.
217, 893, 922, Ag. 182, Cho. 214, Eum. 920, 1016, etc. That
is used in the
paraphrase offered by the -scholia proves nothing whatever; as the next word,
,
shows, the scholiast was working with a text that had already suffered loss. Hence read
in 282-3:

(4)

Persians 548-550 ~ 558-560

548
549
550
558
559

del. Porson
om. GF
codd.:
Blomfield
D+ Lc (et sic codd. omnes in v. 560):
om. Qac(?) GF:
om. Vac(?) :
T
del. Brunck

Tricl.:

cett.

Bruncks deletion in 559 restores both syntax and responsion, and can safely be
accepted11. We are then left with the failure of responsion in 548 ~ 558 and an
11

Blomfields minor emendation in 549 should probably be accepted also, though it is not strictly
necessary. In the three tragic passages in which forms of
are metrically guaranteed (270,
Supp. 547, Eur. El. 315), they are always corrupt in all or part of the ms. tradition; in the two
passages in which forms of
are metrically guaranteed (Eur. Cycl. 443, Ba. 1168) for each
of which only a single manuscript survives they have been preserved unscathed.

6
apparent surfeit of particles in both lines. 558 as transmitted would have been entirely
satisfactory were it not metrically incoherent. Triclinius transposition gives acceptable
metre, but
is not securely attested in tragedy12. Maas proposed
, and as
2
so often his suggestion was adopted by Gilbert Murray ( Oxford, 1955); it has no merit
is too weak a pronoun to carry a load of one adjective, let alone two (
). The omission of
in family , though probably accidental, is likely to be
right; it was inserted because a connective was thought to be needed. Its removal
leaves the stanza beginning with four iambic dimeters, like Seven 989-992; Supp. 808812 ~ 817-821, though corrupt, seems to begin with five.
In that case, 548 must sacrifice one of its particles. H.D. Broadhead (Cambridge, 1960)
ad loc. finds
redundant, saying that the Chorus does not lament because the whole
of Asia is groaning; but
may be explaining, not the preceding sentence as a whole,
but only its last two words the statement that the fate of the departed is truly a
theme for mourning far and wide (
). The really suspect particle in
this line is
, which Denniston (GP 364) and Broadhead both struggle to explain (and
on which Belloni makes no comment at all). I suggest that
be deleted and
transposed to fill its place (
). Either
has wandered here,
via the margin, from 550 (where it must at some stage have been competing with
before the two particles settled down to joint occupancy), or it owes its presence to
directly below. When
arrived in 548 it bumped out
, which was
eventually restored in the wrong place;
would go well with
(cf. Denniston GP
205).
(5)

Persians 858-860

plerique:
plerique:
Isscr O2
ac
ac
O Q:
A :
s
(
M
Tricl.) codd.:
Pallis

L:

Nd2 (
vel
Keiper
West:

etiam Mac)
plerique:
(sic)

numeri 4da 6da, ut docet stropha (852-4)

M

Wecklein)

12

See J.D. Denniston, The Greek Particles2 (Oxford, 1954) [henceforward Denniston GP] 532.
Verralls conjecture on Eur. Ion 1099 has long since vanished from apparatuses. Denniston either
missed or ignored the badly corrupt Eur. fr. 1019 (
).

13
This passage consists of two statements conjoined by
In the first statement West,
Belloni, and E.M. Hall (Warminster, 1996) are highly likely to be right in accepting three
minority variants and reading
I am concerned here
with the second.

I have quoted the scholia in full above to show that, while for the most part the
scholiasts were merely thrashing about in a desperate attempt to make sense out of
nonsense, they were sure of two things. One of these things, but only one, they could
fairly easily have gathered from the context, namely that (as modern scholars also
agree) the whole antistrophe is about Persian military activities under Darius and nothing
else. The other thing they were sure of is that the corrupt words of 859-860 had
something to do with the sacking of cities (note the phrase
which appears in both versions of the note). The capture of cities, to be
sure, is referred to at the beginning of the next strophe (865), but their sacking is not
mentioned anywhere in this entire ode. The presumption must be that this feature of
the scholia originates from a time when there was mention of the sacking of cities in the
poetic text itself at this point. The root
/
is used several times elsewhere in
this play with explicit or implicit punning on
(65, 178, 348, 714, 1056; in 103-8
the punning goes in the opposite direction).
Being coordinate with
, the verb concealed by
must be imperfect
and first person plural, and since
will not fit the purely dactylic metre, it can
hardly be other than
(Pallis). The transmitted
will not fit in before this14,
and must be a corruption or an intrusion; perhaps the whole phrase
was
originally part of a marginal quotation from another play. It will then be necessary to
suppose that a word has been lost at the end of the sentence, and
utterly
would give appropriate sense (cf. Eur. Hec. 887, Pl. Rep. 421a).
As to the preceding words, on the argument here being pursued we need a mention of
cities, and hence
(Peiper). That leaves only
, which seems to be a
hapax; if it is sound, and if the present proposal is otherwise on the right lines, it will
have to be understood as a metri gratia substitute for
Broadhead too has taken seriously the scholiasts interest in city-sacking, but argues
that since Xerxes too sacked cities (or at any rate one very important city) this on its
own would not be a point of contrast between him and Darius. Such a contrast, he
argues, is provided in the interpretation of the scholia: Darius, when he sacked cities,
behaved in accordance with
and did not, like Xerxes, destroy temples15.
Broadhead cannot find a restoration along these lines which will satisfy idiom and metre;
his key phrase,
, occurs (as
) only once before Philo (Pl. Laws
647c), and there it does not mean in conformity with law but with the help of law. If
this was what the chorus meant, we would have expected them to say something like
13

Unless with Page we emend

away (he prints
, citing in support the reading
[sic] of a
ms. ( ) which is most unlikely to have preserved the the truth alone see West Studies 324-330).
West was perfectly justified in making no mention of s reading, or of Pages conjecture, either in
his apparatus or in Studies 90.
14
Unless one sacrifices the dactylic metre which A. Sidgwick (Oxford, 1903) was prepared to do,
adopting the variant
in the strophe and scanning it (a licence which, as it happens, is
attested in tragedy for
, e.g. Eur. Hipp. 170, but never for
).
15
Or tombs, adds the scholiast; but he has taken his eye off the ball Xerxes is nowhere in this
play said to have desecrated tombs. Probably the scholiast is thinking vaguely of the Greek battlecry of 402-5, which calls inter alia for the liberation of
, and of the juxtaposition of
temples and tombs in the denunciation of city-sackers in Eur. Tro. 95-97.

which Broadhead doubtless, and rightly, thought was too far from the
paradosis to be a credible emendation. In any case, the chorus are not professing here
to be illustrating the contrast between Darius and Xerxes; they are professing to
illustrate the felicity of Persia under Darius (852-7), and recent disasters under Xerxes
are not mentioned till the final lines of the final epode (903-7).
(6)

Persians 948

I do not know what

is doing here, and most editors and translators seem not to know
either, since they treat the passage as if
were not there. An exception is Seth
Benardete16, who distinctly over-translates it (Again a wailing filled with tears Ill cry;
that would require
). Rather,
ought to be contrasting the sentence, and its
first word(s) in particular, with something that preceded, and this would be quite
inappropriate; this sentence is in fact saying the same, in different words, as the
previous sentence did. What is needed is , placing emotional emphasis on the verb
(Denniston GP 214-5): the chorus will not merely lament (
944, sc.
), they will lament loudly.
(7)

Seven 274-8

Eteocles vows to all the gods of Thebes

{ }
{

276 del. Ritschl

277
O
Y
Hutchinson
278a habent soli M

277
278a

I
I Rb: om. X
Ms et plerique:
M I:
loco verborum corruptorum infinitivum desiderari monet
I

Pac

aac,

tantum Q

: om. cett.

The text of this passage is well discussed by G.O. Hutchinson (Oxford, 1985) ad loc.,
who concludes that
is, or arises from, an elucidation of
, and has displaced a half-line which included a future infinitive. West thought of
keeping
and adding
; I have not been able to trace this verb. A
compound of
would be appropriate, cf. Eur. El. 898, and one might think of
and to fix up, though I have not then found a good way to complete
the line.
(8)

Seven 345-7
{

16

In D. Grene and R. Lattimore ed. The Complete Greek Tragedies: Aeschylus II (Chicago, 1956)
80.

345
347

(MI:
cett.) codd.:
Hermann:

Hermann:
Wilamowitz
Paley: numeri ut vid. 2cr

For
I suggest
. This adjective occurs otherwise only in Cho. 365, also in
lyrics, and referring to Greeks killed in the Trojan war. It is possible that the sequence
could be vulnerable to omission before

(9)

Seven 363-5

numeri, si fides strophae: tr lec / 2tr / lec

There is no significant variation in the mss., except that Triclinius (in T only) makes the
facile metrical correction
. Hutchinson and West both rightly find suspicious the
similarity between
and
: I find suspicious also the similarity of
meaning between
and
We need a verb, or equivalent, and the M-scholium
(364a Smith),
may well
indicate what that verb should be (hence
West, though this
involves keeping a word which the proposer himself17 has just declared to be probably a
dittography while deleting another,
, which has nothing objectionable about
it). If we remove
, taking it to have originated in one or more glosses or
variants, we must supply : perhaps then
or
will
endure (LSJ
I 4). The tense is future, not present, because at the time being
described the women are just being led away captive (326-335); their forced submission
to actual concubinage still lies in the realm of future expectation (cf.
367) at
present the enemy are much too busy killing the male population (340-1, 346-350),
seizing plunder (351-5) and setting fire to buildings (323, 341-2).

(10)

Seven 576-9

576
B Dsscr
Mazon

vel sim. MA:

I
Wsscr
):

577
Schtz
578 versum del. Murray

vel sim. cett. (

Dsscr,
Aldmg:
Francken:
:
Burges
van Herwerden

There can be little doubt that

is correct at the end of 576; indeed, this word
may well lie concealed in the meaningless
(or whatever) that precedes (on all
17

In his apparatus note to 364.

10
this see Hutchinson). If so, the transmitted letters may be no good guide to what
originally preceded
: but
indicates that Amphiaraus is being said to have
treated Tydeus (571-5) and Polyneices in a parallel way, and Franckens
(cf.
Prom. 595) cannot be far from the sense (
, despite its closeness to one
medieval variant, is unlikely, since it could not form a parallel with Amphiaraus
treatment of Tydeus, to whom he cannot have gone over since Tydeus station at the
first gate was not adjacent to Amphiaraus at the sixth).
What of 577-8? Clearly, in the first place, what Amphiaraus is here doing is drawing
attention to the meaning of Polyneices name and its appropriateness to the great
strife that he has caused. Secondly, it is unlikely that Aeschylus wrote
twice,
merely to serve as object to two conjoined participles; so either the word is corrupt on
one of its two occurrences, or else there has been interpolation. The only remotely
plausible suggestion for getting rid of the repetition of
has been Schtzs
in
577, adopted by West; but as Hutchinson shows, it would not provide an appropriate
sense18. Rather, we must posit interpolation.
It is tempting, with Murray, simply to delete 578, whose omission leaves perfect sense.
But
(cf. Aesch. fr. 350.1, Soph. Tr. 791) is not a word that an interpolator
would be likely to use, and
is suspect because Polyneices name is the
topic of the passage and if it was not originally mentioned in the text it might well soon
come to be written above the line as a gloss (
would have been added from 569,
571, 620 and 641 to fill out the verse, once the name had been mistaken for part of the
text).
On any view,
presents a serious problem. If there has not been
interpolation, what is the true reading of these words, and what do they mean? If there
has been, how did the phrase get into the text? There has been no satisfactory answer
to either question. Ms paraphrase
might be thought to point to a
reading
(Groeneboom), but the scholiasts solution to the enigma of
was
simply to ignore it. As to interpolators, they do not arbitrarily insert nonsensical
phrases; but if a copyist thinks there is a gap in the text, he may well fill it with words or
letters which look, from their position on the page, as though they might be meant as
part of the text, even if he does not understand them and perhaps that is what
happened here.
and
may have been separate scraps of an annotation.
But I am far from happy with any explanation that has been offered of this phrase.
Suppose, then, we do delete
and
, and further assume
that one of the two occurrences of
is a duplicate of the other. We will then have
removed a complete line, and a small adjustment to what remains yields
19
. Aeschylus would thus be saying that Amphiaraus called out to
Polyneices turning his name upside down and dwelling on it (the meaning of
in the two passages cited above). What would turning his name upside
down mean? The scholia gloss
as
unfolding, unrolling, and
18

C. Collard, AC 64 (1995) 185-6, comparing Aeschines 1.132, thinks that haughty and distancing
contempt is an appropriate attitude for Amphiaraus to adopt. But Amphiaraus is not, like the
man of whom Aeschines is speaking, despising his addressee as his social and/or intellectual
inferior: he, the man with the blank shield, the only one of the Seven who utters no boasts, would
be the very last person to do so. Rather, he is condemning Polyneices, as he did Tydeus, on moral
grounds.
19
Another three-word iambic trimeter, in a play that already contains more of them (thirteen) than
any other tragedy. See W.B. Stanford, CR 54 (1940) 8-10; M. Griffith, The Authenticity of
Prometheus Bound (Cambridge, 1977) 91-92. The others are 19, 29, 72, 243, 431, 449, 464,
496, 541, 614, 621, 635 and 798. Prometheus Bound has nine such lines; next, significantly,
comes Euripides Phoenician Maidens with eight.

11
Rose suggests that Aeschylus metaphor is taken from the act of opening out a papyrus
roll and laying it on its back. While there is no direct evidence that
could
bear this meaning, it is not a very obvious one for an ancient commentator to have
dreamed up, and it may well therefore have been current in the language of the
commentators time and perhaps also in the language of Aeschylus time. If the
scholiast is right, what is being said is that Amphiaraus was disclosing the significance of
Polyneices name which is, of course, man of great strife.
While not wishing to exclude this explanation, I would like to put forward another. This
is that
means inverting the name in the sense of reversing its pitchpattern. It is well known that when an oxytone adjective is used as a personal name,
there is a strong tendency for the accent to be moved to the recessive position (i.e. to
be placed as early as the rules of the language allow); thus to the adjectives
20
and
(Cho. 406) correspond the personal names
and
.
I suspect that it may have become a popular jest to address a person with his name
accented as though it were an ordinary vocabulary word, e.g. addressing a Glaucus as
blue-eyes or a Theorus as
tourist with the pitch rising instead of
falling towards the end of the word and that thus here the audience would readily
understand that Amphiaraus had addressed the son of Oedipus not as
but
as
man of great strife21. It is not impossible that in Seven 658, where
Eteocles in his turn emphasizes the appropriateness of his brothers name, the actor was
instructed to pronounce it
; so too maybe at Eur. Phoen. 636.

(11)

Suppliants 207

This was very reasonably obelized by Page22; but I am surprised that no one has
proposed the simple emendation
(though Bothe suggested
). The
meaning is may victory attend this ploy, may our stratagem be successful, the
stratagem being that of occupying the shrine of the
with their suppliant
insignia. For
or
in this sense cf. 459, 462, Ag. 1582, 1609, fr. 373;
and for the form of the sentence, cf. 951

(12)
20

Suppliants 330-2

See P. Probert, Ancient Greek Accentuation (Oxford, 2006) 298-300 and A New Short Guide to
the Accentuation of Ancient Greek (London, 2003) 112-3 (where she actually cites
as
an example of the phenomenon).
21
In Seven as we have it this adjective is in fact used, in the plural, at 830; but Verrall was
probably right (see R.D. Dawe in Dionysiaca [n.3] 88-89) to regard 822-831 as spurious.
22
Various attempts have been made to defend the transmitted text. W.J. Verdenius, Mnemosyne
43 (1990) 429, seems to understand the words as meaning let there be [i.e. make sure you
have] command of a means, referring to the altar at which the Danaids are to assume a suppliant
position; but nothing in the context specifies the end to which this means is to lead, unless
indeed 209, with its
, is to be placed before 207 (not one of the more popular
among the innumerable transpositions that have been suggested). G. Liberman, SemRom 1
(1998) 246-7, takes
to be equivalent to
(citing as a parallel
Eur. Hec. 883 where, however, the presence of the dative
makes all the difference) and
this to mean take possession of your means <of safety>, i.e. of the altar. This is to treat a
script composed for a mass audience as if it were a cryptic crossword. P. Sandin in his
commentary (Gteborg, 2003) tells us that cunning and plans are of little value unless there is
strength to carry them out, but does not explain the force of the genitive
(and in any
case Danaus plan to secure asylum does not depend on strength at all for its effectiveness).

12

331

Robortello, Turnebus:

332

Turnebus

West Studies 142-3 takes

to mean an old family connection23;
but, as was pointed out by H. Friis Johansen and E.W. Whittle (Copenhagen, 1980) ad
loc.,
means a family connection by marriage (whence
, the Attic word for
any male affine), and that is absurd when the Danaids have just spent some thirty lines
proving that they are kin to the Argives by blood. Rather,
must mean a
marriage alliance within the family, i.e. the marriage with their cousins from which the
Danaids are fleeing; and therefore it must be the object, not the subject, of the participle
in line 332.
A subject for
is badly needed, as Johansen/Whittle and West agree, and Schtz
inserted
after the infinitive, but that is not necessarily the only possible place for this
pronoun.
I wish to commend again a proposal considered, but not adopted, by Johansen/Whittle24:

For who ever supposed that I, on such an unexpected flight, should land at Argos,
utterly rejecting with disgust [lit. spitting away] a marriage-tie with my kinsmen through
loathing of the marital bed?
The sense is exactly appropriate, and the corruptions not difficult.
is an
Aeschylean favourite, but is confined to his later plays Supp. (594, 692, 781) and the
Oresteia, in which it is extremely common (as it also is in Prometheus Bound and, to a
lesser extent, in several plays of Sophocles). For the sense in which
would be
used here compare, in Aeschylus, Ag. 1192, Cho. 197, Eum. 191, 303 again, all in late
plays.
(13)

Suppliants 405-6

405

M: corr. Victorius

The oddest thing about this passage is its scholium in M (codex unicus), which seems to
bear no relation at all to the text:
Probably this should
23

Sandin too takes

as subject of
and makes it mean, at least initially, a
blood-related grief or, as one might reasonably paraphrase, your grieving kindred.
Unfortunately, in post-Homeric poetry
means specifically the grief of mourners, and the
Danaids, while they have experienced many afflictions, have never (so far as we are informed)
experienced the affliction of a bereavement. Being bullied by male cousins (Sandin 179) is not a
24

is Johansens conjecture,
and obelise

is Whittles. In their text, they let

stand

13
be disregarded, as being an interpretation rather than a paraphrase. Valckenaer25
managed to come fairly close to it with
but only at the
26
cost of introducing a verb not otherwise found in Attic poetry . Sidgwicks
gives good sense (in what way will doing the right thing cause subsequent grief?),
but quite apart from the fact that an adjective
is not known to have existed, it
is uncomfortable that the subject and complement of a verbless sentence are separated
by a four-word genitive absolute. H. Friis Johansen, SymbOsl 50 (1975) 28-29,
proposed
(Johansen)
(Headlam). In every respect but
one this is fully satisfactory. The Danaids are constantly urging Pelasgus to consider the
consequences of accepting and of rejecting their supplication. He who respects
suppliants will never fall into want (362-3, as restored by Headlam on the basis of the
scholia). Pelasgus should guard against pollution (375) and beware the wrath of Zeus
Hikesios (381-6, 427). The effects of his decision on his children and house will be
lasting (433-7). Hence What future pain will you suffer, if you do what is right? is an
appropriate sense; and since there is no room for a future tense, this would have to be
expressed in the potential optative, here in its short form (this is not otherwise found in
contract verbs in genuine Aeschylus, but it appears in Prom. 978 and Soph. Trach. 1235,
OT 1470, Phil. 895, 1044, OC 507). The one difficulty, which led Johansen and Whittle
in 1980 to abandon the proposal, is that the absence of
with the potential optative
cannot be convincingly defended. But this difficulty is one that is very easily solved:
read

(14)

Suppliants 830

This is the longest piece of continuous text that M offers in the desperately corrupt
passage 825-835, but it makes neither metre nor sense. West (see Studies 156), taking
a hint from Turnebus who saw in
a corruption of some form of
, boldly
prints in his text (I have no doubt that Aeschylus wrote )
here I see the beginnings of troublesome violence for my protector [i.e.
Pelasgus]; but while the Danaids, once promised asylum and protection, certainly
express their gratitude to the Argives and pray for their welfare, they never once
elsewhere show any appreciation of the dangers to which, for their sake, the Argives are
exposing themselves, and from the time when they first learn that the Egyptian fleet is
in the offing (710-733) to the time when Pelasgus comes to their rescue (911) they have
thoughts only for their own peril.
We can get nowhere with the line unless we make, at least provisionally, some
hypothesis about its metre, and West is probably right to take it as dochmiac. We
should also note that M leaves a gap before
, thus treating these words as a
separate verse which tells somewhat against any suggestion of moving these words
(whether or not emended) to an earlier position.
The likeliest explanation of the impossible
is not miscopying but loss. It is clear
that an ancestor of M was badly damaged or partly illegible hereabouts, and a plausible
restoration is
, which gives us both an object for
and a good dochmiac.

25

Before Maas; see P.J. Finglass, GRBS 49 (2009) 195.

Sandin also, at least tentatively, tries to provide a meaning close to that stated by the scholiast;
he suggests that
may be sound and mean hesitate, agonize, comparing Eur. Med. 996,
Hec. 214 for the use of
But he comes nowhere near showing that a person hesitating over a
decision can be said
: he cites only Cho. 1016, where Orestes has no more decisions to
make and is grieving over things that have already happened.
26

14

From the middle of the line we move to the end, where

is rightly objected
to by West because the troubles, though imminent, are not yet actual. The diagnosis
assimilation of endings is easy, the treatment almost equally so: read
(dative of
disadvantage). The sense is now beginning to become clear: I see men who have
performed a prelude to violent sufferings for me. The prelude is the landing of their
pursuers from the ship on the land (
), to use what are probably the
pursuers own words (826a, b: ascribed by West, with their context, to the Egyptians).
There remains
This could just about pass as a dochmiac, but it is of a
type unknown before Euripides last decade (see M.L. West, Greek Metre [Oxford, 1982]
109) and is surely corrupt; the simplest solution is to get rid of
and read

While I would not venture to say that I have no doubt what Aeschylus wrote, I will at
any rate, then, suggest that he could have written

(15)

Suppliants 872-3

How can the Herald be so confident? He knows that the Danaids and their father have
risked the dangers of a long sea-voyage to escape marriage with their cousins; he
thinks, rightly, that it will only be possible to get them on to the ship by the most brutal
use of force; they have repeatedly made it clear to us and the Argives that rather than
accept the marriages they would prefer to die, and while the Herald has not actually
heard them say so, there is no reason why he should suppose them totally incapable of
suicide. And yet he is sure that once in the ship, they will not jump overboard. This can
only mean that they will be physically prevented from doing so; in other words, that they
will be chained or tied up during the voyage. But it is asking a bit much of the spectator
to expect him to make this inference without assistance. I conclude that the Egyptians
accompanying the Herald are in fact holding ropes or fetters, and brandish them at 873.
The Herald is claiming the Danaids as his property, as if they were runaway slaves; cf.
918, 924 (where
alludes to the procedure of
/
27
exemplified e.g. in Lysias 23.9-12 ). Their fear that they might become
, with their cousins in the role of owners (335, 337; cf. 38
), was no exaggeration.

(16)

Suppliants 957-961

958 post 961 transp. Burges

959
Weil:
Turnebus:
27

Porson:

See D.M. MacDowell, The Law in Classical Athens (London, 1978) 80.

15

960
961

M:
Sandin

Blaydes:
Md:

Johansen:

vel

Whittle:

P. Sandin (Eranos 100 [2002] 150-2) has discussed this passage. He rightly argues,
comparing the words of Danaus in 1009-11, that Pelasgus must be offering just two
alternative forms of accommodation: the Danaids can either live with many others in
buildings that are public property, or in quarters reserved exclusively for themselves in
(one of) the kings residence(s)28. The Danaids ask to be allowed to consult their father
first before deciding (968-971)29; Danaus, when he comes, does not actually say which
option he prefers, but his anxiety about the preservation of his daughters chastity,
which takes up the greater part of his speech (991-1009, 1012-3), strongly suggests
that he would like to minimize contact between them and the Argives and that he will
accept the offer of exclusive accommodation in the palace. This in due course will
facilitate a takeover of power by Danaus (who already has a personal bodyguard: 985-8)
and also the plotting of the wedding-night murders.
However, as Sandin says, the text of 957-961 as transmitted fails to make it clear what
options, or how many, are being offered. Indeed the most plausible interpretation of it
is that three alternatives are put on the table, or more precisely two alternatives one of
which is subdivided: (i) public housing, not further specified (957); (ii) accommodation
in the palace (958) which may be either (a) shared (959-960a) or (b) exclusive (960b961). It is also possible, though more difficult30, to take
as referring to both the
options presented in 957-8, in which case the Danaids are being offered the choice of
public or royal accommodation and, in either case, the choice of it being shared or
exclusive. Sandin rightly seeks a text that will reduce this confusion to a clear, straight
choice.
Sandins solution is to take
as a demonstrative pronoun (there) rather than a
relative pronoun (where), and to take it as referring exclusively to 957, i.e. to the
public housing, while emending the slightly suspect 31
in 960 to
referring to the last mentioned place, the house of Pelasgus. This will not work.
Neither the Danaids nor the audience can be expected to divine that
is
demonstrative and not relative (especially when, as Sandin admits, the demonstrative
was rare), or that it refers only to the remoter, not the nearer, of its two available
antecedents, or that
, which normally refers to something relatively distant, here
refers to the last mentioned place. Sandin speculates that Pelasgus might
28

Johansen/Whittle, contrariwise, had argued that the superlatives of 962 (

) imply that more than two kinds of accommodation are being offered to them. This
need not, however, be the case; there is already a multiplicity of choice present (but not spelt out
in detail) within each of the categories of public and private housing, indicated by
(958) and
(959). Of each of the two types of accommodation, much more is available than
the Danaids need, and therefore, whichever category they prefer, they will be able to take their
pick of the best accommodation within that category.
29
Probably 975-6 should be transposed to follow 971 (West), and
(Schwerdt) read in 972; the
Danaids will then be saying that their father will need to consider where we should reside so as to
be well reputed, and spoken of without anger, by the native population, because even if a
country is friendly, everyone is ready to speak ill of people of alien language (cf. 496-8, 994-5).
30
Because if this was what Pelasgus wanted to say, he could have said it more clearly by not using
a linking adverb at all in 959 but instead starting an entirely separate sentence, e.g. with
31

Suspect because, in the words of Johansen/Whittle, pred[icate]s containing an adjectival

combined with a compar[ative] do not occur in Aeschylus, in Sophocles, in Aristophanes, in

Pindar, in the first four books of Herodotus, in Lysias, in Isocrates, or in the first volume of the
Oxford text of Plato: a very restrictive criterion, and a curiously selected corpus (why, for a start,
is Euripides omitted from consideration?)

16
disambiguate his words by gesturing in different directions, but this would not work
either. In this play the two directions that count, presumably represented by the two
eisodoi respectively, are that of the sea (from which the Danaids come at the beginning
of the play, and the Egyptians later on) and that of the city (to and from which Pelasgus,
Danaus, and sundry groups of armed Argives travel at various moments, and to which
everyone departs at the end of the play)32. Particularly with a war imminent, and the
enemy already having landed, any possible safe accommodation for the Danaids must
certainly be in the city, and indeed Pelasgus has just said so (955-6); therefore both
alternatives lie in the same direction, and gesture cannot be used to distinguish between
them.
It is very surprising that Sandin makes no mention at all of Burges transposition of 958
to follow 961, which solves the problem completely. The
of 957 will now find its
answering
not in 958 but in 960. On the one hand, says Pelasgus, there is plenty of
public housing (957) where the Danaids can live in well-prepared accommodation with
many others (959-960a); on the other hand, they can also live in exclusive quarters
(961), for33 he himself is housed on no mean scale (958).
(17)

Suppliants 999-1002
sc.

So M, except that it makes

into one word (corr. apographus Guelferbytanus
c.1495),
at the end of 1000 (corr. Robortello) and that
has been written
above the
of the penultimate word in 1002. Can anything be made out of this mess?
The passage has been examined by West Studies 165-7 and by Sandin, Eranos 100
(2002) 152-4. West takes 1000 as a nominativus pendens (probably rightly, I think34)
and, by some quite simple emendations, makes 1002 emerge as
also, at the same time, softening up the unripe, so as to madden them with
desire. There are some difficulties with this. It makes a distinction between ripe juicy
(female) fruit (the
of 1001), towards which Aphrodite directs the
desire of males, and unripe fruit, in which she implants desire leaving it quite unclear
which of the two corresponds to the Danaids. Nor is
an entirely suitable verb
here: when this verb has a living being as object, it normally refers to a deterioration in
the creatures health or vigour (see T.H. Talboy and A.H. Sommerstein in Sommerstein
et al. Sophocles: Selected Fragmentary Plays I [Oxford, 2006] 313-4, on Soph. fr. 693).
Sandin takes a different line, suggesting (after Scaliger and others) that
may
conceal
=
and prey, and that the meaning of 1001-2 is Cypris
proclaims the ripe fruit and the <
> prey alike to be a care of Love (
). This too is problematic. It makes a distinction between fruit and (animal)
prey which brings in an entirely irrelevant dichotomy between vegetarian and

32

I have analysed the plays significant movements in Aeschylean Tragedy (Bari, 1996) 159-162.
, as scholiasts often say (Denniston GP 169, citing inter alia lines 190 and 651 of
this play).
34
Sandin takes 1000 to be in apposition to
in 999, but he can cite no passage in which
is clearly meant to include birds.
33

17
carnivorous creatures35, when the whole point is that sexual appetite is the same in all
species alike; more importantly,
is not used to refer to a predators potential
victims but only to its actual victims a creature does not become
until the
predator actually takes it (as in 800-1 where
is paralleled by
); more importantly still, there is not one passage in any archaic or classical text
in which
is used in reference to an animal it always refers to a person who
becomes prey either to dogs, birds, etc., or to his human enemies, except once where
it refers to property at risk of being plundered (Od. 13.208).
We know from 999 and from
(1000, 1003) that what is said about animals in
1000-2 must have been broadly parallel to what is said about humans in 1003-5, and
West is right to infer from that passage that the tristich about animals must likewise
describe their susceptibility to mutual sexual arousal. When he goes on to say that
there should be some reference to the vulnerability of the immature in particular, this
is also correct provided that we remember that in the human context, immature
actually means physically nubile but not yet considered by society to be ripe for
marriage. Wests desiderata would be satisfied, at the start of 1002, by
(Wecklein)36, giving the sense Cypris advertises the availability of (
,
see Studies 166) juicy fruits when they are not yet ripe, hindering them from When
the maturing female is very young and tender (
998), she already, says Danaus,
has charms that attract the attention and desire of males; Aphrodite has contrived this,
and it makes it harder for the female to what? If
is the right interpretation of the
last three letters of 1002, the answer must surely be resist desire; and hence an
infinitive meaning resist, governing a dative, and scanning ()x,37 must be sought
for this slot. The transmitted letters assuming that the superscript in M, which does
provide us with an infinitive ending, is a true reading are ( )
Murray
proposed
remain as they are in face of desire, but the expression is
feeble and the use of the dative highly dubious. Going by sense alone, one might think
of
or
(for the former construction cf. Soph. Phil. 1241, for the
latter Eur. Ion 391, Phoen. 1268-9); but how does one explain the corruption? I cannot
find a solution along this line; perhaps others will.
(18)

Agamemnon 214-7

fere MVF (
M

fort. Mac sec. West, sed hoc negat Di Benedetto38):

T
codd.:
Bamberger
codd.:
West

I have omitted all controversial punctuation.

35

Which itself, as Sandin admits, is far from matching the distinction made in 1000 between birds
and beasts. Greeks knew all about birds of prey, and they are prominent in the imagery of Supp.
itself (e.g. 62, 223-6, 510) as are scavenger birds (751-2, 800-1).
36
The corruption of
to
probably had much to do with the fact that three surrounding
lines (1000, 1001, 1003) begin with
37
The optional extra syllable at the beginning caters for the possibility that
was elided.
If it had this extra syllable, the infinitive must have begun with a vowel; if not, with a consonant.
38
V. Di Benedetto, RFIC 120 (1992) 133-4.

18
The logic of Wests argument (Studies 178-181), as it proceeds from point to point,
seems irresistible; and yet he has reached an impossible conclusion. Punctuating after
, and printing in his text the above-mentioned conjecture in place of the next
two transmitted words, he makes Agamemnon say For they [the allies] are furiously
eager for a sacrifice to stop the winds and for a maidens blood; but Right forbids it.
May all be well!
This is the end of a reported soliloquy in which Agamemnon has worked his way to a
decision to sacrifice his daughter, in obedience to Calchas pronouncement that only thus
can Artemis be placated, the contrary winds ended, and the fleet enabled to sail for Troy
(198-202), to which his initial reaction had been to burst into tears (202-4). He is
certainly fully aware of the enormity of the action, which he describes in graphic words
(208-211). And yet he does it. In that case, as more than one scholar has pointed out
since 199039, the last consideration that he takes into account must be one that can
credibly be imagined as tipping the balance in favour of the sacrifice. In Wests text, his
last consideration is one that tells strongly against the sacrifice. Is there a parallel
anywhere in drama, or in Homer either, for a person deliberating on whether to pursue
course A or course B, ending his deliberations by stating plainly an obvious and powerful
argument in favour of course B, and then (with or without a verbal crossing of the
fingers, like
here) plumping for course A? I certainly know of none.
The text printed by Page (who follows M except that he accepts the one-letter
emendation of Bamberger, see above) gives the sense For it is
that they [the
allies] should with great fury desire a sacrifice to stop the winds and a maidens blood.
May all be well! Whatever problems this text may present (and we shall consider these
forthwith), it does at least make Agamemnon end on a note that harmonizes with the
action he proceeds to take, and thus serves, as Wests text does not, to tell us what has
caused him to take that action rather than any alternative: the fact that Agamemnons
allies41 are fiercely eager for the sacrifice (because they are eager for the war: 225-6,
230) and that if he refuses to carry out the sacrifice, and abandons or disbands the
expedition, the league of allies, of which he is the leader, will fall apart42.
What then are the problems supposed to beset the conventional text and interpretation?
(1) The alleged oddity of using
in reference to the armys feelings: the burning
question is not whether it is legitimate for them to feel like that, but whether it is
legitimate for him to do the deed (West Studies 179). But
need not mean
legitimate or right and proper; in Homer it sometimes means no more than natural,
the way of the world, as when Agamemnon says that it is
for men and women to
have sex (Iliad 9.134) it is obviously not legitimate for any man and any woman to do

39

Di Benedetto (n.38) 134; id. in J.A. Lpez Frez ed. La tragedia griega en sus textos (Madrid,
2004) 109; C.W. Willink, QUCC 77 (2004) 52.
40
I leave this word untranslated for a reason that will appear in due course.
41
That is, of course, the contingents (and their leaders) who have joined the expedition from cities
other than Argos. Willink (n.39) takes the reference to be solely to Menelaus; but for one thing
Menelaus in this play is not Agamemnons
but his co-ruler in Argos, and for another we
know, and Agamemnon knew, that Menelaus was not eager for the sacrifice on hearing the
words of Calchas, both the Atreidae burst into tears (202-4).
42
For
(213) does not mean failing in my duty as an ally (E. Fraenkel [Oxford,
1950] ad loc.); it means losing my allies (trans. H. Lloyd-Jones [London, 1979]). I have argued
the case for this interpretation in Aeschylean Tragedy (n.32) 364-5; it has the further advantage
that it does not require a surreptitious change in the meaning of
from the duties of an
ally (in which, on the Fraenkel interpretation, Agamemnon would have failed) to the allies as a
collectivity (who, on any interpretation, are described as eager for the sacrifice).

19
so under any and all circumstances43 or Eumaeus that it is
for a woman to grieve
when her husband has perished abroad (Odyssey 14.130). Similarly when a character in
an unknown play of Sophocles says that it is not
for anyone except the gods to live
without suffering (Soph. fr. 946) (s)he does not mean that it would be wrong to do so,
but that that is not how the world is. If it is natural for the army to be eager for the
sacrifice, then it can safely be assumed that they are eager for it; and if that is so, then
they are likely to be indignant against Agamemnon if he refuses to perform it, and this
may well lead to the untoward political consequences that he fears.
(2)
It has often been noted that this is a common gloss on
(and that
the scholia do in fact here gloss
with
). But that X is a common
gloss on Y does not in itself prove that in any given passage even a passage in which Y
also appears X is a cuckoo in the textual nest. The word itself is blameless: as
Fraenkel pointed out, it appears in Sophocles early Trachiniae (617) and in Euripides
early Alcestis (867).
(3) The variant
which appears with
in M and was adopted by Triclinius in the
copy he wrote himself. It has clearly been in or around the text for a long time, and
must somehow be accounted for and no one has satisfactorily accounted for it (I have
shown that Wests explanation is not satisfactory). Yet a very simple explanation is
available. The variant originally referred, not to line 215, but to line 245, where MV read
(doubtless the paradosis), FT correctly
; it was written to the left of the text
there, and found its way across to the right side of the preceding column in a late
antique or early medieval codex written, as many were, in double columns44.
Thus Agamemnons thought-process becomes clear. He is faced with the choice between
disregarding Calchas prescription (206) and staining his hands with his own daughters
blood (207-211); and he sees these alternatives as about equally bad (
, 206-7;
211). Then a consideration arises which tips the
balance. The army, he is sure, are passionately eager for war, and will not readily
forgive him for denying it to them by refusing to perform the sacrifice45: the alliance will
break up, and he will lose his position as the leader of Greece. He helps himself
overcome any residual doubts by using prejudicial language. He speaks of the
abandonment of the expedition as desertion of the fleet (
212),
making it sound like the act of a coward and a serious crime and commentator after
commentator has been taken in by his spin; in fact it neither is nor ever was cowardly or
criminal for a commander to abandon or discontinue a military enterprise when its
material or moral cost has clearly become too high to justify. (It might well, of course,
be imprudent, as many a prosecuted Athenian general could bear witness46.) And his
use of the word
, even though he is not actually using it in a moral sense (and is
43

And one cannot suppose that Agamemnon, more suo, is arrogantly assuming that for him it is
legitimate, because the skilful speaker Odysseus, who in reporting Agamemnons words to Achilles
suppresses the tactless conclusion of his speech (9.158-161), sees no harm in repeating this line
(9.276) with only a slight change of form (see J. Griffin [Oxford, 1995] on 158-161, 276, and
300).
44
Cf. perhaps Eur. Hipp. 867 where the words
have found their way into the
text (in all mss.) from 821, displacing some genuine words and creating a meaningless sentence.
I have argued in P. Thiercy and M. Menu ed. Aristophane: la langue, la scne, la cite (Bari, 1997)
281-2 = Sommerstein, Talking about Laughter (Oxford, 2009) 188-9 that a much-discussed
scholium on Ar. Clouds 889, which asserts that the Better and Worse arguments were brought on
in wicker cages, fighting like cocks, is actually a displaced and corrupted version of a note on
847 where two domestic fowls were in fact brought on stage probably in wicker cages.
45
In fact, had he taken that course, he would have saved not only Iphigeneias life but those of
thousands of those enthusiastic warriors too especially in Aeschylus treatment, in which
Agamemnon sails with a thousand ships (45) and comes home with one (650-673).
46
Including at least one under whom Aeschylus had fought (Hdt. 6.135-6).

20
applying it to the armys feelings, not his own actions), will still in all probabililty help
him to convince himself that what he is going to do is right; it is not for nothing that he
is made to make this the last word of his deliberations. Prestige and political expediency
triumph; Agamemnon puts on the yokestrap of necessity (219), abandons the
restraints of reason and humanity (220-3), and slaughters Iphigeneia.

(19)

Agamemnon 675-6

677

f:

Toup, cf. Hesych. 553.

As the text stands, the Herald gives no remotely adequate reason why he should be so
confident after a storm in which thousands have perished and which, so far as he
knows, only one ship got through safely that Menelaus in particular will return home47.
His remark about Zeus not yet being willing to destroy the family completely will not fill
the bill: it is only a hopeful guess, and in any case, even if the Herald has divined Zeuss
will correctly, that would not make Menelaus safe return significantly more likely, since
Agamemnon has come home and therefore, in the Heralds mind, the complete
destruction of the family has been averted anyway48.
The Heralds first statement about Menelaus certainly ought to be an optimistic one,
since it is linked by
to the wish that things turn out for the best: proposals like
(Sonny) and
(Hartung) can be ruled out. The best suggestion so far has
been that of Murray, who, following H.L. Ahrens, read
and posited a lacuna before
675 (though perhaps it would have been better placed after that line). To fill the lacuna,
Murray tentatively offered
. But while it will
make a great deal of difference to Menelaus chances of survival and of returning home
whether the gods above are concerned for his welfare, the feelings of the Argive people 49
can have no effect whatsoever on his prospects. (They cannot, for example, send out a
rescue expedition, since they have no idea of Menelaus whereabouts.) The Herald is
more likely to have suggested a reason why the gods can be expected to care first and
especially for Menelaus, and the most obvious reason had already been pointed out in
the Odyssey (4.569): that he was a son-in-law of Zeus. Aeschylus may therefore have
written something like

47

That, as Fraenkel shows, is what the bare

would have to mean.
Of course future events known to the audience but not to the Herald will prove him wrong about
this; but that is a matter of dramatic irony.
49
I presume that Murray intended
to bear the meaning people (as in Eum. 566, 569,
668, 683, 889) rather than army, since most of the Argive army now consists of corpses floating
in the Aegean (659-660); but in Agamemnon the word always means army, with one very
doubtful exception (547, where it refers, very confusingly, to the Argive home population, having
been used in 538 and 545 to refer to the army;
Heimsoeth), and it has been so used no less
than six times during the Heralds report of the disaster at sea (624, 627, 634, 639, 652, 670).
50
In the Loeb edition I proposed this supplement with
at this point; this was wrong, since
is not used with a direct (or even an indirect) object denoting the bride.
48

21

(20)

Agamemnon 838-842

What Agamemnon is trying to say here is clear. He has been discoursing sagely on
jealousy, and here tells the chorus that he knows all about it from his experience in the
war when, he says, Odysseus was the only one of the leaders who was consistently
loyal51. But the text we have hardly makes him say this, as the struggles of two careful
translators may testify:
With knowledge for I am well acquainted with that mirror, intercourse I may
pronounce image of a shadow those who seem most devoted to me. (Fraenkel)
I can speak with knowledge, for I well understand companionships mirror; its
image is a shadows, persons appearing very well-disposed to me. (C. Collard
[Oxford, 2002])
Both these scholars are forced to treat
as if it were
or
, and they also have difficulty with
both in effect making
Agamemnon say that those who seem most devoted to me are an image of a
shadow, when what is unreal is not the men but their supposed devotion 52. Even if the
passage could be understood by an audience, it would give an impression of incoherence
which would not surprise us if the speaker were the Herald, but is quite unlike anything
else in Agamemnons part.
In the Loeb edition I attributed the positing of a lacuna between 839 and 840 to H.D.F.
Kitto53. In fact it goes back at least as far as B.H. Kennedys second edition (Cambridge,
1882), as I should have gathered from Wests supplemental repertory of conjectures
(Studies 391). Kennedy proposed, exempli gratia,
This deals with one, but only one, of the two difficulties mentioned in the
previous paragraph; it still identifies the image of a shadow with the men instead of
with their pretence of loyalty.
Probably, then, more than a single line has been lost. In the Loeb edition I did not
suggest any specific restoration of the Greek text but offered what might be called a
diagnostic translation:
I can say with knowledge for I am very well acquainted with the mirror of social
relations that <the loyalty of friends is> a mere shadowy phantom. <I know

51

J.D. Denniston and D.L. Page (Oxford, 1957) ad loc. pertinently ask what fault Agamemnon had
to find with Nestor, Diomede, and many others, and suggest that the allusion is to the Nostoi
(Arg. 1 West), where these two are said to have sailed for home when Agamemnon wished to
remain at Troy and make sacrifices to Athena. But according to the Odyssey (3.162-4) Odysseus
too was among those who made this early departure, though he thought better of it at Tenedos
and returned to Troy. More likely we are meant to perceive Agamemnons judgement of his
colleagues as being grossly unfair. Fraenkel, for whom the Aeschylean Agamemnon was a great
gentleman (ii 441), skates over the whole problem.
52
As A.W. Verrall (2London, 1904) saw when he translated 840 (his 831) as the hypocrites
semblance of devotion to me as though Aeschylus had written
53
Form and Meaning in Drama (London, 1956) 23 n.2.

22
that many of the leaders of my army were really my jealous enemies,> though to
all appearance they were very friendly to me. Only Odysseus
I now offer a very tentative restoration of the text, in the hope that others may be able
to improve upon it:

(20)

Agamemnon 1005-7

If this had occurred in an astrophic lyric, it would probably have been judged sound.
The antistrophe (1022-4), however, shows that something is missing; its text is itself
uncertain at one or two points, but there is no reason to suspect interpolation, and no
serious doubt that the metrical scheme is

However, while something has certainly been lost from 1005-7, it is almost impossible to
determine where it has been lost. A seven-syllable lacuna could in principle be posited
at any of six places in the sentence (at the beginning, or after any of the first five words)
and has in fact been posited in at least four of these places. H. Weir Smyth (London and
Cambridge MA, 1926), G. Thomson (Cambridge, 1938; 2Prague, 1966), and West have
all favoured, and two of them have actually printed, H.L. Ahrens supplement of
after
This is good, but could, I think, be improved. One feels
that
could do with an additional descriptor of some kind, and that this is needed
more than
is: if the voyage was proceeding smoothly, and the reef is invisible, we
do not need to be told that the ship strikes it suddenly and unexpectedly55. I suggest
that a preferable restoration would be

As in the previous sentence (1001-4) it was the fittest, healthiest man who was
particularly vulnerable to sickness, here it is the rich man who is liable to be ruined at
any moment by the shipwreck of his fortunes and we will gather presently that he is
apparently rich enough to be able to jettison a substantial proportion of his wealth in
54

Cf. Eur. fr. 295 and trag. adesp. 535.

Besides,
is not found in any uncontroversially genuine work of Aeschylus (only in
Prometheus Unbound, fr. 195.4).
55

23
order to save the rest (1008-13), as Clytaemestra proposes to do in the (vain) hope of
buying off the daimon of the house of Pleisthenes (1568-76). The adjective
occurs twice elsewhere in Aeschylus, both times in anapaests (Pers. 3; fr. 96).
The whole image will be reprised by the Erinyes in Eum. 553-565, where the victim is
twice described as wealthy (554
, 563
) and
also as arrogant (561
the one who boasted <it could> never
<happen to him>).
(21)

Agamemnon 1472-4

>.

1474

Tr pc
1473
Stanley
T :
f
1474 disyllabum excidisse docet stropha (1454):
Page:
Kayser: <
ante

Murray:
Risberg: alii alia

The subject is the daimon of the house, which the chorus here half-identify with
Clytaemestra (as she herself will do at 1500-4); she it is, after all, who is standing over
the body. To try and recover the two syllables lost at the end would be useless
guesswork, says Fraenkel; all the same, I venture to suggest
Clytaemestra
has made it very clear that she does rejoice in her murder of Agamemnon (1391-2) and
Cassandra (1446-7), and in one of these passages she has used the verbs
(1391, 1394) and
(1394) in close proximity. This supplement also adds to
the condemnatory force of the choruss words: not only has Clytaemestra murdered her
husband and king, not only does she take pleasure56 in having done so, but she publicly
glories in the killing and in the pleasure she has derived from it.
(22)

Agamemnon 1649-53

1652
1653

Fraenkel:
Lobel:
f

f:

Lobel
Auratus:

1649 Aegisthum dicere inter omnes constat

1650 choro tribuit f: Aegistho
continuavit Stanley: praefecto satellitum dedit Verrall
1651 Aegistho
tribuit Fac: choro GFpc: in T nulla nota: praefecto dedit Thomson
1652
Aegistho, 1653 choro tribuit f: vice versa Stanley
The textual problem here relates not to the words of the script (for which the
restorations printed above are not now seriously disputed) but entirely to the attribution
of lines.

56

Quasi-erotic pleasure, at that; see my discussion of Ag. 1372-1447 in A. Willi ed. The Language
of Greek Comedy (Oxford, 2002) 154-7.

24
West gives the five lines to Aegisthus and the chorus-leader in strict alternation, thus
agreeing with the manuscripts for the first two lines and with Stanley for the last two; he
relies mainly (Studies 225-6) on the arguments of Denniston/Page. So far as lines 1652
and 1653 are concerned, these arguments are indeed decisive. It cannot possibly be
supposed that the chorus wear swords, so 1651 must be addressed not to them but to
the guards confronting them57; from which it follows that 1652 must belong to the
chorus-leader, 1653 to Aegisthus.
But what of 1650 and 1651? On 1650 West adopts Pages argument that
is more
appropriate as an expression of the chorus solidarity than as Aegisthus address to his
subordinates. But it is, and always has been, common for military commanders to
address their men with terms denoting or implying affection or protectiveness (lads,
mes enfants, etc.); in Xenophons Cyropaedia Cyrus can address his whole army as
(2.3.2) and his subordinate Abradatas uses the same formula when ordering
his men to charge (7.1.29). Sophocles has two choruses of humble sailors, both of
which are addressed by their commanders as
(Aj. 349, 406; Phil. 825). In
Aeschylus, the Persian Queen, whom the chorus hail with profound obeisances as wife
of a god and mother of a god (Pers. 157), repeatedly addresses them as
(162,
206, 231, 445, 598, 619). And so far as armed guards are concerned, Pelasgus in
Suppliants (954) tells the Danaids to take confidence from the escort of
Aegisthus in particular has every reason to adopt a studiously friendly attitude towards
the underlings on whose help he now depends (Fraenkel; emphasis mine).
On the other hand, as Medda rightly points out59,
is the word that Aeschylus
chooses in Cho. 768 to denote these same guards, and on its only other occurrence in
tragedy (Soph. OT 751) it likewise refers to the armed attendants of a ruler; the word
appears twice elsewhere in classical Greek (Xen. Anab. 6.6.7, Cyr. 2.2.7), both times as
a military term denoting a soldier or soldiers under the command of a particular
On this evidence,
would be far less inappropriate in the mouth of Aegisthus than
would be in the mouth of the chorus-leader.
Besides, what is
, spoken by the chorus-leader, supposed to mean?
What is the job that is at hand? If this is Aegisthus speaking to his guards, the
meaning is clear; he has just told the chorus, in effect, that if they want to oppose him
they will learn (
, as he might have put it) what are the consequences of
doing so (1649, cf. 1619-23), and he now, inexplicitly but plainly, instructs the guards to
inflict these consequences upon them. But if the speaker is the chorus-leader, there is
nothing to be done yet putting oneself in readiness to resist an attack whenever it may
be launched is not an
unless we are to suppose (which neither Page nor West
does) that the old men are being urged to take the initiative themselves and attack the
armed guards with their staffs.
Aegisthus, then, is the speaker of 1650 as well as 1649. Does he speak 1651 as well?
Stanley thought so, and so more recently has E. Dettori (Museum Criticum 21/22
[1986/7] 28-31). But it would be pointless to make Aegisthus give two successive
orders to the guards, only the second of which has any effect (why not just give the
57

P. Judet de la Combe, LAgamemnon dEschyle: Commentaire des dialogues (Villeneuve dAscq,

2001) 759, and E. Medda, Lexis 19 (2001) 46-50, suggest that in 1651 the chorus-leader is
addressing, not his colleagues, but any Argive capable of fighting: it is not clear how the
audience are supposed to know this (contrast Eur. Or. 1621-4 where Menelaus explicitly calls out
to
), and even if they did manage to work it out, it would
leave them wondering why the appeal finds no response, either now or later.
58
So M; on Schtzs
, still adopted by West, all that is necessary was said by O.P.Taplin, The
Stagecraft of Aeschylus (Oxford, 1977) 233 n.3 and by Johansen/Whittle ad loc.
59
Medda (n.57) 37-38.

25

effective order in 1651?); nor, having attracted their attention once with
he need to do so again immediately afterwards60.

, would

The proposal to introduce in this scene an additional speaking character in the shape of
the captain of the guard was first made by Verrall61, and that in itself has probably cost
it some credibility; Verrall, moreover, gave the captain lines 1650 and 1653, which is
certainly wrong a tyrants
are an instrument in his hands, and should speak
or act only on his orders. But Thomsons proposal to have him speak 1651 62 deserved
better than the almost complete neglect that has been its fate 63. Aegisthus did not have
the courage to kill Agamemnon in person but delegated the task to Clytaemestra (cf.
1633-5, 1643-6); now we see that he does not even have the guts to give an explicit
order for the massacre of the Elders (and that is clearly what is envisaged, as
Clytaemestra perceives at 1654-6). He gives them a vague order which may be
paraphrased as You see what your duty is, and which their captain then translates into
a specific executive instruction. The chorus declare their intention not to yield, even at
the cost of their lives, and Aegisthus grimly assures them that it will cost them their lives
at which point Clytaemestra intervenes and, as always in this play, takes command of
the situation, dominating all the two dozen or so males present.
That the Captain only speaks one line is neither here nor there. In the next play we
shall meet a Doorkeeper speaking one line (Cho. 657), quite likely from behind a closed
door without even coming on stage64; a Servant speaking a total of eleven (Cho. 875884, 886); and, famously, Pylades who is constantly present with Orestes (at least up to
930) but speaks just three lines, though they are of enormous weight (Cho. 900-2)65.
(23)

Agamemnon 1672-3

, unde 1672

Canter, 1673

Auratus

So the last lines of Agamemnon are usually restored; but the absence of an object for
is worrying, and Fraenkel showed that it was indeed abnormal. Since
is dispensable66, he proposed
,
with
as options for filling the gap. This, however, as
Denniston-Page note, requires us to assume a complex and improbable process of
60

See Medda (n.57) 40-41.

Verrall (n.52) ad loc.
62
In his 1938 edition Thomson appears to have been under the impression that this had been
Verralls proposal; in 1966 he silently retracted this and claimed the credit himself.
63
Fraenkel gave it a mention, but clearly did not take it seriously. Denniston/Page, West, and
Judet de la Combe ignore it completely; Medda (n.57) 36 at least gives it the courtesy of a
rejection backed by some sort of argument. The only scholar I know of, other than Thomson and
myself, who has championed it in print is my former teacher A.D. Fitton Brown (CR 1 [1951] 1335).
64
So Taplin (n.58) 341.
65
There are eight speaking characters in Choephoroi, and four of them (the Doorkeeper, Aegisthus,
the Servant and Pylades) speak a combined total of 29 lines. Medda (n.57) 36 n.10 rejects
Pylades as a parallel because he is a much more significant character than the Captain would be,
and rejects the Doorkeeper as a parallel because he is not a significant character at all; he really
cant have it both ways!
66
Though the evidence is not sufficient to warrant Fraenkels claim that it is actually contrary to
Aeschylean usage.
61

26

corruption, and a simpler restoration would be

If
were lost,
would find itself next to
and might
easily be assimilated to its case, afterwards being placed before
either as a
metrical correction (it not being noticed that the line was still two syllables short) or
under the influence of the scholia.
(24)

Choephoroi 71-74

71
73

74

Stephanus:
M
M, quo servato
pro
(
H.L. Ahrens) in strophe
(68) coni. Sier:
Lachmann:
Tucker (hoc si verum
est, glossema erit
):
Risberg
Porson:
M
M:
Musgrave,
post Scaligerum Heath

Given that some corruption is certain here (over and above those which have been
corrected with general consent)67, there may well be suspicion of the two participles with
identical endings, so close to each other and neither coordinate nor in a clear
relationship of subordination. In addition, as Sier (n.67) has noted,
does not
normally mean stream unless words in the context make it clear that this is its sense.
A.F. Garvie (Oxford, 1986) ad loc. gives good reasons for adopting Tuckers
; this verbal root is found twice again in Aeschylus (Eum. 237; fr. 148) and
never in Sophocles or Euripides68. If
is right,
will be a gloss on
it, and may have displaced a word or words that would have provided the required
disambiguation of
, e.g.
(cf. Eum. 452
, also about the
cleansing of blood-pollution). The fact that
and
share five of their six
letters may have contributed to the loss of the phrase.
(25)

67

Choephoroi 160-3

I cannot include
among these consensus corrections, since K. Sier, Die lyrischen
Partien der Choephoren des Aischylos (Stuttgart, 1988) ad loc. rejects
on the ground that a
gnomic aorist (he claims) is not possible here (I do not understand why not).
68
Siers objection that
is ein literarisch unbezeugtes Verbum is utterly pedantic; its
derivative
appears in Eum. 237 and probably also in Aesch. fr. 148 (
cod.,
corr. Stephanus), and the lexical lemma
(Hesychius 678) must come from a literary
(and presumably a lyric) source. V. Citti, Philologus 146 (2002) 210-5, adopts Lachmanns
wetting, retaining the transmitted
in 68; quite apart from the dubious
credentials of
(which is clearly not what the scholiast read, and whose later meaning was
suffering pain, not causing pain), what is the object of
supposed to be?

27

160
161
162
163

anon.:
Robortello:
ante
M
:
M

M
add. Weil
(sscr.
)M
M: del. Paley
Blaydes:

So my Loeb text, which is identical to Wests except that I have adopted

at the end:
the scholium (
) cannot have been written to a text in which
swords as well as arrows were referred to as
An unsatisfactory feature of this text, however, is the juxtaposition (
) of
two unrelated phrases introduced by the same preposition in different applications.
Additionally, it may be noted that one does not brandish (
) a weapon while
actually using it in the work of Ares. I suggest
with a view to <warlike>
action (LSJ
B.III.2).
(26)

Choephoroi 423-455

Abnormis est dispositio et stropharum (7.8.9.9.7.8) et personarum (El. [429-433] Or.

[434-8] non interveniente choro. So Wests apparatus note on 434. Actually neither of
these two abnormalities is readily removable by any transposition.
Up to 422 we have had four cycles of stanzas, each comprising (i) choral anapaests, (ii)
lyrics by Orestes, (iii) lyrics by the chorus, (iv) lyrics by Electra; (ii) and (iv) of each
cycle have been in responsion with each other, and (iii) of the first and third cycles were
in responsion with (iii) of the second and fourth cycles respectively. At 423 this pattern
changes, the chorus leading off the fifth cycle with lyrics; this cycle contains six stanzas,
and these are distributed asymmetrically three to the chorus, two to Electra and one to
Orestes69. There is no way to arrange these six stanzas so that the dispositio
personarum will follow a pattern resembling that of 306-422; moreover, since 434-8
must necessarily have been preceded by at least one stanza referring to the degrading
treatment of Agamemnon after his death, it is certain that the previously regular
alternation between Orestes and Electra is broken at least once.
As to the dispositio stropharum, there are various logically possible ways of arranging
the stanzas that would count as normal, depending on whether we chose to compare
this passage with the earlier part of the kommos, with the regular pattern of tragic
choral odes, or even with the triadic pattern familiar in Pindar, but there are certain quite
stringent limiting conditions. 445-450 and 451-5 certainly belong together, since the
latter continues the advice to Orestes that began near the end of the former. 423-8
surely belongs at the beginning, since we would expect the cycle to begin with a choral
stanza, as all the preceding cycles did, and this is the only choral stanza in the present

69

M hereabouts gives no indication at all of who sings what, but the internal evidence is
unequivocal. The oriental lament of 423-4, accompanied by gashing and head-beating, whether
performed or merely recalled, suits the chorus of war captives (75-77) who sang of similar actions
when they first appeared (22-31). In 429-433 the singer is a child of Clytaemestra who knows
how Agamemnons funeral was conducted; so too in 445-450 where her identity is further
confirmed by feminine adjectives. In 434-8 the speaker is male and is vowing to kill Clytaemestra
with his own hands. In 439-444 he is being addressed, but not by Electra, since in 445 she refers
to the singers of 439-444 in the second person. Lastly, 451-5 picks up and continues 445-450,
but cannot be assigned to Electra, since strophe(s) 7 + 8 (lines 423-433) are divided between two
voices and therefore, in accordance with invariable practice in all parts of this kommos,
antistrophe(s) 7 + 8 (445-455) must be divided also.

28
cycle that is not a response to something said by another 70. And 429-433 must precede
434-8, as noted above, and must also precede 439-443 (which, with its
, adds a
new degradation to another or others previously mentioned), which in turn must precede
445-450 (whose opening shows that someone other than Electra has been singing about
the aftermath of Agamemnons death).
In other words, 423-433 (strophes 7 and 8) must come together, in that order, at the
beginning of the cycle; 445-455 (antistrophes 7 and 8) must come together, in that
order, at some later point; and 439-443 (antistrophe or maybe strophe 9) must
come somewhere between these blocks. The only stanza whose position cannot be
pinned down is 434-8 which is also the one that has always attracted the most
interest, because it is sung by Orestes and because in it he declares more specifically
than ever before his intention of killing Clytaemestra.
Denoting Orestes stanza 434-8 as 9or, and the choral stanza 439-443 as 9ch, the only
arrangements that satisfy the conditions set out in the two preceding paragraphs are the
following; in each case I give both the sequence of stanzas and the sequence of singers.
(a)
(b)
(c)

7.8.9or.9ch.7.8
7.8.9ch.9or.7.8
7.8.9ch.7.8.9or

Cho.El.Or.Cho.El.Cho.
Cho.El.Cho.Or.El.Cho.
Cho.El.Cho.El.Cho.Or.

Of these, (a) is the transmitted sequence; (c) was proposed by Schtz, was
subsequently favoured by Weil, Wilamowitz and Lesky, and has recently again been
argued for by Dawe71; (b) was mentioned in Gilbert Murrays apparatus (2Oxford, 1955)
as the view of unidentified alii72, but has since sunk from view, neither Garvie nor Sier
making any reference to it.
The nearest approach to a normal sequence is (c), if the whole passage is regarded as
one great strophic pair; but then we get a different kind of asymmetry, since the chorus
will have two-thirds of the strophe but only one-third of the antistrophe, and in
addition Orestes, who on this hypothesis sang the last stanza of this fifth cycle, will also
be singing the opening of the short sixth cycle73 (456-465: one strophic pair, each
stanza divided among Orestes, Electra and the chorus in that order). As to sequence
(b), it does nothing at all to cure either of the abnormalities with which we began.
I conclude that we cannot decide on the basis of structural considerations whether (a),
(b) or (c) is correct; our only guide can be the sequence of thought. We will bear in
mind, of course, that the one manuscript (and, at least ex silentio, the scholia) bear
witness for (a); but there are other passages in Aeschylus where it is at least highly
probable that lyric stanzas, or parts thereof, have changed places, sometimes across
more than minimal distances74, so that while, as always, the null hypothesis is that the
70

Sier (n.67) 155-8 nevertheless transposes 423-8 to follow 429-433. Since he retains the
transmitted order of the subsequent stanzas, this yields a strophic sequence 8.7.9.9.7.8, which he
admits is ohne genaue Parallel. Moreover, this emphatic assertion of the servants extravagant
grieving, coming after Electras statement that Agamemnon was buried
, will sound like a contradiction of it; the chorus, unlike Electra (445, 447, 449), never
say that they were shut up in the house at the time. And, as Garvie noted (JHS 110 [1990] 215),
434 cannot directly follow 428; so Siers transposition will necessarily entail at least one more!
71
R.D. Dawe, Eranos 97 (1999) 24-44, at 28-31.
72
I have not been able to track these nameless scholars down.
73
Hence Wilamowitz proposed a lacuna before 456.
74
See Dawe (n.71). Instances accepted by influential recent scholarship include K.O. Mllers
transposition of Pers. 93-101 to follow 114, Preusss interchange of Cho. 623-630 with 631-8, and
Westphals interchange of Supp. 93-95 with 88-90 (each of these being half a stanza only).
Compare also, in epirrhematic passages, Stavrids interchange of Pers. 272-3 with 278-9 (see n.7
above) and Oberdicks of Supp. 872-5 with 882-4.

29
transmitted text is correct, the presumption in its favour cannot be regarded as
irrebuttable.
The key line is 434, which links Orestes stanza to whatever preceded it:
a note which is then immediately struck again in the next line
, the syllables
being repeated in the same position of a verse
of the same structure. As
shows, this must be an immediate and highly emotional
reaction to an account of how Agamemnon was dishonoured. Garvie is right to argue
that it cannot follow 455, at which point two stanzas have passed without any mention of
the dishonouring of Agamemnon, and one without mention of the dishonouring of
anyone at all. Thus (c) is ruled out, and we are left with the choice between (a) and (b).
In this choice the expression
seems to me crucial, and Garvies discussion of the
issue makes no mention of it. Whether the line means your whole story is one of
dishonour (Garvie) or wholly dishonoured, you say (Collard), it is entirely out of
proportion as a reaction to the statement that Agamemnon was buried without the
participation of the citizenry and without mourning from his family. That was, no doubt,
an act of great dishonour, but nothing like as great as the murder itself, and hardly
sufficient to evoke the declaration of 435-8. Immediately after this, moreover,
according to the transmitted text, the chorus tell Orestes of what really is a great horror,
perhaps greater even than the murder the mutilation of Agamemnons dead body,
specifically intended by Clytaemestra, so the chorus say, to make his death unbearable
for you to live with (441-2)75 and to this he does not react at all. This has got to be
the wrong way round.
If we transpose 434-8 and 439-444, we get a coherent sequence. First Electra mentions
Agamemnons unseemly burial; then the chorus cap this by saying that she who
performed this burial (440) also mutilated the corpse; then Orestes, picking up the
choruss last line
(444), says that the whole tale (i.e. what
Electra has said combined with what the chorus has said) is one of utter dishonour and
that his own hands will make Clytaemestra pay for it. Electra then adds the further,
though milder, point that she herself was shut away inside the palace and not allowed to
take part in the funeral, and both she and the chorus urge Orestes to absorb all this into
his mind and to enter the arena with inflexible will (455). All then proceed to make a
joint prayer, first to Agamemnon (456-460) and then to the gods (462).
Does this transposition create any difficulties? Garvies defence of the transmitted
sequence against Schtzs transposition (our (c)) raises one or two that are also relevant
to the present proposal, to which I will presently add another.
(1) The subject of
(435) is more obvious if it comes immediately after [429433]. Garvie himself says that even after 455 the subject would not [be] hard to
supply; and with the sequence proposed here, Clytaemestra has been the subject of a
sentence occupying most of the immediately preceding stanza (439-442).
(2) The anaphora at 436f. echoes that at 431f. (and, we may add, all the four cola
concerned are metrically identical, even though the stanzas themselves are not in
responsion). The echo will still be there, if at a slightly greater distance, if the
transposition is accepted.

75

Does this mean that Clytaemestra hoped that Orestes, on learning of what had been done to his
father, would be unable to live with the thought of it and would commit suicide? (In fact, a
moment later, he will wish for death but on the condition that he has first killed Clytaemestra:
438.) Or are the chorus saying, ironically, that if Clytaemestra had been trying to make Orestes
determined to take revenge at all costs, she couldnt have chosen a surer way to achieve this?

30
(3) A point not raised by Garvie because not relevant to the transposition he was
arguing against: Electras first words in 445,
, are most obviously
taken to refer to 439-444, the only stanza in which the treatment of Agamemnon after
his death has been described by anyone but Electra herself. However, by the end of this
stanza (450) Electra is certainly addressing Orestes (so rightly Garvie), and she can
perfectly well be addressing him already at 445: he too has been talking about our
fathers death the depth of indignity inflicted on him and the certainty that the
perpetrator will pay for it.
I conclude that the best solution to the problems of sequence and coherence posed by
this passage is to transpose 434-8 with 439-444. The metrical pattern is indeed
unparalleled, but so is the composition of this kommos as a whole (Garvie). Like any
other possible arrangement, this one keeps Orestes silent for a considerable time, in this
case through five successive stanzas (antistrophes 5 and 6, strophes 7, 8 and 9); in
effect, he misses one turn to sing. It may be significant that in the last words he did
sing, he said that the remnants of the Atreidae were
and asked
(407-9); that suggests that he may be falling into a state of despair
and depression (which, as Garvie notes, seems to alarm the chorus, 410-4), and we may
be meant to infer that this is what causes him to remain silent when next due to sing
after 428. It is the account of Agamemnons dishonourable post-mortem treatment,
above all the bestial mutilation, that rouses him and makes him specifically confirm his
resolve; and in the next cycle he will be the first to speak (456), taking the lead as the
head of the family ought to. He was grieved that he had not been present to lament for
Agamemnons death (8) or to stretch out his hand at the
(9); what he had not
known was that no one else, not even his sister, had been allowed to do so either, and
that Agamemnon had been taken to his grave wearing a grisly necklace of his own bodily
extremities.
(27)

Choephoroi 785-7

785-6 numeri, ut vid. collata antistropha, 2cr lec76

The passage is thoroughly (if a little confusingly) discussed by Garvie, who offers various
suggestions. A good starting-point for further consideration is his rejection of any
restoration involving the root of
because [such a] prayer is altogether too tame
for an occasion [on] which extreme violence is demanded. This leads Garvie to suggest
(Page)
( ) (Bothe, slightly modified), followed by either
or
: the latter notion, as he points out, is thematic in the Oresteia, and
it provides
with an appropriate (understood) object. It would be preferable,
however, if possible, not to have to shift
and not to have to delete one of the two
occurrences of
This could be achieved if we were to read
(Bamberger)
grant that fortune may fall out well
for the masters of the house, who long to see the light. Garvies and Siers objection
that at 658 the
were Clytaemestra and Aegisthus is, as Garvie at least
evidently recognizes, not a strong one: when Orestes said that, both we and he knew

76

Though Sier (n.67) 246-7, 251-3 makes major transpositions in the antistrophe which change
the metre here to lec 2cr.

31
very well that the real
was himself77. But the corruption may lie deeper:
is a term much used by scholiasts giving what they consider to be the proper or
primary sense of a word78, and part of an annotation may well have been incorporated in
the text.
(28)

Choephoroi 802-5

x>

The recent consensus is strongly in support of Wilamowitzs lacuna here, since the only
alternative would be to make two deletions, of
here and of the whole of 815 (on
which see further below). Garvie is right to reject the usual interpretation of
as a
hortatory interjection on the grounds that Aeschylus does not use the plural form in this
way (neither indeed do Sophocles or Euripides)79 and does not follow up the singular
with an imperative80; but after
in its normal imperative use he finds it hard to see
how the sentence might have continued. How about something like
? It is true that the master Orestes is actually already in the house; but
he is not yet able to function as its master and is effectively still in exile, as he said at
252-4 that both he and Electra were even though Electra had lived all her life in the
palace (cf. also 336).
(29)

Choephoroi 815-8

815
816
817

Wilamowitz
del. Hermann
West
Heyse:

in rasura) M:

Wilamowitz

Bamberger

West discussed this passage briefly in Gnomon 59 (1987) 197, and more fully in Studies
254-5. His interpretation of
as he makes many things appear
different (sc. from what they really are) is not supported by any parallels; more
importantly,
(unlike
) always implies that what is seen corresponds to
reality that is, it means not so much cause to appear as reveal. West quite rightly
points out that the passage relates to Hermes qualities as an ally in an enterprise

77

Siers further objection against

that it would give
two indirect objects (the first
being
in 783), for which he could find no parallel in the language of prayer would not apply
to the text here proposed, in which
is governed by
78
It appears nine times in the Aeschylean scholia of M alone (Pers. 428; Seven 17, 251, 343, 857860; Ag. 65; Prom. 54, 429, 499) and many times more in those of later mss.
79
It is found three times in Aristophanes (Peace 469, Lys. 664, Eccl. 82).
80
Hence Siers supplement
is to be ruled out.

32
involving deception, but the expression is a polar one81 (what Garvie ad loc. calls a foilantithesis). Just as Hermes is good at making manifest what without him would be
concealed (for example in his roles as the guide of travellers [
, cf. e.g. Theocr.
25.4-6] and as the patron of interpreters), so too he is good at concealing what
otherwise would be visible. Hence Wilamowitzs
can safely be accepted. The
asyndeton presents no problem: these lines explain in detail the general statement that
Hermes is
(or whatever should actually be read in 8134), as e.g. Ag. 836-7 explain in detail the preceding statement that one who suffers
while another prospers has a double burden to bear.
Garvie does well to suspect that
may be a gloss on
(814), whose meaning
it inelegantly repeats; if so, then together with that other gloss,
(on
), it
must have displaced a genuine word of the form ; perhaps Aeschylus wrote
something like
I cannot see anything wrong with
, eloquently censured by West.
Hermes pulls wool over peoples eyes by saying things that are hard to see through, like
the
of his great-grandson Odysseus (devised with the aid of
Hermes Dolios, Phil. 133) of which Philoctetes complains in Soph. Phil. 1111-282.
Orestes, under Hermes auspices (cf. 727-9, and Garvie on 583-4), did exactly that in
order to gain entry to the palace83, and may yet need to do more of it in order to
complete his revenge. I do not know why Hermes ability and readiness to lie should be
thought more appropriate to epic narrative than to drama (West). Tragedy is full of
liars84, some of whom, as we have seen, invoke Hermes as their patron (or have him
invoked on their behalf by others).
(30)

Choephoroi 991-6

992
(
993
994
995
996

81

Robortello:
M

s
) Turnebus:
M
A.Y. Campbell
Hermann:
Ms:
M
Robortello, Turnebus
H.L. Ahrens:
West

993
995

om. M, add. Ms

Sier (n.67) 260; Garvie ad loc. calls it a foil-antithesis. L. Battezzato, SCO 42 (1992) 86,
objects against this that 816-8 is not antithetical to 815 in meaning; but this objection rests on a
misrepresentation of the meaning of 816-8 (as di notte non visibile).
82
A parallel noted by Battezzato (n.81) 86. Odysseus maternal grandfather Autolycus was a son
of Hermes according to Hes. fr. 64.17-18 M-W = 65.17-18 Most and Pherecydes fr. 120 Fowler
(though Od. 19.396-8 implies otherwise).
83
As he also does in Sophocles Electra, also with the aid of Hermes (El. 1395-6).
84
And the Oresteia in particular: Atreus (Ag. 1590-3), Agamemnon (Ag. 1522-3, undoubtedly
referring to the pretence that Iphigeneia was to be married to Achilles), Clytaemestra, Orestes
and this chorus too, when, immediately before the ode now under discussion, they encouraged the
Nurse to give Aegisthus a falsified version of Clytaemestras message (Cho. 766-782).

33
Garvie in his note on 993 gives good reasons to be doubtful about almost every
imaginable punctuation/construal of the line as transmitted, and also against A.Y.
Campbells
(as it makes clear by biting her). I suggest
: Clytaemestras child (
formally qualifies
) was once her
friend (in infancy, before she had wronged him by first banishing him and then
murdering his father) but is now, as he has demonstrated, a deadly creature that is her
enemy. The word
will recall the snake to which Clytaemestra in her dream gave
birth (called
in 530), as she herself recalled it in the last words we heard
from her (928), and also the description of Clytaemestra herself as a noxious beast by
Cassandra in the previous play (Ag. 1232-4:
again, 1232), which Orestes will be
reprising a moment later (994-6). That this emendation makes three consecutive lines
end in is not a strong objection to it: so do 291-3 and Seven 58-60, while Pers.
361-3, Supp. 476-8, and the four lines Cho.764-7 all end in , Ag. 634-6 all end in - ,
Ag. 1183-5 all end in , and Cho. 97-99 all end in
These sequences all seem to be
purely casual; evidently Aeschylus made no effort to avoid them.
On 995 Garvie raises three serious problems which are presented by the text of that
line quite apart from the particularly feeble and surely corrupt
: (1) it is an
odd way of emphasizing [Clytaemestras] dangerous nature to say that if she had been a
snake she would have been even more dangerous; (2) the power of killing by a mere
touch, without biting, is elsewhere ascribed to certain other creatures but not to the
or the
; (3) Clytaemestra did bite, i.e. stab, her victims, just as Orestes
bit her (cf. above). There may be an excess of logic in this argument. Clytaemestra
has previously been compared to a snake, but to Orestes at this moment that
comparison seems inadequate to the horror of her character and actions, and he is
fumbling for one that might be adequate as Cassandra too did, as Orestes does when
seeking something to which to compare the robe in which Agamemnon was trapped
(997-1004), as the Pythia does when trying to describe the appearance of the Erinyes
(Eum. 46-52), and as several other characters struggle for the right words on various
occasions throughout the trilogy85. If she were a snake, she would have to be the
exceptionally virulent kind of snake that can kill without biting; if such a snake does bite
(and ps.-Aristotles description of the Thessalian sacred snake, Mir. 845b16-32, shows
that some at least of them were believed to do so), it will be all the more deadly 86.
So we do not need to posit far-reaching corruption, and can concentrate on
and
perhaps also
We could do with a noun in place of
, and if we have to fill the
gap in the phrase a serpent that can rot a ____ without biting, the likeliest candidate
for the position is certainly man. The conjecture
(Groeneboom87) is
thus tempting, with
derived from the nomen sacrum abbreviation
: but we
cant do without
since Clytaemestra is not a serpent and never was, nor was she in
fact ever capable of shrivelling anyone up with a touch. One might think of
: either negative is as good as the other here (to invert a
remark of Garvies, the participle is as likely to be conditional or generic as concessive).
Possibly the first four letters of
were corrupted into
, then
was
reinserted,
dropped as nonsensical and unmetrical, and finally
corrected
into

(31)
85

Choephoroi 1042-3

The what shall I say? theme of A. Lebeck, The Oresteia: A Study in Language and Structure
(Washington, 1971) 103-4; cf. also Ag. 783-7, Cho. 87-99, 315-8, 418.
86
To say that a dog can transmit rabies by a lick is not to say that rabid dogs never bite, nor that
the bite of a rabid dog is no more dangerous than that of a healthy one.
87
In his edition (Groningen, 1949); he also changed
to

34

Thus M ends Orestes last major speech. The sentence lacks a verb (
Weil);
is left unexplained; and the choruss reaction (1044-7) shows that Orestes has
ended his speech on a negative and ill-omened note. There can be little doubt that at
least one line has been lost, in which Orestes specified that the reputation (
) he
was leaving behind him was that of a matricide; the chorus reply in effect no, of a
tyrannicide. But where should the lacuna be? Hermann, followed by (e.g.) Page and
Garvie, put it after 1043; Dindorf, followed by West, after 1042. The latter option must
be rejected for the reason given by Garvie: it leaves 1043 as the conclusion of the
speech, and 1043 is a very weak conclusion. If, as we would have to assume, Orestes
said in the lost line [I] am departing, a matricide (with whatever elaboration was
thought appropriate to fill out the line), 1043 would add only that this would be his
reputation up to and beyond death.
If we place the lacuna after 1043, we can either leave the text of 1042-3 as it stands or
accept an emendation like Weils. In the former case we might restore something like
; in the latter case it becomes easier to include a
reference to Orestes justification for the matricide, as he almost always does elsewhere
whenever he mentions it (923-930, 974, 978-989, 1010-3, 1027-8; Eum. 458-464, 588602), e.g.
This last option has the further
advantage that it enables us to place the reference to the matricide right at the end and
so maximize the motivation for the choruss horrified reaction; for if the extra line
follows 1043 and contains the main verb of the sentence (as it will if 1042 is left as
transmitted), the sentence structure requires that the explanation of
take priority and the main verb come afterwards. I conclude that Weils
is correct
and that the repetition of Ag. 1282 (
) is about as
close as it could possibly be. That line referred to Orestes previous exile, only just
ended; this sentence refers to his next, which is about to begin.
(32)

Eumenides 861-3

862

Stephanus:

Mpc:

fere Macf

I have spread the obeli wide, since it is not agreed how far corruption extends. In my
1989 edition, noting the scholiasts gloss
, I printed Musgraves
,
giving the sense making <their hearts> seethe like the hearts of fighting-cocks; but I
do not now find myself convinced by the single parallel I cited for the order of the words
that follow, and West (Studies 290) bluntly calls this order nonsense88 and makes a
good case that the positioning of
is metrically unacceptable too. He suggests
, which likewise gives full value to the evidence of the
scholium. Can ( )
, though, bear the meaning here required? I cannot find a
parallel for the metaphorical awakening of a heart; nearest is Iliad 5.510

88

Not with specific reference to Musgraves proposal or to my edition, which had appeared too late
for him to use (it was published in November 1989, and the preface to Studies is dated in February
of that year).

35

West considers and rapidly rejects a simpler solution, to delete

without further
change, leaving
: the resulting sentence, he says, is
crude, with nothing to mitigate the apparent literalness of the heart transplant.
Certainly, with 862-3 understood and punctuated as it usually is (and as shown above),
this text of 861 would be intolerable: the galline hearts would be left on the operating
table, without any indication of what was to happen to them, while a propensity to
violence against compatriots was implanted into the Athenians. All that is needed,
however, is for the actor to pause (and the modern editor to insert a comma, as Verrall
and Podlecki have done89) after
, separating it from
and encouraging us to
understand
as its object. Do not, says Athena to the Erinyes, take the heart
out of fighting-cocks and implant it in my citizens, a spirit of internecine violence that
emboldens them to fight each other. By thus treating
as the speakers
gloss on
, it is made clear that the heart transplant is metaphorical.
But are we now ignoring the scholium? Certainly
could never be glossed as
But consider the rest of the note: For the bird is pugnacious, and
whereas other animals respect their kin, this one alone does not spare them. The
scholiast makes no distinction between fowl in a normal state and fowl in an excited
state; the species, according to him, is pugnacious by nature. Hence the understood
object of
is not
but
; it is glossing,
not just the second word of 861, but the entire participial phrase which it introduces;
and it is not in itself evidence that
is corrupt.
And in fact
(or
for that matter) is not what the context requires. It
is true that the owners of fighting-cocks tried to enhance their pugnacity by feeding
them garlic (
): cf. Ar. Ach. 166, Knights 494, 946. But it is also true that the
species was believed to be exceptionally pugnacious by nature, young males being eager
to fight any opponent including their own fathers: cf. Ar. Clouds 1427-8, Birds 757-9,
1347-50. Athenas point is that Athenians are not like that by nature, but may become
so if the Erinyes stir them up; the Erinyes would then be turning them, not from peaceful
cocks into violent cocks (for peaceful cocks do not exist), but from humans into
(inherently violent) cocks.
(33)

Eumenides 1044-6

1044
1045

M:
Musgrave:

f.
codd.

Let us get to the essentials, as West says at the beginning of his discussion (Studies
294). An adjective
does not exist. There is a personal name
, that of the
mother of Peleus and Telamon (Bacch. 13.96; Pind. Nem. 5.12), but the second syllable
of this name is long and its Attic-Ionic form is
( Il. 16.14, 21.184-5; Pind.
Nem. 5.12; Eur. Andr. 687; [Apollod.] Bibl. 3.12.6; Plut. Thes. 10.3; Paus. 2.29.10)90.
And the last seven letters of Ms
are identical to a sequence of seven letters
89

A.W. Verrall, The Eumenides of Aeschylus (London, 1908); A.J. Podlecki, Aeschylus: Eumenides
(Warminster, 1989).
90
There was also a group of nymphs worshipped in Cyprus and known as
(Hsch. 2775).

36
earlier in the line. We can at once infer that these seven letters are an intrusion, and
have displaced the true text.
To make our prospects of restoring the text even worse, some suspicion must also
surround
It is certainly corrupt, as metre shows, but how likely is it that this
Lieblingsphrase of the Oresteia and especially of Eumenides91 should have got here by a
mere copying error? Might it not come from a parallel passage cited in the margin
890-1 perhaps, where Athena invites the Erinyes to become residents of Athens,
honoured
? In that case, these words too could have displaced just about
anything.
Having said that, Wests restoration
would provide a
more or less appropriate general sense, though
is another word not known to
exist (besides which,
torch is a very rare word in tragedy92) and one does not
normally urge the participants in a solemn procession (and Awesome Goddesses at that)
to hurry to their destination. In the Loeb I chose (without much conviction) to
translate this restoration, except that I left the epithet of
open; one might suggest
the home that is justly yours (cf. for the idea 890-1; for the word, 804
and Supp. 776).
Of other lines of approach, the most promising, often tried, has been to accept
in the sense of peace treaty and link it with the unity between Zeus and Moira spoken
of in 1045-6. When this unity is contrasted with the claims made earlier by the Erinyes
that their role as avengers has been assigned by Moira (334-9, 392; at 961-2 the Moirai
are their sisters) and that Zeus and his family have set Moira at defiance (171-2, 723-8),
it is hardly an exaggeration to speak of a war (or at least a stasis) among the gods that
is now ended. Combining this idea with Wilamowitzs and Headlams suggestion that
might conceal some form or derivative of
(cf. 1011, 1018), I proposed
93
tentatively in my 1989 edition
there is a treaty <which provides> that you shall for ever be resident among Pallas
citizens; I now feel that I may have been too modest in suppressing this in the Loeb in
favour of Headlams
which introduces yet another word
whose very existence is a matter of conjecture. For this sense of
in Aeschylus
cf. Ag. 1235

(34)

Prometheus 235

VQ

:
:

M D Lb:
cett.

Iac

:
L

or
( )

OW
cett.

Hutchinson (CR 34 [1984] 2) remarked that to write

would make
less abrupt (Mark Griffith in his edition [Cambridge, 1982] had complained of
91

It occurs eight times in Eumenides (52, 83, 200, 291, 401, 538, 670, 891), three times in the
rest of the Oresteia (Ag. 682; Cho. 684, 939), and once in all other tragedy (Eur. Hcld. 575).
92
It appears only in Eur. fr. 472.13.
93
For the shift from addressing the chorus in the plural (as is done consistently from 1033 to 1042)
to addressing them in the singular ( ) cf. Eur. Alc. 215-7 where a similar shift is made within a
sentence, and Seven 95-99 where the chorus twice ask each other whether they should clasp the
images of the gods in supplication, once using the first person singular and once the first person
plural.

37

the abrupt asyndeton). To write

, by tightening the bond between I
dared and I released to make it almost a hendiadys (= I dared to release), would be
a further improvement.
(35)

Prometheus 354

Attempts at emendation should focus on the superfluous

:
is indispensable94,
and
is far more appropriate in sense than any alternative that has been
suggested. The front runners so far have been
(Headlam) and
(West).
I propose
, signalling that we are going back into the past (in contrast with
Typhons present confinement, mentioned briefly in 353 and described more fully in 3635).
(36)

Prometheus 688-692

.
688
690
691

Wecklein:
om. M I
om. Qac

semel , bis cett.

M Wpc D

om. D

West (Studies 303-4) has taken an important step by showing that Wilamowitz was right
to propose
for
, and if we also adopt Pages
, supposing
to
be due to the influence of 688, we have a good pair of dochmiacs to finish with,
That leaves us with
to make
something of. Omissions of one word or another by individual mss. are not of much
significance; but of the three words, it is
that does not belong. The plural, until
Roman times, belongs almost exclusively to epic (it appears once in tragedy, Eur. Hel.
1271), and denotes something unclean or polluting (sufferings only in h.Orph. 14.14
which may well be of Roman date). Probably
is in origin a variant for
which ended up alongside it in the text. Of the other two words,
is clearly
sound: Ios sufferings are indeed
, hard (for her) to bear and
hard (for the chorus and also for Prometheus and the audience) to be spectators of.
But the two-element asyndeton
cannot be right; read
(yielding
another dochmiac), and construe it with
Ios sufferings, sing the chorus, strike
my soul with a double-pronged goad of terror.

94

(Minckwitz) would leave it temporarily unclear that the narrative was going back in time,
and would not explain how
came to be added. The simple deletion of
(first proposed by an
unknown early nineteenth-century scholar) leaves what Griffith, who adopts it, admits to be a
particularly harsh asyndeton. Butler deleted the whole line, a solution that A.J. Podlecki (Oxford,
2005) seems to favour; but a line built up from a gloss or paraphrase would never have included