Performative Utterances

Performative Utterances and Divine Language in Ugaritic Author(s): Seth L.
Sanders Reviewed work(s): Source: Journal of Near Eastern Studies, Vol. 63, No. 3 (July 2004), pp. 161-181 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/10.1086/424769 . Accessed: 04/02/2013 12:04
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to Journal of Near Eastern Studies.
http://www.jstor.org
This content downloaded on Mon, 4 Feb 2013 12:04:06 PM All use subject to JSTOR Terms and Conditions
PERFORMATIVE UTTERANCES AND DIVINE LANGUAGE IN UGARITIC

Seth L. Sanders, The University of Chicago
I. Introduction ecause of the special way they unite speech and action, the grammatical category of performative utterances has aroused strong interest within Semitic philology, biblical studies, and the history of religion.1 Yet within these disciplines, the mechanism by which performatives unite language and life is still a mystery: the discussion of the performative in these elds has so far failed to produce a coherent grammatical denition or even an agreed-upon corpus of examples. In this article, I attempt to move beyond the current impasse by advancing a denition informed by linguistic anthropology and examining the most important Ugaritic examples in their cultural context. A back-and-forth dialogue between ancient instances and the modern theory of the performative reveals that Ugaritic texts themselves provide mythic models of how language and ritual were thought to work: Canaanite myths imply a Canaanite speech-act theory distinct from our own. Attention to native ideas of how to do things with words renes our understanding of both the performative concept and the texts in Ugaritic because performatives are a primary site where universal grammatical features meet specic native beliefs. This paper thus uses the performative as an entree into a broader issue: the interpretation of language and culture in ancient Near Eastern religious texts. In this case, West Semitic myth, ritual, and historical grammar are shown to be mutually illuminating. II. The Problem of the Performative in Semitic Grammar A perennial difculty facing scholars who interpret ancient cultures is the recovery of context. Thinking we are merely recognizing the effect of the passage of time, we cleanly separate text from context, imagining our texts as springing from a single real situation that died forever with the texts writers. Yet we also know that the very reason our texts are worth reading is that they were creative and powerful cultural tools, written down to
1 This paper was developed in presentations at the Society of Biblical Literature 2000 annual meeting and the University of Chicago Oriental Institutes 2001 Comparative Semitics workshop. I thank Chip DobbsAllsopp, Baruch Levine, Wayne Pitard, and Anson Rainey for their comments and criticism. I am especially indebted to Michael Silverstein for the con-
[JNES 63 no. 3 (2004)] 2004 by The University of Chicago. All rights reserved. 002229682004/63030001$10.00.
ceptual framework and Dennis Pardee for a detailed philological critique. It is extremely interesting to note that the rst twentieth-century discussion of the performative concept (using the analogous German category Koinzidenzfall ), by the German linguist Erwin Koschmieder, was partly inspired by problems of the Biblical Hebrew tense system. See his Beitrge zur allgemeinen Syntax (Heidelberg, 1965), p. 26. Fundamental questions of the Semitic languages and performativity were thus intertwined at the beginning of the modern debate. This paper attempts to reunite them.
161
162
Journal of Near Eastern Studies
be read and reused. These tools were thus used repeatedly in creating contexts: the myth is recited once more in a new performance, the diplomatic letter is pulled from the archive, its statements drawn on in a new political action. Are there ways to study the specics of how texts and contexts join, how each helped produce the other? A promising place to nd the sutures of text and context is the old West Semitic corpus of performative utterances, a linguistic category recognized in many ancient and modern languages that has the peculiar function of explicitly creating its own context at the levels of both language and action. Performatives are one-line verbal rituals2 that seem to accomplish actions by referring to them. In line with current research in Hebrew, Aramaic, and Akkadian, recent treatments of Ugaritic language (by D. Pardee, M. Smith, and J. Tropper)3 and religion (D. Wright)4 have assigned a signicant role to performatives. Interest in this category is well deserved. Grammatically, performatives offer insight into still unresolved problems of Ugaritic (and more generally Semitic) tense and aspect because in ancient West Semitic their canonical form is that of a perfective, sufxing verb representing an action that coincides with the moment of speaking. The latest research indicates that the canonical grammatical form of the performative in West Semitic undergoes a shift in the last centuries b.c.e. from the sufxing form to the participle. Thus discussion of this phenomenon casts long-debated questions of Semitic tense and aspect in a fruitfully pragmatic and historical light. The way performatives link language with action has also made them of interest outside of philology: in the history of religion, performatives offered the promise of understanding how supposedly irrational magical utterances could be effective and thus allowed new ways of interpreting both ritual and myth.5 In contemporary theory, they are part of the recognition in disciplines such as linguistics, sociology, anthropology, and philosophy of the way language both refers to and helps create its own context. The most useful discussions, such as those of P. Bourdieu, J. Butler, S. Cavell, and M. Silverstein, have
2 The phrase is that of Michael Silverstein (personal communication, August, 2000). 3 Dennis Pardee and Robert Whiting, Aspects of Epistolary Verbal Usage in Ugaritic and Akkadian BSOAS 50 (1987): 131; Mark S. Smith, The *qatala Form in Ugaritic Narrative Poetry, in David P. Wright, David Noel Freedman, and Avi Hurvitz, eds., Pomegranates and Golden Bells: Studies in Biblical, Jewish, and Near Eastern Ritual, Law, and Literature in Honor of Jacob Milgrom (Winona Lake, Indiana, 1995), pp. 789803; Josef Tropper, Ugaritische Grammatik, AOAT 273 (Mnster, 2000), p. 714. 4 The Play of Ritual in the Aqhat Narrative in G. D. Young et al., eds., Crossing Boundaries and Linking Horizons: Studies in Honor of Michael C. Astour on His 80th Birthday (Bethesda, Maryland, 1997), pp. 57798. An expanded version appeared as Ritual in Narrative: The Dynamics of Feasting, Mourning and Retaliation Rites in the Ugaritic Tale of Aqhat (Winona Lake, Indiana, 2001). Wrights approach to Ugaritic ritual merits attention. He utilizes Ronald Grimess taxonomy of ritual (which reads modern
colloquial English designations for failure as universal categories of ritual action; thus ritual failures are categorized with such homey terms as nonplays, hitches, ops, etc.). On the problem of using modern folk-theoretical speculation as a basis for understanding other cultures rituals, Joel Robbins has recently emphasized the fundamental point that if conceptions of language and ritual are linked and conceptions of language differ across cultures, it follows that conceptions of ritual will also vary. See his Ritual Communication and Linguistic Ideology: A Reading and Partial Reformulation of Rappaports Theory of Ritual, in Current Anthropology 42 (2001): 592; the response of Ellen Basso on p. 600 is especially pertinent to the present articles concerns. 5 On performativity and performance in Greek epic, see Gregory Nagy, Pindars Home (Baltimore, 1990); on Jewish magic, see Rebecca Lesses, Ritual Practices to Gain Power: Angels, Incantations, and Revelation in Early Jewish Mysticism (Harrisburg, Pennsylvania, 1998).
Performative Utterances and Divine Language in Ugaritic
163
spanned disciplines and advanced our knowledge of the historical contexts of performatives as well as our theoretical understanding of what makes them work.6 But while they are now an accepted part of Ugaritic grammar, it is a strange and significant fact that Ugaritic (and other ancient Semitic) performatives continue to elude coherent grammatical denition. As we will see, previous discussions undermined their own attempts at denition by pointing to Ugaritic texts where utterances perform actions with a confusing variety of linguistic forms. So what makes a performative, verbal morphology? personal authority? If they take on power only in real-time speech contexts (as the inventor of the term, J. L. Austin, seemed to argue), are any of the forms transcribed in ancient texts really performatives at all? The rst step in a solution is recognizing that the fascination of the performatives and its difculty are of one piece. What is interesting about the performative is that it helps create its own context by explicitly bridging the gap between speech and action, text and context.7 And what is unique about the performative is the way it bridges this gap: through reexivity and deixis, by pointing to itself and to what it is doing. Recognizing the reexive and deictic nature of the performative, we will review some denitions and instances in Ugariticcases where texts reach into their contextsand suggest a denition taken from linguistic anthropology that can help explain the different forms the performative takes in the ancient Semitic languages. III. Research on Performatives in Semitic Grammar Scholars of the ancient Near East have long been interested in texts own accounts of themselves. Early twentieth-century German scholars, such as the biblicist Hermann Gunkel and the Assyriologist Adam Falkenstein, inuenced by contemporary currents in the study of folklore and comparative literature, focused on clues in texts that could signal how they should be read. They argued that formal markers of genre and type in psalms, stories, and incantations gave metalinguistic messages, signals about how to read the code that could serve as cues to ancient conditions of production and reception.8 Later scholars found their
6 See Pierre Bourdieus Authorized Language: The Social Conditions for the Effectiveness of Ritual Discourse, in his Language and Symbolic Power (Cambridge, Mass., 1991); Judith Butlers Excitable Speech: A Politics of the Performative (New York, 1997); and Stanley Cavells What Did Derrida Want of Austin?, in his Philosophical Passages: Wittgenstein, Emerson, Austin, Derrida (Cambridge, Mass., 1995). The treatment that is most important for our purposes, because it goes the farthest in specifying the relationship between the grammar of the performative and the realm of cultural action, is Michael Silversteins Language Structure and Linguistic Ideology in P. Clyne et al., eds., The Elements: A Parasession on Linguistic Units and Levels (Chicago, 1979), pp. 193247. 7 Of course, all uses of language are forms of social action. This well-studied dimension of linguistics is designated pragmatics; an intelligent survey is Steven Levinsons Pragmatics (Cambridge, 1983).
What makes performatives unusual is the way their pragmatics work: they act by talking about acting, metapragmatically. To be more specic, performatives expose and reify and thus by a kind of metapragmatic regimentation focus and intensify the already existing element of action in speech. The most profound investigation of this phenomenon, which, almost miraculously, also serves as a lucid introduction, in Benjamin Lees Talking Heads: Language, Metalanguage and the Semiotics of Subjectivity (Chapel Hill, North Carolina, 1996). 8 Gunkels use of genre as a cue to interpretation was of decisive inuence in the history of biblical studies; his groundbreaking work is found in his Genesis, 3d ed. (Gttingen, 1910) and Einleitung in die Psalmen: Die Gattungen der religisen Lyrik Israels, completed by Joachim Begrich (Gttingen, 1933). Adam Falkensteins equally interesting work on functional formulas in Sumerian magical-medical incantations appears
164
work incomplete because we cannot simply read function from form any more than we can read the history of a texts interpretation from the text itself: the text cannot determine the context.9 Ethnography, as well as the study of reader response and the history of reception have shown that the life of a text is in its readers at least as much as in the text itself. Yet there are ways that texts demonstrably do anchor themselves in their contexts. The clearest and simplest examples are those grammatical forms known as shifters: rst and second person pronouns such as I and you, demonstratives such as this and that, and verbal markers of tense that position the time of the event being spoken of relative to the time of speaking.10 While shifters contextualize the speech event in time and space, gender, number, and honoric markers index qualities of physical person and social power. In honorics we see a second feature of shifters: in addition to just reporting context, these indexical features help create the contexts in which they live. For example, every time a subordinate uses an honoric in an utterance to a superior, the subordinates utterance actually acts to contribute to a context of social inequality by once again honoring the higher-status persons position; in this way, the use of shifters is always social action.11 Perhaps the most striking and explicit way language is creative of context is that type of one-line verbal ritual dubbed performative by the philosopher J. L. Austin.12 English performatives are usually described as verbs in the rst-person present active indicative that can take an innitive or relative clausal complement; structurally they are said to work by reporting metalinguistically on what they are doing. Thus, on the witness stand, I swear to tell the truth, describes the witnesss linguistic behavior (swearing), but to make this description is simultaneously to perform the pragmatic legal action of swearing: after taking this oath the witness is bound and liable for the crime of perjury. Similarly, Pharaohs speech to Joseph in Gen. 41:41, nnty tk l kl rs msrym I give you authority over the whole land of Egypt, actually accomplishes the transfer of authority. It also shows the sufx form recognized in Hebrew for the performative and introduces a basic
in his Die Haupttypen der sumerischen Beschwrung literarisch untersucht, Leipziger semitistische Studien, neue Folge, vol. 1 (Leipzig, 1931). 9 There is an old and inuential interpretive tradition for Ugaritic, but unfortunately it is largely non-native and hostile. This is the reading of Canaanite myths and rituals as representing a primitive nature religion, background to the Bible, which has roots in Classical and Patristic polemics. For an alternative view of Canaanite myth that substitutes a contemporary historyof-religions approach for the unquenchable, incestuous Canaanite sexuality of Baal and Anat, see the exemplary work of Neal Walls, The Goddess Anat in Ugaritic Myth, Society of Biblical Literature, Dissertation Series, no. 135 (Atlanta, 1993), and the judicious comments of Jo Ann Hackett in Can a Sexist Model Liberate Us? Journal of Womens Studies in Religion 5 (1989): 6576. 10 The most important analysis of shifters, a concept originating with the linguist Otto Jespersen, is Ro-
man Jakobsons Shifters, Verbal Categories, and the Russian Verb (Cambridge, Mass., 1957). Jakobsons work is signicantly extended by Michael Silverstein in his Shifters, Linguistic Categories, and Cultural Description, in K. Basso and H. Selby, eds., Meaning in Anthropology (Albuquerque, New Mexico, 1976), pp. 1155. For tense markers as shifters, see simply a handbook such as Bernard Comries Tense (Cambridge, 1985). The issue is well put by Douglas Gropp, Tense, as generally understood, is a deictic category. That is, it is an element within the text that points to the speech situation. See his The Function of the Finite Verb in Classical Biblical Hebrew, in Hebrew Annual Review 13 (1991): 53 and n. 20. 11 An excellent treatment is that of Asif Agha, Honorication, in Annual Review of Anthropology 23 (1994): 277302. 12 How to Do Things with Words, 2d ed. (Cambridge, Mass., 1975), based on lectures given at Harvard University in 1955.
165
linguistic question: why should West Semitic perfect forms correlate with the English and German present? The process of identifying performatives has now been extended to all the classical Semitic languages, raising both diachronic and denitional questions. The past few years have seen groundbreaking work on performatives in Ethiopic by S. Weninger and in Aramaic by M. Rogland.13 In the course of this linguistic eldwork, problems with both the theory and the evidence emerged. While Hebrew and other West Semitic languages displayed perfective sufx forms, Akkadian presented a historically complex distributive of forms, and recently the situation has been shown to be more complex in West Semitic itself. Roglands study of Qumran Aramaic, for example, shows that the dominant form is here the participle, and a glance at the Aramaic betrothal formulas found in the Palestinian Talmud and in the Geniza documents (using mqds and mqwdst respectively)14 shows that this phenomenon is not a peculiarity of Qumran Aramaic but rather generalized in later Hebrew and Aramaic. At the same time, while there is a core group of undisputed verb types involving the direct exercise of power in language, such as swearing, granting, adjuring, etc., each scholar has proposed a differentsometimes strikingly differentlist of examples. Divergence is to be expected for a difcult language such as Ugaritic, lacking as it does interpretive traditions, but the most basic questions remain: what is a performative? does the performative encompass all instances of doing things with words, or should it be limited to instances where the rst-person singular of the sufx form is used to do things with words? Is it performative to announce that one is falling on ones face? What about falling on ones face at a distance (mrhqtm)? And what about the most bare-bones metalinguistic act, simply saying that one is saying something? Theory is not epiphenominal to these philological problems: it is at their root. It is precisely the way performative utterances help create their own context that has led to problems in consistently dening them or even, grammatically speaking, dening them at all and that has also led to questions about how they work as part of the ancient cultures that created them. Attention to this sort of creativity in its ancient West Semitic context can also suggest answers to these questions. IV. Dening Ugaritic Performatives: Pragmatics, Metapragmatics, and the Grammar of Performatives Previous denitions of the performative in the ancient Semitic languages have been brief, in the nature of allusions or rough-and-ready guides. While the rst application of the performative concept to a West Semitic language was probably that of the German linguist Erwin Koschmieder in the 1930s, the rst scholar of Near Eastern studies explicitly
13 Stefan Weninger, On Performatives in Classical Ethiopic, Journal of Semitic Studies 45 (2000): 91101, and M. Rogland, A Note on Performative Utterances in Qumran Aramaic, Revue de Qumrn 19 (1999): 27780. 14 P. Qid III.2, 63d and Friedman Geniza text 10a: 1617, cited in Mordechai Friedman, Jewish Marriage in Palestine (New York, 198081), vol. 1, pp. 196 and
200, respectively. The Geniza example, functioning by proxy, is at least as complicated as the long-distance performative I hereby fall from afar! discussed as the last example below. As I will argue, rather than dismissing them as anomalies, the broad attestation of such complex cases should make us reect on the ways performatives, and ritual language in general, register different types of agents and participants.
166
to apply the term performative to Biblical Hebrew was Delbert Hillers.15 I will restrict myself to Hillerss treatment because it is the most in-depth in English.16 Because he wished to avoid the theoretical pitfalls into which it seemed to him that Austin had run, Hillers used the denition Austin started out with: performatives are pronouncements where the uttering of the sentence is not a description of an action, but itself the doing of an action, or part of the doing of an action (p. 758). But as Austin found, this denition itself introduces problems, the rst of which is that it does not dene performatives (I order you to shut the window, I declare this meeting adjourned), at least in distinction to imperatives (shut the window!) or declarations (this meeting is adjourned), neither of which describe actions and both of which perform actions (ordering, adjourning, or at least declaring adjourned). Worse, by this denition performatives actually look less performative than other pragmatic uses of language such as imperatives or declarations, since performatives do seem to describe actions (at least one must explain why descriptions of actions such as I stir the cake mix have precisely the same grammatical form as performatives) at the same time as they perform them. Indeed, Hillers himself provides a different characterization of performatives two pages on, when he says that [the performative] sentence does not refer to a past act but to an action in the present that is accomplished, at least in part, by the speakers pronouncement of the utterance under appropriate circumstances (p. 760). Here, the performative sentence is not just an action, but also a reference to an action. This distinction is not only theoretically important, it also rules out imperatives and declarations. But this second denition now does not work with all of Hillerss examples, which include the declaration bny th you are my son! in Psalm 2:7. As we will see, the problem of dening performatives continues in the literature. Dennis Pardee was the rst to apply the concept to Ugaritic letters in his 1983 article The Epistolary Perfect in Hebrew Letters. While recognizing the essential issues of tense and translation (the forms are to be translated with English presents), Pardee utilized the concept of epistolary tenses from classical grammar, which themselves say nothing about pragmatics.17 Theodore J. Lewis then applied the analogous German concept of the Koinzidenzfall, borrowed from Akkadian grammar, to his interpretation of an Ugaritic ritual. In Lewiss 1986 dissertation, published in 1989, The Cult of the Dead in Ancient Israel and Ugarit,18 he described a sequence of sufx forms as utterances in which the
15 Delbert R. Hillers, Some Performative Utterances in the Bible, in Wright, Freedman, and Hurvitz, eds., Pomegranates and Golden Bells, pp. 75766. The phenomenon has been recognized many times before: see Yochanan Muffs, Studies in the Aramaic Legal Papyri from Elephantine, Studia et documenta ad jura Orientis antiqui pertinentia, vol. 8 (Leiden, 1969), p. 32, n. 2, with references to the work of Driver, Rashbam ad Gen. 23:11, and Rashi ad Gen. 14:22. 16 This is to neglect Andreas Wagners book-length study, Sprechakte und Sprechaktanalyse im Alten Testament: Untersuchungen im biblischen Hebrisch an der Nahtstelle zwischen Handlungsebene und Grammatik, BZAW 253 (Berlin, 1997). My study of performative utterances in the literary context of the book of Genesis (currently in preparation) will contain a sur-
vey of how performatives have been used in biblical studies. 17 Biblische Notizen 22 (1983): 34-40. Pardee already recognized that there were pragmatic problems at hand. See his note on p. 34, n. 7. He used the term performative in relation to these West Semitic forms shortly thereafter in his review of Robert G. Boling and G. Ernest Wright, Joshua: A New Translation (Garden City, New York, 1982), in JNES 44 (1985): 148. 18 Theodore Lewis, Cults of the Dead in Ancient Israel and Ugarit, Harvard Semitic Monographs, no. 39 (Atlanta, 1989), p. 13. Lewis cites the precedent of the Koinzidenzfall as applied by W. R. Mayer, W. Heimpel, and G. Guidi to Akkadian and Hillerss then unpublished work on Biblical Hebrew, now in Pomegranates and Golden Bells.
167
words recited represent the very action to which they refer.19 Lewis suggested that the sequence of qra verbs in KTU 1.161 were to be translated not as perfects or as a hodgepodge of perfects and imperatives, but rather as performatives.20 Similarly, Mark S. Smiths 1995 discussion The *qatala Form in Ugaritic Narrative Poetry cites the performative perfect as one of six basic functions of the qatala form. Like Lewiss, his characterization is brief and potentially ambiguous: is it the words themselves that represent or, perhaps better, perform the action that they name? Smith goes on to cite the detailed study of Franois Recanati on performative utterances.21 Crucially, the passage Smith cites happens to be Recanatis summary of the work of the great linguist Emile Benveniste. A direct look at Benvenistes work provides a real advance over previous denitions because Benveniste emphasized a crucial feature only implicit in Austins descriptionreexivity:22 . . . a performative utterance must name the spoken performance as well as its performer . . . There is no performative utterance unless it contains the mention of the act.23 This distinction, which is also that between Hillerss rst and second denitions of the performative, reveals the crucial difference between performatives and other ways of doing things with words. What sets performatives apart from other pragmatic language is self-reference. Performatives simultaneously say something and do something by talking about saying and talking about doing. Thus in the schema initiated by Roman Jakobson, an imperative is language intended to produce a pragmatic effect and so functions on both linguistic and pragmatic levels. But the imperative does not talk about itself, nor does it talk about what it is doing, so while it acts, it is neither metalinguistic nor metapragmatic. It is the performative that seems to have the unique distinction of being linguistic, pragmatic, metalinguistic, and metapragmatic all at once: to utter I promise is to describe a kind of talking and a kind of acting even as it is also to talk and to act. While this may seem obvious, the fact is that virtually all previous denitions of the performative have bogged down precisely here, characterizing performatives as simply pragmaticlanguage that (like
19 Lewiss curt characterization is based not so much on Austins performative as on the Koinzidenzfall, developed by Koschmieder (see citation in n. 1 above) and at that time known to Semitists through Wolfram von Sodens Grundriss der akkadischen Grammatik, AnOr 33 (Rome, 1952), s80c (signicantly expanded by Mayer in the third [1995] editionsee the new note at s79b and extra material at s80c). 20 The translation as a sequence of perfects involved problems of sense only resolved by an ad hoc appeal to the ritual nature of the language, while translating as a mix of perfects and imperatives involved an unexpected and confusing modal switch. Several interpretations of verb forms in the invocation section of 161 assume rubrics that are not graphically marked. Though we know that in Ugaritic switches in voice within a text can be graphically unmarked (contrast 1.100 [rubrics] 1.22, 1.119 [lacking]), there seems to have been a tendency to use marking; cf. 1.114 obv., 1.4 V 42 43. 21 [Performative verbs of speech] represent the performance of the action that they name; see Smith,
The *qatala Form, p. 795, citing Recanatis Meaning and Force: The Pragmatics of Performative Utterances (Cambridge, 1987). 22 The most recent treatment, that of Josef Troppers Ugaritische Grammatik (p. 714), cites, unmodied, the pioneering work of Koschmeider on the Koinzidenzfall. See his Beitrge, where it has not been widely enough noted that Koschmeider himself insisted that performativity inhered in usage rather than the verb itself (p. 27). 23 The context of the quote is this: An utterance is performative in that it denominates the act performed because [the speaker] pronounces a formula containing a verb in the rst person of the present: I declare the meeting adjourned; I swear to tell the truth. Hence a performative utterance must name the spoken performance as well as its performer. . . . There is nothing like this in the imperative. . . . There is no performative utterance unless it contains the mention of the act; see Emile Benveniste, Analytical Philosophy and Language, in Problems in General Linguistics (1963; Coral Gables, Florida, 1971), p. 327.
168
all uses of language) is also actionand thus failing to distinguish performatives from the large realm of language with some pragmatic function or other. Jakobsons student Michael Silverstein applied Jakobsons insights to the performative in his article on Language Structure and Linguistic Ideology. Here he maps the relation between linguistic form, reexivity, and social action. Silversteins cross-linguistic denition of the performative species how the performative works metapragmatically to accomplish a pragmatic goal. His denition is represented in the simplied diagram below (Table 1):24
TABLE 1 First-person Logical Subject Metapragmatic Verb of Speaking or Giving inected for morphologically minimal, unmarked categories [e.g., in English, simple nonpast tense; in many languages without tense; unmarked Punctual/ Perfective Aspect]25 Second-person Logical Recipient/Benefactee Object Optional Element: Subordinate Clause or Donated Object clause stipulates in nite or nominalized nonnite construction a state-of-affairs causally related to the successful occurrence of the performative
Silverstein explains the peculiar morphology of the performative and its relationships to tense, mood, and aspect through the way it both anchors itself in and alters its context. The key to the working of the performative is the unusual deictic position it establishes for its content, the performative speech act. Deictically, while the rst-person agentive and
24 See Silverstein, Language Structure and Linguistic Ideology, in Clyne et al., eds., The Elements, esp. pp. 209 and 215, where, however, Silverstein describes the performative as predicating in a special way, upon the securing of uptake or, in other words, by implicature. This special mechanism of predication is further specied in his Metapragmatic Discourse and Metapragmatic Function, in John Lucy, ed., Reexive Language (Cambridge, 1993), pp. 3358; on predication and performatives, see pp. 5253. 25 Cross-linguistically, in languages that morphologically mark only aspect, the Punctual/Perfective is less marked with respect to deixis. This is borne out in Ugaritic, where it is used for nonmodal expressions of states and facts (for evidence, see the work of E. D. Mallon, The Ugaritic Verb in the Letters and Administrative Documents [Ph.D. diss., Catholic University, 1982]). The question of whether the sufx form is indeed the unmarked member of the sufx/prex pair in Classical Biblical Hebrew prose is a more complex subject. In a thoughtful and sophisticated analysis, Douglas Gropps The Function of the Finite Verb in Classical Biblical Hebrew, in Hebrew Annual Review 13 (1991): 4562, Gropp argued for a three-way system of marking in Classical Biblical Hebrew prose in which the sufx form is marked as [+ANTERIOR]. While I cannot take up all his arguments here, I can
show why I think a broader picture is necessary. First, Gropp makes clear that he is not describing the entire system of nite verbs, but only that of active verbs. On pp. 5152 he states that in differentiating sufx from prex forms . . . we will consider only the opposition of perfect and imperfect for verbs that are lexically nonstative. Some might consider this a fatal qualication. . . . I must confess that I do. Though space does not permit him to offer arguments for this proposal, Gropp claims that the function of statives in the sufx form falls outside of the system he is describing. But even if one accepts this (and it is difcult to accept for transitive statives), this would still represent a large lacuna in his system. Second, even if one were to accept all of Gropps categories, in his analysis of the evidence he does not show which elements of tense and mood are assigned pragmatically, by implicature. In assigning tense-mood-aspect values to verb forms, his method is to compare the translation values of the most common forms of the active verbs without taking implicature into account. But it is simpler to argue that non-stative verbs in the sufx form receive what he calls their [+ANTERIOR] marking pragmatically, rather than through the inherent semantics of the form. For this analysis of the role of implicature in creating tense values in a West Semitic verbal system, see Shlomo Izreels Canaano-Akkadian (Munich, 1998).
One Line Long
169
second-person patientive pronominal morphology point specically to the persons present in the speech event, the verb itself is minimally marked, so it does not make deictic connection with the speech event, stating any sort of is relationship between its elements. Uniquely, performatives refer but they do not predicate. This is why they are not felt to describe (in Hillerss rst denition), while they are still felt to refer (as in his second denition). As in English simple present forms such as declare, the West Semitic sufx form, descended from the tenseless proto-Semitic stative, by itself does not specify a relation to the moment of speaking. Instead, when it is used in an appropriate context by participants considered authorizedwhen it is, as Austin would put it, felicitousthe performative actually brings about the relationship between the rst-person referent and the second-person referent denoted by the performative verb in the utterance. It does this by virtue of securing uptakethe participants knowing its meaning and accepting it as legitimate, whereupon it becomes true. The performative thus has the feature, owing from the morphological and semantic minimality of its verb form, of referring without predicating within the sentence. Instead, the predication is assigned by implicature, an effect long ago described by the linguist Jerrold Sadock as a sense perlocution.26 We may now refer to some examples to see how this theory works. For the fullest version of the performative sentence, compare English I declare this meeting adjourned and Biblical Hebrew hydty bkm hywm ky htytm27 bnpswtykm . . . I warn you this day that you were deceitful at heart . . . (Jer. 42:1920), where declare and the hiphil of wd are both metalinguistic and metapragmatic verbs and this meeting (is) adjourned and htytm bnpswtykm you were deceitful at heart represent nominalized nite clauses of the stateof-affairs brought about by the successful performative utterance. Indeed, the ritual use of performative verbs in West Semitic is widespread. One may conveniently compare the use of the ntn/ytn give root (noting that it is replaced in later Aramaic by the yhb root) in a corpus of mythical and legal texts spanning about 1,000 years that includes Biblical Hebrew (Gen. 41:41), Early Egyptian Aramaic (Porten and Yardeni B1.1:23),28 Late Egyptian Aramaic (B2.3:3),29 and Ugaritic (KTU 1.100:7576).
26 For this idea, see J. Sadock, Toward a Linguistic Theory of Speech Acts (New York, 1974), p. 53, cited in Silverstein, Metapragmatic Discourse, p. 52. It is neither necessary nor, perhaps, possible, to decide which of the West Semitic verb forms is the unmarked form in all situations. As Comrie notes, it is possible for the markedness relations of a single set to vary by context. What is essential for our purposes is to note that the sufx form does t the criterion of morphological minimality and that, in contrast to the prex form, it lacks modal force, functioning for the simple assertion of facts and states. For the morphological criterion and other, sometimes contradictory, theories, see Comrie, Aspect, p. 111 ff. 27 Reading, with the Masoretes and modern scholars, the Qere form htytm rather than the obviously metathetic Ketiv httym. 28 B. Porten and A. Yardeni choose to translate this and the following example, consistent with their uniform and highly problematic practice for translating operative sufx forms in Aramaic contracts, as a simple past, I gave you; see their Textbook of Aramaic
Documents from Ancient Egypt (Jerusalem, 1986), passim (hereafter TADAE). This reading begs the question of the form of the operative clause in Aramaic law because it merely shifts the burden of the deed backwards in time: on the day specied in the contract, the donor had to have performed the legal act of giving. At that point, all comparative evidence points to the conclusion that the performative tbhy formula would have been used. For difculties with their rendering of another sufx formula, see n. 32 below. For a similar parallel series of performatives in West Semitic, note the Phoenician, Aramaic, and Hebrew instances of the performative blessing formula brkt(k) I bless (you) (cf. the famous instance in Gen. 17:20) gathered by Pardee The Proposition in Ugaritic, in UF 8 (1976): 22122 and his interpretation of these as epistolary perfects in Letters from Tell Arad, UF 10 (1978): 311. 29 The example is taken from T. Muraoka and B. Portens Grammar of Egyptian Aramaic, Handbuch der Orientalistik, 1. Abt., vol. 32 (Leiden and New York, 1998), p. 194.
170
Silversteins denition thus offers the most straightforward and powerful explanation of performative morphology. It explains why performatives select the simple nonpast in English and German but the perfect in ancient West Semitic languages: this is because they select the morphologically and semantically minimal verbal category, the form with sufcient inectional morphology to qualify as a verb but no more. This happens to be the sufx form in Ugaritic and other archaic West Semitic languages.30 And we can see further that what makes performative utterances so confusing and interesting at the same time is the way they laminate all four areas on top of one another into a maximal grammatical domain (the sentence) that is simultaneously a minimal textual domain (again, the sentence is the smallest piece of text that generally qualies as an independent piece of discourse). They are linguistic, metalinguistic, pragmatic, and metapragmatic at once. Performatives are revealing because of their capacity to act as ideological prisms. In a performative utterance one can simultaneously talk, do, talk about talking, and talk about doing. And because every culture has its own folk theories of language and action, it is often around instances of the performative that cultures exemplify and explore their own ideas about how one does things with words. But given these linguistic criteria, an important question remains: how does one tell if a particular verb is performative or not? As Austin already pointed out, linguistic form alone is not a sufcient criterion: to insult is a certainly metapragmatic verb; yet I insult you is not performative in English. Intriguingly, performativity seems to be one of those verbal categories, like transitivity, which are not apparent from the outside; they are culturally dened, or, in linguistic terms, lexical, only becoming apparent in attested cultural contexts in which the verb is used. This basic observation has serious consequences for the way the performative concept has been applied to ancient languages because it suggests that translatability into English or German performatives is not a sufcient criterion for performativity. Since English and German have their own culturally specic rules and lexica of performativity, translatability-as-performative, the cherished hereby/hiermit test whereby scholars have determined whether a verb is really performative, may reect nothing more than the grammatical structures of the target language in question.31
30 For the sufx forms unmarked, semantically neutral role in Ugaritic prose to assert simple facts, the best treatment of the subject is Mallons dissertation The Ugaritic Verb in the Letters and Administrative Documents: [The sufx form] in the letters and administrative documents generally functions as a past tense in opposition to the [prex form]. It is also used to express the pluperfect and future perfect. The [sufx form] of stative verbs is used with a present meaning (p. 55). This contrasts with the more complicated semantics and morphology of the various prex forms. The fact that within certain other genres such as narrative poetry, sufx forms are actually less common than prex forms is irrelevant here. As Trubetzkoy and Jakobson, the inventors of the markedness concept, already noted, it is structure rather than statistical preponderance that is at issue. Greenbergs later attempt to introduce statistical frequency as a criterion for
markedness failed because, among other things, different discourse genres and registers within one language may utilize some forms more than other genres and registers; narrative poetrys preference for prex forms is an example of this. For discussion, see Henning Andersen, Markedness TheoryThe First 150 Years, in O. M. Tomic3, ed., Markedness in Synchrony and Diachrony (Berlin, 1989), pp. 2930, and further, Edwin Battistella, Markedness: The Evaluative Superstructure of Language (Albany, New York, 1990), pp. 3738 and 8991, for a discussion of the markedness relations of English verbs, with the nonpast as unmarked. 31 This is not to say that hereby does not retain its usefulness as a way for translators to unambiguously mark the decision to translate a verb as performative, once they have made this decision on lexical and contextual grounds.
171
The classic West Semitic example of the problem of culturally translating performatives appears in the language of ancient Aramaic contracts illuminated by Yohanan Muffs.32 The state of deliberate good will necessary to form a valid contract is gured through a discourse of love and inner feeling. And here the operative, and performative, divorce formula, known from Elephantine documents and persisting through the records of the Geniza, is I hate you.33 Whether or not this verb passes the hereby test (I hereby hate you is just as counterintuitive a legal statement in English as a simple I hate you), translatability is useless here because English does not use verbs of feeling performatively; to us they merely index an affective state rather than performing any sort of action. Thus, as much as the phenomenon of the performative may be a part of universal laws of grammar, performatives are always also a part of a local and particular culture that we ignore at our peril.34 V. Performatives in Ugaritic Contexts: Mythic Models of How Language Acts I now discuss some examples proposed in the literature of the performative in Ugaritic: two relatively certain and three merely possible. Because of the rich contexts in which
32 Muffs, Studies in the Aramaic Legal Papyri from Elephantine and Love and Joy: Law, Language, and Religion in Ancient Israel (New York, 1992). 33 Public pronunciation of the phrase sunt(k) I hate (you) addressed to the spouse was the central act of divorcing in Jewish-Aramaic tradition, and similar uses of a hate formula for divorce were widespread in the ancient Near East, a fact already well established in Reuven Yarons Introduction to the Law of the Aramaic Papyri (Oxford, 1961), pp. 5455, 1012. Grammatically, the Aramaic hate formula is a sufx-form verbal ritual that works to bring about the state it describes and thus ts perfectly the denition of a performative formula. There remains some needless hesitation on this in the literature (as in the introduction to Porten and Yardenis TADAE 2 [p. xiv], and their pastperfect translation as I hated, which may be an artifact of TADAES general translation policy). But there is no reasonable doubt that the formula ts a wellestablished performative pattern that hatred was a recognized legal state of affairs brought about by its utterance and that its Aramaic contexts require presenttense performative translation in English. First, there is widespread and unequivocal evidence that the snt(k) formula, uttered in proper context, performatively accomplished divorce. Provision for the dissolution of marriage by means of publicly uttering snt(k) comes at the end of the Ketubbah, the marriage contract, in Jewish marriage documents from fthcentury b.c.e. Elephantine down through the Geniza of the tentheleventh centuries c.e. Its absence from the Nahal Hever marriage documents may be an idiosyncracy or may suggest alternative traditions. In the Elephantine documents, it is followed by stipulations for the departure of the woman and her possessions from the mans house; discussion of a sum to be paid
by the party who initiates the divorce, termed ksp snh divorce penalty (literally, money of hatred); and the forfeiture of future legal claims. The situation in the Genizah documents is similar. Proponents of a translation repudiate or the like, implying other components to the divorce, have failed to provide any of these other components. Second, the formula predicated a state of affairs beginning at the moment of utterance and extending into the future. It could be expanded with a more or less abbreviated reversal of the present-future marriage formula (for example, l hwh lk ntt I will not be your wife, TADAE 3.8:25, though this was optional, with no sign of additional legal force). The divorce formula was opposite and parallel to the marriage formula hy ntty wnh b l[h] She is (hereby) my wife and I am (hereby) [her] husband (TADAE 3.8:4, already noted by Yaron, Introduction, p. 55), which similarly both refers to and brings about an event at the moment of utterance. Thus hatred was the native legal term for divorce. In the terms only appearance outside of marriage documents, the grant of usufruct TADAE 2.4, the act dissolving the marriage bond between the daughter and son-in-law, leading to division of property, brings about a state of hatred (2.4:8), and the money to be paid by the spouse initiating the divorce is money of hatred. Like the utterance of the hate formula, the action prohibited in 2.4 brings about a state of affairs designated by the recognized legal and social concept snh. 34 It would seem that legally the hate formula functioned in this cultural discourse by negating the ofcial state of good will upon which the marriage legally depended (for the role of an ideology of goodwill in constructing legally binding relationships in
172
they appear, all ve are evidence of a native Ugaritic folk-theory of how to do things with words. In the text examples below, performatives are in boldface. The rst and most assured case is the use of ytn give in a magical-medical ritual against snakebite, KTU 1.100. As the above examples indicate, because of its widespread use, the y/ntn give root is one of the best examples of a performative verb in ancient West Semitic, covering the legal,35 as well as physical, act of giving. 1.100 is also of interest because of its structure as a historiola, a narrative with a ritual application.36
73 tn.km.mhry.nhsm yhr.tn.km / mhry. w bn.btn.itnny 75 ytt. nhsm.mhrk. bn btn / itnnk Mare: Give as my bride-price serpents, give poisonous lizards as my bride-price, adders as my wife-price.
Horon: I hereby give serpents as your bride-price, adders as your wife-price.
In addition to the well-understood legal and ritual role of y/ntn in West Semitic performatives, this narrative incantation provides its own clear dramatic context. The Mare demands that Horon perform an action of ytn to save her. While the physical actions of the protagonists have been described in detail up to this point (cf. the full elaboration of Horons physical expression and movements in lines 6167), here the incantation simply describes Horons response as a speech about his own action: ytt. Already in his 1978 study of the Ugaritic serpent incantation, Dennis Pardee translated Horons nal speech to the mother of horses as a performative utterance, I hereby give serpents as your brideprice.37 Morphologically a rst-person sufx form, the verb is used here discursively to answer a demand that was made using the same verb in the imperative, give serpents as my bride-price! In the narrative this speech comes at the conclusion of the text as a culminating act that works to resolve the problems of the narrative. Considering the clear parallels in the cognate languages sharing a similar tense-mood-aspect system, this is the best example of Ugaritics use of the sufx form for the performative. Our second example is a verb of speaking, pr, proclaim.38 It was plausibly originally as good an example as the rst, but the context is now broken. In 1.1 IV a passage is framed as a naming ceremony in which El dubs or appoints his son (line 12, tgr.il.bnh), probably Yammu.
ancient Aramaic and Babylonian tradition, see the work of Muffs cited above). The linguist Benjamin Lee Whorf referred to these culturally determined linguistic categories, which manifest themselves in practice rather than grammatical surface structure, as cryptotypes. The problem with cryptotypes is their extreme cultural embeddedness: while native-speakers of a language can freely produce examples, our texts cannot. The intuitions of an English-speaker, when applied to a culturally selected category like performative verbs, will be by denition misleading, since those intuitions will necessarily reect English- (or other standard European) language patterns and ideologies. 35 See, for exmple, KTU 1.3 I 10. 36 On the special relationship between narrative and ritual in ancient Near Eastern historiolas, see my A
Historiography of Demons: Preterit-Thema, Para-Myth and Historiola in the Morphology of Genres, in Tzvi Abusch et al., eds., Proceedings of the XLV Rencontre Assyriologique Internationale: Historiography in the Cuneiform World, vol. 1 (Bethesda, Maryland, 2001), pp. 429 40. The translation of KTU 1.100 is that of Pardee in William Hallo and Lawson K. Younger, eds., The Context of Scripture, vol. 1 (Leiden, 1997) p. 298. 37 N. Wyatts I shall give is another promising form in English but does not express the performative perfect. See his Religious Texts from Ugarit: The Words of Ilimilku and His Colleagues, Biblical Seminar 53 (Shefeld, 1998). 38 Olmo Lete and Sanmartn dene pr as abrir la boca > gritar, proclamar. See G. del Olmo Lete and J. Sanmartn, Diccionario de la lengua ugartica, vol. 1 (Sabadell, Spain, 1996), s.v.

15 16 17 18 19 20 wpr.sm.ym[ tnyn.lzntn [ at.adn.tpr [ ankltpn.il[ l.ydm.prt [ smk.mdd.i[l ] ] ] ] ] ] And he proclaims the name Yam. [ ] They answer . . . [ ] You, O Lord, may you proclaim [ ] I, Benicent El [ ] Upon the hands . . . I proclaim [ ] your name . . . Beloved of E[l ]39
173
While the passage is badly damaged, with about half of each line missing, the general narrative context is known and several key verb forms that syntactically frame the speeches are preserved. First, El holds a feast, inviting the other gods to his house. In a public declaration, El announces something about the name of my son (line 14, sm.bny). The narrator then announces a new event, designated as a proclamation by the verb pr in the sufx form preceded by waw. In a public ceremony, someone implores El to make the proclamation (line 17). In lines 1820, if the broken text is understood correctly by translators such as Smith,40 El says I proclaim . . . your name. Just as in the ritual exchange between the Mare and Horon in 1.100, the text itself sets up, and thus contextualizes, the use of a metapragmatic verb in the rst-person sufx form by rst having a participant request an action in a second-person form (the Mares imperative tn, the unidentied speakers tpr) addressed to the speaker of the performative. In both cases a request for action is answered by speech.41 What is important about this pattern is that it anchors the speech event in a narrative stream, thus providing a model for how gods do things with words in Ugaritic myth. We do not need to rely here on assumptions about the universal nature of performatives or even on comparative evidence for the general grammatical form of performatives in ancient West Semitic. As cultural artifacts that encode Ugaritic notions of how (divine) language works, the incantations and myths provide examples of what the right words in the mouths of the right performers can do. The interpretation of the naming passage above as a ceremony wherein El performatively dubs Yammu with a divine identity is strengthened by a second model of divine proclamation that also appears in the Baal Epic, also using this verb. In the speeches of Kothar-wa-Hasis in 1.2 IV 11 and 18, where the magical craftsman god creates Baals weapons, the verb pr is again used by a divine being for an act of naming. Here each magical creation is introduced by a statement that Kothar fashions the weapons (smdm.ynht) and designates their names (wypr.smthm).42 Remarkably, the weapons names are also actions as well as references to the very story in which they appearthus 1.2 IV 1112 and 19:43
smkat/ygrs. ygrs.grsym You, your name is It-Shall-Drive; It-Shall-Drive, drive Yam out!
42 On the verbal magic of Kothar-wa-Hasis, see briey Mark S. Smith, The Magic of Kothar, the Ugaritic Craftsman God, in KTU 1.6 VI 4950, RB 91 (1984): 37780, as well as Smiths Kothar wa-Hasis, the Ugaritic Craftsman God (Ph.D. diss., Yale University, 1986). 43 These verbs appear in references to divine combat at 1.1 iv 24 and 1.2 i 67, respectively, and may well be being used strategically and ironically here.
39 KTU 1.1 IV 15 ff., translated after Smith in Simon Parker, ed., Ugaritic Narrative Poetry (Atlanta, 1997) p. 89. 40 In fact Smith switches between pronounce (as in lines 15, 19) and proclaim (l. 17). 41 In the Baal passage this speech seems to be accompanied by a laying on of hands (see l.ydm in line 19), typical of a West Semitic ritual between a father and son.
174
smk.at.aymr. aymr.mr.ym

You, your name is Ay! It-Shall-Expel; Ay! It-Shall-Expel, expel Yam!44
The direct discourse of Kothar is not performative by virtue of social conventions but on a higher level: as sheer self-enacting divine language, Kothars speech is performative by cosmic law. The two named weapons that he both dubs and creates are jussive forms of telic, goal-oriented verbs. Immediately after Kothar designates the weapons names in verbless clauses, he invokes those names in imperative forms, grammatically shifting both the verbs and the weapons with the command to attack built into their names. Immediately upon being shifted from the decontextualized third person and deictically inserted into the present context, invoked and addressed through second-person imperatives, the weapons act to meet their goals. For Kothar, to unpack their verbal identity is to detonate the weapons, which proceed to defeat Yammu more or less by themselves. The self-activating verbs stored in the weapons names are thus icons of self-performing actions. Kothar-wa-Hasis, as a kind of divine wordsmith, has the power to craft weapons that work like performatives: being spoken, they act. This passages signicance goes beyond helping dene pr as a performative verb. The weapon-naming scene provides an Ugaritic mythic model of self-enacting divine language, whereby the act of pr, in the mouth of an empowered divine speaker such as Kothar, causes the words framed in divine discourse to jump up out of that discourse into the narrated reality. Because it names the act of speech itself, the semantics and pragmatics of pr offer an ideal model of how divine language was supposed to act in Ugaritic. With all due reservations for the broken context of our second example, the instances discussed exemplify ritual language by contextualizing it and reframing it through shifts in person and mood. These narratives show how ritual language is moved through direct to indirect discourseletting readers see someone say it and then letting them watch how it happens when it is narrated. Thus this use of pr provides a context within the text, a context created by Ugaritic speakers themselves. I will now describe three problematic cases. The rst is another verb of speaking, qra.45 It appears in a passage at the beginning of the famous liturgy KTU 1.161, to summon the ancestor spirits known as Rephaim from the underworld. The text is difcult because the passage admits of several valid but mutually exclusive readings. After considering some prominent suggestions, I will argue that there are reasons to prefer a performative reading. One interpretation that has contributed to our understanding of the Ugaritic and Near Eastern cultural context of KTU 1.161 is that of B. Levine and J.-M. de Tarragon.46 They read the passage as a series of imperatives, thus:
2 qritm.rpi.a[rs] 3 qbitm.qbs.d[dn] 4 qra.ulkn.rp[u] You summon the Rephaim of the netherworld; You command the Council of the Didanites! Summon Ulkn, the Rapha!
44 The literal translation is mine; the clearest and most pungent English rendering is still that of Theodore Gaster: Thou, thy name is EXPELLER; Expeller, expel Sir Sea, expel Sir Sea from his throne. . . . Thou, thy name is DRIVER; Driver, drive Sir Sea, drive Sir Sea from his throne . . . , in his Thespis: Ritual, Myth, and Drama in the Ancient Near East (New
York, 1961), pp. 163, 167. 45 Olmo Lete and Sanmartn (Diccionario, s.v.) expresses the wide range of use as llamar, gritar cry out, invitar; invocar, evocar; proclamar. 46 Dead Kings and Rephaim: The Patrons of the Ugaritic Dynasty, JAOS 104 (1984): 64959.

5 6 7 8 9 10 11 12 qra.trmn.rp[u] qra.sdn.w.rd[n] qra.tr. llmn [ ] qru.rpim.qdmym qritm.rpi.ars qbitm.qbs.ddn qra.mttmr.m[l ]k qra.u.nqmd[.]mlk. Summon Trmn, the Rapha! Summon Sdn-w-Rdn Summon Tr-llmn (All) summon the most ancient Rephaim! You summon the Rephaim of the netherworld; You command the Council of the Didanites! Summon Ammishtamru, the king! Summon, as well, Niqmaddu, the king!
175
Levine and Tarragons overall interpretation is well documented, and their translation reads felicitously in English; however, their grammatical interpretation, following that of Pope, assumes a pattern not widely attested in Ugaritic and ancient Near Eastern ritual. They read the rst eleven verbs as imperative instructions to unnamed participants. While ritual texts are often elliptical, assuming a level of prior knowledge, this choice of readings produces an especially redundant and ambiguous text. The participants are repeatedly instructed to summon various cthonic beings from the underworld, but they are not told how to summon them or what to say. And this is a signicant gap because some of the cthonic beings47 are mentioned only in this passage: why are they named but not explicitly summoned? Read this way, they are not actually summoned; someone is merely ordered to summon them.48 After this series, the text is then understood to move into direct discourse, not commands about the ritual but the ritual itself; except for a single unmarked rubric in line 19, the rest of the text would be a liturgy. Here the performers and divine beings act out a ritual drama of mourning, descent into the underworld, and sacrice. The resulting text, which switches from a series of rubrics to the text of a liturgy, appears incoherent. Next we consider Lewiss interpretation.49
2 3 4 5 6 7 8 9 quratum rapima a[rsi ] qubatum qabusi d[idani ] quraa ulkn rap[u] quraa trmn rap[u] quraa(a?) sdn-wa-rd[n] quraa tr llmn qurau rapima qadmiyyima quratum rapima arsi You are summoned, O heroes of the underworld, You are invoked, O gathered ones of Didanu! NN, the hero, is summoned, NN, the hero, is summoned, NN is (are?) summoned, NN is summoned, The heroes of old are summoned!50 You are summoned, O heroes of the underworld,
47 Namely, Ulkn, Trmn, Sdn-w-rdn, tr. llmn, and possibly qbs.ddn (if this last is different from the Council of Ditanu, mentioned in Kirta). 48 It is possible to argue here that the mere act of commanding an unnamed participant to summon these beings could itself effect the summoning. But do we have examples of this? Surely our texts are often allusive and ambiguous (witness the Genealogy of the Hammurapi Dynasty, edited by J. J. Finkelstein, JCS 20 (1966): 95188, to which KTU 1.161 is undoubtedly kin). Here we could simply ll the gap this reading creates with the ad hoc assumption that unnamed participants would be expected to perform an unscripted part of the ritual involving the actual summoning. Justifying this would require argumentation and parallels
within Ugaritic or other contemporary cuneiform cultures. In this case, Levine and de Tarrgon offer the intuition that alternative readings would be less dramatic than actively exhorting those assembled to summon them (Dead Kings, p. 652). This may seem true to us, given the tense-mood-aspect system of modern European languages, but begs the question of what the ancient West Semitic rules would have been. 49 The vocalization is mine, while the translation is based on that of Lewis, Cults of the Dead, p. 7. 50 Reading, with Tropper, these plural nouns as subjects with oblique case-marking, a phenomenon well attested in Ugaritic. See Tropper, Morphologische Besonderheiten des Sptugaritischen, UF 25 (1993): 39091. Several alternative, nonperformative readings
176

You are invoked, O gathered ones of Didanu! King Ammishtamru is summoned, King Niqmaddu is summoned as well!
10 qubatum qabusi Didani 11 quraa Ammittamru m[al ]ku 12 quraa u Niqmaddu malku
From a grammatical point of view, Lewiss solution is more elegant. Reading all of the verb forms as G passives produces a pattern consistent with the structure of lines 212, beginning after the rubric of line 1 and marked off by the -u- of 12. Similarly, the vocalization of the subjects, the Rephaim, in the vocative (with oblique morphology) is perfectly consistent with the second-person verb forms that address them.51 But the reading of Pope, Levine, and de Tarragon is also grammatically correct. How can we decide between them? Our discussion so far has moved directly from the level of isolated letters and words to general intuitions about what the text should be saying. But there are crucial details of the texts language and genre within Ugaritic that can make this connection between form and content both more explicit and more rigorous. While Lewiss solution is inherently plausible, two philological questions remain: why passives (you are summoned) rather than actives (I summon you), and is there further evidence within Ugaritic justifying qras performative character? The question of the passives in 1.161 should not be seen in isolation from the passivity of the text as a whole. That is, there are no named rst- or second-person agents in the entire ritual. None of the pragmatic language in the text, whether Lewiss proposed performatives at the beginning or the commonly accepted imperatives in the following section, comes from a specied source. This is in contrast to ritual texts such as 1.23 and poetic texts such as 1.24, both of which begin with a rst-person reference to the speaker. Wyatt has recently made a striking proposal on this subject, connecting the passive form to the inherent danger of dealing with underworld beings.52 These dangerous forces, whom one usually wants to keep underground, are here summoned up in the polite or remote passive
of line 8 are possible: the most grammatically straightforward, if contextually difcult, reading is that of Pardee, Les textes rituels, Ras-Shamra-Ougarit, vol. 12 (Paris, 2000), p. 821 (cf. his English version in Ritual and Cult at Ugarit [Atlanta, 2002], p. 87, They have called the ancient Rapauma). Also possible is a jussive, as in qarau rapima qadmiyima May they (together) summon the heroes of old. As this vocalization indicates, the last qru rpim could be interpreted as a precative perfect, known from other Ugaritic incantations and well in place in ancient Near Eastern incantation traditions of summoning gods to summon other gods. Note the rhetoric of Maql I, incantation 1, or the incantation edited by Heinrich Zimmern in Beitrge zur Kenntnis der babylonischen Religion, Assyriologische Bibliothek, vol. 12 (Leipzig, 1901), p. 95: 47 with parallels in BBR 8990+K 3654+: 29 Nisaba the pure, exalted daughter of Anu, who summons the great gods, who summons the divine judges, who convenes the great gods, who convenes the divine judgesmay she convene the great gods! May she convene the divine judges! For a partial reedition of this text with new joins, see W. G. Lambert and A. R. Millard Atra-Hasis: The Babylonian Story of the
Flood (Oxford, 1969), p. 154. 51 While this was the subject of some debate at the 2000 SBL presentation of this paper, the evidence is clear that vocatives with oblique case marking were used in Ugaritic. For oblique (syntactically accusative) morphology on a vocative, see the one clear example of a ritual address to the Rephaim in the vocative call of El, lk bty rpim set off to my house, O Rephaim in KTU 1.21 II 9. This is the only completely preserved reference to the Rephaim in this text that is clearly being spoken in the rst person in direct address; other instances such as 1.21 II 3 4, which receive nominative morphology, are spoken in the third person in the narrators voice and may not directly address the Rephaim. There is no need, grammatical or contextual, to translate the latter forms as vocatives. Treating them as such produces unnecessary inconsistencies (cf. Lewiss resulting uncertainty as to how to reconstruct the alephs in Els vocative addresses in 1.21 II 2 56, 1.22 II 3 8, in Ugaritic Narrative Poetry, pp. 199201). I do not understand the readings in the second edition of KTU, which reconstruct the forms as nominative in 1.21 II 56 and I.22 II 8 and oblique in 1.22 II 3. 52 Wyatt, Religious Texts from Ugarit, pp. 43233, n. 8.
177
voice, with no marked agent. This denial of personal agency, and thus responsibility, is well known in Mesopotamian ritual. Parallels in Sumerian incantations declare, This is not my spell, it is the spell of Enki (lord of magic), a statement termed a prophylactic formula by Falkenstein in his typology of Sumerian incantation forms. In English the passive performative of bureaucratic proclamations has a similar effect, diffusing or redirecting responsibilitywe will return to the relationship between politeness, power, and participant responsibility in performatives at the end. In this view, the absence of the rst person and the restriction of the second person to ritual instructions to anonymous priestly participants in lines 2730 results in all of the explicit agents and none of the objects of human action 1.161 being divine. Pardee suggests that this in itself may have been the deliberate rhetorical strategy of the text.53 Turning next to internal Ugaritic evidence for the verb: while clearly a verb of speaking, qra in Ugaritic seems to be conned to a poetic register. The only other instances of the root occur in ritual and myth: 1.100 2, 57, our serpent incantation again, where gods in distress cry out to each other (as well as parallel passages in 1.107 9, 15). In 1.4 VII 47, qra is used in a divine-royal summons or message to Mot, and in 1.5 II 22 Mot himself explicitly refers to it as an act of qra with the social effect of summoning him to a meal. A nal mythic example, KTU 1.23, The Birth of the Gracious Gods, is a ritual text designed for performance at a feast with invitations to the participants to eat bread and drink wine. In the rst, explicitly ritual, half of this text, the speaker twice announces his wish to summon the twin protagonists of the text to the banquet: Let me invoke the gracious gods (iqra.ilm.n[mm.] in line 1; similarly, iqran.ilm.nmm[.] in line 23). Of course, the use is not grammatically performative and is not represented as producing any immediate effect; while the two invocation passages contain addresses in the second person, the gracious gods continue to be addressed in the third person as if they are not present. Yet the narrative nonetheless ends with the gods successful arrival. How does this invocation work? The second half of the text contains a narrative myth, a historiola, which describes the birth, ravenous hunger, and exile in the wilderness of the gracious gods. It concludes with a mythic being, the Guardian of the Sown (as Lewis renders nfr.mdr), allowing them to enter a place where their request to eat bread and drink wine is nally granted. The place to which the guardian invites them fuses with the occasion at which the myth is narrated; the gracious gods arrival is framed in narrative rather than imperative direct address. Thus, while the qra-invocation of the gracious gods is not grammatically performative, it is performed by being narrated in the ritual. The form of the myth of the gracious gods is more elaborate than that of Kothar-wa-Hasiss speech; yet it is structured similarly: both act metapragmatically, calling their goals into being by talking about them. Ugaritic myths and performatives both accomplish action iconically, by modeling it through language. All of these examples are of both philological and theoretical signicance because they represent Ugaritic, rather than English-language, concepts of how words are supposed to work in ritual. Because the narratives display correct uses of language and their outcomes, the myths actually function as working examples of how one invokes gods at a distance, whether in the sky as Shapshu, in the underworld as Mot, or at the banquet, using the root
53
Pardee, personal communication, April 2001.
178
qra. Thus, these ritual examples of qra using sufxing forms, as well as a historiola, serve as instances where the text represents itself as acting on a divine being and provides divine models for it. The second difcult example is yet another verb of speaking, rgm. There are cases in letters where this just means tell, say, inform.54 Similarly, in the myths, it is used mostly for transmitting messages (1.1 III 4, 1.2 I 16, etc.). Yet there may also be a legal use declare.55 More prominently among recent treatments, the usage in 1.2 IV 7, lrgmt/ lk.lzbl.bl Indeed, I tell you, Prince Baal has been generally taken as performative.56 This raises the basic question of whether verba dicendi are automatically performative: is any metonym of say translatable as I [hereby] say? According to the simple, roughand-ready denitions of the performative most Semitists have used, all verbs of saying would qualify. But Ugaritic evidence is dubious here: intriguingly, the parallel verb tny, repeat, is never taken as performative. Yet it is also used to introduce direct discourse. The immediate problem is one of nding examples of the consequences of using rgm: does it produce any special effects? In 1.2 I 45, an.rgmt.lym . . . hwt.gmr.hd . . . has been translated by Pardee as I, for my part, hereby say to Yamm but also as mere I say to Yamm . . . word of the Avenger Haddu (as Wyatt, similarly Smith). Clearly this frames a message from Baal, but the context is too broken to see what the message does. The larger problem is whether one can performatively introduce direct discourse. According to some of the denitions we have seen, all verba dicendi automatically are performative, since they clearly do something (announce, declare etc.) by reporting on it. But to say I say X does not really perform the act it names: it is metalinguistic, but is it metapragmatic? According to English-language work on performatives, the answer is clearly nosimple verbs of speaking are not automatically performative. Austin himself describes verbs of speaking such as I cite and I quote as dubious. He ends up excluding them from his category of expositives.57 Later work makes explicit that the clausal complement of a performative verb of saying must be indirect discourse, a that clause or a to clause but never a quotation. The reasons for this are essential to the way the performative works. The clausal complement of a performative must describe an objective state of affairs to be brought about (in terms of the philosophy of language, the performative requires a de re rather than a de dicto reading of the clausal complement).58 But since this theoretical description is based on English, there may well be entire different linguistic ideologies underlying Semitic performatives. Promising foundations for research on
54 Olmo Lete and Sanmartn (Diccionario, s.v.) decir, anunciar, comunicar, informar; contestar; recitar. 55 Pardee renders rgmt in KTU 2.33:25, in Further Studies in Ugaritic Epistolography, AfO 31 (1984): 216 as a Gp, 2000 horses have thus been declared against me. Similarly in 2.42:6, Epigraphic and Philological Notes, UF 19 (1987): 205: I indeed do speak (to Baal Saphon and other deities). Unfortunately there is no way to distinguish from a past use here, and if the addressee of his speech is indeed a group of gods, it is unclear how he is hereby speaking to them. 56 Pardee may read a performative here (I hereby announce to you, Prince Balu, and I repeat, Cloud-
Rider; see Context, p. 248), while both Wyatt (Religious Texts, p. 65) and Smith (Ugaritic Narrative Poetry, p. 103) do not. 57 Austin, How to Do Things with Words, pp. 85, 16163, where on p. 161 he explicitly classies quotation and citation as dubious because they may just be describing practice. 58 Compare the de re (and acceptably performative) We the jury nd the defendant innocent with the de dicto (and unacceptable) We the jury nd, youre innocent. On de re and de dicto, see further Lee, Talking Heads, p. 83.
One Line Short
179
these ideologies have already been laid by recent work on direct and indirect discourse in Ethiopic and Hebrew.59 As a nal example, I will allude to the unpublished study of F. W. Dobbs-Allsop on the so-called epistolary performative I hereby fall, which occurs in both Akkadian and Ugaritic letters from subordinates to superiors.60 Two Ugaritic examples are:
KTU 2.13 56 l.pn.umy/qlt I hereby fall before my mother
KTU 2.12 69 l.pn/adty/svd/w. sbid/ 1011 mrhqtm/qlt I hereby fall before my mistress, seven and seven times from afar.
The subject has been thoroughly treated by Pardee and Whiting, who also classify these occurrences as epistolary performatives, a special type of speech act that can only take place in letters. In his reexamination of the subject, Dobbs-Allsop observes that if the letters are indeed written to be performed in front of the superior as if the inferior is speaking to them, then the forms that appear in Ugaritic letters as qlt I hereby fall may be taken as ordinary performatives. The presence of an intermediary does not affect the status of the performative, since all of the other features are in place. To show that one may fall performatively without the need for a physical gesture, Dobbs-Allsop cites the parallel of 2 Sam. 16:4, where Ziba, Mephibosheths steward, presents David with gifts. David rewards Zibas loyalty by giving the steward all of his masters possessions. Ziba replies: hsthwy ty ms hn bynyk dny hmlk I do obeisance; let me nd favor in your sight, my lord the king (NRSV). But according to Pardee, this still begs the question of the speaker: the problem Pardee and Whiting recognized is that these sentences look like ordinary explicit performatives, but they do two things differently from the explicit performative. The performative is a specic action that happens under a specic set of circumstances. But when and under what circumstances is the act of falling understood to occur? Furthermore, the performative has only one logical (agentive) subject and one logical (patientive) object. So who is the speaker who performs the fall from afar? As in the case of identifying performatives in myth, we may be helped out of a modern theoretical impasse by attending to the way ancient writers and speakers expressed themselves. Note that at least four of the Ugaritic letters add the qualication qlt mrhqtm I fall
59 On Direct Speech and the Hebrew Bible, in Gideon Goldenberg, Studies in Semitic Linguistics: Selected Writings (Jerusalem, 1998), pp. 197214, and Cynthia L. Miller, The Representation of Speech in Biblical Hebrew Narrative: A Linguistic Analysis (Atlanta, 1996). In fact, Silverstein has argued (Language Structure, pp. 21112) that our understanding of performatives as bridging a putative gap between speech and action itself ows from an Englishlanguage ideology based on structural differences between direct and indirect discourse in English. Since
Goldenberg and Miller have demonstrated that this difference does not hold, or at least not in the same way, in Biblical Hebrew and other Semitic languages, the performative ideology as expressed in English should be expected to be different in Semitic. 60 See F. W. Dobbs-Allsopp, I Hereby Fall: A Performative Utterance in Western Peripheral Akkadian, Ugaritic, and Hebrew (unpublished MS, 2000), which disputes the conclusions of Pardee and Whiting in Aspects of Epistolary Verbal Usage.
180
from afar.61 This suggests that the lack of physical copresence may have been salient for them as well as us: the acknowledgement of distance suggests the author himself did not consider his a straightforward fall. The resulting fall from afar is an act of physical obeisance performed in a space and time that seems peculiar to us. But given the evidence of its widespread use and conventional nature, there is no doubt it was an ordinary ritual occurrence. On the other hand, even if the author explicitly acknowledges that he is not physically present, he feels himself able to utter the performative without acknowledging that he is also not speaking it. The physically present speaker of the text, the messenger or scribe who read the text out loud, is treated as transparent. The term speaker is therefore not appropriate here because it does not jibe with the evidence we have of the participants awareness. Instead, Ugaritic awareness is more precisely described in terms of participant roles that distinguish between the author, who is both the composer of the text and the agent responsible for it (the I to which the text refers) and the animator of the text, who performs it (in the nontechnical sense).62 Perhaps a better question, then, is not whether the fall from afar is an ordinary or epistolary performative, but what kind of Ugaritic performative it is. That is, the question is how this ancient culture indexed different participant roles in a situation more complicated than face-to-face communication. At least in the high Ugaritic culture of diplomacy, it appears that the author who dictates, and has someone else read, the performance of a physical gesture that the author cannot physically perform in the addressees presence is a different sort of participant, with a different sort of culturally dened presence and responsibilities, from someone doing the same thing face to face. And because this sort of performative can only be addressed by a highstatus person to a higher-status person far away, it is at least as strongly linked to differences in power as it is to the medium of communication. Further research on ancient performatives will attend to the relationship between performativity and honorication to nd ways that different performatives may be possible only in different power relationships.63 The work of D. Manahlot on Tigrinya performatives provides a striking parallel. Here there are two classes of verbs used, with two different complementation structures, depending on the status relationship between the speaker and hearer. Furthermore, the more polite verbs are used differently in writing, appearing in the imperfective form at the end of formal business letters.64 Whether or not these imperfective forms turn out, on examination, to be explicit performatives, the point is that in both ancient and modern Semitic, language and social power are not opaque to each other; we can attend to the specic ways that each helps create the other.
61 2.12 10, 2.64 15, 2.68 5, 2.24 7 (broken but reasonably certain). For a well-documented opposing argument that from afar represents the respectful distance kept by an inferior in the presence of a superior, rather than the distance between two correspondants, see S. E. Loewenstamm, Prostration from Afar in Ugaritic, Accadian and Hebrew, BASOR 188 (1967): 41 43. 62 This helps explain the sort of violation involved in shooting the messenger, when an animator who bears no agency is nonetheless assigned responsibility for the contents of the message. The assumptions of a whole system are being attacked, a tactic that can eas-
ily backre. For these categories and their application to ritual, see William F. Hanks, Exorcism and the Description of Participant Roles, in Michael Silverstein and Greg Urban, eds., Natural Histories of Discourse (Chicago, 1996), pp. 160200. 63 On honorics and related sociolinguistic processes, see again Agha, Honorication. 64 Some Notes on Amharic Performative Verbs, in Taddese Beyene, ed., Proceedings of the Eighth International Conference of Ethiopian Studies, University of Addis Ababa, 1984 (Addis Ababa, 198889), vol. 1, pp. 62328.
Performative Utterances and Divine Language in Ugaritic VI. Conclusion
181
In this paper I have tried to show how our understanding of the category of performative in Ugaritic is improved by a dialogue between theory and philology. The most important result of these investigations is to show that Ugaritic texts present their own models of how language, and thus culture, works. Investigating the theory of the performative helped us see this in the texts because a theoretically sound view of the performative pays close attention to native metapragmatics, what the culture itself says about how language acts. This can act as a corrective to our own impulse to project our own native notions of language and culture onto those of others. Investigating metapragmatic discourse in ancient Near Eastern cultures opens up native thought about myth, magic, and religion. In Ugaritic, by contrast with the many theories of magic foisted off on ancient pagans, we can now hold up a native Ugaritic theory, expressed in myth. This theory holds that divine language is self-enacting. In the Kothar-waHasis passage examined above, gods words are icons of things. Things are endowed with identity by dubbing them with names, but divinely endowed names themselves retain their wordhood in the fullest sense: they are semantically and pragmatically transparent and alive, open to grammatical change that effects physical action. They represent the ideal of fully productive pragmatic language. Human language, even magic and ritual language, does not work the same way. Instead of being directly self-enacting, it must work through modeling. Gods cannot be invoked directly but must be invoked by narrating the actions of other divine agents invoking them, as in the case of the gracious gods, or by drawing on unnamed agents as in the case of the passive qra sequence in 1.100. I will conclude by indicating another, more strictly philological, dividend of a more sophisticated theoretical understanding of the performative. This is that understanding metapragmatics helps us understand historical grammar. We have seen that the grammatical form for the explicit performative in ancient West Semitic is universally the sufx form, corresponding to the expected cross-linguistic pattern. But as Rogland has demonstrated, the canonical form in Syriac (and, as we have seen, in Rabbinic Hebrew and Aramaic as well) is the participle.65 As he has also demonstrated, this shift dates to around the time of Qumran Aramaic, with the exception of the frozen form qblt, which dates back to the Elephantine papyri. What explains this shift? One should recall that at precisely this time there is a shift in the tense-mood-aspect system of both Hebrew and Aramaic (probably due not to Aramaic inuence so much as a mutual Sprachbund 66) from a two-part to a three-part system. In Mishnaic and later Hebrew, as well as in Aramaic, the perfect form becomes marked for past tense, the imperfect for future, and the participle an unmarked, and morphologically minimal, present. This ts Silversteins pattern whereby performatives select for both morphologically and semantically minimal forms. The later history of the performative in West Semitic is thus predicated by our cross-linguistic denition.
65 Performative Utterances in Classical Syriac, JSS 46 (2001): 24350. 66 A good characterization of the conditions of
Sprachbund is available in Sarah Grey Thomason and Terrence Kaufman, Language Contact, Creolization, and Genetic Linguistics (Berkeley, 1988), pp. 9597.

Performative Utterances

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Performative Utterances

Uploaded by

Copyright:

Available Formats

Performative Utterances and Divine Language in Ugaritic Author(s): Seth L.

PERFORMATIVE UTTERANCES AND DIVINE LANGUAGE IN UGARITIC

Journal of Near Eastern Studies

Performative Utterances and Divine Language in Ugaritic

Journal of Near Eastern Studies

Performative Utterances and Divine Language in Ugaritic

Journal of Near Eastern Studies

Performative Utterances and Divine Language in Ugaritic

Journal of Near Eastern Studies

One Line Long

Performative Utterances and Divine Language in Ugaritic

Journal of Near Eastern Studies

Performative Utterances and Divine Language in Ugaritic

Journal of Near Eastern Studies

Horon: I hereby give serpents as your bride-price, adders as your wife-price.

Performative Utterances and Divine Language in Ugaritic

Journal of Near Eastern Studies

Performative Utterances and Divine Language in Ugaritic

Journal of Near Eastern Studies

Performative Utterances and Divine Language in Ugaritic

Pardee, personal communication, April 2001.

Journal of Near Eastern Studies

One Line Short

Performative Utterances and Divine Language in Ugaritic

Journal of Near Eastern Studies

Performative Utterances and Divine Language in Ugaritic VI. Conclusion

You might also like