Bryn Mawr Classical Review 1999.09.26
June Allison, Word and Concept in Thucydides. Atlanta: Scholars Press, 1997. Pp. xvi, 278. ISBN 0-7885-0363-4. $27.95.
Reviewed by Susan Prince, University of Colorado, Boulder (Susan.Prince@Colorado.edu)
Word count: 2955 words
Most scholars of ancient thought believe that serious, self-conscious reflection about the nature of language and the relations among language, thought, and the sensible world began before Plato and Aristotle and followed lines of inquiry broader than those pursued by the characters in Plato's Cratylus. The problem comes in pointing to a clear instance of this serious thought in any surviving text which is itself undigested by Plato, Aristotle or their followers. June Allison, developing intimations of previous scholars, invites us to look to Thucydides as a primary contributor in this realm, indeed as the earliest surviving primary contributor to the development of a metalanguage for theoretical discussion about language and thought. Word and Concept in Thucydides promotes the thesis that Thucydides was an original contributor to the development of abstract language in the fifth century B.C., not just an inheritor of developments achieved by the Pre-Socratics and Sophists. As in her previous book on Thucydides,1 A. strives to build a broad argument about Thucydides and his intellectual outlook from the countable and statistically analyzable items in his lexicon, including their deployment in context and across contexts. In its attempt to address, and presumably to cross-pollinate, two distinct audiences, historically minded philosophers of language and thought on the one hand and scholars of Thucydides on the other, this book is boldly original. Although classicists working in ancient thought often wish for better communication across the divide between modern analytic philosophy and classical studies with its cultural emphases, few scholars have dared to risk the possibilities of anachronism inherent in A.'s project: to approach Thucydides' work from modern points of views about language, thought and concepts, namely those of R. Carnap, W. V. Quine and W. Sellars.2
A. is well aware of the risks she takes, and she justifies her project and method in the preface and introductory chapter. The Platonic bias is so well embedded in all of us, according to A., that "it is not entirely possible even to examine pre-Platonic Greek with an unprejudiced, un-Platonized eye" (7). However, according to A. there is one standpoint from which we can try to view the pre-Platonic past without Platonic lenses: "at least recent thinkers from Wittgenstein on have provided a way to talk about words, syntax, cognition, and existence," sc. in non-Platonic terms. Modern metalinguistics, A. states, "can serve as an explicans and yet not contaminate the explicanda". In fact, the climax of A.'s book is the identification of a group of terms which, she argues, constitute Thucydidean metalinguistics; although A. comes close to identifying some of these with terms in modern metalinguistics, her arguments for taking these terms as metalinguistic are based on Thucydides.
The book consists of a preface, five chapters, a conclusion, an appendix of abstract nouns which are hapax legomena, either in T. or in Greek, organized by suffix in -σις, -ια, or -μος, and two bibliographies and three indices. Chapters three to five are themselves subdivided into five, three and three parts, respectively, "loosely connected" by A.s own account. Space does not allow for a thorough examination of each part. A brief summary of the highlights will be followed by further evaluation of the project and central thesis.
A brief preface (pp. ix-xvi) and the introductory chapter (pp. 1-18) describe the intended audience, introduce the methodology and the thesis ("The History must be reckoned as the earliest extant text from which one can isolate a definition of a concept by means of metalinguistic vocabulary," xi), address the definition of "concept" and the criteria for locating self-consciousness about concepts, and promote T.'s History as a place to look for self-consciousness about concepts. A. does not take a precise position on the nature of "concept": this is "whatever relationship exists between word and thought" (xi). She identifies F. Solmsen and A. Cook as primary scholars of Thucydides who have denied that T. was self-conscious about concepts (xi n.2); later (7) she adds the more recent T. Cole.3 In general she enthusiastically endorses E. Havelock's recognition of the importance of writing in the development of self-conscious thought and his passing identification of Thucydides as an important figure in this movement.4
A.'s professed methodology has three strands: listing and cataloguing the abstract nouns that occur throughout the work, as a way to get an objective handle on the text; "straightforward philological and stylistic analyses," that is, interpretations of short individual passages; and analyses of T.'s language, that is, his articulations of ideas as complexes of particular words, in "philosophical" and ontological terms rather than "linguistic" or literary-critical terms. In laying out the philosophical and ontological terms that are to be useful, A. provisionally adopts Sellars' developmental framework for the capacity of a language to articulate abstract concepts, then rejects it in favor of Quine's more fluid theory of semantic ascent, "the shift from talking in certain terms to talking about them." It is talk about words that Allison recognizes as the critical use of abstract language which will open the door to identification of a Thucydidean metalanguage.
The third chapter, "Thematic Uses of Abstraction" (pp. 35-103), consists of two studies in word repetition ("Repetition and Shorthand" and "Artistic and Rhetorical Emphasis") and three readings of extended passages ("The Plague," "Diodotus," and "Book 8: The Strategy of Composition"), all demonstrating aspects of the effects T. achieves in his use of abstract nouns. In "Repetition and Shorthand," A. examines repetitions of the root χωρ- "proceed" at 4.127 ff. and of τειχ- "wall" in clusters of passages in books 1,2,6 and 7 and finds that words in these lexical families are deployed in an order which, as it were, forces the process of semantic ascent. This is an interesting idea; A.'s proof suffers somewhat from the few cases noted and from the untidy patterns in the examples she does explore. Part C, "The Plague," aims to demonstrate how T. "moves from unbiased, observable features to the account of the event" (66). The progress from thing to word and concept is interestingly illuminated here because T. is unable to complete it: he can describe the symptoms of the plague, but he does not know its nature or cause. T. defers the final articulation of what the plague is to individual doctors and laypersons, who should provide the account (λεγέτω) "as each knows" (ὡς ἕκαστος γιγνώσκει, 2.48.3). On the strength of this passage A. attributes to T. a sort of subjective epistemology, which she contrasts with that attributed to Protagoras by Plato. In Part E on Book 8, A. argues that the relative lack of abstract nouns after 8.4 reflects an unfinished composition, still essentially oral instead of literary. A summary of the chapter refers to Havelock's theory of the connection between abstract thought and the technology of writing as a partial explanation for T.'s complexity and intellectual progress; it then characterizes Thucydides' activity with abstract words as semantic ascent in Quine's sense. A.'s use of Quine here at the end of her lengthy chapter is a good illustration of her general method, to examine T. first on his own terms, then compare him to modern philosophical work. Since T. is allegedly practicing semantic ascent, not talking about it, as Quine is, it is certain that T. is far less of a linguistic philosopher than Quine; the relevance of Quine's terminology is nevertheless well worth entertaining.
In the fourth chapter, "Linguistic Considerations" (pp. 105-161), A. embarks on more technical issues: "Nominative abstracts and periphrasis," "Choices: -μος or -σις nouns," and "Comparison." "Nominative abstracts and periphrasis" examines the use of abstract nouns as subject of predication, especially of the verbs εἶναι and γίγνεσθαι. Acknowledging the previous work in this vein by Porzig and Freundlich,5 A. argues that this usage tends not to be a form of personification, as it is often in Homer and the tragedians, but the reification of a process, that is, an abstraction. Some of the predications, she argues, come close to category statements (122), the kind of metalinguistic statement of which Aristotle is the master. Part C, "Comparison," introduces the first items in the metalinguistic vocabulary A. proposes for T.: In T.'s comment on why Homer specified the capacity of the Boeotian ships and those of Philoctetes but not others (δηλῶν, ὡς ἐμοὶ δοκεῖ, τὰς μεγίστας καὶ ἐλαχίστας, 1.10.14), A. finds an embedded process of concept formation through comparison (145): in a process of δόκησις, a perceiver (here T.) takes in evidence, then he formulates, by means of comparison, what seems likely or εἶκος (a term used in the previous sentence, 1.10.3); when a perceiver (here Homer) states a judgment, he participates in a process of δήλωσις. A. may be overreading this passage; nevertheless she here begins to identify the set of terms which in the following chapter she will set out as T.'s metalinguistic vocabulary, and this in its cumulative force is more persuasive.
In the fifth chapter, "Thucydides on Language" (pp. 163-238), the climax of the book, A. discusses the most explicitly metalinguistic passage in the History and posits T.'s metalinguistic lexicon. This chapter again has three parts, "The ἀξίωσις of words," "Words on Words," and "ἀλήθεια". "The ἀξίωσις of words" is a reading of 3.82-83 in light of the fifth-century debate about language and thought: A. undertakes to examine the meanings of the key terms in the key passage, καὶ τὴν εἰωθυῖαν ἀξίωσιν τῶν ὀνομάτων ἐς τὰ ἔργα ἀντήλλαξαν τῇ δικαιώσει (3.82.4). A. understands the relationship between ἀξίωσις and δικαίωσις as the core of this statement, rather than the relationship between either of these terms and ὀνόματα or ἔργα, since ἀξίωσις and δικαίωσις are the words for the acts of people; this is convincing. What changes during civil strife, she emphasizes, is not the (theoretical) meaning of the words, but the value of the words, that is, the normative appropriateness of a word for a thing. The speaker's formulation of an utterance is a complex process: first, within his or her own mind the speaker performs a δικαίωσις or judgment (of the thing), then in the act of ἀξίωσις or assessment of reputation (of the language) he or she refers to the public, social norms of language use. ἀξίωσις refers to the standard of correct language: this standard is not ontological, as in Plato's system, but social and normative. A.'s argument here requires that T. is being precise in his use of terminology: in her examination of further passages in which one or more of these terms occurs (e.g. 4.86.6, 2.61.4, 5.26.2) she is sometimes at pains to maintain a consistent account of their interrelation. Nevertheless, her demonstration that T. is talking about more levels than words and objects is valuable and convincing. In this she builds on the work of others, including N. Loraux.6 A. concludes the section by positing δικαίωσις and ἀξίωσις as metalinguistic terms for two higher levels in the process of formulating speech. On the strength of her previous discussion of the plague passage (ch. 3.C), and further discussion of passages in book 6, A. proposes that the bottom level in this process is that signified by the term δόκησις: this is the direct, prelinguistic evaluation of a sense perception.
Part B, "Words on Words," is a glossary of forty-three terms or families of terms A. claims constitute T.'s metalinguistic vocabulary, excluding the basic Greek vocabulary for speaking, such as λέγειν. As she acknowledges, A.'s glossary reflects her own reconstruction of T.'s views of language and thus sometimes seems an arbitrary selection. The glossary shows the extent and variety of T.'s vocabulary for the processes of converting things and events into words and back again. T. borrows the words for his new vocabulary about language mainly from the realms of social interaction and sight: he is not original in using these realms metaphorically, but he is original in using them metalinguistically. A. concludes the section with a comparison between T.'s terminology and that of Plato, who, she proposes, selected from T.'s vocabulary; she discusses δήλωσις, a key term in Cratylus, at length (195-198). The survival rate of fifth-century literature must give us pause here: the stem δηλο- happens to be used of language also by Gorgias, if the pseudo-Aristotelian de Melisso, Xenophane, Gorgia preserves his original words, and by Antisthenes, if Diogenes Laertius is quoting him at 6.3. It is not impossible that T. was nevertheless the writer who first used this term metalinguistically, or who handed it down to Plato, but there are several rival candidates. Part C, "ἀλήθεια," is a discussion of the nature of Truth according to T.: ἀλήθεια will turn out to be T.'s quintessential concept.
The Conclusion (pp. 239-250) ties Thucydides' innovations in metalinguistic vocabulary to the growth of literacy in classical Athens and the highly literary quality of T.'s work and then summarizes the way T. achieves semantic ascent. A final chart (249) illustrates the ascent, on the part of a speaker, from δόκησις to δικαίωσις to ἀξίωσις to ἀλήθεια, which, according to A., would in modern terms be called a process of abstraction: the λόγοι which embody the ἀλήθεια then cause δήλωσις for the recipient of the message.
My general criticisms of the book are two. First, in seeking a point of view on thought and language unimpeded by Plato's Theory of Forms, A. might have found one right within the ancient world: the considerable work that has been done on Stoic views of language in the last two decades, by, among others, A. A. Long, J. Rist and M. Frede,7 has not been taken into account. The only published scholarly work on ancient theory of language cited in her bibliography which post-dates Classen's 1976 survey of the sophists is A. Silverman's 1992 article on Cratylus.8 Although the Stoics did succeed Plato, their views on speaking, thinking and perceiving are non-Platonic; indeed, their observation of these three levels, each divided from the next by some gap or potential slippage, is quite close to A.'s account of Thucydides. The views in Aristotle's de Interpretatione might also be relevant, and, if the pseudo-Aristotelian MXG preserves any record of fifth-century thought, we find the same levels already in Gorgias, who may have gotten them from Parmenides, e.g. fr.B6 (A. does discuss Parmenides repeatedly, but only as advocate of a two-level system). Ironically, given her goals, I found A.'s own view of ancient philosophy to be excessively Platonic. For example, she identifies the players in the fifth-century debate over the nature of language and truth as Parmenides vs. Protagoras. Parmenides is unquestionably central, but one might more plausibly understand the other side as anti-Parmenides, represented by many figures, including Democritus and Gorgias as well as Protagoras. A.'s poles are in fact Plato's, as found in Theaetetus and other key dialogues, and in the modern history of interpretation of these dialogues. Although A.'s recourse to philosophers since Wittgenstein for a non-Platonic point of view turns out not to contaminate her project, the book might have made a richer contribution to the study of ancient thought if it had engaged itself more with Aristotle and the Stoics. A. seems to target modern philosophers of language as part of her audience; however, her bibliography is outdated also in this field (her key theorists wrote in the sixties or before), and therefore the reception of this project among modern philosophers of language will probably be slight.
Second, in many parts of the book, the conclusions drawn about Thucydides' awareness of words and concepts could be asserted almost as successfully of a great poet, such as one of the tragedians or even Pindar and sometimes Hesiod, or alternatively of many Pre-Socratics. A. indeed compares T. to poets, especially Aeschylus (54). That is to say, in relaxing the criterion for self-consciousness about concepts to one which cannot be stated outright but becomes apparent through illustration, A. risks losing the criterion all together. She is aware that her case for T. is a cumulative argument, not an especially linear one, which in the end depends on acceptance of the higher level of communication T. is establishing between himself and his reader. It is really A.'s interpretation of the complex of terms in 3.82.4 that points T. out as more self-conscious about words and concepts than Aeschylus. And here we must consider many accidents, such as which of the perceptions about language and thought that occurred in the minds of fifth-century writers happened to get committed to the page, and which of their pages survive today, before drawing sharp conclusions about originality. A. protects herself somewhat from this criticism by posing her thesis as the claim that the Histories "must be reckoned as the earliest extant text" which displays the originality in question, not that T. himself is necessarily original. But, still, the text to which she refers is the text according to her own reading, and one is not convinced that earlier texts could not be read, by A. or by others, with comparable results.
Regardless of what she does or does not prove, Allison's work is enlightening about Thucydides and interesting in itself. In particular, her description of a fifth-century version of "relativism" or nominalism which is more nuanced and plausible than the accounts Plato hands down is a welcome contribution to the history of thought. Her emphasis on language as a social medium, and her assumptions about the asymmetry between the speaker and the audience of a verbal messages, are useful additions to the standard accounts of words and objects. Although Allison has not uttered the definitive word on Thucydides' contribution to language theory, she has made clear progress in showing how the question can be approached. This book is a good contribution to Thucydidean studies and to Greek prose and rhetoric studies, if not to language philosophy in itself.
1. June Allison, Power and Preparedness in Thucydides. Baltimore: Johns Hopkins 1989.
2. R. Carnap, Logical syntax of language London (1937) 1951 and Meaning and Necessity Chicago 1947; W. V. Quine, Word and Object Cambridge 1960; W. Sellars, "Abstract Entities," Review of Metaphysics 16 (1963) and Essays in Philosophy and History Dordrecht 1974 and Science and metaphysics: variations on Kantian themes London 1968.
3. F. Solmsen, "Thucydides' treatment of words and concepts," Hermes 99 (1971) 385-408; A. Cook, "Particular and General in Thucydides," Illinois Classical Studies 10 (1985) 23-51; T. Cole, The Origins of Rhetoric in Ancient Greece Baltimore 1991.
4. E. Havelock, Preface to Plato Cambridge 1963 and esp. The literate revolution in Greece and its cultural consequences Princeton 1982.
5. W. Porzig, Die Namen für Satzinhalte im Griechischen und im Indogermanischen, Untersuchungen zur Indogermanischen Sprachwissenschaft und Kulturwissenschaft 10, Berlin 1942; R. Freundlich, Verbalsubstantive als Namen für Satzinhalte in der Sprache des Thukudides Frankfurt 1987.
6. N. Loraux, "Thucydide a écrit la guerre du Péloponnèse," Metis 1 (1986) 139-161.
7. In general see the bibliography in S. Everson ed., Language, Companions to ancient thought vol. 3, Cambridge 1994.
8. A. Silverman, Plato's Cratylus: the naming of nature and the nature of naming," Oxford studies in ancient philosophy 10 (1992) 25-71.