Bryn Mawr Classical Review

Bryn Mawr Classical Review 2009.08.12

Nicoletta Marini (ed.), Demetrio, Lo Stile. Pleiadi 4.   Roma:  Edizioni di storia e letteratura, 2007.  Pp. x, 353.  ISBN 978-88-8498-352-7.  €48.00 (pb).  

Reviewed by Casper C. de Jonge, Leiden University (
Word count: 2617 words

Demetrius's On Style is one of the most valuable works of ancient rhetoric and literary criticism for several reasons. First, it offers a very detailed and instructive account of ancient stylistic theory that shows us the development of rhetorical teaching after Aristotle. Second, it informs us about several topics for which we do not have many other sources, in particular musical theory, prose rhythm, and the ancient theory of letter writing. Finally, Demetrius cites and discusses a great number of classical models and thereby preserves some important fragments from prose writers including Ctesias, Demades and Sophron, and poets such as Anacreon, Archilochus and especially Sappho. His observations on these authors are in many cases illuminating for modern readers.

The most recent editions with translations of Demetrius are those of Chiron (1993, Budé) and Innes (1995, Loeb).1 The latter translation is based on the earlier one by Rhys Roberts (Cambridge 1902, revised for the Loeb series in 1927), whose original notes are still worth consulting.2 Nicoletta Marini has now produced an elegant and useful introduction, translation and commentary in Italian. Her text (without apparatus criticus) is based on Chiron's edition, from which she departs in 36 places. The introduction deals with the date and authorship of the work, the title, structure and contents, the rhetorical system that On Style represents, Demetrius's place in the tradition, the modern reception and the textual transmission of the work. As far as a non-native reader of Italian can judge, the translation is clear and correct. But this is the third Italian translation of the work to appear in a decade, so it will have to compete with Lombardo (1999) and Ascani (2002); the former translation has detailed annotations, but no Greek text; the latter comes with a Greek text and a fine introduction by Schenkeveld.3 Marini's book, on the other hand, has the advantage of a rich commentary, which contains many helpful observations on textual criticism, linguistic difficulties and ancient rhetorical theory. It is therefore a welcome addition to the annotations that accompany earlier translations, especially those of Rhys Roberts, Chiron and Lombardo.

One important predecessor of Marini has not yet been mentioned: in 1968, Doreen Innes completed an introduction and commentary on Demetrius that has, unfortunately, never been published. This Oxford dissertation (by one of the most learned specialists in the world) can still be consulted in the Bodleian Library.4 It is both understandable and regrettable that it does not appear in Marini's otherwise comprehensive bibliography.

The most remarkable aspect of Marini's book is her view on the date of the treatise. Scholarship on the treatise On Style has always concentrated on its date and authorship. It was traditionally ascribed to Demetrius of Phaleron (ca. 360-280 BC). This is the attribution in the superscription of the important manuscript P (Parisinus gr. 1741, tenth century), but the simpler version in the subscription of the same manuscript is 'Demetrius On Style' (Δημητρίου περὶ ἑρμηνείας), which is probably correct. The author was thus a certain Demetrius, but according to most scholars it was not Demetrius of Phaleron. Various Demetrii have been suggested, including Demetrius of Tarsus (first century AD: Rhys Roberts 1927) and Demetrius of Syria ('an experienced and famous teacher of eloquence' who taught Cicero around 80 BC in Athens, as we learn from Brutus 315: Chiron 1993). But the author may also be a Demetrius not known from other sources.

Concerning the date, many possibilities have been suggested, from the early third century BC to the second century AD. In the beginning of the twentieth century, Radermacher (1901) and Rhys Roberts (1902 and 1927) argued for a late date, in the first century AD.5 In recent scholarship, however, there was a growing consensus that the work belongs to a relatively early period, in the first century BC (Schenkeveld), second or early first century BC (Chiron, Innes) or even third century BC (Morpurgo Tagliabue).6 Against this tendency, Marini now argues that On Style belongs to the first century AD, thus giving yet another turn to this ongoing debate.

Why does Marini disagree with the recent specialists who favor an earlier date? Having discussed the traditional arguments concerning the language and contents of the treatise, she attaches great importance to three arguments in particular, which she borrows from Rhys Roberts (1927) and Calcante (2000, 2004).7 First, she argues, following Calcante (2000), that Demetrius's theory of four styles is an extension and adaptation of the earlier system of three styles that we find in Rhetorica ad Herennium, Cicero and Dionysius of Halicarnassus. Demetrius would have inserted a fourth style into the traditional system of three styles by dividing the middle style into the elegant style on the one hand, and the forceful style on the other. According to this reconstruction, Demetrius would come at the end of the development of stylistic theory, after rhetoricians like Cicero and Dionysius. But other scholars have convincingly argued that Demetrius's system of four styles belongs to a period in which the system of three styles had not yet become canonical. So the difference between the unusual system of four styles and the common system of three styles can in fact be an argument for an early rather than a late date.

Second, Marini draws attention to Demetrius's discussion of figured speech or innuendo (par. 287-295), which is to be used in addressing tyrants. Following Rhys Roberts (1927: 277), she argues that this passage, which can be compared with Quintilian's discussion of emphasis (9.2.64-99), reflects a period in which the freedom of speech was limited: this would fit the second half of the first century AD. However, Demetrius's advice on this topic seems useful for orators in many periods, not just in the context of imperial Rome. His examples are concerned with tyrants of earlier periods (e.g. Plato addressing Dionysius in par. 290 and Demetrius of Phaleron addressing the Macedonian Craterus in par. 289). The same passage has even been used in support of the authorship of Demetrius of Phaleron himself (third century BC), who was clearly interested in methods of addressing both tyrants and assemblies.8 In other words, scholars have employed this evidence for completely opposite claims, consequently it cannot decide the matter. A link between Demetrius's ideas on innuendo and the situation in the Roman Empire remains possible, but not at all necessary.

Marini's final argument for a date in the first century AD, which I will discuss at more length, concerns her interpretation of Demetrius 36: here she again follows the ideas of Calcante (2004: 111). Having introduced the four styles (the plain, the grand, the elegant and the forceful), Demetrius points out that most of these four styles can be combined, except for the grand and the plain. He then adds that 'some writers maintain that only these two styles (i.e. the grand and the plain styles) exist, and the remaining two (i.e. the elegant and the forceful) are between them (τοὺς δὲ λοιποὺς δύο μεταξὺ τούτων)'. According to Calcante, Demetrius here refers to Dionysius of Halicarnassus (On Demosthenes 33), who distinguishes three styles, namely the elevated style, the plain style and the style 'between those styles' (τὸν μεταξὺ τούτων). Marini adopts Calcante's interpretation: when Demetrius rejects the idea of some predecessors who treat the elegant and the forceful styles as 'between the other two', he is thinking of Dionysius's presentation of the mixed style as intermediate between the elevated style and the plain style. We would then have a terminus post quem for Demetrius's On Style, namely 30 BC, the year in which Dionysius came to Rome and started to work on his history of Rome and his rhetorical treatises.

This argument is in fact rather problematic, for two reasons. First, Dionysius of Halicarnassus distinguishes between three styles (the elevated, the plain and the middle or mixed), whereas Demetrius refers to a theory of two styles ('some writers maintain that only these two styles exist', i.e. the grand and the plain), which he opposes to his own theory of four styles (grand, plain, elegant, forceful). Dionysius does not say that only two styles exist. Second, μεταξὺ τούτων (Demetrius 36) does not mean that the elegant and forceful styles together formed one middle style between the grand and the plain styles (this would be the meaning that is required for the supposed reference to Dionysius). Demetrius himself explains what he means by μεταξὺ τούτων: his predecessors do not give the elegant and the forceful styles a separate status, that is, they believe that the elegant belongs to the plain and the forceful to the grand. In other words, Demetrius does not mention Dionysius's system of three styles (grand, plain, intermediate), but he objects to a system of two styles, in which the elegant is covered by the plain style and the forceful by the grand style. My explanation of the passage agrees with the interpretations of both Chiron (1993: 14 n. 55: 'Démétrios ne mentionne pas la division du style en trois types') and Innes (1968: 51: Demetrius 'combats only a theory of two styles'). Her translation of Demetrius 36 (in the 1995 Loeb) is illuminating: 'For this reason some writers maintain that only these two styles exist, and the other two are subsumed within them; and instead they assimilate the elegant to the plain, and the forceful to the grand, as though the first contained something slight and refined, the second something massive and imposing.'

It is clear that in this passage Demetrius opposes his own system of four styles to an earlier system of two styles, thereby ignoring the system of three styles that we find in Dionysius of Halicarnassus. In my view, therefore, Demetrius 36 does not contain any reference to Dionysius, and there is no 'rilettura critica dello schema dionisiano' (Marini 2007: 37 citing Calcante 2004: 111). Consequently, there is no terminus post quem, so this argument for dating Demetrius in the first century AD seems untenable. There is no reason to believe that Dionysius and Demetrius would necessarily have known each other's work. In general, arguments based on supposed references between the two authors (as well as arguments based on the silence of one author about the other) are to be distrusted.9

Apart from the three arguments mentioned, Marini also points to linguistic aspects of the treatise, but this is very uncertain evidence. It is true that Demetrius shows an incipient interest in linguistic Atticism (e.g. in the use of the dual), but recent scholars have rightly emphasized that our knowledge of Hellenistic prose language is too limited to draw sweeping conclusions. The few terms that Demetrius explicitly mentions as usages that had recently been introduced in his own time form the most interesting category (λόγιος, κακόζηλος, etc.), but they do not preclude a date in the second or first century BC.10

Why is it that other specialists have argued for an early date? The main argument remains the fact that the treatise On Style contains many ideas that are traditional and relatively early, whereas it ignores a number of later rhetorical theories. Demetrius's treatment of figures of style (which ignores the category of tropes), the strong influence of Aristotle and the early Peripatos, and Demetrius's familiarity with authors of the fourth and third centuries all point to a relatively early date. I would add two further arguments, which are not usually mentioned. First, Demetrius does not employ the terminology of the sublime (ὕψος), which we do find in Dionysius of Halicarnassus and later authors. Second, Demosthenes is an important model in the treatment of the forceful style, but Demetrius does not give him the supreme status that he will have in the works of Cicero, Dionysius and Hermogenes. These points may not be decisive, but I agree with recent scholars that it is on the whole more likely that Demetrius precedes Dionysius. In short, Marini's arguments for a late date (first century AD, after Dionysius) have not convinced me.

Leaving the complex issue of the date, I would like to emphasize once more that Marini's commentary contains numerous useful observations, even if one can disagree on details. A few examples will be sufficient. In her discussion of Demetrius par. 106 (on 'epiphoneme'), Marini gives relevant parallels from Hermogenes and Quintilian, and she discusses the uncertain authorship of Demetrius's example (Sappho fr. 105b Voigt). Sappho is Demetrius's model for the elegant style, whereas the epiphoneme (and its illustration traditionally attributed to Sappho) is treated under the grand style. Therefore Marini concludes that the attribution of this citation to Sappho may be wrong. Her doubts may be justified, but I would argue that the 'epiphoneme' is closely related to the elegant style because it is 'decorative' (Demetrius uses the verb ἐπικοσμεῖν three times in par. 106). 'Decoration' or 'adornment' (κόσμος) is especially characteristic of the elegant style (see par. 164-165), so Sappho might be the right model for the epiphoneme after all.

In her comment on Demetrius par. 199 (on natural word order in the plain style), Marini helpfully refers readers to similar discussions in Dionysius (Comp. 5) and Longinus (22.1). According to Marini, Stoic influence on this passage is possible. Stoic thought can indeed be traced in Dionysius's chapter on natural word order. However, Demetrius's theory ('mention first what it is about, then what it is') seems more pragmatic and rather related to Aristotelian ideas: Aristotle (Rh. 1415a12-21) already recommends beginning a text by mentioning 'what it is about'. Demetrius applies the same principle to the order of words.11

One of the most fascinating aspects of On Style is the rewriting or rearrangement of texts, so-called metathesis. In many cases Demetrius analyzes his examples by changing the word order of the original text, so that he can compare two different versions of one text. In this way, he is able to illustrate the stylistic qualities or shortcomings of a text. The same technique can be found in the works of Philodemus, Cicero, Dionysius and Longinus. This intriguing instrument of literary analysis, which also has a pedagogical function, does not receive much attention in Marini's commentary. Some important studies of this phenomenon should at least have been mentioned.12

Where Marini departs from Chiron's text, she gives instructive discussions of the textual problem and the solutions that have been suggested, and she provides sound arguments for her reading. Among her interesting and convincing choices I should mention her readings of 'Herodorus's instead of 'Herodotus's in par. 66 (an emendation by Orth) and Ἰλλυρικῶν ('Illyrian') instead of λαυρικῶν ('of Laurium') in par. 121 (an emendation by Gärtner, supported by Schenkeveld). In both cases Marini's text differs from the editions of Chiron and Innes, and in both cases I find her arguments convincing. In par. 66, the citation of 'Herodotus's is absent from our text of Herodotus's Histories (Herodotus 1.203.1 has been suggested, but this passage is in fact rather different). However, the same text does fit the interests of the historian Herodorus of Heraclea: thus, 'Herodorus's might indeed be the correct reading here.13 In par. 121, the hills 'of Laurium' would be called λαυρειωτικῶν or λαυρεωτικῶν rather than λαυρικῶν, so the emendation 'Illyrian' is indeed attractive.14

Although I do not agree with Marini's arguments concerning the date of the treatise, I think that her book is an important contribution to the scholarship on Demetrius. The strongest part of this book is the commentary: when reading On Style, everyone will profit from her comments, which are more extensive than the notes by Rhys Roberts, Chiron and Lombardo. This book will be useful for many scholars, including those who are interested in the history of rhetoric. We may hope that Marini's work will stimulate other scholars to think further about this fascinating text.


1.   Pierre Chiron (ed., trans.), Démétrios, Du style, Paris 1993; Doreen C. Innes (ed., trans.), Demetrius on Style, based on W. Rhys Roberts, in Aristotle, Poetics. Longinus, On the Sublime. Demetrius, On Style, Cambridge, MA / London, England 1995 (reprint with corrections: Cambridge, MA / London, England 1999).
2.   W. Rhys Roberts (ed., trans.), Demetrius On Style. The Greek Text of Demetrius De Elocutione, edited after the Paris Manuscript, Cambridge 1902 (reprint Hildesheim 1969); W. Rhys Roberts (ed., trans.), Demetrius On Style, London / Cambridge, MA 1927 (Loeb, based on 1902 edition). Another good translation is G.M.A. Grube, A Greek Critic. Demetrius on Style, Toronto 1961.
3.   Giovanni Lombardo (trans.), Demetrio, Lo stile, Palermo 1999; Alessia Ascani (trans.), Demetrio, Sullo stile, introduzione di Dirk M. Schenkeveld, Milano 2002. Ascani also wrote two unpublished dissertations on Demetrius: Ricerche sul Peri Hermeneias di Demetrio Ps.-Falereo, Diss. Bologna 1994; and De sermone figurato quaestio rhetorica. Per un'ipotesi di pragmatica linguistica antica, Diss. VU Amsterdam 2006.
4.   Doreen Innes, Introduction and Commentary on the Peri hermeneias ascribed to Demetrius of Phaleron, Diss. Oxford 1968.
5.   Ludwig Radermacher, Demetrii Phalerei qui dicitur de Elocutione libellus, Leipzig 1901.
6.   G. Morpurgo Tagliabue, Demetrio: dello stile, Roma 1980; D.M. Schenkeveld (trans.), Demetrius. De juiste woorden, Groningen 2000; D.M. Schenkeveld, 'The Intended Public of Demetrius's On Style: The Place of the Treatise in the Hellenistic Educational System', Rhetorica 18 (2000), 29-48. In his earlier work, Schenkeveld argued that On Style belongs to the first century AD but reflects theories from the second century BC. See D.M. Schenkeveld, Studies on Demetrius On Style, Amsterdam 1964, 135-148. G.M.A. Grube, 'The Date of Demetrius "On Style"', Phoenix 18 (1964), 294-302 argues for a date around 270 BC.
7.   M.C. Calcante, Genera dicendi e retorica del Sublime, Pisa / Roma 2000; M.C. Calcante, 'Il De elocutione di Demetrio e il De Demosthene di Dionigi di Alicarnasso', RIL 138 (2004), 99-124.
8.   Grube (1961) 52-53. The fact that Demetrius of Phaleron is mentioned in par. 289 does not exclude the possibility that this Demetrius is himself the author (as some scholars have claimed): many ancient authors refer to themselves in the third person: cf. Grube (1961) 39-40.
9.   In par. 179, Demetrius points out that no previous writer has discussed the elegant composition type. This remark implies ignorance of Dionysius of Halicarnassus, Comp. 23 and Dem. 40. Scholars have therefore argued that Demetrius predates Dionysius. Marini (pp. 246-247), however, again following Calcante (2000: 140-142), argues that Demetrius merely means to say that no earlier writer has discussed one specific aspect of elegant composition, i.e. rhythm. With this interpretation, Demetrius could still be later than Dionysius. But Calcante misrepresents Demetrius's words: he does not say that one aspect of elegant composition has been neglected. He states that 'no one has spoken about elegant composition'. We can only conclude that Demetrius did not know Dionysius's treatment: either Demetrius was earlier, or, if he was later, he just did not know Dionysius's work. Note that this is another argument against Calcante's idea that Demetrius par. 36 refers to Dionysius, Dem. 33.
10.   Innes (1995) 320-321.
11.   Demetrius's ideas on this topic remind us of modern discussions of word order. See now C.C. de Jonge, 'From Demetrius to Dik. Ancient and Modern Views on Natural Word Order', in R. Allan & M. Buijs (eds.), The Language of Literature. Linguistic Approaches to Classical Texts, Leiden / Boston 2007, 211-232.
12.   C. Damon, 'Aesthetic Response and Technical Analysis in the Rhetorical Writings of Dionysius of Halicarnassus', Museum Helveticum 48 (1991), 33-58; C.C. de Jonge, 'Dionysius of Halicarnassus and the Method of Metathesis', Classical Quarterly 55 (2005), 463-480. See also R. Janko, Philodemus On Poems Book 1, Oxford 2000, 227 n. 2, who lists the instances of metathesis in Demetrius.
13.   See E. Orth, 'Ein Fragment des Herodorus. Zu Demetrios par. 66', Philologische Wochenschrift 45 (1925), 778-783.
14.   See D.M. Schenkeveld, Review of Chiron 1993, Mnemosyne 50 (1994), 403. λοκρικῶν has also been suggested.

Comment on this review in the BMCR blog
Read Latest
Index for 2009
Change Greek Display
Books Available for Review
Bryn Mawr Classical Commentaries

HTML generated at 08:52:34, Thursday, 06 August 2009