Bryn Mawr Classical Review

[Disclaimer: the reviewer knew the author slightly in graduate school.]

[Table of Contents is listed at the end of the review.]

This is a specialized monograph, primarily of interest to scholars of Plautine metrics but valuable to anyone interested in archaic Latin. The central argument is that apparent metrical anomalies, including departures from Greek practice, are not simply poetic license but reflect the actual speech of Plautus' day; that is, "prosody" in a linguistic sense ("the phonological organization of natural speech," p. 4 n. 11) can explain such phenomena as violations of metrical laws, when e is aphaeretized in est, or how and when words underwent iambic shortening.

The book, a revised Ph.D. dissertation (Harvard, 1996), is organized by topic. After preliminary remarks and an overview of the basics of Plautine scansion, there are chapters on violations of Luchs', Meyer's and Jacobsohn's laws, enjambement, aphaeresis, iambic shortening, and a conclusion applying the book's findings toward a reconstruction of Plautine speech. The findings of each chapter are very helpfully summarized in a separate section at the end. The study is, not unreasonably, limited to the iambo-trochaic sections in the Varronian canon (there is the odd example of a rare phenomenon in a canticum, but no fragments or poets other than Plautus) and relies on earlier scholarship (e.g., Ceccarelli, Drexler¹) for some of the data collection. The focus is on the synchronic picture. "Diachronic" in the title alludes to occasional references to later authors such as Vergil or earlier/later developments in the Latin language (from Proto-Indo-european to the romance languages), but this is not a diachronic study like Soubiran (1988), for example, who traces iambo-trochaics from Euripides to the Pervigilium Veneris²). Fortson is particularly interested in phonology-syntax interface: the effects of syntactic processes on phonology, particularly when these leave traces in the metrics (e.g., clitic groups filling slots normally reserved for single words).

Written by an indo-europeanist with impeccable credentials in both linguistics and classics, this is a genuinely interdisciplinary study. It draws on textual criticism and classical philology as comfortably as on X-bar theory and pragmatics, approaching Plautus with new questions (e.g., what linguistic information may be embedded in loci Jacobsohniani) and shedding new light on old ones (e.g., using linguistic arguments to defend manuscript readings that violate metrical laws). This very strength may be a stumbling block for readers. The book assumes linguistic background not universal among classicists (sandhi rules, wh-movement, syntax trees, concepts from pragmatics such as topic and focus, and specialized vocabulary, such as rheme, adjunct, and matrix verb). It is influenced by, and frequently cites, Devine and Stephens (1994)³. It also requires more metrical knowledge than is provided in the overview chapter. For example, there is no mention of prosodic hiatus, apocope of final -s, or synizesis, but readers need to know these to scan Men. 550 (p. 38), Bac. 220 (ibid.), and Mil. 1104 (p. 39), respectively. They need to know that the e in Philippeo (Bac. 200) must be short and the second syllable light (only the latter is mentioned, and not until p. 208 n. 80) and that a resolved "C" violates Meyer's law in the same way as a single heavy "C" (this is not spelled out in the introduction but appears in the very first example cited, Aul. 55, p. 55). Boldface is used to single out the phenomenon in question in each cited line, but lines are not usually scanned and the meter is not identified. A few "ia6's" and "tr7's" would have been helpful for non-specialists, who may be understandably confused. (To take one example: Meyer's and Luchs's Laws are introduced as applying to senarii and septenarii (p. 34), but the first example cited under "Violations" (p. 37) is an octonarius.) None of this will pose a problem to specialists, but it does limit the audience. The author concedes at the start that "the study of ancient metrics is an inordinately complicated affair" (p. 3), and it seems only fair to note the topic is not simplified by the application of linguistics. The Latin is translated where the argument requires understanding the meaning of a citation, but in general it is assumed the reader will be able to recognize parts of speech and phrase structure without assistance, while the frequent arguments about prosodic domain presume some familiarity with vocabulary (e.g., knowing whether a word is a function word or a lexical one).

It is impossible to do justice to the kind of argument this book makes in summary form. The phenomena under scrutiny are shown, roughly speaking, to reflect properties of Latin prosodic organization. Both expertise and long, patient care went into constructing the book's carefully considered, nuanced and detailed arguments. What follows is a brief summary of the conclusions on enjambement, aphaeresis, and iambic shortening, followed by a more detailed discussion of the analysis of violations of Meyer's law. Concerning enjambement, Fortson argues that there are indeed restrictions which reflect Latin prosody: prepositions and their objects are never split by line breaks, non/nec are never line-final except when they occur at the end of a phrase, and it is probable that conjunctions, especially monosyllabic ones, are "proclitic or prosodically subordinate" (pp. 132-3). Aphaeresis of est is constrained by syntax, but lack of aphaeresis is not; it is allowed "at the right edge of a noun or adjective phrase, between a noun or adjective and a following adjunct or appositive", (p. 174), after at least one adverb (intus), and after the pronoun ipsus and the quantifiers nullus and totus (pp.157-60). Iambic shortening (brevis brevians, abbr. BB) is "an optional phonological rule that affected only words or sequences of syllables in acoustic dips"; it is not caused by underlying word accent and does not require close syntactic connection with what follows (but its rarity before caesurae or diaereses "suggests that any prosodic break following ... would have had to be quite weak for BB to be licensed" (p. 256). Its only preconditions are "the lack of a strong accent, potentially also a lower pitch ... together with a tight connection between the two syllables making up the iambic sequence" (ibid.).

In alphabetic notation (with the letters A to D representing heavy (upper case) and light (lower case) syllables as, for example, "ABcDABcDABcD"), Meyer's law rules out unaccented word-ends of the form ..CD/ or ...ccD/, i.e., it requires a single light syllable in the "c" position. Fortson examines violations of this "law." He agrees with Soubiran and Ceccarelli that it is not strictly enough observed to be called a "hard-and-fast rule," pp. 74-5⁴ and concludes that it describes at least two separate phenomena. The monosyllables that follow violations in the first metron are "phonologically left-leaning so as to form a close prosodic unit (C D A) before the caesura" (p. 74), e.g., the nunc in apscede etiam nunc (scanned ABccDA), Aul. 55; those following violations in the second metron are right-leaning, so as to make the final unit in the line self-contained (e.g., the si in ...dicat. Si dixerit (scanned CDABcD), As. 800. Scholars' tendency to use the word "monosyllable" in describing allowable exceptions to the law has obscured the distinction between left- and right-leaning clitics, and thus differences between first and second measure violations. To account for this observed difference, Fortson suggests that the left-leaning clitics produced a phonological reduction of the preceding word: the long syllable in the "C" position "would have been shorter and less prominent than if the word were alone or standing at the end of its phrase" (p. 61). He adduces a cross-linguistic argument, drawing on the concept of overlengthening (which makes long vowels longer and more prominent in stressed than in unstressed syllables): the left-leaning words either formed a clitic group with reaccentuation (e.g., ULtro fit becoming ulTROfit) or simply made the stressed syllable "shorter relative to the length it would have had if it had stood in isolation" (p. 62, on the grounds that this kind of shortening -- called "refooting" -- can happen inside groups larger than the word or clitic group).

The discussion of Meyer's law illustrates both the strengths of the book and some of the challenges presented by the data. Because Fortson focuses on explaining the violations, he notices a pattern of differences between first and second measure violations evident in Questa's examples (most of the violations followed by a monosyllable are in the first measure, and of those followed by quadrisyllables, in the second measure). Questa does not mention this pattern⁵, nor does Ceccarelli, who notes the low frequency of word breaks at the first D (and relatively high frequency of violations of Meyer's law) but explicates the imbalance only in terms of aesthetics (Plautus sought rhythmic variety).⁶ Fortson is also the first to note and attempt to explain why monosyllables tend to lean leftward after CD/, but rightward after cD/. Because he looks for a linguistic basis for metrical phenomena, he seeks explanations in the unconscious behavior of native speakers, rather than the conscious application of a complicated set of rules (drawing, for example, on concepts such as prosodic domains and phrase structure rather than coining terms like "metrical word" (for propter me, erga me, Questa p. 401) or "quadrisyllabic group" (for labes popli, idem, p. 387)). However difficult archaic Latin phonology may be to recover, this kind of explanation has an inherent plausibility. Mastering highly complex rules is part of normal language acquisition but far more difficult, if not impossible, for a system that is learned later, such as a second language or iambo-trochaic metrics (particularly, one might add, since the latter was still being reinvented for Latin).

Fortson takes a conservative approach to the text, trying to explain rather than emend violations of Meyer's law, partly out of principle ("the readings of P are part of the historical record just as surely as A is and must be explained," p. 89) and partly out of necessity (if we conceded, with Laidlaw, that "there is no theory under the sun for which an array of support could not be collected from the scribal errors in the P-archetype", then "Plautine philology would virtually grind to a halt", p. 89). Many will sympathize. There will, however, be disagreement about the textually problematic cases. For example, one potential first-measure violation of Meyer's law, quid vos maestos tam tristisque esse conspicor (scanned ABCDABCDaBcD, Bac. 669), requires what Fortson concedes is a "very unusual positioning of tam following the adjective it modifies", not found in the standard dictionaries or Lexicon Plautinum (p. 56). Editors have emended the passage and even disputed whether it is iambic. Fortson ultimately dismisses the textual problem ("whether the passage is correct as transmitted is not of great moment for any of my arguments; either way tam would be prosodically most closely grouped with maestos" p. 57). This is true, but if the order is reversed there is no breach of Meyer's law and even if the text is sound it is hard to see how this passage preserves evidence of the phonology of ordinary speech. The one example I can find of a postponed tam is at Ecl. 5.83, nec percussa iuuant fluctu tam litora. Its very occurrence in Virgil and not in a colloquial context would argue against its being an "allowable rhythm of natural speech" (p. 4) and cast some doubt on the idea that the metrics here really is providing a "window on the prosody of [Plautus'] Latin" (p. 5).

Fortson is disarmingly frank about the problems of the text and recognizes that not all manuscript readings can or should be defended (e.g., he rejects on textual grounds two examples of iambic shortening before -que, p. 252 n. 58). He faces textual problems squarely and objectively: anything that could affect the argument is discussed in a note. These problems do mean that different editors will identify different instances of a phenomenon (the breach of Meyer's law at Bac. 669 discussed above, for example, or possible aphaeresis after a long vowel plus -s, e.g., haecst, rest (< res est), which "almost all scholars have emended...out of existence" p. 134 n. 4). Similar problems arise when the scansion of a line is disputed. As Fortson admits, "a number of these lines can be read in more than one way, each way a violation of a different metrical 'law'" (p. 81). These cases are treated carefully, however, and the book's arguments are made on the basis of less problematic examples. Samples are relatively small, in some cases because the phenomenon is rare and in others because the author excludes a subgroup from a total number of examples (e.g., grammaticalized expressions such as opus est are excluded from the pool of -us nouns followed by aphaeretized es/est on the grounds that aphaeresis in the former case cannot have been linguistically motivated) or simply decided to reduce the sample to a manageable size (e.g., taking the first 76 lines of Menaechmi to test for skews in the distribution of stressed heavy syllables (pp. 64-5), essentially a question about the coincidence of ictus and accent, although Fortson is agnostic about the existence of a verse ictus in Plautus and therefore does not discuss it).

Fortson does not make claims about anything with fewer than a dozen data points, so samples are large enough to generate skews that are probably not due to chance. For example, in the 76-line sample from Menaechmi, he finds a ratio of 28 stressed to 8 unstressed heavy syllables in second-metron B, and 5:35 in second-metron C. It is not unreasonable to conclude from this that phonetic differences between stressed and unstressed heavy syllables are relevant to the metrics ("it is possible that this reflects a desire to avoid placing stressed heavy syllables in the second C," p. 65) but it is surprising that there are no quantitative statements about the likelihood that this is due to chance. This would not have been difficult to calculate (Fortson himself notes that "further statistical tests would be needed for this to be firmly established" p. 65) and is becoming the norm in this type of research (for an example in Plautine studies, see De Melo 2007, q.v. p. 15 n. 24, for references on statistics and language research generally⁷). There are some problems in the choice of samples. For example, Fortson collects examples of second metron loci Jacobsohniani (second or third metron "D" position syllables filled by a single light syllable, i.e., d, not D or dd) from the 1132 iambo-trochaic lines of the Miles (89 times or 7.9%) but uses the entire Plautine corpus for examples in the first metron (100 times or 0.6%, these examples having been collected by earlier scholars). The choice of a single play as a sample assumes no change over time, but this seems unlikely, given that at least one metrical phenomenon -- the ratio of cantica to iambo-trochaics -- is known to have changed over Plautus' career and a poet's use of iambic trimeters, in particular, can change (witness the case of Euripides). Since the Miles is probably an early play, it is less likely to be typical, and so choosing it as a sample may have introduced a bias that could have been eliminated (for example, by choosing 1000 random lines out of the total number of iambo-trochaic lines).

Fortson's strengths are as a linguist and philologist. He cites cross-linguistic parallels from an array of languages that rarely appear in books on Plautus: Latvian, Serbo-Croatian, Walayalam, Ponapean, Nez Perce, Menomini, Goroa, Bhojpuri. Some will be more comfortable with arguments based on romance language developments (e.g., the French impersonal clitic on as evidence for reduction of the word homo in phrases such as illic homo, p. 41, or French/Italian ma, ta, as evidence that synizesis in meus, tuos, etc. reflects vowel coalescence, rather than iambic shortening) than the assumption that tone-sandhi in Fuzhou can tell us something about intonational phrasing in Latin (p. 108). But Fortson points out that there are solid linguistic grounds for stressing the commonalities in human language ("One of the greatest contributions that the linguistic sciences have made, especially in more recent times, is the demonstration that superficially very different languages are in fact astoundingly similar at practically every level of grammar" p. 6). He cites the example of function words, which typically become part of the prosodic domain of nearby stressed words, to explain some apparent exceptions to the law of the split anapest: "ut o-", in ut opinione is allowed, but not "-cit e-", in dicit elephans. Given the enormous difficulties of recovering the phonology of archaic Latin (the epigraphic evidence, which Fortson cites frequently, often brings more problems than answers), cross-linguistic evidence may be the best place to turn. Fortson is careful about pushing this evidence too far. For example, for first-measure Meyer's law violations followed by a left-leaning clitic, he concludes only that "some process like refooting operated on these long C's that explains the word distribution patterns" (p. 63, citing Devine and Stephens⁸). Here, as elsewhere, the book shows caution and very careful thought about its underlying assumptions.

The style is appropriate to the subject. There are some minor inconsistencies arising from attempts to strike a balance between precision and accessibility (e.g., arguing against, but still using, traditional language such as law, violation, breach, using "prosodic" in a metrical, not linguistic, sense on p. 55, or referring to a "lengthened" syllable on p. 76, alongside "heavy" vs. "light" syllables). There is a lot of cross-referencing and deferring of arguments that can make the book hard to follow (I had trouble identifying the Gustafson's "different analysis" of line-end monosyllables preceded by syntactic break, mentioned at p. 91, or locating where it was argued that the univerbation of clitic groups is "at best unproved and at worst unlikely" p. 71). A subject index would have been extremely useful (there is an index locorum). The longer lists of examples are useful for reference but do not make lively reading (e.g., the 43 examples of loci Jacobsohniani after short C's, organized alphabetically by play, pp. 79-81). In this, the book is more in the tradition of metrical than linguistic studies.⁹

Meticulous care went into every stage of this project, from the research to the final proofreading. Errors are few and minor, and the author's breadth of knowledge is impressive. One small example: a beautifully learned note argues that the rare examples of feminine nominative singulars in long -a in Plautus are true archaisms, comparing these with the equally rare -osio genitives preserved in a few archaic Latin texts (in both cases, the preponderance are personal names) and drawing an illuminating parallel with Lycian, which also preserves Proto-Indo-European -oso in names (p. 85 n. 109). Students or scholars interested in an early Latin project should read Fortson's encouraging note (p. 109 n. 30) about the work that needs to be done in sacral language in Plautus.

This is a welcome contribution to a field that has, with the notable exception of C. Questa (2007), seen relatively little recent work, even though metrical issues are crucial to anyone who works with the text of Plautus (currently being re-edited in the Editio Plautina Sarsinatis series). It is the nature of this type of study to have an impact primarily on a small circle, but its indirect effects are potentially much wider. It offers a new line of argument to those who want to see Plautus as an original Italian artist who drew on native traditions, rather than a derivative importer of Greek culture. It explains differences from Greek metrics in terms of Latin phonology, and even suggests that an oral tradition may lurk behind phenomena such as the metrical regularity of final feet ("While Plautus is no longer operating in an oral-formulaic tradition, some of the poetic forms he implemented are descended from ones used in such a tradition and carry over many of the same properties from it" p. 96). The book's most important contribution, however, is to re-examine metrical phenomena through the lens of linguistics. It shows how evidence of archaic Latin phonology may be recovered from observed metrical behavior in Plautus, and offers, in particular, new insights into what constitutes a phonological "word" and arguments for the prosodic reduction of verbs. It shows how evidence preserved in the manuscripts can be obscured by editorial decisions based on "laws", some of which are no more than "tendencies" (and comparative studies, such as Soubiran, show significant differences between Plautus and other authors of iambo-trochaic verse). In its integration of philology and modern linguistics, the book contributes to an important broadening of the boundaries of the field.

Table of Contents

1.   Ceccarelli, L., 1988. La norma di Meyer nei versi giambici e trocaici di Plauto e Terenzio. Rome: Ediun Coopergion. Drexler, H., 1969. Die Iambenkürzung. Hildesheim: Olms.
2.   Soubiran, J. 1988. Essai sur la versification dramatique des Romains: sénaire iambique et septénaire trochaïque. Paris: Centre national de la recherche scientifique.
3.   Devine, A. M. and Stephens, L. D. 1994, The Prosody of Greek Speech. New York and Oxford: Oxford University Press.
4.   Soubiran (1988) p. 340. Ceccarelli (1988) p. 17.
5.   Questa, C. 2007, La Metrica di Plauto e di Terenzio. Urbino: QuattroVenti. pp. 386-9.
6.   From Ceccarelli's (1988) figures (p. 21), word breaks at the first D position occur in 4.3% of Plautus' senarii, with an 18% rate of violation of Meyer's law; at the second D, in 18.2% of the lines, but with only a 12.3% violation rate.
7.   De Melo, W. C. de 2007, The Early Latin Verb System: Archaic Forms in Plautus, Terence, and Beyond. Oxford: Oxford University Press.
8.   Devine and Stephens (1994), pp. 273-4.
9.   Cf. De Melo (2007), who limits the examples cited in the actual text and provides brief context and translation for each, or Devine and Stephens (1994) and (2006), who also provide translations.