Bryn Mawr Classical Review

Bryn Mawr Classical Review 2006.08.27

Evangelos Karakasis, Terence and the Language of Roman Comedy. Cambridge Classical Studies.   Cambridge:  Cambridge University Press, 2005.  Pp. xiv, 309.  ISBN 0-521-84298-0.  $80.00.  

Reviewed by Ariana Traill, University of Illinois at Urbana-Champaign (
Word count: 3005 words

Terence's language has probably received more attention than any other aspect of his dramatic art. There has been a steady stream of articles and monographs for over a hundred years, including two recent German dissertations and now Karakasis' revised Cambridge dissertation, directed by Richard Hunter.1 Karakasis (hence K.) sets out to investigate a basic contradiction: ancient commentators saw linguistic differentiation by character in Terence but many modern scholars see a uniform Latin. K. argues that both observations have some truth: there is differentiation, primarily by age and status, but at the same time Terence's plays are stylistically distinct from the mainstream tradition of the fabulae palliatae, togatae and atellanae. The only exception is the Eunuchus, where Terence catered to popular taste ("after the failure of Hecyra ... our author needed a success," p. 15). These two arguments are made in the first and second half of the book, respectively. The intended audience is primarily specialists and advanced students. K. assumes readers know the major characters of the plays and are familiar with a wide range of grammatical and linguistic terms (gnomic aorist, syntagm, lexeme, accusativus cum infinitivo abbreviated a.c.i., etc.). Many words and constructions will be unfamiliar to those who do not work on early Latin, especially in the second half, where K.'s thesis dictates a focus on non-classical forms.

The bulk of each chapter is a commentary-style list of lemmata organized by linguistic category (morphology before syntax, nouns before verbs, etc.), with brief discussions and parallel passages. K.'s findings are helpfully summarized at the end of each chapter. Chapter headings in the first half indicate the features by which characters are differentiated: colloquialisms, archaisms, senilis μακρολογία and περισσολογία ("long-windedness"), hellenisms, elevated language and miscellaneous features treated under "idiolect". One chapter is devoted to the special case of the Eunuchus. K. demonstrates that colloquialisms and hellenisms are found in speeches by and to low-status or rural types, with the exception of prostitutes (who use the former) and the maid Pythias (who uses the latter). Archaisms and long-windedness, i.e., pleonasms and accumulated synonyms, are largely confined to speeches by and to older characters. High status characters, particularly senes, use "elevated language," that is, words found more often in tragedy and epic than in comedy (e.g., satias instead of satietas, quadrupes instead of animal) or borrowed from legal, official or ritual language.

None of this is entirely new. As K. admits in the book's introduction, "the basic ideas at its heart are already to be found in ancient scholarship" (p. 1). For colloquialisms, K. relies heavily on the research of Cooper.2 Despite some sharp criticisms ("even though some of [Maltby's] conclusions will prove to be correct, his data are often false and his approach over-simplistic," p. 12) K. also cites Maltby throughout and much of the first half is devoted to extending his work.3 K.'s demonstration is more thorough, however, than his predecessors', his synthesis of previous work is quite useful, and he makes some new contributions. The chapter on idiolect presents new evidence for linguistic differentiation by individual and type (e.g., adsum in the sense of "pay attention" is used only by old men, p. 104). The chapter on hellenisms focuses on syntactic borrowings rather than loanwords, territory covered by Maltby. K. also makes a more detailed case than Maltby for the uniqueness of the Eunuchus, applying both linguistic and stylistic criteria: e.g., the image plenus rimarum. . .perfluo, which is "prolonged" over several lines after the style of Plautus (p. 132) or the use of di immortales without an introductory morpheme -- such as pro -- which is frequent in Plautus but found only at Eun. 232 in Terence (p. 139).

The second half of the book compares Terence with Plautus and early Roman drama. K. develops Wright's argument that Terence is anomalous within the tradition of early Roman drama, offering a broader and more systematic demonstration by examining linguistic features (Wright concentrates on stylistic elements) and including the fabulae togatae and atellanae. K. takes a leaf from Bertil Axelson and looks at what's not in Terence: morphology, syntax, lexical items and stylistic features -- such as figurative expressions, iuncturae or word order -- that can be found in the mainstream tradition (i.e., at least once in Plautus and once in another playwright) but not in Terence. K. is also careful to list the contradictory evidence, items aligning Terence, rather than Plautus, with this tradition, although these are not discussed in detail. Scholars will be grateful for the meticulous compilation of evidence. K. has scrutinized the extant body of palliatae, atellanae and togatae word-by-word. Although he is not the first to draw on this material (see, for example, Molsberger and Bagordo4), his is certainly "the first complete comparative analysis of the language of Roman comedy" (p. 145). The cumulative case for Terence's innovation is persuasive and K. strengthens Maltby's corrective to the argument that the language simply changed between Plautus and Terence. In this case one would expect an even distribution of non-classical formations, rather than a concentration in one play, and one would not expect Turpilius and Afranius, who post-date Terence, to use early Latin forms closer to Plautus'.

The book draws on a broad range of sources, both literary and epigraphic, including post-classical material. Electronic databases have made this kind of work easier but K.'s research goes far beyond word searches. He notices, for example, that Plautus and Caecilius use efficio with an ut clause (Pl. Mil. 936-7, Caec. com. 148-9) whereas Terence uses the verb absolutely, with an accusative direct object, or with qui plus the subjunctive (p. 173). The book is full of the kind of fascinating discoveries that arise from attentive reading. Plautus uses equidem with second and third person verbs quite often; Terence does it only once (Eun. 956, p. 141); Terence uses derivo figuratively (Ph. 323, iram derivem senis), whereas Plautus and Titinius use it literally (Tit. tog. 130, Pl. Truc. 563, p. 228); and the same distinction holds for ferveo (used literally at Tit. tog. 125 and Pl. Ps. 840, figuratively at Ter. Ad. 534, p. 228). K.'s note on personal nouns ending in -o, apropos of Naev. ex. inc. v. 20 ganeo lustro aleo, is typically informative: whereas Plautus revels in non-classical forms (bucco, calcitro, congerro, esurio etc.), Terence uses only those that become common in classical Latin (helluo, homuncio, nebulo, verbero, with the single exception of gerro, HT 1033). Abundant attestations from Afranius, Pomponius, Laberius, Naevius and Caecilius, along with Catullus 29.2 (vorax et aleo), show the colloquial character of these formations (p. 189). K. is an extremely competent Latinist with a broader knowledge of linguistics than many classicists. Despite an early recusatio (p. 29), he ventures profitably into pragmatics and discourse analysis (greeting formulas, oaths, affirmative/negative responses, pp. 116-18) and repeatedly argues for what sociolinguists would call accommodation (characters modifying their speech for an addressee, e.g., increased use of archaisms to older, higher status figures, pp. 80-1).

I have some questions about the selection of data. K. presumes continuity in the colloquial language, since he draws on sources spanning over 700 years to prove the colloquial character of items in Terence, although he admits that some words changed in meaning (e.g., scortor, p. 39). The sheer number of items in the table on pp. 41-43 (proposed by other scholars but excluded by K. as uncertain5) leaves a reader with less confidence that colloquialisms really are as scarce in Terence as K. argues. There is also some subjectivity in deciding what count as accumulated synonyms. No one will dispute cupio and desidero (HT 425), but as a "synonym" for periit and abiit, navem escendit at Ad. 703 is surely intended as a joke. On the one hand, one feels that magnifica and nobilis are not the only items contributing to the effect of μακρολογία at HT 227 (potens, procax, magnifica, sumptuosa, nobilis) and that docui, pravom and largiendo ought to be marked along with the other two items in each of the following examples: docui monui bene praecepi Ad. 963, pravom ineptum absurdum, Ad. 944, adsentando indulgendo et largiendo Ad. 988. On the other hand, it would not be hard to pin down differences between recte and commode or dolia and seriae (p. 68) and argue that they add something of significance. K. concedes that "synonymy depends on context and subjective judgement" and admits to casting the net wide because even near-synonyms "still impart a rambling or long-winded character to the speech of old people" (p. 80), which K. attributes to "the feebleness of old age" (p. 5). This stereotype is often true in a genre like Roman comedy, but it makes K. overlook other expressive effects of amplification and redundancy. K. himself argues that a young man's amoque et laudo et vehementer desidero (Hec. 488) is not "long-winded"; it merely expresses "strong feelings" (p. 67). To explain why the prostitute Bacchis, with her archaisms and accumulated synonyms (pp. 51, 72), exhibits "linguistic behaviour ... often similar to that of old people" (pp. 81-2), K. suggests that the good character of this bona meretrix is reflected in her "dignity of speech" (p. 119). One would, however, expect to see the same features in Thais' language. I suspect a slightly padded compliment from a prostitute to an ex-lover (tu ecastor morem antiquom atque ingenium obtines Hec. 860, p. 72) rang more of blanditia than dignitas to Roman ears.

The relative size of the corpus of Plautus versus that of Terence also needs to be taken into account more consistently. For example, K. states that capitulum at Caecilius com. v. 90, found twice in Plautus, "appears only once" in Terence (p. 175). Two occurrences in Plautus, however, constitute a lower frequency than one in Terence since the Plautine corpus is more than twice as large. In the case of infortunatus (Caec. com. v. 163), which also appears "only once" in Terence (p. 176), the two Plautine instances appear in successive lines, with one speaker echoing another's words. Small numbers sometimes pose a problem. Many items appear only once or twice in Plautus and once in the fragmentary poets, and so claims that an item is "common in Plautus and other comic dramatists" are a bit overstated (this claim, found on p. 187, is based on two instances of sermonem serit in Plautus and one in Caecilius). Likewise, more than one example would be helpful to show that verbs in -issare are "very common in comedy" (p. 33), and more than two examples of commoditas in the sense commoda to show a "distributional pattern" (especially since Simo, at And. 569, is playing off incommoditas in 567, so commoda is arguably not a "submeaning" here, p. 103). Some of the pairs of statistics cited on pp. 80-82 are close enough that a few examples more or less would significantly change the results (e.g., accumulated synynoms make up 1.87% of the speech of old people in the Hecyra vs. 1.80% of other characters' speech). And there is room to dispute individual items. For example, K. identifies gula as a word Terence avoids, citing HT 673, crucior bolum tantum mi ereptum tam desubito e faucibus as evidence that Terence "uses only fauces in the sense of 'gullet'" (p. 177), but this is an idiomatic use of fauces in the sense of "jaws" (v. OLD s.v. 1e). The seven syntactical hellenisms K. identifies in this chapter are by no means universally agreed. Eri semper lenitas verebar (And. 175-6, cited p. 86), for example, is a hellenising construction only if semper is taken to modify lenitas rather than verebar. It is a pity K. did not consult Bagordo, who argues extensively for Greek borrowings in Terence. Bagordo's examples might have modified K.'s conclusion, from the "surprisingly few entries" he found, that Terence absorbed the Scipios' disdain for Greek and limited his hellenisms to low-status characters.6 It is a little difficult to square this argument with the syntagms K. actually cites, since these mostly come from higher registers (gnomic aorists, est with the infinitive). I was a little surprised that S. Goldberg's Understanding Terence (Princeton, 1986) was not cited because Goldberg also discusses the influence of legal language and oratory and argues that Terence anticipates classical Latin in his syntactical innovations. A few other disputable items are discussed below.7

Items in the second half of the book were selected only if a) they appeared in Plautus and at least one other dramatist and b) Terence "had the opportunity to use the word in question, but instead systematically preferred a synonymous or nearly synonymous alternative" (p. 146). Since K. does not cite these synonymous alternatives in most cases, it is hard to assess the claim that Terence "consciously opted not to" use the item because he preferred what would become the classical form. In effect, this amounts to excluding specialized or obscure items: not using singularius, meritissimo or vulgarius is stylistic choice; not using vesti(s)pica or cumatilis is lack of opportunity ("thematically, Terence did not need a word with this meaning" p. 205, 223). It is not, however, clear that these are thematic, rather than stylistic, decisions. Vesti(s)pica and cumatilis occur in list jokes (Trin. 252, Epid. 233), and the first is a Romanism, both typically Plautine devices K. elsewhere admits are rarely found in Terence. In this case, I would have accepted a looser screen (these obscure words do seem to align Plautus with other palliata writers, against Terence). This would of course only strengthen K.'s already persuasive case.

K.'s arguments would also have been strengthened by greater use of statistics. He presents figures for long-windedness on pp. 80-2, but the following claims also beg for numbers: "such formations [sc. adjectives in -inus] abound in comedy" (p. 30), "the colloquial character of [frequentatives] is evident by their frequency in comedy, especially in Plautus, in contrast to Terence, who is here again more reserved" (p. 33), "old people and other characters use almost the same amount of synonyms [sc. in HT]" (p. 70). The evidence might also have been presented in a way that better supports K.'s conclusions. There is sometimes a mismatch between the emphasis in discussions of individual items and the overall conclusion. The chapter on archaisms, for example, concludes that "comparatively rare [early Latin] features characterise mainly the speech of old people" (p. 61) but the examples cited seem to support the thesis "Terence anticipates classical Latin". Organization in a commentary-style format by lemma is useful because it allows K. to discuss each item and list comparanda but it obscures the larger picture. The claim that the Eunuchus is Plautine in language is asserted many times before the chapter devoted to arguing this case. This evidence is scattered, however (it includes items discussed in in the chapters on colloquialisms and archaisms), so it is hard to judge how compelling it is. It would be useful to have a summary, for example, a table like the one K. uses for colloquialisms.

The same may be said for the argument that T. is most Plautine when he "deviates" from his model (e.g., in scenes with a fourth speaker, the added scene from Diphilus in the Adelphoe, or the extra pair of lovers in the Andria). To identify "deviating scenes" K. usually relies on the testimony of Donatus but he also makes the circular argument that "divergence from purity of diction and style can, along with other factors ... be an indication for Terence's altering his model" (p. 16). It is quite possible K. is right but again it is hard to judge this because the evidence is never assembled in one place. I felt "deviation" was sometimes used to explain away the awkward evidence of non-classical forms appearing in plays other than the Eunuchus and there is some special pleading, e.g., the non-classical coeperet at Ad. 397, which is not found in a "deviating" scene, "seems to be used for stylistic effect in the speech of an old person, Demeas" (p. 171). K.'s deviation argument also raises a lot of questions: why would Terence become more Plautine because he was borrowing from Diphilus for a scene instead of Menander? Why was he able to write the entire Eunuchus in a different style? And why did he resume the stylistic trajectory K. traces (towards classical forms) for the Phormio, the Adelphoe and both revisions of the Hecyra (which K. wants to place stylistically with the later plays, p. 81 n. 84) after this taste of popular success?

The book is written in a style appropriate to a technical subject. Comments on lemmata are clear and concise, although it is sometimes difficult to follow the surrounding text. For example, passages like the following would have benefited from revision for clarity: "No pleonastic combination occurring in the speech of other, non-old characters is absent from senile language. When there is restriction of specific pleonastic categories to the speech of a character type, this type is always the old people" (p. 77). This is a complicated way to say that X is a subset of Y. It is also a little confusing that "character type" means a demographic category such as old/young or high born/low born, rather than a stock type, and I would prefer that items in the many numbered lists K. uses (of aims, central arguments, defining features of a linguistic category, deficiencies of other scholarship) grammatically complete the "lead", as in multiple choice questions. Organization by author in the second half results in excessive repetition. Two of K.'s theses are repeated for every major author treated (in the introductions on pp. 150, 168-9, 188-9, 205, 223 and 235, and in many of the conclusions). An index verborum would be useful in book that is written to be consulted, rather than read cover-to-cover. It would be helpful, for example, to know there is a discussion of attat(ae) on pp. 130 and 194 or prodigere, on pp. 178 and p. 193. There are some typographical errors. I found an average of one every three and half pages.8


1.   The dissertations are: A. Bagordo, Beobachtungen zur Sprache des Terenz mit besonderer Berücksichtigung der umgangssprachlichen Elemente, Vandenhoeck & Ruprecht, Göttingen 2001; R. Müller, Sprechen und Sprache: Dialoglinguistiche Studien zu Terenz, Bibliothek der Klassichen Altertumswissenschaften, Neue Folge, 2 Reihe, Bd 99, Heidelberg 1997.
2.   F. T. Cooper, Word Formation in the Roman Sermo Plebeius. New York, 1895.
3.   R. Maltby, A Comparative Study of the Language of Plautus and Terence. Diss. Cambridge, 1976; "Linguistic Characterisation of Old Men in Terence," CP 74 (1979) 136-47, "The Distribution of Greek Loan-Words in Terence," CQ35 (1985) 110-23.
4.   H. J. Molsberger, Abstrakter Ausdruck im Altlatein: Form und dramatische Funktion abstrakt-begrifflichen Sprechens in der altlateinischen Bühnensprache , Athenaeums Monografien, Altertumswissenschaft, Beiträge zur klassichen Philologie, Bd. 193, Frankfurt am Main, 1989; Bagordo, cited n. 1.
5.   Some have already been challenged by Bagordo, cited n. 1. E.g., quid + part. gen. of person (pp. 67-9), adverb + esse (pp. 77-80), instrumental ablative with facio, fio (pp. 83-4), combinations of ille. . .ille (pp. 84-7). Bagordo also disputes one colloquialism which K. accepts: emungo (pp. 62-4).
6.   Bagordo, cited n. 1, pp. 97-146. Against K., Bagordo pp. 17-18, 28-30 also argues that Terence spoke the language of the educated circles of his day, circles that regularly used Greek expressions.
7.   Eun. 1009, vidi nec videbo, does not seem terribly "elevated" in context (p. 99): the first part of the line reads numquam pol hominem stultiorem. Pl. Men. 758 is not, strictly speaking, an example of mala aetas as a periphrasis for senectus because mala is predicative here, although Plautus is certainly punning on this sense (p. 163). Eam ... dormitum in Caec. com. vv. 94-5 (mihi suadet ut eam quisquam dormitum) is not functioning as "a simple periphrastic future" because it is embedded in an ut clause dependent on suadet (p. 172). A few of K.'s parallels between Plautus and the palliata tradition clearly go back to Greek models (e.g., morbus "used figuratively for amor, libido," p. 182 or the iunctura, si linguas habeam decem, found in Plautus and Caecilius, p. 184).
8.   Potentially confusing errors: p. 6, line five, read "Bacchis" for "Thais"; p. 131, lines 1 and 5, read 37.5 for 37:5, 247.3 for 247:3; p. 177, v. 269 vallata gulla (read gula for gulla): this fragment, which is not in Warmington, should be identified as Ribb. inc. XXXIII (v. 283); p. 208, v. 192 line 9, the "equals" sign should be an "is not equal to" sign. Small corrections: on p. xii, editions of Plautus (Lindsay) and Terence (Lindsay and Kauer) should be identified with the others; p. 55, clarify that And. 558-60, Phorm. 897-8 and 1045, and HT 584 are not being cited to show the particular use of the subjunctive under discussion here; p. 64 n. 9, Cic. Or. 184 would be better support for the claim that senarii approximate everyday language than Arist. Poet. 1449a21ff; p. 73 dolor at Ad. 602 cannot be "childbirth pain" because they are talking about the mater virginis (598); p. 119, four lines from the bottom, Chaerea is an ephebe, not a soldier (see line 824); p. 192, ex inc. v. 1, line 8, clarify that advance means "a sum of money" here; p. 200, potes in Juv. fr. 1 is probably from possum, not potare.

Read Latest
Index for 2006
Change Greek Display
Books Available for Review

HTML generated at 13:31:14, Friday, 03 April 2009