SOURCES FOR ATHENODORUS

From the Neoplatonic commentators on the Categories, we know that Athenodorus was a ‘Stoic’ (Porph. in Cat. 86.22; cf. Dexipp. [but only in chapter heading] 1.9) who interpreted the Categories as an attempt to classify and describe ‘verbal expressions qua verbal expressions’ (λέξɛις καθὸ λέξɛις). He criticized the treatise as insufficiently comprehensive about this subject, since it has nothing to say about parts of speech such as conjunctions and articles (Porph. in Cat. 59.5–14; cf. 88.22), or about simple sentences (such as ‘Dion walks’), or about various other kinds of words (retailed by Dexipp. in Cat. 12.3–11; cf. 32.18–29). His criticisms were embodied in a book, according to Simplicius (who at in Cat. 62.26 references Ἀθηνόδωρος ἐν τῷ Πρὸς τὰς Ἀριστοτέλους μὲν Κατηγορίας ἐπιγɛγραμμένῳ βιβλίῳ). Athenodorus must have expressed these views before another Stoic, L. Annaeus Cornutus, composed a rebuttal or expansion entitled Πρὸς Ἀθηνόδωρον καὶ Ἀριστοτέλην (Simpl. in Cat. 62.26). As Cornutus' identity is reasonably well established and his lifetime is firmly attached to the reign of Nero,Footnote ⁷ we can date his book around the second half of the first century ce, and date Athenodorus' critical activity on the Categories sometime before this.

The tradition offers several historical candidates for identification with the Athenodorus described by Porphyry and Simplicius (who name him) and Dexippus (who does not name him, but reports views similar to the Athenodoran positions in Porphyry and Simplicius). We know about multiple Athenodori of the Stoic school, all of whom appear to have lived prior to L. Annaeus Cornutus, and at least two of whom hailed from the city of Tarsus. One of the two best-documented candidates is the Athenodorus called ‘Cordylion’, a contemporary of Marcus Cato who became librarian at Pergamon and was accused of bowdlerizing Zeno's Republic (Diog. Laert. 7.34); and the other is Athenodorus ‘Calvus’, identified as the son of ‘Sandon’ and a tutor of Augustus and correspondent of Cicero. Strabo (14.5.14) informs us that these two Athenodori are best distinguished by their connections with Cato and Augustus, respectively; thus Cordylion lived before the suicide of Cato in 46 bce, while the son of Sandon died in Tarsus at the age of eighty-two, after advising Augustus well into his reign as emperor (cf. [Lucian], Macr. 21).

Of the son of Sandon, we know a good deal historically because of his association with the young Octavian and the mature Augustus.Footnote ⁸ In addition to his candidacy as the author of a book criticizing the Categories, Athenodorus Calvus is credited with a ‘history of his fatherland’; with a work dedicated to Octavia Minor; with a work πɛρὶ σπουδῆς και παιδɛίας; and with a treatment of at least eight books on ‘Peripatetics’ (e.g. Diog. Laert. 9.42.4, although it is also plausible that this is Cordylion's). As the later sources were primarily interested in Calvus' association with Augustus, many anecdotes about his mentorship have come down to us, several of which may carry hints about his ethical doctrine.Footnote ⁹

Of the Athenodorus called Cordylion we know markedly less. Strabo mentions him as a Tarsian Stoic alongside Antipater, Archedemus, Nestor and Athenodorus Calvus (Strabo 14.5.14, a list which bears comparison to the roster of Stoic logicians at Diog. Laert. 7.68).Footnote ¹⁰ During his tenure as librarian at Pergamon, according to the rhetorician ‘Isidorus of Pergamon’, Cordylion attempted to excise passages from the Republic of Zeno: Isidorus presents the consequences in a rather unflattering light (ϕωραθέντος τοῦ Ἀθηνοδώρου καὶ κινδυνɛύσαντος, Diog. Laert. 7.34).Footnote ¹¹ By 67 bce Cordylion had won renown as a Stoic and had also developed a reputation for refusing the patronage of kings and politicians (Plut. Cat. Min. 10.1–3). But a determined M. Porcius Cato succeeded in winning him over, and subsequently Athenodorus went to live with Cato in Rome (ibid. 16.1; Strabo 14.5.14).

Richard Goulet in his Dictionnaire exercises caution in assigning ambiguous fragments to any one Athenodorus, noting the wide and poorly distinguished field of candidates. But there is some consensus among modern scholars, including Goulet, Moraux and Hijmans, that the Athenodorus who commented on the Categories is the son of Sandon – chiefly because of Calvus' late dating, taken in association with the date traditionally assigned to the publication of Andronicus of Rhodes' publication of Aristotle (see e.g. Moraux [n. 1], 1.150). Hijmans presents a list of arguments in favour of this identification, and is followed in the main by Goulet. Hijmans narrows the candidates to the two Tarsians: the alternative candidates, such as the ‘Athenodorus of Rhodes’ mentioned by Quintilian (Inst. 2.17.15), are said to be too obscure, and the earlier student of Zeno and brother of Aratus, ‘Athenodorus of Soli’, is hardly mentioned in the literature.Footnote ¹² Second, the Athenodorus cited by Diogenes Laertius (7.68) as a logician interested in propositions could well be identified with the son of Sandon (Hijmans [n. 5], 105, citing Zeller's History, 607). Third, it appears probable that Athenodorus Cordylion died before Andronicus' edition of the Categories was published.

But the strength of these arguments in favour of the candidacy of Calvus is questionable. The dating of Andronicus' edition might not be relevant to the problem at hand. Given valid doubts regarding the true extent of Andronicus' influence and the muddy waters surrounding that edition's date of publication, it is problematic to use this event as a definite terminus post quem for critical activity on the Categories. Certainly the candidacy of the earlier Stoic Cordylion and of other ‘pre-Andronican’ candidates should not be rejected solely on these grounds.Footnote ¹³

DIOGENES LAERTIUS AND DEXIPPUS

I believe it would be more fruitful to focus attention upon the ‘Athenodorus’ mentioned by Diogenes Laertius as a pioneering Stoic logician at 7.68, and to seek reasons internal to the text to identify this Athenodorus either with Calvus or with any of the other Athenodori in our record. At 7.68, Diogenes' ‘Athenodorus’ is said to have concerned himself with the distinction of simple propositions from complex ones, and to have included ‘assertoric’ or categorical propositions within the rubric of ‘simple’ propositions (7.68–70, ed. Long):

Diogenes Laertius' Athenodorus, then, was a particularly memorable or influential exponent of the distinction between simple and non-simple propositions, a basic division articulated and defended by Stoic logicians. Athenodorus maintains that a categorical or ‘assertoric’ (katêgorikon) sentence such as ‘Dion walks’ represents a simple proposition, whereas a sentence composed from several such propositions – such as the protasis and apodosis of a conditional joined by ‘if’ – represents a complex proposition.

This discussion of propositions falls within Diogenes' general exposition of Stoic dialectic, beginning at 7.63. Dialectic is a branch of knowledge dealing with sayables (lekta), those items which we ‘mean’ or ‘point out’ (sêmainein) by our verbal expressions, and specifically with those sayables that may be true or false. One subdivision of the subject matter of dialectic is propositions (axiômata). Propositions are complete (autotelê) in the sense that they contain both a subject and a predicate, forming a complete sentence: ‘writes’ (graphei) signifies an incomplete lekton because it begs the question ‘Who?’ (7.63.8), whereas the sentence ‘Socrates writes’ signifies a complete lekton that begs no questions. Thus a lone predicate (katêgorêma) such as writes, which could be ‘said of’ something (kata tinos agoreuomenon), when taken by itself is ‘incomplete’ or ‘defective’ (ellipes, 7.64). On the orthodox Stoic position outlined by Diogenes, then, an example of the simplest kind of verbal expression that has meaning and is complete is the sentence ‘Dion walks’, a subject and a predicate. Such propositions are the elementary building blocks of Stoic logic, the simplest units that are not incomplete (ellipes).

Aristotelian logicians after the first century bce took another view, namely, that the elementary building blocks of logic are terms and not propositions. (This position is briefly sketched below.) In adopting this view, Peripatetics were influenced by the Categories and perhaps by Andronicus' ordering of the Organon to reflect a progression of increasing complexity from terms (Categories) to propositions (Int.) to syllogisms (An. pr.) to demonstration (An. post.). According to the position that became Peripatetic orthodoxy, the Categories was concerned to explain and classify ‘meaningful, simple verbal expressions’ (it is πɛρὶ ϕωνῶν σημαντικῶν ἁπλῶν, e.g. Porph. in Cat. 58.5), where these are understood as terms – in the language of Categories 2 (1a16–19), as things ‘said without combination’ (tr. Ackrill):

Suppose a Stoic reader, who read this passage of the Categories in light of the views maintained by Stoic logicians such as Athenodorus in Diogenes Laertius. Such a reader might object that, if the Categories is really trying to deal with ‘meaningful, simple verbal expressions’ (as the Peripatetics maintain), it fails, because its classification scheme only accounts for a fraction of such expressions, namely a subset of certain incomplete or defective (ellipes) verbal expressions. It omits simple propositions such as ‘man runs’ or ‘Dion walks’, the simplest complete units of logic. It also omits many parts of complex propositions, such as conjunctions.

That is, in fact, roughly the line adopted by the critics of the Categories described as ‘Stoics’ in the chapter headings of Dexippus' commentary on the Categories, and the line taken by ‘Athenodorus’ in Porphyry and Simplicius. These critics appear to have complained that the Categories omitted sentences such as ‘Dion walks’, as well as various parts of speech such as conjunctions and articles. To this line of attack, the Peripatetics (as reflected in the later Neoplatonic commentaries) replied that a sentence such as ‘Dion walks’ is in fact a complex expression constituted of terms; the Categories deals only with simple, meaningful expressions, and therefore rightly omits the sentences that the Stoics wrongly regard as ‘simple’ (because they are complex) as well as various parts of speech that are not ‘meaningful’ (tr. Dillon, slightly modified).

As Dillon points out, the Neoplatonists' ‘Athenodorus’ must be one of the critics meant here, to whom Dexippus is replying; likewise he is named at Porphyry, in Cat. 59.5–14 and in Simplicius as levelling similar criticisms, albeit more briefly.Footnote ¹⁷

This passage, which embodies a tradition of Stoic and Peripatetic debate about whether the Categories is complete or deficient in classifying its proper subject matter, also reflects the well-known controversy concerning whether propositions or terms are the basic building blocks of logic. As discussed above, the basic units of the Categories and later Peripatetic logic are simple terms, whereas the basic units of Stoic logic are simple propositions. That distinction appears to have been played upon for polemical ends by both critics and defenders of the Categories, as Dexippus testifies. It may be extremely suggestive of the motives that led Stoic logicians such as Athenodorus and Cornutus to focus a series of works on the problems raised by the Categories. J. Barnes points out, for example, that the later Peripatetic Alexander of Aphrodisias' essay on negation (in A. Pr. 402.1–405.16) focusses on the simplicity or composite status of sentences such as Καλλίας πɛριπατɛῖ.Footnote ¹⁸ For Alexander, a negation sign will modify the verb πɛριπατɛῖ alone; but for his opponents (apparently Stoic logicians), a negation sign modifies the entire propositional sentence, which should itself be viewed as simple. It is generally assumed that Alexander's opponents here are ‘old Stoics’, but Barnes makes a case that ‘they may … have been later Stoics – Athenodorus and Cornutus, say – who determined to defend the old Stoic doctrine against its resuscitated Peripatetic rival’.Footnote ¹⁹ I think we may support and develop this identification of Alexander's opponents as plausible, at least in the case of Athenodorus, based on the texts considered above.

For our purposes here, the connection to Alexander's opponents is relevant because it again links the professional interests of ‘Athenodorus’ at Diogenes Laertius 7.68 with the criticism broadly ascribed to ‘Athenodorus’ by Porphyry, Simplicius and (implicitly) by Dexippus: namely, that the division or classification scheme of the Categories is incomplete (e.g. ɛὑρίσκοντɛς ἐλλιπῆ ϕασιν ɛἶναι τὴν διαίρɛσιν, Porph. in Cat. 59.13–14) because the text does not deal with many kinds of expression which it ought to include under the heading of simple, meaningful verbal expressions. There are two key samples of its failure: (a) sentences such as ‘Dion walks’, and (b) a raft of other kinds of words such as conjunctions (mentioned by Porphyry) and adjectives, coined vocabulary and so on (mentioned by Dexippus). In reply, the Peripatetics grant that (a) sentences are meaningful, but not simple; (b) conjunctions and other parts of speech listed by Athenodorus may be simple, but they are not (in and of themselves) meaningful. Thus the Peripatetics respond that only the items classified by the Categories are both simple and meaningful, and that the Categories succeeds in classifying these completely. This debate provides at least a textually and philosophically grounded association between the Athenodorus who commented on the Categories and the logician of Diogenes Laertius 7.68.

Drawing the connection between Diogenes' Athenodorus and the Neoplatonists' critic of the Categories may suggest another perspective on Athenodorus' motives in criticizing the Categories. The word used by the Stoics to characterize incomplete lekta such as a predicate (katêgorêma) is ‘deficient’ or ‘incomplete’ (ellipes, e.g. Diog. Laert. 7.64). The Neoplatonists also suggest that Athenodorus and other Stoic critics used the word ellipes to denote the ‘deficiency’ or ‘incompleteness’ of the classification scheme in the Categories. It seems possible that Athenodorus regarded the λɛγόμɛνα κατὰ συμπλοκήν of the Categories as vocalizations that signified incomplete, ellipê lekta, and that it had nothing to say about vocalizations which signified complete or autotelê lekta.

If Athenodorus encountered the Categories prepended to the Topics – that is, as a propaedeutic to the study of ‘dialectic’ – and not yet in its Andronican position at the head of the curriculum, he might have regarded the treatise as a poor introduction to dialectic, that branch of knowledge which, on the Stoic view, deals with complete lekta that admit of truth and falsehood. In response to such a complaint, a Peripatetic defender of the text's usefulness as an introduction to logic, such as Andronicus of Rhodes, could have replied that logic should begin with terms, that only terms are properly ‘simple’ as well as ‘meaningful’, and that the Categories is a natural starting point for the progression from terms to propositions and syllogisms. Detaching the Categories from the Topics, entitling it Categories, and pulling it forward to the outset of the logical curriculum in a significant publication of the Aristotelian corpus, would serve to emphasize that point.

IDENTIFICATION AND IMPORTANCE

So far, I hope to have suggested some motives for linking our Athenodorus to the figure described in Diogenes Laertius 7.68. Does that passage offer any useful implications about his historical identity? Notably Diogenes Laertius' logician is primarily associated with Chrysippus at 7.68, and secondly with a number of second-century bce philosophers of the Stoic school. In itself, this grouping might cast doubt on the later, first-century date required by Athenodorus Calvus, who does not fit the temporal associations of the group at 7.68. Indeed, the philosophers mentioned by Diogenes Laertius at 7.68 are presented as doing important early work in Stoic logical doctrine. They are, in order, Chrysippus himself (third century bce), Archedemus (second century bce), Athenodorus (unknown), Antipater (second century bce) and Crinis (probably second century bce, and later than Archedemus: cf. Epictetus, Disc. 3.2.15); compare also the list of notable Tarsian Stoics at Strabo 14.5.14. As far as we can tell, this group follows a rough chronological order. If our ‘Athenodorus’ is the son of Sandon, then he would appear in this list out of order, that is, much earlier than we should expect if he lived in the first century bce. While again not decisive, this might weigh against the traditional identification. Indeed, Athenodorus Cordylion (cf. 7.34) would be a better fit for Diogenes Laertius 7.68 in terms of chronology.

There may be an incentive to entertain an even earlier date for the Athenodorus named at Diogenes Laertius 7.68. At 7.190 Diogenes provides a detailed list of the works of Chrysippus, where one book is in fact entitled Πɛρὶ τῶν καταγορɛυτικῶν πρὸς Ἀθηνόδορον. (Katagoreutika, as we know from Diog. Laert. 7.70, are sentences which consist of a demonstrative in the nominative case and a predicate.) The Athenodorus to whom Chrysippus dedicated this book could plausibly be identified with the Athenodorus of 7.68, who also worked on categorical propositions, and is mentioned in association with Chrysippus in this context. Such an association could date our logician to the fourth or third century bce. In fact, there is a recorded match for this chronology: Athenodorus of Soli (mentioned earlier at Diog. Laert. 7.38 = SVF 1.38), who lived slightly earlier than Chrysippus and joined his brother Aratus as a pupil of Zeno (as recounted in the Anonymous Life of Aratus). An interesting argument could be made, if not for Athenodorus of Soli in particular, then at least for a contemporary of Zeno and Chrysippus. It is tempting to speculate that the Athenodorus of 7.68, himself associated with Chrysippus, was both the Athenodorus to whom Chrysippus dedicated a work Πɛρὶ τῶν καταγορɛυτικῶν (7.190) and the Athenodorus depicted by Porphyry, Dexippus and Simplicius as criticizing the Categories for failing to take into account certain ‘simple’ categorical propositions. Such an identification might be challenged on the grounds that Simplicius mentions ‘Athenodorus and Cornutus’ in the same breath as if they were near contemporaries; but Simplicius elsewhere mentions ‘Xenocrates and Andronicus’ in the same way (63.22), although they, like Athenodorus of Soli and L. Annaeus Cornutus, are separated by some three centuries.

It might be reasonable to situate the earliest Stoic criticism of the Categories, not within the context of the first century bce as a response to Andronicus' publication, but rather in a prior environment of Stoic studies of logic and dialectic in the circle of Chrysippus. The Categories itself was dismissed as a somewhat crude treatment, as Dexippus explains, because it omitted many parts of grammar, including the simple categorical proposition (such as ‘Dion walks’). The Peripatetic reply that it is terms that are simple, but propositions that are complex, may have motivated the first-century bce catalogue or curricular order traceable to Andronicus,Footnote ²⁰ which locates the Categories (as dealing with simple terms) before the De interpretatione (as dealing with propositions), the De interpretatione before the Prior Analytics (as dealing with syllogisms) and so on, progressing from the simpler to the more complex. Thus, as I have suggested here, we might regard Andronicus as replying to such early Stoic criticisms by stressing the simplicity of terms, anticipating the response that became Peripatetic orthodoxy.

Again, we might even speculate that Andronicus relocated the Categories to the outset of the logical curriculum to emphasize this principle and this progression. In fact, the opening sentence of Andronicus' paraphrase of the Categories, which we can reconstitute as τῶν λɛγομένων τὸ μὲν ἄνɛυ συμπλοκῆς λέγɛται …,Footnote ²¹ suggests that he wished to highlight just this subject matter from the start. Prior to Andronicus, as I have mentioned above, there is no evidence that the Categories appeared at the start of any catalogue of Aristotle's work that we know about (in fact, it does not appear at all in the Hellenistic catalogues reported at Diog. Laert. 5.22–7 and 42–50, except as a later interpolation); Andronicus seems to have found this text with the title Before the Topics (cf. Simpl. in Cat. 379.9), perhaps located immediately before the Topics in whatever manuscripts he worked from, and presented as a kind of propaedeutic to that larger work. And it may have been in this context that early Stoics read and criticized the short treatise, whatever title they knew it under. If Andronicus, like later Peripatetics, read the Categories as a basic treatment of simple, meaningful terms – πɛρὶ ϕωνῶν σημαντικῶν ἁπλῶν καθὸ σημαντικαί ɛἰσι τῶν πραγμάτων, according to the formula that became traditional (Porph. in Cat. 58.5–6) – then his framing of the Categories through his paraphrase and catalogue and his relocation of this treatise to the head of the logical curriculum may have formed part of his response to Stoic criticism of the work for its omission of simple sentences and of non-significant or co-significant parts of speech. Andronicus' relocation of the Categories may reflect a wish to rehabilitate the foundations of a truly distinct Peripatetic logic, not merely a crude forerunner of Chrysippus' ‘dialectic of the gods’.

I hope to have expressed some reasons to consider the Athenodorus mentioned by Diogenes Laertius 7.68 as a proponent of views about logic that are directly relevant to the criticisms of the Categories ascribed to ‘Athenodorus’ by Porphyry and Simplicius. The concerns of Diogenes' Athenodorus might, then, help to explain and motivate the recorded Stoic criticism of the Categories, and point towards a motivation for the Andronican ordering of the Aristotelian Organon.