Bryn Mawr Classical Review 98.6.24


Kenneth Dover, The Evolution of Greek Prose Style. Oxford: Clarendon Press, 1997. Pp. xxii, 198. $65.00. ISBN 0-19-814028-2.


Reviewed by David Sansone, Classics, University of Illinois, dsansone@uiuc.edu.

A new book by Kenneth Dover is always welcome. His style is lucid and lively, and his prose is spiced with the occasional autobiographical detail that makes his work especially endearing. The subject of Greek prose style has actively engaged Dover for nearly half a century, his preface to the second edition of Denniston's The Greek Particles being dated "September 1950." During that half-century he has made valuable contributions to the study of several prose authors (and, of course, poets as well), including Thucydides, Lysias and Plato. The Evolution of Greek Prose Style thus bears the fruit of long acquaintance and intense involvement with its subject-matter.

Inevitably a book on style will appear in the form of capita selecta, as it is impossible to catalogue exhaustively the elements that characterize the style of even a single author. Nor is it necessary to do so. Often the really distinctive features of an author's style amount to only a handful. And so we have discussions of vocabulary, rhythm, sentence-structure, repetition, and so on, but not word order or asyndeton, which were treated in Denniston's Greek Prose Style (1952). Word order, on which Dover published an important book in 1960, is excluded (p. vi) on the grounds that, while further study "would contribute something to our understanding of the Greek language as a whole it would not necessarily be equally helpful to the study of style, where difference between texts is crucial" (Dover's emphasis in each instance). While the insistence on difference between texts as crucial for determining style is valid (see also p. 1, "choice between alternatives"), elimination of word order from consideration would make sense only if Greek word order were rigidly fixed. That it is not fixed, and that different texts display comparable constituents in different sequences, is confirmed (if confirmation were necessary) by some of the examples given by Dover himself on pp. 14-16 and 157.

While one would perhaps have preferred to see included a discussion of word order (and some other features of style that are passed over), there can be no doubt that the topics that are treated are given very satisfactory coverage. Especially valuable, for example, are the characteristically clear-headed comments on the different levels of style (pp. 2-4), the questioning of the usefulness of the category "sentence" and the introduction of the category "main-clause-finite-verb unit," or MCF (pp. 27 ff.), eminently sensible general principles regarding the use of statistics (pp. 43-44) and an absolutely revelatory demonstration of the degree to which genre is as important a determinant of style as the individuality of the author (pp. 46-48; cf. also 154).

This last comprises material that originally appeared at ICS 19 (1994) 83-87. Dover analyzes three texts, Thuc. 3.82-83 ("Text I") and two other texts not immediately identified, by quantifying the appearance of a select group of stylistic features in each of the texts. The passage from Thucydides, the account of the effects of stasis on political morality, is shown to bear a striking stylistic affinity with "Text III," while differing markedly from "Text II." It turns out, however, that "Text II" is the immediately following military narrative (3.85-91) from Thucydides, while "Text III" is a "generalizing description of moral and political behaviour," related in character and style to "Text I," from Isocrates. This impressive demonstration of the influence of genre on style finds confirmation of sorts in a recent study of Latin prose authors, B. Frischer et al., "'Sentence' Length and Word-Type at 'Sentence' Beginning and End," in Research in Humanities Computing 5 (Oxford 1996) 110-142, where it is proposed that generic expectations and rhetorical theory are aspects of what is delightfully characterized as the ancients' "anxiety of originality" (p. 135). But, as always with such matters, one must be especially careful not to claim too much. In referring back to his own analysis, Dover speaks (p. 53) of the "demonstration ... that Thucydidean generalization [i.e. "Text I"] and Thucydidean military narrative [i.e. "Text II"] are, statistically speaking, different populations in respect of style." Rather, what Dover has demonstrated is that "Text I" and "Text II" belong to different populations in respect of style, depending upon which stylistic features have been selected for analysis. The possibility remains that, if we were to choose different stylistic features, it might be shown that, in respect of these other criteria, "Text I" and "Text II" resemble one another more than they resemble "Text III."

The title of Dover's book and his preface ("My primary concern is the history of linguistic differences between texts," p. v) proclaim the author's emphasis on change and development. But that development is traced only as far as the time of Plato and Demosthenes, when, according to Dover, Greek prose "attained its proper nature" (p. vi). Our acquiescence in that judgment appears to be taken for granted. It is a judgment, however, that ought to have been examined, for it rests as much upon the subsequent history of Greek prose as upon the development that produced the major figures of fourth-century Greek (mostly Attic) prose. That development is examined in detail and to great effect in chapters 4 ("Speeches, Stories, and Talk") and 5 ("Time and Place"). Chapter 4 investigates the models, both formal and informal, available to "the first Greek ... who committed a prose composition to papyrus" (p. 57). Chapter 5 seeks to show that early Greek prose is a largely Ionian phenomenon, and that a number of the features of even late fifth-century Attic prose can be traced ultimately to the influence of the Ionic vernacular, rather than, as has sometimes been argued, to poetic influence or an archaizing tendency. In this Dover is likely to be correct, but it should be noted that the issue is not so clear-cut as may appear. Dover does not cite the remarkable opening chapter, "The Growth of the Attic Dialect," of W. G. Rutherford's The New Phrynichus (London 1881), which long ago showed, with copious examples, that "the basis of the language of Tragedy is the Attic of the time when Tragedy sprang into life" (p. 3), a time when the Attic and Ionic dialects were much less sharply differentiated from one another. Thus, a feature like the word KA/RTA is Ionic and poetic and archaic, so that it is difficult indeed to assign it to one category or another. For this word, see Rutherford, pp. 8-9, Dover, pp. 80-82, and H. Thesleff, Studies on Intensification in Early and Classical Greek, Commentationes Humanarum Litterarum 21.1 (Helsinki 1954) 80-92. Thesleff's monograph is one of a number publications that one is surprised to find no mention of in Dover's book. Another is D. M. Clay, A Formal Analysis of the Vocabularies of Aeschylus, Sophocles and Euripides I (Minneapolis 1960), II (Athens 1958). Rather than searching "the first 200 adjective-lexemes of Soph. Phil." and "the first 200 of Eur. Orestes" (p. 118) for adjectives in -IKO/S, Dover could have found a complete list and discussion in Clay (I pp. 76-78 and II pp. 48-49). Nor does Dover cite F. Solmsen, Intellectual Experiments of the Greek Enlightenment (Princeton 1975), whose extended discussion (pp. 110-125) of the neuter adjective used substantively in place of an abstract noun, especially in Thucydides, is relevant to Dover's treatment of substantival phrases (pp. 32-36). Note that Dover refers (p. 34) to an "experiment" on the part of Antiphon and Thucydides to attach dependent genitives to neuter adjectives; Solmsen had quoted (p. 111) from the 440s Soph. Aj. 964 TA)GAQO\N XEROI=N and Ant. 365 TO\ MHXANO/EN TE/XNAS.

Particularly successful and illuminating is chapter 6, "Special Languages," which concerns itself with poetic language, technical vocabulary and metaphor. These sections reveal Dover's great linguistic sensitivity and his talent for identifying just the right features for analysis and selecting just the right examples. In this chapter and throughout the book Dover's presentation is clarified by a series of intelligent and intelligible tables and figures. My students in Greek prose composition will be much more responsive to "Table 6.1: 'Die' and 'kill' in poetry and prose" than to my feeble exhortations that they use A)POQNH|/SKEIN and A)POKTEI/NEIN rather than the uncompounded verbs. The section on metaphor reveals in a particularly telling manner that the study of the Greek language cannot be divorced from the study of Greek thought. In his discussion of levels of style (i.e. linguistic style, philosophical style, etc.) Dover had earlier made the excellent point (p. 4) that "some degree of interaction between levels is to be expected," and nowhere is this more apparent than in the case of metaphor. Dover examines the metaphorical use of the word POIMH/N (pp. 124-5) and says, of the occurrence at Pl. Plt. 274E, "Here we may suspect that (as so often) there is a non-linguistic determinant of the data: in this case, the tradition of Homeric exegesis and the educational need to find wisdom in Homer's language." That is to say, Dover takes this instance, with reference to Xen. Mem. 3.2.1 and [Pl.] Minos 321C, as a reflection of the Homeric formula POIME/NA LAW=N. And that is undoubtedly correct. But it is also a reflection of the Socratic analogy with the crafts; cf. Xen. Cyr. 1.1.2, Pl. Gorg. 516A, Rep. 343B, and elsewhere, where the figure of thought, but not the figure of speech, occurs. Dover recognizes (p. 129) that extended imagery, for example, belongs "to a stylistic level higher than the linguistic," but is that not the case with metaphor as a whole? If, for whatever idiosyncratic reasons I might choose to do so, I refer to the giraffe as "the emissary of affection" or to the city of Toronto as "an open door to self-improvement," my decision to express myself in this manner is a decision taken at a much different (if not necessarily a higher) level than the decision to choose one rather than another synonym or one order of words than another. Even the decision to express myself metaphorically rather than in some less flamboyant fashion is taken more at the level of style of thought than of linguistic style.

Dover's chapter on rhythm records the results of a "simple experiment" (p. 163) designed to discover the extent to which prose authors favored the use of "poetic rhythms." This clever experiment consists of manipulating texts by means of transposition, and then examining the texts to see whether there is a significant difference, with regard to the frequency of poetic rhythms, between the manipulated text and the text as originally produced by the author. In his "Poetic Rhythms in the Myth of the Soul," in L. Ayres (ed.), The Passionate Intellect (New Brunswick, NJ 1995) 13-22, Dover had carried out the same experiment on Plato's myth of the soul in the Phaedrus, a passage that Socrates himself characterizes as having been delivered in poetic language. In that case, the results clearly showed a noticeably higher incidence of poetic rhythms in Plato's text than what might have arisen by chance (the manipulated text). Once again, genre is revealed to be a significant stylistic determinant. For, in the book under review, another passage from Plato (a portion of Book 2 of the Republic; see p. 164) is subjected to the experiment and is found to display no such predilection for poetic rhythm. The same is true ("Table 8.2," p. 168) of passages that Dover analyzes from Thucydides and Isocrates; a passage from Demosthenes' On the Crown, however, appears to display a significant preference for poetic rhythms. Further investigation is clearly to be desired, and Dover has provided us with a useful tool for carrying out the necessary work.

There is one detail, however, in Dover's method of analysis concerning which I have some reservations. Like many scholars who study prose rhythm, Dover regards a short syllable followed by pause as equivalent to a long, as it would be counted in poetry (brevis in longo). I believe this to be illegitimate. One has only to compare the first four notes of Beethoven's fifth symphony or of Strauss' Elektra with the first four notes of Lutoslawski's Symphony no. 3. In each instance a four-note motto is played fortissimo and is followed by a rest. But, whereas the Beethoven has paeonic rhythm (three eighth notes followed by a half note) and the Strauss is ionic (two sixteenth notes followed by two eighth notes-the "Agamemnon" motif), the Lutoslawski consists of four notes of equal duration (four eighth notes). The rhythm of the Beethoven and the Strauss (underlined by their descending pitch and by the fact that the downbeat falls on the long notes) marks them as self-contained and even, for all their opening position, as "clausular." The Lutoslawski, by contrast, with its steady pitch and its downbeat on the opening note, and for all that this motto closes as well as opens the symphony, is manifestly anticipatory; cf. C. M. Rae, The Music of Lutoslawski (London 1994) 167. This perception, that a short before pause does not have the same rhythmic character as a long, is confirmed by explicit ancient testimony. At Rhetoric 3.8 Aristotle prescribes that prose diction should be rhythmical, but not metrical (DEI= MH/TE E)/MMETRON EI)=NAI MH/TE A)/RRUQMON), for metrical speech appears contrived (PEPLA/SQAI GA\R DOKEI=). I take this to mean that speech should create its own rhythmical patterns and should not have artificial metrical constraints placed upon it. Aristotle goes on to recommend a long syllable for final position, for a short syllable, by being inconclusive (A)TELH/S), curtails the ending. Aristotle is, of course, speaking of genuinely long syllables, and not of short syllables artificially "counted as long" by virtue of the following pause, for in the latter case all final syllables would count as long, and Aristotle would be talking nonsense. (I am aware that D. F. McCabe, The Prose-Rhythm of Demosthenes [New York 1981] 119-130, has convincingly demonstrated that Demosthenes treated final short syllables as having the character of brevis in longo, but he was not able to demonstrate the same for Isaeus or for the spurious works in the Demosthenic corpus. Although further work still needs to be done, McCabe's findings appear to suggest that Demosthenes was the odd man out, and we ought not to assume that what holds true for Demosthenes holds true for Greek prose authors generally.)

In conclusion, then, this is a wonderful book, one that will be read and reread and consulted with profit for years to come. Progress in the study of Greek prose style is slow, as it depends upon a small number of scholars who have extensive experience with both prose and verse, who are sensitive to nuance and, increasingly, who are capable of carrying out and appreciating statistical analysis. Dover certainly possesses all those qualifications, and The Evolution of Greek Prose Style will both facilitate and encourage further research. The book itself is very well produced, and there are fewer typographical errors than in some of Dover's other recent publications, perhaps because he had the benefit of help from some admirable young proofreaders (p. vii). But they did not save him from offending in his bibliography the two American members of the BMCR editorial board whom he cites, by misspelling the name of one (p. xii) and by omitting the title (and part of the subtitle) of an article by another (p. xix). Otherwise, only the weird intrusion on p. 39 will cause readers to take notice, and only a couple of errors in the tables will mislead: In "Table 7.3" (p. 146) the figures for Gorgias, Helen and the "scaled up" figures should be in the same columns, and in "Table 7.4" (p. 151) the last figure in the left-hand column should be "+1."