SOURCE: "The Date of Demetrius on Style," The Phoenix, Vol. XVIII, No. 4, Winter, 1964, pp. 294-302.

[In the following essay, Grube reasserts his position regarding the date of On Style, responding to opposing arguments made by Goold and Rist (see two previous excerpts).]

Since the publication of my A Greek Critic, Demetrius on Style,¹ two of my Toronto colleagues have published articles challenging my dating of the treatise in 270 B.C. or not much later; both G. P. Goold's "A Greek Professorial Circle at Rome," TAP A 92 (1961) 168-192 and J. M. Rist's "Demetrius the Stylist and Artemon the Compiler" in Phoenix 18 (1964) 2-8, argue for a much later date.

I

Goold has a double aim: he seeks to reinforce the view that Demetrius was a contemporary of Dionysius of Halicamassus in the first century B.C. in Rome, and then to identify him with the Peripatetic philosopher whose contention (that Demosthenes learned his rhetorical art from Aristotle) Dionysius disproves in the first letter to Ammaeus, and also with the Demetrius to whom Dionysius had sent an essay on Imitation, as he tells us in the letter to Pompey (chapter 3 ad init.). Pompeius himself is then further identified with the author of On the Sublime, but that does not concern us at present.

Goold writes well, but I fear that attractive rhetorical phrases not infrequently take the place of argument. Like Horace, he says what others have said before but says it better. These parts of the article I must ignore, for I cannot here repeat all I have already said in many pages. To give one example: I showed that there are in the treatise an unusually large number of references to and quotations from persons who are known to have lived in the late fourth and early third centuries, and I listed and discussed these. Goold replies (p. 181) that "they have not seemed so to any other scholar," which proves nothing. Indeed it was my reason for pointing this out. That kind of argument I must ignore, and try to restrict myself to points that are new, or where something is added.

Nor am I a lone objector to "the orthodox opinion which puts it three hundred years later." This was the orthodox view sixty years ago, but a number of scholars have more recently argued for earlier dates, varying from the late third century to 100 B.C.²

I agree that the passage in Philodemus' Rhetoric³ which states that "long periods are bad for delivery, as we read in Demetrius about those of Isocrates …" refers to Demetrius of Phalerum. In fact I pointed this out. My argument was and is that Philodemus is clearly referring to the Phalerean, and that if the reference is to our treatise (which Hammer believed and Rhys Roberts thought possible) then, obviously, Philodemus thought the Phalerean to be its author. Goold maintains there is no such criticism of Isocrates in the treatise On Style, that "one cannot infer from our Demetrius that he disapproved of Isocrates' periods or even regarded them as long at all" (p. 179). This is not quite Philodemus' point, but Goold is too hasty when he says that "Isocrates is mentioned only three times, once (12) to cite him with Gorgias and Alcidamas as an example of the periodic style, and twice to comment on his avoidance of the hiatus." At 12 Demetrius goes on to say that "in their work" -(i.e., the work of Isocrates, Gorgias, and Alcidamas) "one period follows another just as surely as one hexameter follows another in Homer" and 15 proves this to be adverse criticism since it says that no logos should consist of a string of periods; 16 adds that no period should have more than four clauses. And then, at 25-28, quotations from Isocrates and Theopompus, among others, are followed by strong comments that the rhetorical devices of their style are incompatible with force or passion and are bad art. …

But Philodemus' "bad for delivery" would naturally refer to the section on the forcible style … (240-303) since this depends on delivery more than any other. We have already been told (193) that the style which omits connectives is more suited to debate, more histrionic …; at 271 forcefulness is all but equated with the disconnected style (the opposite of the Isocratean). Then we find that the forceful style requires short phrases (241), brevity (242), avoids antitheses and balanced clauses (247, where an example of these is given from Theopompus, the famous pupil of Isocrates); if one uses periods, they must be short (252); connectives should often be omitted (269); and again (299): "smoothness in the arrangement of words, as affected by the school of Isocrates … with its avoidance of hiatus, is not suited to forceful speech." Finally we are told (303), that "long, continuous periods, which make the speaker pant for breath, not only surfeit but repel him."

It is crystal clear that Demetrius did dislike the long, balanced, antithetic periodic style of Isocrates and his school. In fact he condemns it in no uncertain terms, both in his introductory discussion and especially in connection with delivery and the forceful style. The precise words of Philodemus are not found in any particular section, but the idea is expressed several times. Even if Isocrates' name were not mentioned at all, every one of the above passages would be understood by any reader acquainted with rhetoric to refer to him and his school, and there is in my opinion a strong probability that Philodemus is referring to our treatise.

There is another passage of Philodemus which I did not use because, as always, the text is very doubtful, but Goold uses it to prove that Philodemus knew only of a three-style formula (185, n. 37); but not of any four-style formula such as we find in Demetrius. On the contrary, the most probable interpretation of the sentence seems to prove that Philodemus at least knew a four-style formula.…

However, modern scholars did not like the idea that Philodemus knew of a four-style formula, and so Radermacher proposed [mesoteta] instead of [megesthos].⁴ This is legitimate, since only the is certain; but this reading requires us to take the third XXX in a completely different sense from the other two; it signifies equivalence where the other two are enumerative; you cannot, of course, have a mean style in a four-style formula. I find this reading quite unconvincing. Goold himself suggests (185, n. 37) [mesen tina] which avoids the difficulty of the third [e], which it deletes. But this use of [tina], i.e., [mesen tina giaphuroteta] "or (thirdly) a kind of intermediate elegance," is very awkward and unusual in a general description of the different kinds of [plasma]. As is the case with Radermacher's, the only purpose of this restoration is to make the text conform to the preconception that Philodemus must be referring to three styles. I think that Sudhaus' restoration, [megethos e] is much more likely, and that Philodemus is referring to a formula of four styles. And if we can take [hadrographia] in the sense of forcefulness, a natural meaning, though later critics use [hadron] as a synonym for the grand style (here [megethas]), then these are indeed the four styles of Demetrius. However, I did not, and do not, press this argument, for the text is much too fragmentary. It may be noted, however, that Wilamowitz also thought that Philodemus here refers to four styles (Hermes 35 [1900] 30, n. 4).

Goold (p. 180) makes much of a reference in Syrianus (c. 400 A.D.) who is speaking of formulae of styles:⁵

One of these is Dionysius, for he says that there are three styles: the simple, the mean, and the grand. Hipparchus adds the picturesque and the flowery. Demetrius rejects the picturesque, being satisfied with four.

We are told that Syrianus must be referring to these three critics in chronological order. Demetrius is therefore later than Dionysius and the unknown Hipparchus. This, however, is far from certain, especially so with the use of the present tense. Dionysius, moreover, did not use [grand] for the grand style, and Demetrius is left with four styles with a mean! I do not think much importance can be attached to this late and somewhat confused passage.

On the language of Demetrius, Goold indulges in a false analogy (pp. 183-184): he says quite rightly that very often perfectly normal derivatives do not appear for a long time, and he illustrates this by English examples; whereas "help" and "force" are attested in 1300, yet "forcefully" does not appear till 1774 and "helpfully" till 1832. He then adds: "it is not through lack of written English" that we can be sure that certain adverbs are not found "from Chaucer to Hemingway." But therein lies precisely the falsity of his analogy. My point was and is that we do not have written texts (and no other critical text) which would enable us to determine whether certain forms were or were not in use in the third century B.C. And we do have an earlier critical text, the Rhetorica ad Alexandrum, firmly dated in the late fourth century where we find quite a number of expressions and constructions which are also "late."⁶

There are two other difficulties which Goold has to face: he admits that Demetrius' attitude to Demosthenes is very different from that of later critics: to them the great orator was a paragon of all the virtues and styles; Demetrius uses him only to illustrate forcefulness. So Goold has to imagine his Demetrius as objecting "to the perpetual hosannas everywhere raised at the name of Demosthenes" in the first century, and then further as inventing a forceful style specifically in order to be able to "demote" the orator into it, thus fashioning "an effective weapon against the teaching of Dionysius" (p. 185). This fanciful picture is not supported by any evidence; indeed, the whole tone of On Style is against it. It is not in the least polemical, and shows no hostility whatever against Demosthenes, nor does the orator in any way monopolize the chapter on forcefulness. The absence of any reference to oratorical mimêsis or emulation in our treatise is explained by a similar imaginative flight: Demetrius, as a "staunch Peripatetic" is here also in conscious opposition to Dionysius (p. 187).

But the description of Demetrius as a staunch Peripatetic is somewhat startling. I discussed at some length (pp. 32-38) the unusual attitude of our author to Aristotle, whose works on rhetoric he obviously knew and used for his purposes, simplifying his terminology at times, sometimes improving on him, in a way unparalleled in later times. This is Goold's only reply (186):

The relationship is, on the contrary, similar to that which might exist between a Roman Catholic theologian and Thomas Aquinas. On fundamental matters a chronological proximity might appear indicated, but differences of emphasis in the application of doctrine and a greater precision of technical minutiae, not to mention general linguistic discrepancies, would reveal the lapse of centuries between them.

My knowledge of theology is inadequate to enable me to appreciate the nature of the relationship here indicated, but the simile does not apply to Demetrius and Aristotle, for the differences between them are in essential matters as well as in details—the whole basis and structure of the treatise is un-Aristotelian: the four styles, distinguished by difference of subject-matter, diction, word-arrangement, figures, and the rest. Demetrius has none of Aristotle's philosophical approach and logical distinctions. Whatever Demetrius may have been, he was no staunch Aristotelian.

I shall but briefly refer to Goold's identifications of Demetrius in the first century. He looks for a "literary professor" who is "a Peripatetic teacher of rhetoric." He finds the anonymous Peripatetic philosopher of the letter to Ammaeus and promptly identifies them, though one doubts that any rhetorician would have said that Demosthenes owed his art to Aristotle (who never mentions him). Goold then finds that Dionysius says he addressed essays on Imitation to one Demetrius (a common enough name) and declares this Demetrius to be the same man. As we know nothing more about either of these men the identification can certainly not be disproved. But this is pure play of fancy—a pleasant fancy if we are sure on other grounds that Demetrius wrote in Rome in the time of Dionysius; it makes no contribution whatever to the problem of date.

II

I will now deal with Rist's article on Artemon, whom Demetrius (223) mentions as the editor of Aristotle's letters. This is a very thorough and careful hypothetical structure which concludes that Artemon cannot have lived before the end of the second century B.C.—so that Demetrius should not be dated till 100 B.C. and probably later. I am not convinced, for two reasons: first, because in this type of structure where each hypothesis serves as a basis for the next, the degree of probability is obviously less at each step, and, moreover, a quite different conclusion can be drawn from the same evidence.

Rist's case may fairly be summarized as follows: We have three lists of Aristotle's works. The first two, that of Diogenes Laertius and the second, usually connected with Hesychius, are generally believed to derive either from Hermippus or Ariston, who both lived in the late third century, so either will do for our purpose. Now Hesychius mentions twenty books of Aristotle's letters, while Diogenes knows a number of books which add up to twenty but also some other letters to Philip about the Selymbrians. It follows that only twenty books were known to their common source, whether Hermippus or Ariston.

The third list is preserved only in Arabic, in two versions of the 13th century. These are generally believed to be ultimately derived from the compilation of Andronicus of Rhodes who, we know, worked on the Aristotelian works of the library of Apellikon which Sulla brought to Rome early in the first century. This library of Apellikon contained most of the library of Neleus, who, as the well known story of Strabo tells us, had received the library of Theophrastus at his death.⁷

The Arab versions speak of the library of Apellikon as containing two lots of letters. The first lot is described in one version as (I quote from Rist's article, p. 4): "a book in which a man named Artemon collected a number of letters by Aristotle in eight books"; and the other version says: "a large book, a compendium containing a number of letters in eight parts." The second lot is described in both Arab lists as: "other letters which Andronicus found in twenty parts." From this Rist makes the following deductions: the twenty books of letters are the twenty books mentioned by both Diogenes and Hesychius, and these were known to their third-century source, be it Hermippus or Ariston. The eight books could then be the same as Diogenes' letters to Philip. In this connection, Rist notes some confirmation in that Elias, among works of Aristotle on concrete topics speaks of "the letters which a certain Artemon, who lived after Aristotle (my italics), collected in eight books," and which "were written [pros hera]." These must, Rist believes, be the letters to Philip.⁸ Olympiodorus speaks of Andronicus and Artemon in that order as compilers of Aristotle's letters (my italics: the chronological order again), and he speaks of letters [pros tinas idia], which describes both their compilations. Elias means the letters to Philip (as in Diogenes), Olympiodorus all those to individuals, including the collection of Andronicus.

So we have twenty books throughout, and eight books, the collection of Artemon, unknown in the late third century, but appearing in Diogenes (from somewhere) and in the Arab version. As Artemon's collection was not known in the third century, it cannot have been in the original collection of Apellikon. The Arabs may have got his name from Andronicus, and he would then at best, be "a man who lived some time between the original acquiring of the library by Apellikon and its eventual publication by Andronicus," i.e., between the late second century B.C. and the second half of the first. And so Demetrius cannot be earlier than 100 B.C., since he knows the collection of Artemon (i.e., the letters to Philip) as well as other letters.

This longish summary was necessary if the following criticisms were to be intelligible.

1. It is true that Hermippus is often accepted as the common source of the lists in Diogenes and Hesychius. Moraux does not accept Hermippus but plumps for Ariston.⁹ But it need be neither one nor the other. Let us remember that Diogenes Laertius wrote about the beginning of the third century A.D., Hesychius in the fifth (when the actual compiler of the second list wrote is totally unknown). Diogenes at least does not (pace Düring, p. 79) say Hermippus is his main source, and does not mention him in connection with the list. These exercises in Quellensuche are very uncertain, even those that are generally accepted.
2. That both Diogenes and Hesychius mention twenty books, and that one of them mentioned other letters as well, does not really mean that their source knew only the twenty books. Even if Hesychius is generally believed to be more faithful to his source, in this case it might be Diogenes, especially as he is more detailed about the twenty books also. Yet Rist's reasoning leans strongly on his assumption that Hermippus (or Ariston) did not know Artemon's collection.
3. Rist speaks of the extra letters in Diogenes as "letters to Philip about the Selymbrians," but the text actually is [epistolai pros Phillippon Selumbrion epistolai pros Alexandrou epistolai d'] … By putting a stop after the first [epistolai] and inserting [peri] before [Selumbrion], Düring gets this: "Letters: Letters to Philip about the Selymbrians …," but this use of the first [epistolai] as a heading is unique in the list. A much more natural translation is: Letters to Philip; Letters of (or about) the Selymbrians; i.e. two different lots of letters. Then Rist's identification of Elias' letters [pros hera] (to one person) with the Letters to Philip loses all its force, and so does the identification of the extra letters with Artemon's collection as being to Philip only. The result is confusion, but Aristotle's correspondence seems to have been precisely in that state till Andronicus, except for Artemon's collection.
4. From the Arabic versions above, which mention both Artemon and Andronicus by name, Rist suggests that it looks at first sight as if these two men's collections were in the original library of Apellikon. As this is obviously impossible for Andronicus, "it is not at all unreasonable to doubt whether it contained a book in eight parts edited by Artemon" (p. 6). But even at first sight one surely sees that in the Arabic version, as quoted by Rist himself, what is said of the two men is not at all the same thing. It is categorically stated that there was in the library a collection of eight books of letters edited by Artemon (a large collection, the other version says) and then also twenty parts or books "which Andronicus found" (or read) and these were no doubt the material on which his own edition was based. It does not say or imply that Andronicus' name was connected with these letters before he worked on the library about the middle of the first century. This is the obvious meaning.

However, one might accept all Rist's evidence and most of his assumptions (except those mentioned in the last paragraph) and draw from them quite different conclusions, as follows. The Arab version specifically states that Artemon's collection was in Apellikon's library, and since we have no evidence to the contrary, the natural assumption is that it was there when he acquired it, that it therefore came from Neleus' library and ultimately from that of Theophrastus when he died. In that case the collection must have been made soon after Aristotle's death and before that of Theophrastus (i.e., between 322 and c. 285). This gets slight support from Elias' reference to Artemon "who was born after Aristotle," i.e., a younger contemporary, for it would be a singularly otiose thing to say if he lived a century or so later.¹⁰ Obviously no editor of Aristotle's letters could have lived before him! These were probably letters to individuals, some of them to Philip.… If this collection was unknown to Hermippus and/or Ariston, and generally at the end of the third century B.C., it can only be because this collection was among the books which Strabo tells us went to Neleus and were lost even to the Lyceum.

How does this situation affect Demetrius? He knows Artemon (223) as the man who edited Aristotle's letters … and it is a fair inference that this is the only collection he knows. The quotation from a letter to Antipater which follows immediately shows that the collection contained letters to others than Philip. Demetrius in fact shows an unusual familiarity with Aristotle's correspondence which most probably derived from Artemon's collection, of which he may have possessed a copy. It was, we know from the Arab version, a "large book." Now if, as we have shown, the collection disappeared into Neleus' library, it follows that Demetrius must have known it before Theophrastus' death. He clearly knew and had read Artemon whose views he also quotes on the right epistolary style generally (223-224). This does not mean that all the letters included in Artemon's collection were necessarily lost. He no doubt had copies made for his edition: the originals or other copies may have remained in the Lyceum, been collected in Alexandria, and may even have found their way also into Apellikon's library. The edition of Artemon and the twenty books or bundles no doubt overlapped—but the particular collection of Artemon with his name on it as editor was hidden away with the other Aristotelian manuscripts by the successors of Neleus till Apellikon came along and bought them two generations later.

All these assumptions being taken as fact, "reason cries out" (if I may borrow a phrase from both my colleagues) that Demetrius cannot have written his treatise "On Style" much after 270 B.C.

To prevent misunderstanding, however, let me hasten to make clear that there are far too many uncertainties in this last piece of argumentation of my own for me to attach much weight to it, but its degree of probability is at least as great as that of the conclusions of Professors Rist and Goold. And if I have stated as fact what is only assumption, this of course is one of the rigid rules of the genre.

Notes

¹ University of Toronto Press, 1961.

² See A Greek Critic, p. 22, n. 26; also G. A. Kennedy, The Art of Persuasion in Greece (Princeton 1963) 285-286 and H. Gärtner's review in Gymnasium 70 (1963) 61-63.

³ S. Sudhaus, Philodemi Volumina Rhetorica (Leipzig 1892) vol. 1, p. 198.…

⁴ In RhMus 54 (1899) 361, n. 1.

⁵ Syrianus In Hermogeni Commentaria, ed. H. Rabe (Leipzig 1892) vol. 1, p. 99, 18.

⁶ These are listed in the second appendix of A Greek Critic, pp. 159-163.

⁷ See Strabo 13.54 (608-609) and Plutarch Sulla 26. The story is discussed by P. Moraux, Les Listes anciennes des ouvrages d'Aristote (Louvain 1951), who views it somewhat sceptically, see especially pp. 1-6, 311-321.

⁸ Elias In Cat. Prooemium, Berlin edition of Commentaries on Aristotle, vol. 18, ed. Busse (1900) p. 113, 25-26.…

⁹ Moraux, 221-233 and 245. Cp. I. Düring, Aristotle in the ancient bibliographical Tradition (Göteborg 1957) 67-68.

¹⁰ See note 8.