Bryn Mawr Classical Review 2000.04.05
Stephen Colvin, Dialect in Aristophanes and the Politics of Language in Ancient Greek Literature. Oxford Classical Monographs. Oxford: Clarendon Press, 1999. Pp. xii, 347. ISBN 0-19-815249-3. $75.00.
Reviewed by Joshua T. Katz, Princeton University (email@example.com)
Word count: 3127 words
It is not easy to imagine a subject simultaneously so grateful and so ungrateful as that of Stephen C(olvin)'s book. On the one hand, a study of what at least one linguistically sensitive Athenian thought about dialectal differences must count as a major contribution to the fast-growing shelf of works on Greek attitudes about ethnicity. On the other, the text of Aristophanes is famously messy, with scores of problems in the reading and interpretation of "normal" Greek, and many will doubt that it is possible to say anything conclusive about, say, the Theban merchant's use of ει rather than η in Acharnians, much less draw larger conclusions from this about the politics of language and dialect in the Classical period. But C. is aware of the difficulties of his enterprise, having already some years ago published an interesting article entitled "Aristophanes: Dialect and Textual Criticism" (Mnemosyne 48  34-47), and this suitably cautious monograph deserves attention from a range of scholars even though few will wish to read it in all its detail.
The book has six chapters: "Investigating Language Attitude" (1-38), "The Language of Foreigners in Greek Literature" (39-89), "The Text of Aristophanes" (90-118), "Non-Attic Dialect in Aristophanes" (119-263), "Old Comedy: Shorter Passages and Fragments" (264-95), and "Dialect in Old Comedy: Recapitulation" (296-308); in addition to the standard prefatory material, bibliography, and indexes, there are also an Appendix, "Sources for Laconian, Megarian, and Boeotian Dialect" (309-12), and a "Short Glossary of Linguistic Terms" (313-15). Perhaps the biggest surprise, given the title of the book, is that there is almost no explicit discussion of politics. One of C.'s major (if tentative) conclusions is that there is no evidence that the Athenian audience of Aristophanic comedy would have found non-Attic speech ungenteel, undesirable, or in itself humorous; that we might expect otherwise is the result of the "political and literary background of Western European thought on the uses of non-standard language" (305) and a prejudice of which to be wary. The word "political" shows up rarely, appearing, for example, in the heading of only a single section, an extremely short summary at the close of the final chapter (306-8). It would seem that the phrase "the politics of language" in C.'s title has as its principal function to help sell copies. Those readers who are deeply knowledgeable (as I am not) about the sociopolitical context of Athenian drama have, then, the task of testing C.'s linguistic analyses against their own view of social dynamics in the fifth and fourth centuries. But C. has persuaded me that the function of dialect speakers in Aristophanes, quite unlike in much English literature, "cannot be identified by their accent; neither can their moral character or any other personal characteristics" (300).
Some readers may wonder about the wisdom of "investigating language attitude" -- that is, of conducting sociolinguistic research -- on a stylized form of a dead language. C. addresses such concerns in Chapter 1, both from a theoretical point of view and then with specific reference to English literature (especially Dickens) and Old Comedy. In fact, recent years have seen considerable interest in Ancient Greek sociolinguistics (the work of Claude Brixhe stands out: see, e.g., CRAI 1997, 391-414), and C.'s task has to some extent been made easier by the appearance of Eleanor Dickey's Greek Forms of Address from Herodotus to Lucian (Oxford, 1996), an extraordinarily ambitious and (in my view: see Language 74  644-47) successful sociolinguistic study of Ancient Greek that opens with an extended discussion of method (note, however, that Dickey deliberately excludes poetry from her main corpus, though evidence from Aristophanes and Menander does play a role). It is no accident that C.'s and Dickey's books are both based on University of Oxford dissertations directed by Anna Morpurgo Davies, whose writings on both particular matters of Greek dialectology and the very notion of "dialect" (see, e.g., Verbum 10  7-28) have been very influential. At any rate, C. makes an excellent case for the enterprise ahead and ends the chapter with the useful idea of employing the currently popular study of the "representation of the speech of non-Greeks ... [as a] foil against which to judge the representation of the speech of non-Athenian Greeks" (36).
Chapter 2 gives a representative catalogue of passages in Greek literature before and contemporary with Aristophanes that represent or comment on speech that is somehow not the norm. It is a comparatively long chapter and an important one, whose function is to determine what sorts of models the comedian may have had. As C. observes, "It would be hard to make judgements about his portrayal of foreign Greeks -- and the differences between this and his portrayal of barbarians -- without knowing what the norms were in other types of literature. For we are not conducting an investigation simply into the literary style of Aristophanes; we are hoping to use information from comic drama to illuminate a particular area of Greek social thought, namely thought about language and language variation" (39). The sources C. considers are Homeric epic, lyric poetry, Herodotus, Thucydides, Xenophon, and the three major tragedians; in addition to some good observations on pidgin Greek in Timotheus, he has especially interesting things to say about allusions to foreign languages and Greek dialects in the historical works and tragedies, most notably those of Xenophon (who occasionally spices up his dialogues with dialect) and Aeschylus (whose characters make a few particularly noteworthy metalinguistic remarks, but who seems, like also Sophocles and Euripides, to be barred by convention from actually employing dialect). Perhaps because of limitations of space, C. often comments on his chosen passages somewhat telegraphically and without mentioning important recent scholarship (on the speech of Achilles' horse Xanthus in Il. 19.407 [45, with n. 10], see Sarah Iles Johnston, TAPA 122  85-98 and Hayden Pelliccia, Mind, Body, and Speech in Homer and Pindar [Göttingen, 1995] 62, 105-08, and 167f., with n. 104; for the tone of the vocative πάτερ in Timotheus [56, with n. 31], see Dickey, 78-81; etc.) The bottom line is that since there are few actual "odd" forms in the quoted speech of outsiders -- just an (unsurprising) awareness that dialectal differences exist -- we have evidence only for a "tradition of representing snatches of foreign dialect in Attic literature" (89). In other words, on the basis of the information that has come down to us, Aristophanes did not have a literary model for his own extended representations of dialect, Laconian in Lysistrata, Megarian and Boeotian in Acharnians.
How would an Athenian of the fifth and early fourth century B.C. have expressed dialect on the page, as it were? This is the question C. seeks to answer in Chapter 3, perhaps the part of the book that I enjoyed the most. It seems reasonable to suppose that when Acharnians and Lysistrata were first performed in Athens in 425 and 411 respectively, Aristophanes told the actors to utter non-Attic lines with marked speech; unfortunately, the details are irrecoverable. There remains, however, the interesting textual question of how Aristophanes would have indicated in writing the marked features he wished to convey. Anyone who reads (and, I assume, writes) novels in English in which dialect plays a role, even a small one, is used to stumbling over the representations of non-standard speech (which are typically improvised and, it would seem, almost necessarily crude). For an Athenian writer (and reader) of Aristophanes' time, though, there must have been the additional complication that two alphabets were in use, the Attic and the Ionic, and that these differ notably in how they record the e- and o-vowels: in the old Attic alphabet, there was no H or W, so E and O stood for both "long" and "short" sounds. Thus, small but crucial vocalic distinctions (whether a character would say something like the sound represented, in what we today call the "standard" alphabet, by E or H, EI or H, OU or W, etc.) are a linguistic matter that hits exactly up against the weak spot of an orthographic system in flux. Three things make matters worse for us today as we try to determine Aristophanes' original alphabet: first, the plays that have come down to us are generally textually corrupt; second, it is not obvious that the playwright would have been consistent in his rendering of dialect; and third, post-Aristophanic alterations are especially to be expected in passages with non-standard Greek -- and note that these changes could in principle go in both directions since some scholars and copyists might have tried to get rid of non-Attic forms while clever other ones could have worked (as have certain modern editors) to make the dialect of each non-standard passage consistent in its non-standardness. There is, as C. notes, no communis opinio about the alphabet Athenian authors employed in the second half of the fifth century: the Ionic alphabet was officially introduced in Athens in 403/2, but there is epigraphic and literary evidence that it was in use for some decades before (Euripides fr. 382 Nauck, which dates to before 422, is particularly fascinating: Theseus [Θησεύς, with an eta], an illiterate shepherd, vividly describes the letters in his name). As for Aristophanes in particular, some scholars have claimed that he employed the Ionic alphabet, others that he wrote in the old Attic alphabet and that his plays were then recopied sometime after 400. C. gives plausible arguments for the former view and explains in some detail how Aristophanes would have been able to use the alphabet to represent non-Attic speech. Just as important, he argues convincingly that, as a general rule, neither Hellenistic scholars (most notably Aristophanes of Byzantium) nor editors of the Byzantine renaissance sought to alter Aristophanes' dialectal forms. The chapter ends with a simplified stemma (following A. H. Sommerstein) of the manuscripts of Lysistrata and Acharnians and with some brief remarks on C.'s own views about what constitutes a reasonable dialectal emendation.
Chapter 4 is the heart of the book, a grammar of Laconian, Megarian, and Boeotian as they are found in the extended examples of dialect in the extant plays. After a clear introduction to the Aristophanic passages in question and to what we know about the three dialects from epigraphic and other sources, the chapter is divided into sections on phonology, morphology, syntax and usage, and lexicon and idiom (there is also a final comment on the Spartan and Athenian lyric passages at the end of Lysistrata). Despite the fact that the total number of lines that C. has to consider is by no means large -- he notes (298) that there are approximately 70 of Laconian (excluding lyric) in Lysistrata and 68 of Megarian and 31 of Boeotian in Acharnians -- the chapter is nearly 150 pages long. This is philological rigor at its finest, but even a linguist will find it tough going, especially the 50 pages on phonology. The many diacritical marks that are crucial to linguistic analysis, coupled with the (current) limitations of BMCR's electronic format, would make it foolhardy for a reviewer to comment in any detail on C.'s specific conclusions. Suffice it to say that those who want to know about any linguistic feature of non-Attic dialect in Aristophanes will find thoughtful discussion. Take, for example, the oath ἴττω Δεύς 'May Zeus be my witness!,' found at Ach. 911 (some manuscripts have Ζεύς): C discusses the possibility that ἴττω would have been pronounced with an initial digamma (possible; 178), considers the evidence for Boeotian TT vis-à-vis Attic ST (unclear; 168f.), shows that the initial delta of Δεύς is justified (165), and comments (with reference to ἴττω Ἡρακλῆς in Ach. 860, as well as the Theban Cebes' phrase ἴττω Ζεύς in Plato, Phaedo 62a) on the Boeotian nature of the entire asseveration (Attic has ἴστω and Ζεύς, but does not idiomatically combine the two -- note that oaths are a convenient stereotypical marker of dialect and may not actually be typical, witness, e.g., Mamma mia! in English; 23, 142, 257, 298, and passim). My principal criticism of the chapter has to do not with content, but form: the cramped layout is unattractive and remarkably unconducive to allowing the reader to see easily which forms C. considers to be dialectally appropriate and which are examples rather of "doubtful" or "bad dialect."
If C. has one overall conclusion, it is that Aristophanes represents Laconian and Megarian (both Doric dialects) more accurately than Boeotian (Aeolic, but with affinities to Northwest Greek), an observation in keeping with what is otherwise known about Athenians' reactions to Aeolic speech (note, e.g., the comment about 'barbarous Lesbian' in Plato, Protagoras 341c; see 12, with reference to J. Werner). I agree with C. that the distinctly unusual vowels of actual Boeotian provide one good reason why Aristophanes might be least successful at imitating this dialect, but if I may be allowed a tentative and rather different suggestion, note that in Acharnians, the Boeotian-speaking merchant from Thebes comes on stage (line 860) almost immediately after the departure of the Megarian (line 835), with only a choral song in between. With this in mind, consider the merchant's first list of Boeotian wares: ὀρίγανον, γλαχώ, ψιάθους, θρυαλλίδας, ́ νάσσας, κολοιούς, ἀτταγᾶς, φαλαρίδας, ́ τροχίλους, κολύμβους 'marjoram, pennyroyal, rush-mats, wicks, ducks, jackdaws, francolins, coots, wrens, dabchicks' (874-76). The form νάσσας 'ducks' (= Attic acc. pl. νήττας), with SS, is clearly an example of "bad dialect" (by contrast with ἀτταγᾶς, with Boeotian TT), and C. (who takes the usual line that the Greek word for 'duck' is cognate with Latin anas, -atis, which cannot be correct) has a few words to say about this (167). What is interesting, however, and what C does not point out, is that νάσσας would be appropriate for Megarian, and the same holds true, as far as I can tell, also for all the other forms (including ὀρίγανον, whose γ makes this "doubtful" Boeotian; see 163f.), and most strikingly for the four o-stem accusative plurals in -ους: ψιάθους 'rush-mats,' κολοιούς 'jackdaws,' τροχίλους 'wrens,' and κολύμβους 'dabchicks' (the last is elsewhere an i-stem, however; see 243). This is a famous quartet of forms: all the manuscripts have -ους (as does Athenaeus 9.395e-f, who quotes Ach. 875f.), but Bentley conjectured the more properly Boeotian -ως, an emendation that most editors (e.g., Sommerstein) accept. In view of the problem of Aristophanes' original alphabet (how would he have written the Boeotian equivalent of the Attic "spurious diphthong" ου?; see 100-3, esp. 102), it is hard to be sure about the matter of ου vs. ω, but it certainly is curious that the second list of wares, just a few lines later, consists of χᾶνας, λαγώς, ἀλώπεκας, ́ σκάλοπας, ἐχίνως, αἰελούρως, πικτίδας, ́ ἰκτίδας, ἐνύδριας, ἐγχέλιας Κωπαί̈δας 'geese, hares, foxes, moles, hedgehogs, cats, badgers (?), martens, otters, Copaic eels' (878-80), where all but one of the manuscripts have the o-stem accusative plural forms ἐχίνως 'hedgehogs' and αἰελούρως 'cats,' with -ως. (I cannot consider here the rest of the words in this passage; see esp. 186f., with comments on Elmsley's and Blaydes's generally accepted emendations to the words for 'otters' and 'eels.') On the problem of -ους vs. -ως, C. writes, "Although the tradition is not coherent, modern editors are justified in printing one ending (usually -ως) throughout the accusative plural, since it is scarcely credible that Aristophanes would have been inconsistent in representing this morphological category in the space of a few lines" (155). But this begs the question of how the inconsistency arose in the first place, an inconsistency that, I should add, affects how we view Aristophanes' ability to represent the Boeotian dialect since, as C. notes, we find in Acharnians ω for spurious ου only "in 45% of cases" (101), a figure based largely on the accusative plural forms in the pair of lists just cited. A few years ago, Helmut Rix, in a paper that C. does not cite, asked the following question about pseudo-Boeotian νάσσας: "Kann Aristophanes dem Boioter eine megarische Sprachform in den Mund gelegt haben (kurz zuvor tritt ein Megarer auf: Acharn. 729-835)?"; and Rix continued, "Auch ψιάθους, κολοιούς, τροχίλους, κολύμβους 874-876 (von Bentley nach ψιάθως eines Scholions in -ως korrigiert) könnten neben attisch auch megarisch sein" (Historische Sprachforschung 104  186 n. 1). I would suggest instead that the accusative plurals in -ους (all but one for kinds of animals) did not originate with Aristophanes (for this could not explain the two forms in -ως), but are rather a later (but probably still early) emendation on the part of a copyist who remembered that just a few lines earlier the conversation had been all about 'piggies' and that the Megarian who had dressed his daughters up as χοῖροι ('piglets,' but with an obvious obscene secondary sense) had spoken of these animals with the properly Megarian accusative plural in -ους: χοίρους ἐγώνγα μυστικάς 'I've brought porkers for the Mysteries' (764, trans. Sommerstein).
Chapter 5 is an extension of Chapter 4, examining Old Comedy (both Aristophanic and not) for isolated instances of and references to non-Attic dialect, non-standard Attic speech (e.g., Alcibiades' lisp, alluded to in Wasps 44f., and the actor Hegelochus' famous 'weasel'-error, which Aristophanes makes fun of in Frogs 303f.), and the language of barbarians (e.g., the pseudo-Persian [?] and pidgin Greek of the Persian ambassador Pseudartabas in Ach. 100-04). C. is clearly right that Old Comedy has a "lively interest in representing different forms of language variety" (295) -- this is, of course, quite different from the genres examined in Chapter 2 -- and he arrives at a fascinating conclusion: with the exception of slaves, "there is no example in extant Old Comedy of a non-Athenian Greek or barbarian whose speech is not marked as foreign in some way" (ibid.).
The final section, Chapter 6, provides a convenient summary of C.'s results, three of which are as follows. First, Aristophanes was familiar with the Laconian, Megarian, and Boeotian dialects, which he represents very carefully, pretty carefully, and somewhat less carefully (but still rather well), respectively. Second, quite unlike the "remarkably undifferentiated" speech of barbarians, "[t]here is no implication that the other [i.e., non-Attic] Greeks speak 'bad Greek'" (299f.). And third, dialect in Old Comedy is not added simply for humor, but is rather part of the realism associated with the comic stage. The notion that dialect is not in itself amusing will surely be controversial: it is, after all, no easy thing to separate the comic -- often negative -- attributes with which a dramatist endows his characters from the main mechanism by which the audience gets to know these same personages, namely their speech. But C. has made a persuasive case, and his important book will be of much interest and use to linguists and literary critics alike.