SOURCE: "Demetrius the Stylist and Artemon the Compiler," The Phoenix, Vol. XVIII, No. 1, Spring, 1964, pp. 2-8.

[In the following essay, Rist attempts to determine the date of composition of On Style by investigating Artemon, an editor of Aristotle's Letters mentioned by Demetrius.]

The appearance of G.M.A. Grube's book¹ on Demetrius the Stylist has revived interest in the date of his work. Grube dates it at about 270 B.C. whereas G. P. Goold holds² that it was written in the Augustan Age. Such a discrepancy is disturbing; two hundred and fifty to three hundred years is a wide margin of error. This note therefore is intended to reduce the gap by an investigation of the Artemon who is described by Demetrius (223) as the editor of Aristotle's Letters. It seems that some progress may be possible here, although the matter has been quickly passed over by both Grube³ and Goold.⁴ More in fact can be discovered about the date of Artemon than either of these scholars has indicated. To attain such knowledge, it is necessary to examine the traditional accounts of the contents of the Aristotelian corpus.

There are three basic lists surviving of the Aristotelian writings. The first is that of Diogenes Laertius (3.22-27); the second is by an unknown hand and is often associated with Hesychius;⁵ the third, not preserved in Greek, derives from the list of a certain Ptolemy. Versions of this appear in the writings of two thirteenth-century Arabic scholars: al-Qifti and ibn abi Useibia.⁶ Of the general nature of these catalogues little need be said here. The information is readily available in the works of Moraux and Düring.⁷ Well-grounded opinion exists that the list of Ptolemy, which differs considerably from the others, is the direct representative of a list of the Aristotelian writings compiled by Andronicus of Rhodes in the first century B.C.⁸ Andronicus, according to Porphyry,⁹ arranged Aristotle's works according to subject-matter. The lists of Ps.-Hesychius and Diogenes Laertius, which are derived ultimately from a common source, differ greatly from that of Ptolemy. They are in places affected by the work of Andronicus, but their origin is very much older and is attributed by the vast majority of modern scholars to Hermippus, the pupil of Callimachus.¹⁰ Moraux, who dissents from this view, proposes Ariston,¹¹ the head of the Lyceum after the death of Lycon in 226 B.C. His arguments however have been successfully countered by Düring,¹² and it can be safely assumed that Hermippus is the source of the lists as we have them. For our present purposes, however, it is not important whether Ariston or Hermippus is the source, provided that one of them is. The two men were more or less contemporary. The list must have been composed in the last quarter of the third century B.C..

Let us now consider certain entries on the lists that we have. Almost at the end of Diogenes' version (5.27), we find a group of Aristotle's letters listed as follows: letters to Philip about Selymbrians;¹³ to Alexander, 4 books; to Antipater, 9 books; to Mentor, 1 book; to Ariston, 1 book; to Olympias, 1 book; to Hephaestion, I book; to Themistagoras, 1 book; to Philoxenus, 1 book; to Democritus, I book. We are not told how many books were written to Philip, but the total of the others is twenty books. Thus if we were to attempt to reconstruct the list of Hermippus from Diogenes only, we might suppose he knew of twenty books of letters and a number of extras.

Let us now turn to Ps.-Hesychius. His list contains a reference to letters in twenty books (p. 16 Rose). Thus if we compare Ps.-Hesychius and Diogenes, we find that they both know twenty books of letters, and that Diogenes has some knowledge of a further group, though he does not seem to know how many letters this group contained. What is the most likely deduction from this? Surely that the common source of Diogenes and Ps.-Hesychius knew the twenty-book collection and no more. Thus it was a twenty-book collection which was in circulation about 200 B.C.

We can now turn to our third list of Aristotle's writings—that preserved by Qifti and Useibia from the [pinakes] of Ptolemy and Andronicus. The version of the relevant part of the list that is given below is taken from the text of Qifti.¹⁴ Relevant alternative readings of Useibia are added in square brackets, additions in pointed brackets.

The books which were found in the library of a man called Apellikon. Namely the following books.

A book in which a man named Artemon collected a number of letters by Aristotle in eight parts (juz').

[a large book, a compendium containing a number of letters in eight parts.]

A book by him (Aristotle) on the ways of cities.… Two discourses (maq la).

And other letters which Andronicus found in twenty parts.

And writings in which there are memoranda such that men cannot discover their number.

And the earlier parts ("names," Useibia) of them are to be found in the fifth discourse of the book of Andronicus in the index of the books of Aristotle.

Here we find two collections of the letters of Aristotle: one by Artemon in eight parts, the other "found" by Andronicus in twenty. Even the numbers are significant. The twenty-part collection "found" by Andronicus can hardly be other than the twenty books known to Ps.-Hesychius, Diogenes, and their source, Hermippus. What then about the eight-part collection of Artemon? These could be the same as the unnumbered group which appears in Diogenes, namely the letters to Philip. Some confirmation of this may be found in the Aristotelian commentator Elias.¹⁵ Elias is describing those writings of Aristotle that deal with a specific concrete topic and are addressed [pros hena]. Such writings are called [merika] and letters are an example of them. By "letters" here we must assume that Elias means Artemon's collection, for he writes: "The letters which a certain Artemon, who lived after Aristotle, collected in eight books, … were written [pros hena]." [Pros hena] in these two passages could mean either "to one single person" or "to individuals." Grube, in a letter to me, takes the latter interpretation. If the former is possible, however, we could guess, by comparing the passage of Diogenes, that the single person was Philip. We should recall in addition that, as in Qifti, the number eight is associated with Artemon's collection.

Elias' passage is closely paralleled and probably much influenced by the work of Olympiodorus on the same theme.¹⁶ Olympiodorus brackets Andronicus and Artemon (in that order) together as compilers of Aristotle's letters. He too is analyzing Aristotle's work. He too marks off a class of writings as [merikon]. Such a class contains "everything written privately to certain individuals such as were the letters which Andronicus and Artemon compiled." There is an obvious difference between this text and that of Elias. Elias, speaking only of Artemon's collection, says that the letters were written [pros hena]; Olympiodorus, referring to the collections of both Artemon and Andronicus, says that they were written to certain individuals. … Is it not possible that knowing that Artemon's collection of letters were all written to one single individual, namely Philip, Elias gave up Olympiodorus' [pros tinas] and substituted [pros hena]? However it must be admitted that the identification of the eight-book collection with the letters to Philip is only a conjecture.

Since the source of the Arabic writers is ultimately Andronicus, we must assume that Andronicus knows of two collections: his own in twenty parts, and the eight-part volume of Artemon. Now Diogenes and Ps.-Hesychius are basically dependent on Hermippus, but Diogenes was aware not only of the twenty-book collection, but of other letters as well. We cannot tell where this further information comes from, but since it does not include the number of books, it can hardly come directly from Andronicus' [pinakes]. Now Ps.-Hesychius does not use any additional Andronican material, but rather sticks to the original list of Hermippus. We must repeat that it is the twenty books of Hesychius that Diogenes knows in more detail and that this fact suggests that the common source spoke of twenty books and twenty only.

So far then we can say that Artemon's collection of letters was not known to Hermippus (or for that matter to Ariston) in the late third century B.C. At that time the catalogue of Aristotelian works contained no mention of Artemon. May we not conclude that in 225 B.C. Artemon's collection had not been made?

It is now time to look at the Arab list derived from Ptolemy in more detail. As we have seen, a number of works are there listed as coming from the library of Apellikon. There is, fortunately, a good deal of evidence about the nature of this collection. It seems that on the death of Theophrastus (288 B.C.) his books, together with the writings of Aristotle himself, which had passed into his possession, were left to Neleus of Skepsis.¹⁷ Neleus may have sold some of these books to Ptolemy Philadelphus,¹⁸ but the rest were handed on to his descendants, who took little care of them and later hid them in a cave to prevent their falling into the hands of the bibliophile Attalids of Pergamum. In the late second century B.C. they were sold, in bad condition, to Apellikon of Teos, who had them copied and published, though still inadequately restored and full of errors. Apellikon's library was at Athens, and its owner was raised to high estate in that city by Aristion, the general of Mithradates VI of Pontus. On the capture of Athens by Sulla in 86 B.C., Apellikon's library was seized and taken to Rome. Here Tyrannio worked on it. Later Andronicus acquired copies of the books from Tyrannio, which he published along with the [pinakes] of Aristotle's writings which were circulating in the time of Plutarch.¹⁹

Now we know from Ptolemy through Qifti and Useibia that there is some connection between Artemon and Andronicus, and the library of Apellikon. The two collections of letters are given as among the contents of the library. Have we therefore any means of dating Artemon more precisely? We have already concluded that he is unknown ca. 225 B.C., the time of Hermippus. Can we bring him down later still? The entries in Qifti's list relevant to Artemon and Andronicus are, we recall: (a) a book in which a man called Artemon collected a number of letters by Aristotle in eight parts; and (b) and other letters which Andronicus found in twenty parts. At first sight we might suppose that the names of Artemon and Andronicus were to be found with their respective groups of letters in Apellikon's library when he first acquired it. We cannot say at the moment whether this was true of Artemon; we can be certain, however, that it was not true of Andronicus, for Andronicus (as we have observed above) had no connection with Apellikon's library until the middle of the first century B.C.,²⁰ that is, until long after Apellikon was dead and his books had passed into the hands of Tyrannio. Consequently since, despite what seems an obvious interpretation of Qifti, Apellikon's library could not have originally contained a volume of Aristotle's letters edited by Andronicus, it is not at all unreasonable to doubt whether it contained a book in eight parts edited by Artemon. A possible, and, as we shall see, likely explanation of the text of Qifti is that when Andronicus published the Aristotelian letters (as part of his new edition of the Aristotelian corpus) he associated the name of Artemon, which he had found attached to an eight-book collection of letters, with a part of the work. This would mean that people who knew of Artemon's collection of letters derived their knowledge from Andronicus' great edition of Aristotle and from the [pinakes] he compiled to go with it. Artemon would, on this reasoning, be a man who lived sometime between the original acquiring of Neleus' library by Apellikon and its eventual publication by Andronicus.

There is a minor problem which should be cleared up before we proceed further. If the letters in twenty books of Diogenes and Ps.-Hesychius are the same as those in the edition of Andronicus, why do the texts of Qifti and Useibia say that Andronicus found these letters and imply that knowledge of them derives from him alone? It would seem rather that they were already available to Hermippus—a fact which Andronicus would have known. The explanation of this is probably quite simple and was suggested to me by Rose's attempt to restore the original Greek of Ptolemy. Rose's Greek version runs: [epistolai allai ais enetukhen Anbronikos]. The Arabic word which he renders [enetuchen] is wajada, he ordinary word for "find." If [enetukhen] therefore was the reading of Ptolemy, we should expect to find wajada as an Arabic equivalent if the original Syriac translator took [enetukhen] in its normal sense of "found."²¹ Yet there is a less comrnmon but well-attested extended meaning for [enetukhen], which could easily have eluded the translator. This is simply "read."²² By keeping this in mind, we can free ourselves of difficulties. Andronicus recorded on his [pinax] that he read a twenty-part collection of letters in Apellikon's library. These letters, however, were not those collected by Artemon, which Andronicus had already listed, but the group known to Diogenes, to Ps.-Hesychius, and to their source Hermippus. This twenty-part collection was then not only known by Hermippus but was familiar to one of the original owners of Apellikon's library, that is either to Neleus or to Theophrastus himself. Thus Koskenniemi's view²³ that there was a collection of Aristotelian letters current in the time of Theophrastus may well be correct. This collection, however, was not the compilation of Artemon, but the well known twenty-book collection of Diogenes, Ps.-Hesychius, and Andronicus.

Although it has already been demonstrated that Artemon, on the evidence of Ptolemy, need not be earlier than Apellikon (since Andronicus is certainly later) there still remains the possibility that he might have been earlier. Could Apellikon have procured for his library a volume of Aristotelian letters already labelled with Artemon's name? The answer to this question must be No. First, Artemon's collection is apparently unknown to Hermippus. This is of very considerable importance, since Hermippus made use of the library at Alexandria. (If the source of Diogenes is Ariston, the position is the same. Then it would be the head of Aristotle's own Lyceum who is unaware of Artemon's work.) Secondly, if there was a collection made by Artemon, when in fact could it have been compiled? It seems that before the publication of Apellikon's library this collection was unknown. Therefore if Artemon lived before Apellikon and later than Neleus, he must have worked with certain letters which were uncopied and eventually passed into Apellikon's hands.

We know the history of Apellikon's library. It was in the hands of the descendants of Neleus, who neglected the books. Such careless individuals could hardly have included a compiler of letters. Earlier owners were Neleus himself and Theophrastus. If either of these had known Artemon, their knowledge must have been quite common knowledge since they were both members of the Peripatetic school. And had the knowledge been public, we should hardly know about these particular letters merely through Apellikon's library. We are driven to suppose that Artemon himself lived long after Neleus and Theophrastus, and that his collection of letters is either a new find, unknown even to the earliest followers of Aristotle, or a later forgery. In any case it was added to the original collection some time between the acquiring of Theophrastus' books by Apel-likon and their publication by Andronicus. If the library of Neleus, say, contained both the eight- and the twenty-part collection, how could it come about both that these collections were separate in the library, and that one was widely known while the other was known to no one? Reason cries out that before the time of Apellikon no one knew anything about either Artemon or his collection.

We must therefore conclude that Artemon cannot be dated before the late second century B.C. at the earliest, and that Demetrius the Stylist, who refers to him, must certainly be later than ca. 100 B.C. and probably later than the publication of Andronicus' Aristotle at Rome in the years between 40 and 20 B.C.²⁴ We should observe that Demetrius not only refers to Artemon but also quotes letters other than those in Artemon's collection of letters to Philip. In section 225, for example, he refers to a letter to Antipater. Thus Demetrius probably knew both the eight- and the twenty-book collections. It is likely that he read them in that place where they were to be found side by side, namely the edition of Andronicus.

So far the discussion has been limited to Artemon, the compiler of Aristotle's letters. This is as far as we can go with certainty. Nevertheless it seems quite likely that the view which identifies our Artemon with Artemon of Cassandreia (probably late second century B.C.) is correct. Our Artemon clearly worked on the books from Apellikon's library. Such a man may well have been a bibliophile. Artemon of Cassandreia wrote two works which look like the products of a bibliophile.²⁵ …

Notes

¹ G. M. A. Grube, A Greek Critic: Demetrius on Style (Toronto 1961).

² G. P. Goold, "A Greek Professorial Circle at Rome," TAPA 92 (1961) 168-192.

³ Grube (note 1), who on p. 111 writes that nothing is known of the Artemon who edited Aristotle's letters, mentions on p. 42 the suggestion of H. Koskenniemi, "Studien zur Idee und Phraseologie des griechischen Briefes," Annales Academiae Scientiarum Fennicae B.102 (Helsinki 1956), that the Artemon mentioned by Demetrius may have been a contemporary of Theophrastus.

⁴ Goold (note 2) 181.

⁵ For bibliography on the authorship of this anonymous catalogue (which I shall refer to as "Pseudo-Hesychius") see P. Moraux, Les Listes Anciennes des Ouvrages d' Aristote (Louvain 1951) 195. The text is printed in V. Rose's Aristotelis Fragmenta (Leipzig 1886) 9-18 and (with discussion) in I. Düring's Aristotle in the Biographical Tradition (Göteborg 1957) 82-93.

⁶ Rose has given us a "version" of the Greek of Ptolemy, compiled by an amalgamation of the evidence of the two Arabic sources (pp. 19-22). He is not very accurate. For the whole problem see Düring (note 5) 208-246 and Moraux (note 5) 288-309. On the Arabic evidence (or lack of evidence) about Ptolemy himself, see During 208. I am grateful to Professor G. M. Wickens of the Department of Islamic Studies of the University of Toronto for examining the texts of Qifti and Useibia.

⁷ See note 5.

⁸ Moraux (see note 5) 306-309. For the date of Andronicus' arrival in Rome (50-40 B.C.), see Düring (note 5) 421.

⁹Vita Plotini 24. Cf. Düring (note 5) 414-416.

¹⁰ Cf. Moraux (note 5) 215.

¹¹ Moraux, ibid. 237-247.

¹². During, "Ariston or Hermippus?" Classica et Mediaevalia 17 (1956) 11-21.

¹³ For the correctness of this emended text see During (note 5) 50 and (note 11) 16.

¹⁴ Cf. During (note 5) 230. Professor Wickens has very kindly translated the passages of Qifti and Useibia.

¹⁵ Elias, In Cat. 113. 25-26 Busse.

¹⁶ Olympiodorus, Prol. et in Cat. 6.12 Busse.

¹⁷ D. L. 5.52.

¹⁸ The evidence for the fate of the books derives basically from Strabo 13, 608-609, whose account, with some variations and reductions, reappears in Plutarch, Sulla 26 and Athenaeus 3AB, 214A-F. We should recall, as testimony to Strabo's accuracy, that he was himself a pupil of Tyrannio (12.548). This should be borne in mind when considering the justice or injustice of Moraux's suspicions about the story (note 5, 312-313). For what is probably the best account of the affair, see R. Shute, On the History of the Process by which the Aristotelian Writings Arrived at their present Form (Oxford 1888) 29-34.

¹⁹ Plutarch, Sulla 26.

²⁰ For the dating of Andronicus' work on Aristotle (40-20 B.C.) see Düring (note 5) 421.

²¹ That Ptolemy's writings, like many other Greek philosophical works, passed through a Syriac stage before reaching Arabic, is almost certain. Cf. Moraux (note 5) 291, Düring (note 5) 184-185.

²² Cf. Plutarch, Sulla 26; Plato, Lysis 214 AB, Symp. 177B; Strabo 1.1 etc.…

²³ H. Koskenniemi (note 3).

²⁴ See Düring (note 5) 420-421.

²⁵ I would particularly like to thank Professor G. M. A. Grube for his detailed comments on the original draft of this paper. Our views of course differ widely but without his help this paper would be very much more unpolished than it is.