The Library of Pergamon as a Classical Model

Gregory Nagy
[[Originally published as Nagy, G. 1998. “The Library of Pergamon as a Classical Model.” In Pergamon: Citadel of the Gods (ed. H. Koester) 185-232. Harvard Theological Studies 46. This online edition (2011) contains sight modifications, which do not affect the content. The original page-breaks are indicated within brackets containing the original page-numbers: for example, “the Library {815|186} at the Mouseion” indicates the break between p. 185 and p. 186]]
This inquiry concerns the Library of Pergamon, not so much as a place or institution, but as an idea or concept, a classical model. This model, I hope to show, is a historical reality in its own right. I start by looking at the foundational principles that led to the conception of this library. Then I focus on Crates of Mallos, head of the Library at Pergamon in the middle of the second century BCE. My main question concerns the Library’s role or nonrole in the textual transmission of Aristotle and Homer.


In using the expression classical model, I have in mind the prestige of the classics. By classics here I refer not to any current general definition but to ideas that took shape in the specific historical contexts of centers of learning that flourished in the Greek-speaking world from the fourth through the second centuries BCE. The primary points of reference are:
  1. the Lyceum or Peripatos in Athens, as shaped by Aristotle and his successor, Theophrastus, in the fourth century;
  2. the Library {185|186} at the Mouseion or Museum (‘the sacred precinct of the Muses’) in Alexandria, supported by the dynasty of the Ptolemies (Lagidai) in the third and the second centuries BCE; and
  3. the Library in Pergamon, supported by the dynasty of the Attalids (Attalidai), especially during the reign of Eumenes II (197-158 BCE).
For historical background on the concept of the classics around the fourth through the second centuries BCE, I rely primarily on Rudolf Pfeiffer’s History of classical Scholarship. [1] He takes note of a key word for this concept: krisis, in the sense of ‘separating’, ‘discriminating’, ‘judging’ (verb krinein) those works and authors deemed worthy of special recognition, and those not. [2] Those that were ‘selected’ in this process were the enkrithentes, a term that corresponds to the later Roman concept of the classics (the classici), who were authors of the ‘first class’ (primae classis ). [3] This classical principle of selectivity, where some things have to be excluded in order for other things to be included, is the basis for the modern usage of the word canon. [4] The Greek word for those who were engaged in the process of making these critical selections was kritikoi ‘critics’.
A prime example of the kritikoi is Philitas of Cos, a prototype of the poet-scholars of the Library of Alexandria: Strabo describes him as ποιητὴς ἅμα καὶ κριτικός ‘a poet and at the same time a kritikos’. [5]
This term kritikoi was superseded by the alternative terms grammatikoi and philologoi as the choice self-designation of the scholars {186|187} of the Library of Alexandria. [6] By contrast, the scholars of the Library of Pergamon, most notably Crates of Mallos, preferred the earlier designation kritikoi. [7] According to Crates, the kritikos had to be master of the entire ‘science’ (epistēmē) of logikē, while the grammatikos was confined to explaining questions of vocabulary and prosody. [8]
This formulation of Crates serves here to distinguish the Library of Pergamon as a classical model, in direct comparison with the alternative classical model of the Library of Alexandria. Crates’ generalized claim that the kritikoi of Pergamon stand for a more holistic approach to scholarship than the grammatikoi of the Library of Alexandria is connected, I will argue, to his specific claims about Pergamene approaches to the study and even the reception of Homer. It is also connected to the rivalry between Pergamon and Alexandria over the intellectual legacy of Aristotle’s Peripatos. To begin, I quote a key formulation, to which I will return for a reexamination at the very end of my inquiry:
οἱ δὲ αὐτὸ τοῦτο τὴν διάνοιαν ἐξηγούμενοι, οὐ μόνον Ἀρίσταρχος καὶ Κράτης καὶ ἕτεροι πλείους τῶν ὕστερον γραμματικῶν κληθέντων, πρότερον δὲ κριτικῶν. καὶ δὴ καὶ αὐτὸς Ἀριστοτέλης, ἀφ' οὗ φασι τὴν κριτικήν τε καὶ γραμματικὴν ἀρχὴν λαβεῖν,
And those who make exēgēsis of the meaning (dianoia ) of Homer - not only Aristarchus and Crates and several others of those who were later called grammatikoi but who had earlier been called {187|188} kritikoi, and especially Aristotle himself, from whom they say that kritikē and grammatikē have their origin. [9]

The Central Thesis, with Two Clarifications

Before proceeding, it is essential to clarify two points in order to avoid major misunderstandings. The first has to do with Crates’ actual formulation of the distinction between the kritikoi of Pergamon and the grammatikoi of Alexandria. I wish to emphasize that his claim to a holistic approach is not at odds with a parallel claim of adherence to a principle of selectivity, which is essential to establishing a canon. This principle of selectivity is actually built into Crates’ appropriation of the term kritikos. The ideal of establishing a canon, as the designation kritikos conveys it, was the common property of Pergamene and Alexandrian scholars alike. Moreover, not only the Pergamenes but also the Alexandrians promoted models that combined claims to selectivity with claims to a holistic approach. Such a combination is the essence of what I call the classical model, as conveyed by all three words krisis, enkrithentes, and kritikoi.
Although the Alexandrian scholars eventually abandoned the term kritikos in favor of grammatikos, they preserved and, in fact, perfected the principles that shaped the concept of kritikos in the first place. Their work reveals most clearly the combination of selective and holistic perspectives:
The canon as conceived by the Alexandrian scholars is not to be confused with the actual collection of works housed in the great library of the Museum at Alexandria. The Pinakes or ‘Tables’ of Callimachus, in 120 books, was intended not as a selection but as a complete catalogue of the holdings of the Museum, generally organized along the lines of formal criteria, including meter. [10] {188|189}
The Alexandrian classical model, as formalized in the very concept of the Pinakes of Callimachus, makes it explicit that a holistic perspective is a prerequisite for the application of the principle of selection. [11] To this extent, then, the formulation of Crates is unfair to the Alexandrians, since it implies that only the Pergamene classical model is truly holistic. As I will show, however, it may indeed be fair to say that the Pergamene classical model was more holistic than the Alexandrian, at least in the era of Crates. A case in point is the difference between Pergamene and Alexandrian approaches to Homeric verses suspected of being non-Homeric.
Which brings me to the second point that needs clarification, in order to avoid yet another major misunderstanding. This clarification has to do with Pfeiffer’s own overall view of the historical links between the scholarship represented by the Library of Aristotle at the Lyceum in Athens and that perpetuated by the Library at the Museum in Alexandria. For Pfeiffer, these links were minimal, and the scholarship of the Museum was for all intents and purposes “a new beginning.” [12]
This view, which has held sway for many years, has more recently given way to an opposing view as exemplified by Nicholas Richardson, who argues that the Alexandrian scholars were for the most part direct continuators of the Peripatetic traditions of Aristotle and Theophrastus. [13] As Richardson argues, the evidence for direct continuity from Lyceum to Museum is clearest in the fields of antiquarian and documentary studies, {189|190} literary and historical biography, and ‘grammar’ in the broad humanistic sense of grammatikē as the Aristarchean linguist Dionysius Thrax defined it in the later half of the second century BCE. [14]
Richardson concedes gaps in philosophy and the sciences, which he attributes mainly to the fact that “Aristotle’s followers in the Hellenistic period from the mid-third century onwards did little to advance his own work in the major fields of abstract philosophy or of physics,” [15] but even so he stresses the Alexandrians’ filling in of various gaps in various specific branches of the sciences, most notably in medicine, mathematics, mechanics, and astronomy. [16]
My point, however, is not simply that Pfeiffer’s sketch of a radical new beginning of a new kind of scholarship in the Alexandria of the Ptolemies needs redrawing in light of the ever-accumulating evidence for a continuation of Peripatetic scholarly traditions at the Museum. Rather, I concentrate on the historical reality of the classical model or models that this scholarly continuity represents. This reality has to do with the prestige of what exactly the scholars of the Lyceum and the Museum studied. Earlier, I had spoken of the prestige of the classics, which now leads me to my central thesis: it is this prestige that primarily motivates the overall scholarship of the Library of Alexandria, as well as of its rival, the Library of Pergamon. It is the prestige of the classics that drives the continuum from, say, the Lyceum to the Museum.
My clarification of this point seeks to head off a major misunderstanding about the libraries of both Alexandria and Pergamon. It may be easy to intuit, even in hindsight, the prestige of philosophy and the sciences in the era of the Ptolemies of Alexandria and the Attalids of Pergamon. It may be much harder, however, for us to keep firmly in mind, from a historical perspective, what scholars then considered to be {190|191} the centerpiece of all higher learning: the study of the literary canon in general and of Homer in particular. It is perhaps the hardest thing of all to envision today an era of intellectual history when the prestige of all higher learning centered on the study of Homer.
It is precisely at this point that we run into a most serious problem: not enough direct information about the study of Homer in the Lyceum survives, and most of what little we can know is reconstructed from indirect information stemming primarily from the Libraries of Alexandria and Pergamon. Despite this dearth of information, however, I argue that the prestige of Homeric studies can indeed be traced back primarily to the Lyceum.
In making this argument, I have to disagree on one point not only with Pfeiffer, but also with Richardson. Although Richardson disagrees in general with Pfeiffer’s view of a discontinuity between the Peripatetic tradition of the Lyceum and the new age of scholarship at the Museum in Alexandria, he agrees with Pfeiffer in rejecting continuity between the Lyceum and the Museum in one particular field, the editing of Homer in particular and of classical texts in general:
Aristotle never seems to have actually edited [his emphasis] a classical text: for Pfeiffer was most probably right to argue that there is no basis for the tradition that Aristotle actually produced an “edition” of the Iliad, revised or corrected by himself. [17]
The specific piece of tradition that Richardson cites here comes from the Vita Marciana, which lists among Aristotle’s works not only his Homeric Questions (Ὁμηρικὰ ζητήματα) but also something called the ekdosis of the Iliad (ἡ τῆς Ἰλιάδος ἔκδοσις ἣν δέδωκε τῷ Ἀλεξάνδρῳ). [18]
Let me reexamine for a moment the meaning of ekdosis. To begin, it is vital to stress that it is anachronistic and even reductionistic to understand ekdosis in the modern sense of ‘edition.’ Accordingly, I will hereafter use “edit” and “edition” only within quotation marks in referring to ancient scholars’ efforts to establish a definitive text. Elsewhere, I have argued at length for an evolution in the meaning of the technical term ekdosis, and of its essential correlate, diorthōsis, in the historical {191|192} context of the three great “editors” of Homer at the Library of Alexandria: Zenodotus of Ephesus (early third century BCE), Aristophanes of Byzantium (early second century), and Aristarchus of Samothrace (mid-second century). Of the three, Aristarchus comes closest to the modern notion of establishing a text with an apparatus criticus (though his was apparently embedded in volumes of hupomnēmata ‘commentaries’ that were separate from the volumes of the actual text of Homer). [19] Unlike modern editors, however, even Aristarchus would not have placed any of his own conjectured readings into the Homeric text that he established, restricting himself instead to choices based on variants actually attested in the manuscript traditions that were available to him. [20] Tracing backward from Aristarchus, we may detect further differences between modern and Alexandrian concepts of establishing an ancient classical text, especially in the case of the earliest Alexandrian “editor” of Homer, Zenodotus. [21]
Still, we may note that some aspects of modern editorial techniques do indeed overlap with the meanings of the Greek terms ekdosis and diorthōsis. There are traces, not only in the era of Zenodotus but even earlier, in the era of Aristotle himself, of ongoing scholarly research centering on such “editorial” activities as commentaries on textual variations with a view toward establishing the definitive text, and the marking of signs in the margins of definitive texts, keyed in to corresponding definitive commentaries, with special attention to such details as the pronunciation of different accents over different words. [22] {192|193}
For now, however, I focus not on the mechanics of ekdosis but on its prestige. In line with my central thesis, I maintain that the key to the prestige of ekdosis is the concept of the library as shaped by the larger concept of the classics. This concept, as it became established at Alexandria and at Pergamon, was already taking shape in the Lyceum. I submit that we cannot understand the attribution of an ekdosis of Homer to Aristotle unless we first come to terms with the concept of a concurrent paradosis, by which I mean the transmission of the prestige of the classics from the Lyceum to the Museum. [23]
At this point I need to elaborate on a central concept, what I have been calling the prestige of the classics. From here on, I propose to develop a more specific usage of this word prestige, precisely in terms of a historical perspective on that other central concept, the classics. {193|194}
In speaking practically of prestige, we cannot separate it from power and wealth. From a sociological vantage point, we may think of social status in terms of this triad of power, wealth, and prestige, viewing all three factors as contiguous and interconnected. [24] Throughout the history of ancient Greece, these factors had shaped the very concept of aristocracy as it evolved from a wide variety of sociopolitical realities into a single unified cultural ideal. [25] The medium that conveyed such a seemingly monolithic concept or ideal was the classics. [26] To restate the central thesis, then, in more precise terms: the classical model of the library was interconnected with the power, wealth, and prestige of its founders.

The Library of Alexandria

Here we see the clearest illustration of interconnections between the idea of a library and the political realities of its foundation. Witness the following description of the Library as founded by the dynasty of the Ptolemies, that is, the Lagidai, in Alexandria. The primary physical setting was the sacred precinct of the Muses, the Museum or Mouseion: [27]
ἅπαντα μέντοι συναφῆ καὶ ἀλλήλοις καὶ τῷ λιμένι καὶ ὅσα ἔξω αὐτοῦ. τῶν δὲ βασιλείων μέρος ἐστὶ καὶ τὸ Μουσεῖον, ἔχον περίπατον καὶ ἐξέδραν καὶ οἶκον μέγαν ἐν ᾧ τὸ συσσίτιον τῶν μετεχόντων τοῦ Μουσείου φιλολόγων ἀνδρῶν. ἔστι δὲ τῇ συνόδῳ ταύτῃ καὶ χρήματα κοινὰ καὶ ἱερεὺς ὁ ἐπὶ τῷ Μουσείῳ τεταγμένος τότε {194|195} μὲν ὑπὸ τῶν βασιλέων νῦν δ' ὑπὸ Καίσαρος. μέρος δὲ τῶν βασιλείων ἐστὶ καὶ τὸ καλούμενον Σῶμα, [28] ὃ περίβολος ἦν ἐν ᾧ αἱ τῶν βασιλέων ταφαὶ καὶ ἡ Ἀλεξάνδρου· ἔφθη γὰρ τὸ σῶμα ἀφελόμενος Περδίκκαν ὁ τοῦ Λάγου Πτολεμαῖος κατακομίζοντα ἐκ τῆς Βαβυλῶνος καὶ ἐκτρεπόμενον ταύτῃ κατὰ πλεονεξίαν καὶ ἐξιδιασμὸν τῆς Αἰγύπτου·
All [the buildings] are connected to each other and to the harbor and what lies outside the harbor. The Museum is also part of the royal complex. It has a walkway [ Peripatos ] and an Exedra [29] and a great building that houses the place where the philologoi who take part in the Museum dine in common. Property, too, is held in common by this assembled group, and at their head is the priest who is put in charge of the Museum, who used to be appointed by the kings, but now by Caesar [Augustus]. Another part of the royal complex is the so-called Sōma . This is an enclosure where the tombs of the kings and of Alexander are located. For Ptolemy, the son of Lagos, took his [Alexander’s] body [ sōma ] away from Perdiccas, thus edging him out. [30] Perdiccas had been bringing it [the body] from Babylon and had detoured toward Egypt, moved by greed and by the ambition to make that country his own.
Strabo goes on to report that Perdiccas was killed by his own men, and that his retinue thereupon departed for Macedonia. The narrative then adds this detail about Ptolemy’s next move: τὸ δὲ σῶμα τοῦ Ἀλεξάνδρου κομίσας ὁ Πτολεμαῖος ἐκήδευσεν ἐν τῇ Ἀλεξανδρείᾳ ὅπου νῦν ἔτι κεῖται (‘but Ptolemy brought home [31] the body [ sōma ] and gave it the proper ritual in Alexandria, where it still lies’). [32] Strabo gives {195|196} further details concerning a glass display case in which the body of Alexander was kept during the geographer’s lifetime: this case was a replacement for an original golden sarcophagus stolen in the era of Ptolemy XI.
This whole description of the Museum or Mouseion, the sacred space that contains the Library, abounds in traditional metonyms that reveal the connections of power, wealth, and prestige. By metonym here I mean, as a working definition, the expression of meaning by way of connection (as opposed to metaphor, by which I mean the expression of meaning by way of substitution). The first and most obvious metonymy to notice in Strabo’s description is the idea that the Library is “part” of the Museum or Mouseion, that is, of the sacred precinct of the Muses. This is an example of the specialized metonymy of synecdoche, in that the Library is viewed as a part of the whole, an aspect of the totality that is the Museum. [33] Strabo situates the statement of this contiguity in the larger context of his statement that all buildings are contiguous in Alexandria.
Second, there is the metonymy of contiguity between the Museum and the royal precinct, between the sacred and the political. Within this contiguity is the body of men who are an assembled group (sunodos) that eat together and share their wealth as if they were one. The synecdoche of the body in the context of a Library recalls the narrative in the Letter of Aristeas about the assembled group who simultaneously “translate” into Greek the sacred books of the Jews. [34]
Third and most important for my argument is the metonymic contiguity between the Museum and the royal Ptolemaic tombs, especially the tomb of Alexander the Great. By extension (where “extension” itself becomes a metonym), there is the contiguity between ‘tomb’ (sēma) and ‘body’ (sōma), a contiguity that is embodied in the synecdoche of naming the tomb of Alexander as the body par excellence. [35] The Library {196|197} is envisioned as contiguous with the body of the king, as if it were one overall corpus. The notional and even physical contiguity of the library with the sōma of the king is, I suggest, a traditional concept inherent in the very idea of the library. The corpus of books is coextensive with the corpus of the king. [36] Furthermore, the preservation of the king’s body coincides with the preservation of the books. Even the aetiological narrative recorded by Strabo, concerning the ‘bringing home’ of Alexander’s body to the Museum, is driven by a metonym: the verb komizō in that narrative, which I have translated in the specialized sense of ‘bring home’, conveys the preservation of either a hero’s body [37] or a body of literature. [38]
Fourth, there is an even deeper level of metonymy inherent in the name Museum or Mouseion ‘sacred precinct of the Muses’, which links the preservative phase of the canon, as embodied by the library, with the creative or productive phase of what we know as the Classical era of Greek literature, especially around the fifth century. From a classical point of view, the Muses preside primarily over the production of {197|198} belles lettres, and only secondarily over their preservation. [39] Between production and preservation, of course, is the concept of continuation: for the scholar-poets of Alexandria, continuity is the preservation of the old canonical literature and the production of new non-canonical (or, better, meta-canonical) literature. [40] I should add that such a concept of continuation is perhaps the most ambitious metonym of them all.
Strabo’s description of the Library of Alexandria reinforces my central thesis: the classical model of the library is tied to the power, wealth, and prestige of its patrons. For the model to be classical, moreover, it must establish a continuity with the most prestigious earlier models. For the Library of Alexandria, that earlier model was the Lyceum. Here it is relevant to emphasize that Strabo’s description of the Museum of Alexandria includes the term Peripatos, referring to one of the salient physical features of the Library. This physical feature is also a notional feature metonymically linking the Museum to the Lyceum. [41]
The Museum’s claim to continuity with the Peripatetics is retrojected ideologically to Aristotle himself, as if he had personally intended the Library of Alexandria to become the fulfillment, the teleology, of his Lyceum:
Ἀριστοτέλης ... πρῶτος ὧν ἴσμεν συναγαγὼν βιβλία, καὶ διδάξας τοὺς ἐν Αἰγύπτῳ βασιλέας βιβλιοθήκης σύνταξιν
Aristotle ... was the first that we know of to collect books, and he taught the kings in Egypt how to organize a library [42]
The Alexandrian Library’s appropriation of Aristotle as their very own proto-Peripatetic, as it were, became attenuated and modified with the passage of time. From around two hundred years after Strabo (born around 64 BCE) comes the following description of the prestige of the great libraries of the world:
ἦν δέ, φησί, καὶ βιβλίων κτῆσις αὐτῷ ἀρχαίων Ἑλληνικῶν τοσαύτη ὡς ὑπερβάλλειν πάντας τοὺς ἐπὶ συναγωγῇ τεθαυμασμένους, {198|199} Πολυκράτην τε τὸν Σάμιον καὶ Πεισίστρατον τὸν Ἀθηναίων τυραννήσαντα Εὐκλείδην τε τὸν καὶ αὐτὸν Ἀθηναῖον καὶ Νικοκράτην τὸν Κύπριον ἔτι τε τοὺς Περγάμου βασιλέας Εὐριπίδην τε τὸν ποιητὴν Ἀριστοτέλην τε τὸν φιλόσοφον <καὶ Θεόφραστον> καὶ τὸν τὰ τούτων διατηρήσαντα βιβλία Νηλέα· παρ' οὗ πάντα, φησί, πριάμενος ὁ ἡμεδαπὸς βασιλεὺς Πτολεμαῖος, Φιλάδελφος δὲ ἐπίκλην, μετὰ τῶν Ἀθήνηθεν καὶ τῶν ἀπὸ Ῥόδου εἰς τὴν καλὴν Ἀλεξάνδρειαν μετήγαγε.
Athenaeus says that he [Larensis] had possession [ ktēsis ] of so many ancient Greek books that he surpassed all who have been admired for their collections, including Polycrates of Samos, Peisistratos the tyrant of Athens, Eucleides, likewise an Athenian, Nicocrates of Cyprus, the kings of Pergamon, Euripides the poet, Aristotle the philosopher, Theophrastus, and Neleus, who preserved the books of the two last named. From him [Neleus], he says, our own King Ptolemy, surnamed Philadelphus, purchased them [the books of Aristotle and Theophrastus], and he transported them, along with the books from Athens and Rhodes, to our fair city of Alexandria. [43]
This passage comes from a portion of the author’s work that survives only in epitomized form. Athenaeus, who flourished around 200 CE, was from Naukratis in Egypt, and his provenience signals a Hellenized Egyptian point of view in this listing of libraries. Athenaeus’ gallery of great men who possessed great libraries is an exercise in prestige by association, with a predictable climax: he reports that the collection of Aristotle was teleologically absorbed by the all-encompassing collection of the Ptolemies of Alexandria.
There is a difference here in intellectual orientation: unlike Strabo’s formulation, which highlights the principle of suntaxis ‘organization’ that produces a collection of higher learning in the form of a library, Athenaeus’ account concentrates on the possession of the books themselves. {199|200}

The “Vanished Library” of Aristotle [44]

We are now faced what appears to be, at least on the surface, a major contradiction. [45] On the one hand, Athenaeus claims that Neleus of Skepsis inherited Aristotle’s own collection of books and then sold it to the Ptolemies. Strabo’s version, on the other hand, seems to contradict Athenaeus’ claim. In the same context where Strabo says that Aristotle ‘was the first that we know of to collect books, and he taught the kings in Egypt how to organize a library’, he adds that Aristotle had left his library to Theophrastus, who in turn left his own library, including that of Aristotle, to Neleus:
Θεόφραστος δὲ Νηλεῖ παρέδωκεν· ὁ δ’ εἰς Σκῆψιν κομίσας τοῖς μετ’ αὐτὸν παρέδωκεν, ἰδιώταις ἀνθρώποις, οἳ κατάκλειστα εἶχον τὰ βιβλία οὐδ’ ἐπιμελῶς κείμενα· ἐπειδὴ δὲ ᾔσθοντο τὴν σπουδὴν τῶν Ἀτταλικῶν βασιλέων ὑφ’ οἷς ἦν ἡ πόλις, ζητούντων βιβλία εἰς τὴν κατασκευὴν τῆς ἐν Περγάμῳ βιβλιοθήκης, κατὰ γῆς ἔκρυψαν ἐν διώρυγί τινι· ὑπὸ δὲ νοτίας καὶ σητῶν κακωθέντα ὀψέ ποτε ἀπέδοντο οἱ ἀπὸ τοῦ γένους Ἀπελλικῶντι τῷ Τηίῳ πολλῶν ἀργυρίων τά τε Ἀριστοτέλους καὶ τὰ τοῦ Θεοφράστου βιβλία· ἦν δὲ ὁ Ἀπελλικῶν φιλόβιβλος μᾶλλον ἢ φιλόσοφος· διὸ καὶ ζητῶν ἐπανόρθωσιν τῶν διαβρωμάτων εἰς ἀντίγραφα καινὰ μετήνεγκε τὴν γραφὴν ἀναπληρῶν οὐκ εὖ, καὶ ἐξέδωκεν ἁμαρτάδων πλήρη τὰ βιβλία. συνέβη δὲ τοῖς ἐκ τῶν περιπάτων τοῖς μὲν πάλαι τοῖς μετὰ Θεόφραστον οὐκ ἔχουσιν ὅλως τὰ βιβλία πλὴν ὀλίγων, καὶ μάλιστα τῶν ἐξωτερικῶν, μηδὲν ἔχειν φιλοσοφεῖν πραγματικῶς, ἀλλὰ θέσεις ληκυθίζειν· τοῖς δ’ ὕστερον, ἀφ’ οὗ τὰ βιβλία ταῦτα προῆλθεν, ἄμεινον μὲν ἐκείνων φιλοσοφεῖν καὶ ἀριστοτελίζειν, ἀναγκάζεσθαι μέντοι τὰ πολλὰ εἰκότα λέγειν διὰ τὸ πλῆθος τῶν ἁμαρτιῶν. πολὺ δὲ εἰς τοῦτο καὶ ἡ Ῥώμη προσελάβετο· εὐθὺς γὰρ μετὰ τὴν Ἀπελλικῶντος τελευτὴν Σύλλας ἦρε τὴν Ἀπελλικῶντος {200|201} βιβλιοθήκην ὁ τὰς Ἀθήνας ἑλών, δεῦρο δὲ κομισθεῖσαν Τυραννίων τε ὁ γραμματικὸς διεχειρίσατο φιλαριστοτέλης ὤν, θεραπεύσας τὸν ἐπὶ τῆς βιβλιοθήκης, καὶ βιβλιοπῶλαί τινες γραφεῦσι φαύλοις χρώμενοι καὶ οὐκ ἀντιβάλλοντες, ὅπερ καὶ ἐπὶ τῶν ἄλλων συμβαίνει τῶν εἰς πρᾶσιν γραφομένων βιβλίων καὶ ἐνθάδε καὶ ἐν Ἀλεξανδρείᾳ. περὶ μὲν οὖν τούτων ἀπόχρη.
Theophrastus bequeathed it to Neleus, who brought it home [ komizein ] to Skepsis and bequeathed it to those who came after him, nonaffiliated men, who kept the books locked up but not carefully stored. When they found out about the ambition of the Attalid kings (the city was under their control) in searching for books for the collection of the Library in Pergamon, they hid the books underground in a kind of hollow. Later, after the books had already been damaged by moisture and moths, their descendants sold them—both the books of Aristotle and those of Theophrastus - to Apellicon of Teos for the price of a large sum of money. Apellicon was more of a bibliophile than a philosopher, and so, seeking a restoration [epanorthōsis] of the parts that had been eaten through, he transferred the writings into new copies [antigrapha], filling in the lacunae not very well, and the books that he published [that is, made an of] were full of errors. So this is how it came about that the earlier members of the Peripatos, the ones that came after Theophrastus, had no books at all, except for a few, which were mostly exoteric, and thus they were able to engage in philosophical discourse about practically nothing, but only to talk frivolously about commonplaces. As for the later members [of the Peripatetic school], starting from the time when these books appeared in published form, it was easier to philosophize and “aristotelize,” but they were forced to call most of their statements verisimilitudes because of the large mass of errors. Rome was in large part responsible for this: after the death of Apellicon, Sulla, who had captured Athens, had carried off Apellicon’s library. After it was brought home [komizein] here [to Rome], Tyrannion the grammatikos, who was a pro-Aristotelian, took it in hand, having successfully cultivated the man in charge of the library. Booksellers also took it in hand, but they used inferior scribes and did not collate—something that also happens to other books too that are copied for sale, both here and in Alexandria. [46] {201|202}
The teleology of Strabo’s account is this: the library of Aristotle ends up in the capable hands of Tyrannion, described as a grammatikos and a ‘pro-Aristotelian’, who just happens to have been a teacher of Strabo’s. [47] This Tyrannion, who became a protégé of the likes of Caesar, Cicero, and Atticus, had studied in Rhodes under the distinguished Aristarchean scholar Dionysius Thrax. [48]
For Strabo, the “happy ending” of the story of Aristotle’s library is that the collection falls into the hands of an Aristarchean scholar, a grammatikos. Still, the ending is not altogether happy (as I continue to paraphrase, I must stress that the value judgments that I include are those of the narrative, not mine): even after the text of Aristotle is soundly “edited” by Tyrannion, the dissemination falls into the hands of booksellers who fail to maintain the correctness of the editorial work. By implication, the “new” Peripatetics have access to the augmented text only through booksellers. Consequently, although their textual repertoire is relatively fuller than that of the “old” Peripatetics, their control of the philosophical content is still incomplete because they do not control the source of the text.
Plutarch, by contrast, adds one further stage to the story, and the new closure of the narrative transforms the “happy ending” into one that is no longer Aristarchean but overtly Peripatetic. In the account of Plutarch (Life of Sulla 26.1-2), a genuine Peripatetic named Andronicus takes over from Tyrannion the task of “editing” Aristotle: [49]
καὶ μυηθεὶς ἐξεῖλεν ἑαυτῷ τὴν Ἀπελλικῶνος τοῦ Τηΐου βιβλιοθήκην, ἐν ᾗ τὰ πλεῖστα τῶν Ἀριστοτέλους καὶ Θεοφράστου βιβλίων ἦν, οὔπω τότε σαφῶς γνωριζόμενα τοῖς πολλοῖς. λέγεται δὲ κομισθείσης αὐτῆς εἰς Ῥώμην Τυραννίωνα τὸν γραμματικὸν ἐνσκευάσασθαι τὰ πολλά, καὶ παρ' αὐτοῦ τὸν Ῥόδιον Ἀνδρόνικον εὐπορήσαντα τῶν ἀντιγράφων εἰς μέσον θεῖναι καὶ ἀναγράψαι τοὺς νῦν φερομένους πίνακας. οἱ δὲ πρεσβύτεροι Περιπατητικοὶ φαίνονται μὲν {202|203} καθ' ἑαυτοὺς γενόμενοι χαρίεντες καὶ φιλολόγοι, τῶν δὲ Ἀριστοτέλους καὶ Θεοφράστου γραμμάτων οὔτε πολλοῖς οὔτε ἀκριβῶς ἐντετυχηκότες διὰ τὸ τὸν Νηλέως τοῦ Σκηψίου κλῆρον, ᾧ τὰ βιβλία κατέλιπε Θεόφραστος, εἰς ἀφιλοτίμους καὶ ἰδιώτας ἀνθρώπους περιγενέσθαι.
After being initiated [at the Piraeus] he [Sulla] expropriated for himself the library of Apellicon of Teos. In this collection were most of the books of Aristotle and Theophrastus, which were at that time not yet known to most people. And it is said that, after it [the library] was brought home [komizein] to Rome, [50] Tyrannion the grammatikos furnished the majority of the books with an apparatus and that Andronicus of Rhodes procured copies from him, which he published [made public, put into the public domain] along with a listing of catalogues [pinakes] that are even now in circulation. The earlier Peripatetics, though they seem intrinsically sophisticated and scholarly [philologoi], did not have access to many of the writings of Aristotle and Theophrastus, and what they did have was not accurate—all on account of the legacy of Neleus of Skepsis, to whom Theophrastus had bequeathed the books. Because of that legacy, they [the books] had been handed down to men without any ambition or affiliation.
According to this even happier ending to the story, the “new” Peripatetics now have their own source of authority because they also possess their own “edited” text. In Strabo’s account, they had become less incomplete philosophers than the “old” Peripatetics, thanks to the “editing” by the Aristarchean scholar Tyrannion. In Plutarch’s account, by contrast, the “new” Peripatetics become complete philosophers thanks to the “editing” of the Peripatetic scholar Andronicus. The pinakes ‘catalogues’ of Andronicus must have superseded those of Callimachus—at least with reference to the corpus of Aristotle. [51]
The “editorial” procedures of the authoritative school in charge (in this case, of the Peripatetics) define the validity of the given collection {203|204} of texts. The concept of library depends on the concept of ekdosis, for which I now offer this working definition: ‘publication (that is, making public), controlled by an authoritative source, of a corrected body of texts’. To the extent that our word “edition” fits this working definition, I will hereafter use it without quotation marks in referring to the editorial authority of large-scale libraries like those of Alexandria and Pergamon—as also of smaller-scale collections like the corpus of Aristotle as reconstituted by Andronicus.
In the narratives of both Strabo and Plutarch, the concept of Aristotle’s library as something purchased by Apellicon from the descendants of Neleus represents a “newer” phase of that library, which now claims access to a better and fuller set of Aristotle’s own writings than what had been available during the “older” phase. In the narrative reported by Athenaeus, by contrast, the concept of Aristotle’s library as something purchased by the Ptolemies directly from Neleus remains the official Alexandrian tradition. We may note with interest that Athenaeus, in another context, both acknowledges and slights the role of Apellicon:
Ἀπελλικῶντα τὸν Τήιον, πολίτην δὲ Ἀθηναίων γενόμενον, ποικιλώτατόν τινα καὶ ἁψίκορον ζήσαντα βίον· ὅτε μὲν γὰρ ἐφιλοσόφει [καὶ] τὰ περιπατητικά, καὶ τὴν Ἀριστοτέλους βιβλιοθήκην καὶ ἄλλας συνηγόραζε συχνὰς (ἦν γὰρ πολυχρήματος) τά τ' ἐκ τοῦ Μητρῴου τῶν παλαιῶν αὐτόγραφα ψηφισμάτων ὑφαιρούμενος ἐκτᾶτο καὶ ἐκ τῶν ἄλλων πόλεων εἴ τι παλαιὸν εἴη καὶ ἀπόθετον. ἐφ' οἷς φωραθεὶς ἐν ταῖς Ἀθήναις ἐκινδύνευσεν ἄν, εἰ μὴ ἔφυγεν. καὶ μετ' οὐ πολὺ πάλιν κατῆλθε, θεραπεύσας πολλούς· καὶ συναπεγράφετο τῷ Ἀθηνίωνι ὡς δὴ ἀπὸ τῆς αὐτῆς αἱρέσεως ὄντι.
Apellicon of Teos, who had been made an Athenian citizen. He led a life filled with varied fancies and excesses. When he was in his Peripatetic philosopher phase, he bought up the library of Aristotle and many other books, for he was very rich, and he began to acquire surreptitiously the original written records of the ancient decrees stored in the [Athenian] Mētrōon as well as anything else in the other cities that was old and apotheton [stored in secured places]. Caught in the act of stealing the records in Athens, he would have been in real danger if he had not fled into exile. Not much later, however, he was reinstated, having successfully cultivated many people. [52] {204|205}
These conflicting accounts, of course, leave many things unclarified. One thing, however, is quite clear in the account of Athenaeus that I have just cited: the Alexandrian Library’s initial holdings of Aristotle’s collection did not contain some of the originals or copies of Aristotle’s works that Apellicon later acquired. Athenaeus (5.211d) names Posidonius the Stoic as his source. [53] It is also clear from this account that Apellicon’s new acquisitions came from apotheta, ‘texts stored away in secured places’.
Those who claimed to be “new” Peripatetics, in terms of the ideologies represented by the narratives of both Strabo and Plutarch, must have acknowledged the new reality of an augmented corpus of Aristotle’s works, one that incorporated the textual additions and variants acquired by Apellicon. To justify this purportedly improved Aristotelian corpus, the “new” Peripatetics would have had to rethink the whole concept of the library of Aristotle in terms of the one feature of the collection that was most important to them, that is, the newly reconstituted corpus of Aristotle. By way of metonymy, this concept of a newly augmented corpus of Aristotle, this “Aristoteles auctus,” could lead to a new way of thinking of the whole library of Aristotle. As has become clear from the versions reported by Strabo and Plutarch, it was now the descendants of Neleus and not he himself who supposedly sold the library to Apellicon, that key figure in the evolving concept of an augmented Aristotle.
The official version of the Alexandrian Library, by contrast, could continue to feature an unaugmented narrative about an unaugmented Aristotle: it was Neleus himself who had sold the Library of Aristotle to the Ptolemies. Athenaeus duly reports this version, but he also allows himself at a later point to transmit the conflicting story of an augmented Aristotle conveyed by an augmented library, although he portrays the actual augmentation as a tainted one. That Athenaeus could allow such a contradiction in his work should not be surprising to anyone who simply takes another look at the genre of his work: it is a compilation of reportage that is generally more faithful to the accuracy of its varied sources than to its own internal consistency.
In short, I am persuaded by the argument that the Ptolemies did indeed buy the Library of Aristotle from Neleus. [54] As for the rival idea that {205|206} Apellicon bought it from the descendants of Neleus, I view this augmented narrative as a historical reflex, but not as a historical event. The real events that underlie that narrative have to do with the Apellicon’s acquisition of apotheta that ended up changing the concept of Aristotle’s library by changing the concept of Aristotle’s corpus of works. To say that Apellicon purchased Aristotle’s “library,” not merely a set of important Aristotelian texts, is an exercise in metonymy. Still, this metonymy is historically explainable.
The question still remains: did Apellicon really acquire these texts from the descendants of Neleus? There is a clue in a detail provided by Strabo, who asserts that these texts were being hidden from the Attalids of Pergamon, who would have dearly loved to possess them in their own Library.

“Virtual Libraries”

As we saw in the first report of Athenaeus, [55] the Attalids of Pergamon figure prominently in the gallery of great men in history who possessed great libraries. Athenaeus associates the Attalids, along with the Ptolemies of Alexandria, with archaic figures like Polycrates, the tyrant of Samos, and the Peisistratids, tyrants of Athens. Such an association between great men of the archaic and Hellenistic periods is indicative of the classical models at stake.
Athenaeus uses a key word in this regard: ktēsis ‘possession’. A salient illustration is the case of the Peisistratids of Athens: for this dynasty, possession of the performing arts was a key expression of their power, wealth, and prestige. [56] For example, Herodotus (5.90.2) uses the verb kektēmai ‘possess’, from which the noun ktēsis is derived, in reporting that the Peisistratids ‘possessed’ (ἐκέκτηντο) oracular poetry by way keeping it locked up - that is, in written form - on the acropolis. In this way, this dynasty of tyrants could control the precise time and place of any unlocking, as it were, of the text for occasions of performance. [57] This pattern of political-artistic control as the Peisistratids exercised it extended to the premier performance tradition as ultimately realized in the premier text of Hellenic civilization, the Homeric Iliad {206|207} and Odyssey: here we see the germ of a construct that modern scholars recognize as the “Peisistratean Recension.” [58]
A central passage is the Platonic Hipparkhos 228b-c, which describes the activities of the Peisistratean tyrant Hipparkhos of Athens as a patron of the arts. He is said to be the first to have introduced the epē ‘poetic words’ of Homer to Athens, and the verb used is komizein ‘bring home’ (καὶ τὰ Ὁμήρου ἔπη πρῶτος ἐκόμισεν εἰς τὴν γῆν ταυτηνί). The Peisistratids attempt to “possess” not only epic but lyric as well: in the same context, Hipparkhos is said to have the year-round presence of the lyric master Simonides at his court, having ‘persuaded him with great fees and gifts’ (Σιμωνίδην δὲ τὸν Κεῖον αἰεὶ περὶ αὑτὸν εἶχεν, μεγάλοις μισθοῖς καὶ δώροις πείθων). [59] Furthermore, this Peisistratid makes a special effort to import another such master, Anakreon, by sending a ship to fetch him from overseas: literally, the tyrant ‘brings home’ to Athens the lyric virtuoso himself, and again the verb used is komizein (καὶ τὸν Ἀνακρέοντα τὸν Τήιον πεντηκόντορον στείλας ἐκόμισεν εἰς τὴν πόλιν). The poetic virtuosity of Anacreon is also the possession of the tyrant Polycrates of Samos. Herodotus 3.121, for example, portrays the lyric master participating in a symposium at Samos organized by his patron, the tyrant Polycrates. [60] The joint possession, as it were, of the lyric virtuosity of Anacreon by Polycrates and by the Peisistratid Hipparkhos is a function of these tyrants’ political possession of the performing arts. [61]
Here we return to Athenaeus’ joint mention of Polycrates and the Peisistratids in the larger context of his gallery of great men who possessed great collections of books. [62] It is possible to trace the idea of a political possession of the performing arts by way of a “recension,” as in the case of the construct that we know as the “Peisistratid Recension,” forward in time from the Peisistratids of Athens to two later statesmen, {207|208} also of Athens: Lycurgus (middle of the fourth century BCE) and Demetrius of Phaleron (late fourth century BCE). In both these cases, the possession of the performing arts correlated in some specific way with the regulation of performance by way of the written text. In the case of Demetrius, the regulation had to do with the performing of the Homeric poems themselves. [63] For Lycurgus, the regulating (and regulated) text in question was the “State Script” of the tragedies of Aeschylus, Sophocles, and Euripides, as we read in the Plutarchean Lives of the Ten Orators 841F. [64] Galen of Pergamon also mentions this Athenian canonical text. [65] He reports that Ptolemy III Euergetes (246 to 221 BCE) had borrowed it from the Athenians, who accepted a deposit of fifteen talents for the ostensible purpose of having it copied for the Library of Alexandria; Galen goes on to say that this Athenian state text of three canonized tragedians was never returned to the Athenians. [66]
By possessing the canonical script of the tragedies of Aeschylus, Sophocles, and Euripides, the Ptolemies could claim notional control over the official text of one of the two most prestigious genres of the Athenian performing arts, tragedy. As for epic, the other of these most prestigious genres, there are traces of even earlier Ptolemaic appropriation or possession of the text: a key figure, as we consult the opaque report of Athenaeus 620b-c is Demetrius of Phaleron, reportedly ‘the first to have introduced the so-called Homēristai [Homer-performers] into the theaters’ (τοὺς δὲ νῦν Ὁμηριστὰς ὀνομαζομένους πρῶτος εἰς τὰ θέατρα παρήγαγε Δημήτριος ὁ Φαληρεύς). [67] In view of the fact that this same Demetrius was centrally involved in the foundation and even the organization of the Library of Alexandria after his regime in Athens fell in 307 BCE (and after his patron Cassander died in 297 BCE), the correlation in this case between controlled performance and controlling text is especially noteworthy. [68] {208|209}
The Museum of Alexandria is the embodiment of this kind of correlation: as we have seen from the description of Strabo, the Library was contiguous with the sacred precinct of the Muses, the goddesses of the performing arts. [69] In this case, the very idea of the Library relates to the political impetus to control the application of the spoken word by controlling the text and keeping it in a secured place. [70]
The Library of Alexandria makes it possible to visualize as “virtual libraries” all previous collections of texts recording the spoken word, including the archaic models of Polycrates and Peisistratos. [71] The physical reality of the Library of Alexandria, that is, a holistic collection of the classics as contained in scrolls stored on the shelves of its bibliothēkē ‘book-repository’ [72] and catalogued in the one hundred and twenty pinakes ‘tablets’ of Callimachus, [73] becomes the virtual reality of the Library as a concept that can now subsume all earlier patterns of canonization or classicism. [74]
Vitruvius, who flourished in the second half of the first century BCE, gives an illustrative anecdote in this regard. [75] It concerns one of the aforementioned editors of Homer at the Library of Alexandria, Aristophanes of Byzantium. [76] Although the historical perspective of Vitruvius seems blurred, his reportage is valuable for reconstructing the history of attitudes about the role of the Library as a control for the performing arts. According to the anecdote of Vitruvius, Aristophanes of Byzantium ‘was studiously and with the greatest diligence reading from beginning to end, day after day and in sequence, all the books’ of the Library of Alexandria (qui summo studio summaque diligentia cotidie {209|210} omnes libros ex ordine perlegeret). [77] As the story begins, King Ptolemy (it is not specified which one) selects Aristophanes to become one of seven judges in a competition of poets who are to ‘recite’ to an audience of ‘the assembled public’ the poetry that they had written (cum recitarentur scripta, populus cunctus). Interestingly, the sequence of reciters is set in advance (primo poetarum ordine ad certationem inducto). [78] The other six judges watch for audience approval as their primary criterion for choosing the winners of the contest, but Aristophanes exposes this criterion as fraudulent: he demonstrates to king and public that the least popular of the competing reciters, the one who was therefore not chosen by the judges, had been in fact the real poet, whereas the other contestants had ‘recited the works of others’ (unum ex his eum esse poetam, ceteros aliena recitavisse). The moral of the story, as Aristophanes makes explicit to the public, is that those who ‘judge’ should approve poems that are ‘written’, not ‘stolen’ (oportere autem iudicantes non furta sed scripta probare).
This story serves as an ex-post-facto justification of the scholar-poets of the Alexandrian Library, whose written works exist independently of the demands of the performing arts and are at the same time all-dependent on a notionally total recall of all the classics stored in the Library. [79] Yet, the irony is that the classics themselves stem from the performing arts.
The story expresses the idea of Aristophanes’ total recall of the classics in terms of a canonical sequence or order: the librarian’s ambition had been to read from beginning to end (perlegere) and in sequence (ex ordine) all the books on the shelves of the Library. Furthermore, when he exposes the false poets, he does so by recalling the canonical sequence of the books as they were stored on the shelves:
fretus memoriae certis armariis infinita volumina eduxit et ea cum recitatis conferendo coegit ipsos furatos de se confiteri. {210|211}
relying on his memory, he had countless scrolls brought out from their respective shelves [armaria], and then, by comparing them with the recited texts, he compelled the men to admit about themselves that they stole them.
Ironically, canonical sequencing was also a vital aspect of the performing arts, as we can see even from the incidental mention of a ‘fixed order’ (ordo) of competing reciters.
Earlier, I had mentioned the construct that modern scholars recognize as the “Peisistratean Recension.” The various sources that report the possession of Homeric poetry by the Peisistratids contain a strong emphasis on the fixed order of performance and even of performers. For an illustration, I quote here a passage to which I have already referred earlier:
Ἱππάρχῳ, ... ὃς ἄλλα τε πολλὰ καὶ καλὰ ἔργα σοφίας ἀπεδείξατο, καὶ τὰ Ὁμήρου ἔπη πρῶτος ἐκόμισεν εἰς τὴν γῆν ταυτηνί, καὶ ἠνάγκασε τοὺς ῥαψῳδοὺς Παναθηναίοις ἐξ ὑπολήψεως ἐφεξῆς αὐτὰ διιέναι, ὥσπερ νῦν ἔτι οἵδε ποιοῦσιν
Hipparkhos [the Peisistratid] ... who made a public demonstration of many and beautiful accomplishments to manifest his expertise [sophia], especially by being the first to bring home [komizein] to this land (Athens) the poetic utterances [epē] of Homer, and he forced the rhapsodes [rhapsōidoi] at the Panathenaic Festival to go through [diienai] these utterances in sequence [ephexēs], by relay [hupolēpsis], just as they (the rhapsodes) do even nowadays. [80]
The idea of order in archaic Greek oral poetics finds expression by way of many metaphors: the order or integrity of a composition-in-performance is comparable to such artifacts as the weaving of a weaver, [81] the joining of a joiner or carpenter, [82] and even the text of a writer. [83] This {211|212} is not the place to examine at length the various applications of the last of these metaphors, that of the written text, a task that I have undertaken elsewhere. [84] Suffice it to observe here that the metaphor of performance as written text stems from cultural settings where oral and written poetics coexist. In such settings, the idea of order in the performing arts can be expressed in terms of an integral text or an integral body of texts, or even a reintegrated body of previously disintegrated texts. [85] An example of the last type of metaphor appears in Cicero’s account of the “Peisistratean Recension”: ‘he [Peisistratos] was the first to put in order the scrolls of Homer, which had previously been in disorder’. [86]
There are more elaborate versions of such a metaphor in the scholia to Dionysius Thrax, where the “Peisistratean Recension” culminates, anachronistically, in what the text describes as the most successful of all recensions, the diorthōsis of Homer by Aristarchus. [87] These scholia present a garbled outlook on the history of Homeric scholarship, but the central message is straightforward: we should visualize Peisistratos’ collection of scrolls, culminating in the corpus of Homeric poetry, in terms of an ultimate collection of scrolls that is the Library of Alexandria. The comprehensiveness of the Library as a collection becomes a model for analogous principles of comprehensiveness in earlier collections, which are rethought as virtual libraries. [88] {212|213}

Crates of Mallos and the Library of Pergamon

The comprehensiveness of the Library at Pergamon, in general terms as well as in specifics, is based on principles that closely parallel those of the Library of Alexandria.
To be sure, differences in degree exist between the two great libraries of Alexandria and Pergamon. The development of new technologies in library science involving the media of parchment in place of papyrus and the codex instead of the scroll or roll was, for example, more pronounced in Pergamon than in Alexandria. [89] For the moment, however, my concern is not with the physical reality of “books” - papyrus or parchment, scroll or codex - nor with questions about where and how the books were stored and the degree to which the libraries secured and preserved them. Instead, I propose to highlight the culture and learning represented by the books, which were a function of the prestige inherent in the Library as a center of learning, as a sort of Institute for Advanced Study or Wissenschaftskolleg. [90]
There is a strong parallelism here with the culture and learning represented by art, especially monumental art, as a visible expression of power, wealth, and prestige. Polybius 22.8.10 reports that Attalus I bought the whole island of Aegina from the Aetolians for thirty talents; one of his motives seems to have been the desire to possess the archaic art. [91] Such a move on the part of the Attalids of Pergamon is parallel to their active acquisition of books in general and their founding a library in particular.
The Attalids of Pergamon, building on the capital investment of the hoard of Lysimachus, the uncle founder, have been compared to that ultimate dynasty of “princely bankers,” the Medici. [92] In this context, we may view the formation of the Library of Pergamon as closely parallel with the Attalids’ cultivation of the major centers of learning at Athens (as represented {213|214} by the three main philosophical schools there, the Academy, the Lyceum or Peripatos, and the Stoa). [93] We may point to the story about Lacydes of Cyrene, a head of the Academy, whom King Attalus I (241-197 BCE) invited to Pergamon. [94] Much is made of the words Lacydes reportedly uttered in declining the offer: ‘Pictures should be viewed from a certain distance’. [95] Eumenes II (197-158 BCE) invited Lycon, a Peripatetic philosopher, but he also declined. [96] Finally Crates the Stoic accepted. [97] As one summary puts it, “The Stoics came, and so the Stoa accepted what the Academy and Lyceum had declined.” [98]
Thus it was that Crates became the head of the so-called “Pergamene school” in the middle of the second century BCE, the same era that witnesses the apogee of the “Alexandrian school” under Aristarchus. The Library of Pergamon, founded by the Attalid king Eumenes II, [99] was the historical context of his scholarly activities. A conventional view of Crates is that he was a “good Stoic,” introducing an allegorical interpretation of Homer. [100] Another conventional view, inspired by Varro, De lingua latina 8.23 and 9.1, is that Crates consistently argued for the principle of Anomaly (anōmalia) in seeking to define ‘correct’ Greek, while Aristarchus stood for the opposite principle of analogy (analogia). [101] I shall return to this ancient debate over analogy/anomaly, with specific reference to the extension of this debate to Homeric criticism. For now, however, I simply note that such reductive formulations, by implying a symmetrical rivalry between Alexandria and Pergamon, tend to blur the ongoing rivalry of Pergamon with centers of learning other than Alexandria, especially Athens. In this context, we may look to Crates as the head of an important new school, parallel to and rivaling the schools of Athens in the prestige of culture and learning. {214|215}

The Reception of Homer in the Libraries

The legacy of Crates as head of the Library of Pergamon can best be appreciated by reassessing the role of the Library in the reception of Homer. The Libraries of Pergamon and Alexandria each had their own distinct programs of research on Homer. They also had their own distinct texts of Homer and even their own distinct editions. The first of these two statements is a matter of common agreement. The second, however, goes against currently prevailing views and will have to be defended.
This time, I start with the Library of Pergamon and with the work of its head, Crates of Mallos. [102] The Vita Romana of Homer describes an ‘archaic Iliad’, arkhaia Ilias, the first line of which is quoted:
ἡ δὲ δοκοῦσα ἀρχαία Ἰλιάς, ἡ λεγομένη Ἀπελλικῶντος [απελικωνος ms., corr. Nauck], προοίμιον ἔχει τάδε·
Μούσας ἀείδω καὶ Ἀπόλλωνα κλυτότοξον
ὡς καὶ Νικάνωρ μέμνηται καὶ Κράτης ἐν τοῖς διωρθοτικοῖς. [103]
But the archaic Iliad, the one that is called ‘Apellicon’s Iliad’, has this prooemium:
I sing the Muses and Apollo, famed for his bow and arrows.
This is the way Nicanor also mentions it, and so too Crates in his Diorthōtika.
Let me anticipate my conclusions about this testimony: the reference to ‘Apellicon’s Iliad’ in this context goes to show that the same Apellicon {215|216} whose acquisition of select texts of Aristotle had led to the production of an “Aristoteles auctus” deserves credit for having also acquired a “Homerus auctus,” that is, a Homer text that contained extra verses not found in the texts available to, say, Aristarchus in Alexandria.
Since Crates is said to have mentioned the extra verse at the beginning of ‘Apellicon’s Iliad ’, we may suspect that the source of Apellicon’s acquisition was actually the Library of Pergamon. We may even begin to suspect that the source of Apellicon’s acquisition of Aristotle’s texts was the same Library. The suspicion centers on a detail in Strabo’s narrative: it concerns the reported motives of the descendants of Neleus, who supposedly hid the books of Aristotle from the Attalids. Perhaps they did. What is left unclarified, however, is this related question: if the descendants of Neleus purposely hid the texts from the Attalids, why would they not hide them also from Apellicon? Perhaps Strabo’s narrative has more to do with the motives of Apellicon himself than with those of the descendants of Neleus.
In the case of the Aristotelian corpus, it may be easiest to work backward from the “happy ending” of Plutarch’s account, as I analyzed it above. The edition of an “Aristoteles auctus” by Andronicus the Peripatetic depended ultimately on Apellicon’s acquisition of Aristotelian texts. As I have noted, Apellicon supposedly acquired these texts from the descendants of Neleus, who had supposedly hidden them from the Attalids of Pergamon. Yet, if indeed Apellicon had access to Homeric texts that were available to Crates at the Library of Pergamon, perhaps the same is true of his access to Aristotelian texts. In this connection, we may recall the negative views of Posidonius the Stoic (by way of Athenaeus 5.214d-e), who suspects Apellicon of unauthorized acquisition of old texts from the state archives of the Mētrōon in Athens. There may well be a parallel pattern in Apellicon’s acquisitions of old texts of both Homer and Aristotle.
Pursuing the question of Crates’ edition of the Homeric corpus, let me take a closer look at the testimony of the Vita Romana of Homer, where the implications of a disagreement and even of a point of rivalry between the schools of Pergamon and Alexandria are visible. As I will indicate later, moreover, the variant reading in the quotation amounts to a challenge even to the Homeric scholarship of the Peripatetics in Athens.
The Vita Romana specifies that Nicanor memnētai ‘makes mention’ of the verse in question and that Crates also makes mention of it in his {216|217} Diorthōtika. [104] The title of this work by Crates, which is cited as a source for the quotation of the verse, makes it clear that he was engaged in the diorthōsis of Homer. This term conveys the idea of establishing the ‘correct’ text of Homer.
The Vita Romana further specifies that the quoted verse is a prooemium or ‘proem’, a label that makes sense in terms of the internal evidence of archaic Greek poetics. [105] Since we know that Crates athetized the prooemia of the Hesiodic Theogony and Works and Days, we may infer that he likewise athetized this prooemium of the Iliad. [106] By athetize (the corresponding noun is athetesis), I mean to decide that a verse is spurious by marking it as such in the margin of the definitive text. [107] In this working definition of athetesis, I draw special attention to my use of the word decide, in the ancient Greek sense conveyed by the critical words krisis, enkrithentes, kritikoi. As for the Alexandrians, their Homer text seems to have omitted altogether this single-verse prooemium to the Iliad.
The distinction between the Pergamene and Alexandrian editorial approaches to this single verse becomes clear upon returning to Strabo’s use of the word apotheta in the sense of ‘texts stored away in secured places’. It appears that the ‘archaic Iliad’ mentioned by Crates is an example of such apotheta. What counts as an atheton for Crates is an apotheton as far as Aristarchus is concerned, if indeed he has no access to it. Crates athetized but included the verse in his official text of the Iliad, while Aristarchus omitted it altogether.
This is no trivial matter. The crucial fact remains that the Homeric Iliad and Odyssey lack any such hymnic prooemium in the medieval manuscript tradition. [108] The opposite situation evolved in the medieval manuscripts of the Hesiodic Theogony and Works and Days, both of {217|218} which survived with such hymnic prooemia: they are the first 115 and the first ten verses of the respective poems in our modern edited texts. [109] Crates athetized both these prooemia as well. [110] Aristarchus athetized the prooemium of the Hesiodic Works and Days. [111]
There are still other instructive examples of such critical decisions. It appears that both the Alexandrians and the Pergamenes had inherited a text of the Hesiodic Theogony that included the Catalogue, and that “it was the Alexandrians, in all probability, who decided that the Theogony should end at line 1020, and the Catalogue begin there.” This decision prevailed into the medieval manuscript tradition. Similarly, Apollonius of Rhodes decided that the Works and Days ended at verse 828, rejecting as spurious the verses that followed; this decision has prevailed, and in this case the rejected verses have vanished altogether. [112] Also, Aristophanes of Byzantium decided that the Odyssey ended at verse 296 of Book xxiii; this decision, however, did not prevail. [114]
Again, all these critical decisions concerning athetesis or omission are no trivial matter. Even in cases of athetesis, let alone outright omission, such decisions have occasionally led to the permanent loss of significant portions of the text. It is important, therefore, to outline the emerging patterns of distinctions between the editorial approaches of Crates and Aristarchus: [115] {218|219}
A) Both men athetized verses, but both included in their texts the verses that they athetized;
B) Aristarchus athetized more verses than Crates;
C) Aristarchus also omitted verses that Crates included, or, to put it differently, Crates had access to texts containing some verses that Aristarchus could not or would not verify on the basis of the texts to which he had access. A conventional term used nowadays for verses omitted by Aristarchus is “plus-verses”; [116]
D) Both men tracked variant readings within the verses, recording them and commenting on them in their formal commentaries (hupomnēmata / diorthōtika). Needless to say, the “same” verse containing different variant readings was a “different” verse as far as Aristarchus and Crates were concerned; [117]
E) Aristarchus used a system of signs, affixed at the left-hand margins of his text, to indicate his editorial differences with Crates and with his own Alexandrian predecessors (especially Zenodotus and Aristophanes of Byzantium). [118] One of these signs, the diplē (shaped “>“), was specifically used to mark verses where the athetesis of Aristarchus contradicted the non-athetesis of Crates. [119] {219|220}
The editorial differences between Crates and Aristarchus can be summed up in the microcosm of the following example: [120]
ἀλλὰ σύ, τὸν Ἀρίσταρχον ἀγαπῶν ἀεὶ καὶ θαυμάζων, οὐκ ἀκούεις Κράτητος ἀναγινώσκοντος
Ὠκεανός, ὅσπερ γένεσις πάντεσσι τέτυκται
ἀνδράσιν ἠδὲ θεοῖς, πλείστην <τ᾿>ἐπὶ γαῖαν ἵησιν.
But you are always so enamored of Aristarchus and so impressed with him that you do not hear [akouein] [121] Crates as he reads out loud [anagignōskein]: [122]
the Okeanos, genesis for all
men and gods, and it flows over the whole Earth.
The first of these two verses as quoted by Plutarch corresponds to Iliad XIV 246 (with verse-initial Ὠκεανοῦ, continuing the syntax of verse 245 as we have it), while the second, “XIV 246a,” has been omitted from the text proper of standard modern editions of the Iliad. [123]
For Crates (fragments 32-33 Mette), this additional verse provided evidence for the cosmic theory that the Okeanos was the salt sea that {220|221} covered the Earth, which was supposedly spherical; this theory was opposed by Aristarchus, who argued that the Homeric Okeanos was to be visualized as a fresh water river that surrounded a round and flat Earth. [124]
What was at stake, as far as Crates and Aristarchus were concerned, was, again, no trivial matter. In this case, the stakes were of cosmic dimensions. I find it ironic that I am writing this in an era when it appears fashionable to dismiss Homeric textual variations as “trivial,” “banal,” and even “boring.” [125]
Further ironies lie in the fact that modern editions of Homer ignore such textual variations as those that Crates reported and interpreted. Modern scholarship may well have singled out Crates for his scientific foresight in envisioning a spherical earth instead of a circular one, had he not based his reasoning on the text of Homer. As I have already observed, modern readers find it most difficult to envision an era of intellectual history when the prestige of all higher learning centered on the study of Homer. The modern association of Crates with Homeric criticism has even diminished his potential status as a literary critic. Yet there is enough evidence from what little survives of Crates’ Homeric criticism to acclaim him as a perceptive and sensitive literary critic, one whose interpretations equal, and perhaps even surpass, those of “Longinus’” On the Sublime. [126]
Even from the standpoint of Homeric criticism, the editorial decisions of Crates reflect a solid grounding in the textual evidence. We may often wish to disagree with his specific points of interpretation, but the textual variants that he adduces cannot be dismissed as mere inventions. From an empirical analysis of the formulaic composition of Iliad XIV 246a, for example, there is nothing untraditional about the form of this verse as {221|222} Crates adduces it. Nor is there anything untraditional per se about the contents. From a formulaic point of view, moreover, the verse does not even necessarily convey a vision of the salt sea—let alone a spherical earth, Crates argued. Where the verse does differ from other verses has to do with the immediate context of Okeanos. If we follow the sequence of Iliad XIV 246-246a, the inherent idea is that the Okeanos generates not ‘everything’ (which would be the theme expressed by XIV 246 minus XIV 246a), but, more specifically, all gods and men. There is nothing untraditional about this idea. In fact, it seems just as archaic in conventional outlook to picture the Okeanos as an anthropomorphic father of gods and men instead of a cosmic generator of the universe. [127]
In taking stock of the intellectual legacy of Crates as a textual critic of Homer, it is important to revisit the distinction that Varro made in De lingua latina 8.23 and 9.1 concerning the principles of “anomaly” and “analogy” in the thought of Crates and Aristarchus. This distinction has led to the assumption that Crates applied the principle of anomaly to admit practically any reading that suited his allegorizing theories. [128] A careful reexamination of the variants attributed to him, as in the case of the Iliadic passage on the Okeanos, leads to a more balanced perspective. [129] Similarly, the principle of analogy as Aristarchus practiced it has led to the assumption that Aristarchus would level out anomalous variant in the name of regularity. [130] Yet, it is possible to show that Aristarchus was scrupulous in avoiding regularization of variants at the expense of the manuscript evidence. [131]
In the long run, it is more important to appreciate the convergences rather than the divergences between the approaches of Crates and Aristarchus to Homeric textual criticism. In fact, the intellectual histories of the Aristarchean / Alexandrian and Cratetean / Pergamene “schools” reveal their own tendencies of eventual convergence. An ideal case in point is the Stoic Panaetius, who considered himself a pupil of Crates, [132] {222|223} and who at the same time praised Aristarchus as a mantis ‘seer’ who knows the dianoia ‘train of thought’ of Homer. [133] On the other hand, the Aristarchean pupil Dionysius Thrax was strongly influenced by the Cratetean approach to “the variety of forms in the spoken language, the sunētheia.” [134] Crates’ preeminence in analyzing linguistic “irregularity” was generally acknowledged in the ancient world. [135] His holistic approach to the Homeric corpus in all its potential “irregularities” can be viewed, I suggest, in the same light.
Crates’ approach allows for a “Homerus auctus,” a more expansive and therefore more unwieldy corpus. But we should not assume, however, that he made Homer be this way. Rather, I would prefer to say that the approach of his rival Aristarchus made Homer a less expansive and therefore less unwieldy corpus.
To be sure, Crates, like Aristarchus, athetizes, in some instances daringly, as when he athetizes the prooemia of both the Theogony and the Works and Days of Hesiod. Athetesis does not by itself, however, make a difference in the expansiveness of the Homeric text, since both Crates and Aristarchus leave in whatever they athetize. The difference is in the “plus-verses,” which Aristarchus leaves out but Crates leaves in. Again, we must keep in mind that Crates simply leaves things in: it is not that he puts things in. Conversely, however, we may say that Aristarchus not only leaves things out: he takes things out. I have already noted an illustration in Iliad XIV 246 as Aristarchus read it and in Iliad XIV 246-46a as Crates did.

Crates’ Homer and the “Peisistratean Recension”

I may note with interest the traditional orientation of Iliad XIV 246-46a as Crates read it: it appears Orphic, reflecting the poetic and hermeneutic traditions that the ancient world ascribed to Orpheus. [136] I {223|224} will now argue that Crates derived such verses from a “Homerus auctus,” a Homeric tradition augmented by Orphic traditions; I will argue, further, for the existence of “plus-verses” emanating from Homer editions that had been contaminated, as it were, by these Orphic traditions. As far as the Pergamene scholars were concerned, however, such Orphic accretions in the Homeric textual tradition were not “contaminations” at all, since Orpheus (as well as Musaeus) supposedly lived before Homer. [137] Pergamene editorial practice apparently included “Orphic” plus-verses in the Homeric text, as contrasted with the Alexandrian practice of excluding them—that is, omitting them altogether from the text, instead of merely athetizing them. Such conflicting editorial practices seem reflected in the following piece of ancient witticism:
annales evolvam omnium gentium et quis primus carmina scripserit quaeram? Quantum temporis inter Orphea intersit et Homerum, cum fastos non habeam, computabo? Et Aristarchi ineptias, quibus aliena carmina conpunxit, recognoscam et aetatem in syllabis conteram?
Shall I unroll the annals of the world’s history and try to find out who first wrote poetry? Shall I make an estimate of the number of years that separate Orpheus and Homer, although I do not have the records [fasti]? And shall I investigate the absurd writings of Aristarchus, wherein he ‘skewered’ [conpunxit] other men’s verses, and wear my life away on syllables? [138]
The ‘skewering’ refers to the Alexandrian procedure of marking with an obelos ‘skewer’ those verses that are to be athetized. [139] The mention of syllables may refer to the proverbial Aristarchean obsession with the minutiae of monosyllabic words, an obsession that some followers of Crates ridiculed. [140] Most important of all, the theme of an absence of {224|225} written recordings (fasti) of Orpheus and Homer reflects what appears to be an ongoing dispute between the followers of Crates and Aristarchus concerning the provenience of Homer.
It is easier to start with Aristarchus: he believed, it appears, that Homer was an Athenian who lived around 1000 BCE, [141] and that he wrote down the Iliad and Odyssey. [142] There is no indication, by contrast, that Crates deemed Homer to be an Athenian, let alone a writer of his own poems. [143] Instead, Crates seems to have followed the construct known as the Peisistratean Recension, according to which there were major lacunae in the written tradition of Homer until the Peisistratids reassembled the corpus.
The idea of a Peisistratean Recension, according to one expert, was “unknown in the heyday of Alexandrian scholarship.” [144] I would rather say “deliberately ignored” or at least underrepresented, not “unknown.” There is a parallel pattern in the underrepresentation of Crates. As Pfeiffer points out, the main surviving source for Crates is not the body of scholia of ms A of the Iliad (Venetus 454), which is our main source for Aristarchus, but the so-called exegetical scholia of ms B (Venetus 453), of ms T (Townley = British Museum Burnley 86), and the Genavensis, as well as the related scholia in P.Oxy. 221 and the scholia in mss HM of the Odyssey; “and particularly Eustathius, who was able to excerpt scholia lost to us.” [145] So also with the idea of a Peisistratean Recension: {225|226} the absence of references to this construct in the Iliadic A scholia, reflecting primarily the Aristarchean tradition, is remarkable. In the Homeric scholia, mentions of the Peisistratean Recension surface very rarely: only in scholia T at Iliad X 1 and scholia H at Odyssey xi 604. [146]
Scholia H to verse 604 of Odyssey xi report that this Homeric verse (xi 602-4) was the creation of one Onomacritus. According to Tzetzes, [147] Onomacritus was one of the four redactors responsible for the Peisistratean Recension. [148] Elsewhere, as in Pausanias 9.35.5, Onomacritus is a cover-name for the author of Orphic poems. [149] In general, the ultimate form of the Orphic poems “is unmistakably connected with the Pergamene account of the Pisistratean recension of the Homeric poems.” [150] The division of the Orphic corpus into 24 rhapsōidiai ‘rhapsodies’ seems to be an indication of this connection. [151]
Various scholars have argued that the “inventor” of the whole construct of the Peisistratean Recension was a Pergamene scholar, Asclepiades of Myrlea (fl. ca. 100 BCE), whose purpose was to provide a counterweight to the Alexandrian theories of Homer’s provenience. [152] By contrast, I have counterargued for a much earlier dating of the construct, roughly contemporaneous with the era of the Peisistratids themselves. [153] In support of this counterargument, we may note that {226|227} Onomacritus is mentioned already by Herodotus [154] as a collaborator of the Peisistratids, serving as a diathetēs ‘arranger’ of oracular poetry for their political purposes. In this context, moreover, Herodotus accuses Onomacritus of fraudulently making additions to the oracular poetry of Musaeus. [155] Such additions, of course, would have been athetized by the Pergamene scholars, although these atheteses would have remained in the text.
In view of the fact that Aristarchus in his own publications carried on an extended debate with Crates concerning the textual criticism of Homer, [156] it seems to me more likely that it was the Aristarchean construct of an Athenian Homer, who was supposedly writing his poems around 1000 BCE, that provided a counterweight to the construct of the Peisistratean Recension as accepted by Crates and the Pergamenes. [157] Asclepiades of Myrlea is simply one of the more visible later representatives of this earlier construct. [158]
If indeed Aristarchus’ construct of an Athenian Homer is an innovation meant to counter the older construct of the Peisistratean Recension, as Crates and the Pergamenes accepted, we can better understand the reasoning of Josephus when he says that the songs of Homer were not written down but remembered over time, then scattered, and finally reassembled:
ὅλως δὲ παρὰ τοῖς Ἕλλησιν οὐδὲν ὁμολογούμενον εὑρίσκεται γράμμα τῆς Ὁμήρου ποιήσεως πρεσβύτερον, οὗτος δὲ καὶ τῶν Τρωϊκῶν ὕστερος φαίνεται γενόμενος, καί φασιν οὐδὲ τοῦτον ἐν {227|228} γράμμασι τὴν αὑτοῦ ποίησιν καταλιπεῖν, ἀλλὰ διαμνημονευομένην ἐκ τῶν ᾀσμάτων ὕστερον συντεθῆναι καὶ διὰ τοῦτο πολλὰς ἐν αὐτῇ σχεῖν τὰς διαφωνίας·
In general, no commonly recognized writing is found among the Greeks older than the poetry of Homer. But he too seems to have been later than the Trojan War, and they say that not even he left his poetry in writing, but it was preserved by memory [diamnēmoneuomenēn] and assembled [suntethēnai] later from the songs. And it is because of this that there are so many inconsistencies [diaphōniai] in it. [159]
Evidently, Josephus was using the older construct of the Peisistratean Recension in order to undercut the ideology of his Aristarchean opponent, Apion. [160]

Homer and the Library of Nysa

The intellectual legacy of Crates of Mallos at the Library of Pergamon illuminates a central point about the agenda of ancient libraries in general: to possess an augmented text, even if the augmentation consists of athetized readings, is a mark of great prestige. This point is evident on a much smaller scale in the case of the Library of Nysa.
Let us consider the following autobiographical remark of Strabo, who recounts his having attended the lectures of Aristodemus, described as the son of a pupil of Aristarchus, at Nysa in Caria:
ἄνδρες δὲ γεγόνασιν ἔνδοξοι Νυσαεῖς Ἀπολλώνιός τε ὁ στωικὸς φιλόσοφος τῶν Παναιτίου γνωρίμων ἄριστος, καὶ Μενεκράτης Ἀριστάρχου μαθητής, καὶ Ἀριστόδημος ἐκείνου υἱός, οὗ διηκούσαμεν ἡμεῖς ἐσχατόγηρω νέοι παντελῶς ἐν τῇ Νύσῃ· καὶ Σώστρατος δὲ ὁ ἀδελφὸς τοῦ Ἀριστοδήμου καὶ ἄλλος Ἀριστόδημος ἀνεψιὸς αὐτοῦ ὁ παιδεύσας Μάγνον Πομπήιον ἀξιόλογοι γεγόνασι γραμματικοί· ὁ δ' ἡμέτερος καὶ ἐρρητόρευε καὶ ἐν τῇ Ῥόδῳ καὶ ἐν τῇ πατρίδι δύο σχολὰς συνεῖχε, πρωὶ μὲν τὴν ῥητορικὴν δείλης δὲ τὴν γραμματικὴν σχολήν· ἐν δὲ τῇ Ῥώμῃ τῶν Μάγνου παίδων ἐπιστατῶν ἠρκεῖτο τῇ γραμματικῇ σχολῇ. {228|229}
Distinguished men born at Nysa include: Apollonius the Stoic philosopher, best of the acquaintances of Panaetius; also Menecrates, pupil of Aristarchus; also Aristodemus, his son, whose entire course I took at Nysa when I was a young man; by then, he was at an extreme old age; also Sostratus, the brother of Aristodemus; also another Aristodemus, his cousin, who educated Pompey the Great. They were all grammatikoi of distinction. But my teacher [Aristodemus] also taught rhetoric. He ran two schools in Rhodes and two in his native land. He would teach rhetoric in the morning and grammatikē in the evening. At Rome, however, when he was in charge of the children of Pompey, he was content with teaching only grammatikē. [161]
It would be misleading, I submit, to translate grammatikē, the subject taught by Aristodemus, simply as ‘grammar’. In view of the intellectual confluences that this list of notables represents, including figures like Apollonius, a Stoic acquaintance of Panaetius, pupil of Crates, and Menecrates, pupil of Aristarchus, and his son Aristodemus, teacher of Strabo, it stands to reason that the prestige of higher learning as conveyed by grammatikē must have concentrated on the classics, with primacy of place reserved surely for Homer. [162]
Here it is relevant to adduce a fragment from the Kestoi of Julius Africanus (third century CE), in P.Oxy. 413: he tells of the Library of Nysa, specified as Nysa in Caria, [163] which features a Homer text containing a curious set of plus-verses, as do other Homer texts in other libraries listed by Julius Africanus, including the arkheia ‘archives’ in Jerusalem. These plus-verses appear in the Nekyia of the Odyssey, a set of lines where Odysseus invokes an Egyptian pandemonium: Isis, Anubis, and Phtha. [164]
It would be difficult to make a case for the value of such accretive verses for Homeric studies per se. What is of far greater value, however, is the {229|230} fact that these reports provide contextualized evidence for the ideology of an augmented Homer, with a library as the source and guarantor of the augmentation.

Epilogue: The “Vanished Ekdosis” of Aristotle’s Homer

Earlier, I considered the “vanishing” and then the subsequent reappearance of a corpus of Aristotle’s esoterica. But now I must also reckon with the “vanishing” of Aristotle’s ekdosis of Homer. Mentioned in the Vita Marciana, this ekdosis fails to resurface in the form of a continuing textual tradition that claims Aristotle as editor.
Here it may be useful to reconsider the standpoint of the brief report in the Vita Marciana: in the context of a listing of Aristotle’s own works, the mention of an ekdosis of Homer implies merely the existence of an edition produced by the Lyceum as a library, not necessarily a work authored by Aristotle as an individual. Here I return to the evidence for a massive transfer of both learning and texts from Lyceum to Museum in the era of Demetrius of Phaleron, who was closely associated with both the Lyceum and the Museum.
I have written extensively about this transfer, as also about the likely metamorphosis of a Peripatetic Homer text in the context of ongoing scholarship at the Library of Alexandria. [165] Here I offer simply an outline: once the ekdosis of the Lyceum becomes that of the Museum, the Homer text stops being Aristotelian per se. It becomes reshaped by the successive editorial practices of Demetrius, Zenodotus, Aristophanes of Byazantium, and, finally, Aristarchus. While earlier stages of editions may have promoted the idea that Aristotle himself had worked on the Homer text, as we see in the story about the narthex copy supposedly owned by Alexander the Great, the involvement of Aristotle seems to depend on his connection with Alexander himself, not so much on his Homeric research. [166] In the end, however, the focus of prestige for the Homer text of the Alexandrian Library is not on the mediation of Aristotle but on the authorship of Homer. Aristotle’s Homer is not all that compatible with Aristarchus’ ekdosis of Homer.
In one respect, however, the link between Aristotle and Aristarchus is unbroken. Cicero in De Natura Deorum 1.107 says that he believes {230|231} the claim of Aristotle (F 7 ed. Rose) that Orpheus never existed. [167] In rejecting Orpheus, Aristotle is more radical than the school of Pergamon: two centuries later, Crates’ Homer is still predicated on the old-fashioned construct of a Peisistratean Recension, apparently presupposing the earlier existence of Orpheus and Musaeus. According to this construct, the corpus of Homer disintegrates in Asia Minor, only to be reintegrated by Peisistratos in Athens.
Even if Aristotle may have resisted the idea of a Peisistratean Recension, I doubt that he would have approved of Aristarchus’ idea of an Athenian Homer. I doubt even more that Aristarchus would need as a model a fourth-century version of a Homeric ekdosis, even if the editor were Aristotle himself. This brings me back to Richardson’s skepticism about any ekdosis of Homer by Aristotle. [168] It seems that the word ekdosis itself is the stumbling block. Still, Richardson does acknowledge the relationship between Lyceum and Museum, [169] as well as that between the actual libraries. [170] He even mentions Demetrius of Phaleron as “the crucial link.” [171] Then, starting from the death of Demetrius, Richardson proceeds toward a sort of Phase II, embarking on his valuable outline of Lyceum-Museum continuities as surveyed at the beginning of this inquiry. [172] The text of Homer, however, has vanished in the survey. After the survey, Richardson returns to the topic of Homer, this time in terms of Homeric scholarship in general. [173] Then we finally return to the text of Homer:
Given the fluidity of the Homeric text during the Classical period, it is not so surprising that Zenodotus, the pioneer of Homeric textual work [emphasis mine], should have exercised such a free hand in preparing his edition of the poems. [174] {231|232}
The very idea of an Aristotelian ekdosis of Homer seems to have vanished in Pergamon as well: if this idea was not of much use for later Alexandrians like Aristarchus, it was of even less use to the Pergamenes, whose holistic approach to the text of Homer seems to have embraced even Orpheus and Musaeus.
In two ways, however, the scholars of the Library of Pergamon lived up to the ideals of Aristotle in their analysis of classical texts, even if not in their editing of Homer. First, they were systematic, as Suetonius reports explicitly. [175] Second, they were comprehensive, to the point that their cataloguing system surpassed the Library of Alexandria in the popular imagination. Witness the following remark of Athenaeus 8.336d, on a comedy of Alexis: ‘Certainly neither Callimachus nor Aristophanes has catalogued it, nor have even those who compiled the catalogues in Pergamon’. [176]
I end where I began, with Dio’s description of the scholars of both Alexandria and Pergamon:
Those who make exēgēsis of the meaning (dianoia) of Homer—not only Aristarchus and Crates and several others of those who were later called grammatikoi but who had earlier been called kritikoi, and especially Aristotle himself, from whom they say that kritikē and grammatikē have their origin. [177]
The holism of the Library of Pergamon, as represented by Crates of Mallos, surely measured up to the standard set by Aristotle. [178]


Allen. R. E. 1983. The Attalid Kingdom: A Constitutional History. Oxford.
Allen, T. W. 1924. Homer: The Origins and the Transmission. Oxford.
Apthorp, M. J. 1980. The Manuscript Evidence for Interpolation in Homer. Heidelberg.
Apthorp, M. J. 1995. Review of J. R. Tebben 1994. Concordantia Homerica, Pars I: Odyssea. A Complete Concordance to the Van Thiel Edition of Homer’s Odyssey 2 vols. (Zurich and New York NY). Classical Review 45:221-222.
Baratin, M., and Jacob, C. 1996. Le pouvoir des bibliothèques. Paris.
Barnes, J. 1997. “Roman Aristotle.” Philosophia Togata II: Plato and Aristotle at Rome (ed. J. Barnes and M. Griffin) 1-69. Oxford.
Bloch, M. 1977. “The Disconnection Between Power and Rank as a Process.” Archives européennes de sociologie 18:107–148.
Blum. R. 1991. Kallimachos: The Alexandrian Library and the Origins of Bibliography. Madison.
Bourdieu, P. 1972. Esquisse d’une théorie de la pratique. Paris.
Boyd, T. W. 1995. “Libri Confusi.” Classical Journal 91:35-45. This is Boyd 1995b in my master bibliography.
Canfora, L. 1990. The Vanished Library: A Wonder of the Ancient World. Berkeley and Los Angeles. Translated by M. Ryle from La biblioteca scomparsa (1987).
Canfora, L. 1996. Il viaggio di Aristea. Bari.
Davison, J. A. 1955. “Peisistratus and Homer.” Transactions of the American Philological Association 86:1-21.
Ford, A. 1992. Homer: The Poetry of the Past. Ithaca NY.
Fraser, P. M. 1972. Ptolemaic Alexandria. I-III. Oxford.
Hansen, E. V. 1971. 2nd ed. The Attalids of Pergamon. Ithaca.
Heath, M. 1997. “Polymorphous Homer.” Classical Review 47: 241–242.
Helck, H. 1905. De Cratetis Mallotae studiis criticis quae ad Iliadem spectant. Leipzig.
Heubeck, A., West, S., and Hainsworth, J. B. 1990. eds. A Commentary on Homer’s Odyssey. Vol I: Introduction and Books I-VIII. Oxford.
Hoek, A. van den. 1989. “The Concept of σῶμα τῶν γραφῶν in Alexandrian Theology.” Studia Patristica 19:250-254.
Irigoin, J. 1994. “Les éditions de texts.” In Montanari 1994:39-82.
Jacob, C. 1996. “Lire pour écrire: navigations alexandrines.” Le pouvoir des bibliothèques (ed. M. Baratin and C. Jacob) 47–83. Paris.
Janko, R. 1992. The Iliad: A Commentary IV (general ed. G. S. Kirk). Cambridge MA.
Jensen, M. Skafte 1980. The Homeric Question and the Oral-Formulaic Theory. Copenhagen.
Kassel, R. 1973. “Antimachos in der Vita Chisiana des Dionysios Periegetes.” Catalepton: Festschrift für Bernhard Wyss zum 80. Geburtstag (ed. Ch. Schäublin) 69-76. Basel.
Keaney, J. J., and Lamberton, R. 1996. eds. [Plutarch] Essay on the Life and Poetry of Homer. APA American Classical Studies 20. Atlanta GA.
Kuttner, A. 1995. “Republican Rome looks at Pergamon.” Harvard Studies in Classical Philology 97:157-178.
Leppert R. and Lincoln, B. 1989. “Introduction” to the special issue “Discursive Strategies and the Economy of Prestige.” Cultural Critique 12:5–23.
Lindsay, H. 1997. “Strabo on Apellicon’s Library.” Rheinisches Museum 140: 290-298.
Ludwich, A. 1884 / 1885. Aristarchs Homerische Textkritik nach den Fragmenten des Didymos I / II. Leipzig.
McNamee, K. 1981. “Aristarchus and Everyman’s Homer.” Greek, Roman and Byzantine Studies 22:247-255.
McNamee, K. 1992. Sigla and Select Marginalia in Greek Literary Papyri. Brussels.
Monro, D. B., and Allen, T. W. 1920. ed. Homeri Opera. ed. 3. Oxford.
Montanari, F. 1994. ed. La philologie grecque à l’époque hellénistique et romaine. Entretiens sur l’antiquité classique XL, Fondation Hardt. Vandoeuvres-Geneva.
Nagy, G. 1989 (revised 1993). “Early Greek Views of Poets and Poetry.” In Cambridge History of Literary Criticism I (ed. G. Kennedy) 1-77. Cambridge MA.
Nagy, G. 1989b. “The ‘Professional Muse’ and Models of Prestige in Ancient Greece.” Cultural Critique 12:133-143. Rewritten as part of Ch.6 in Nagy 1990.
Nagy, G. 1990. Pindar’s Homer: The Lyric Possession of an Epic Past. Baltimore MD. <<>> (1997).
Nagy, G. 1996a. Poetry as Performance: Homer and Beyond. Cambridge.
Nagy G. 1996b. Homeric Questions. Austin TX.
Nagy, G. 1996c. “Aristocrazia: Caratteri e stili di vita,” I Greci: Storia, Cultura, Arte, Società (ed. S. Settis) 577-598. Torino.
Nagy, G. 1997. “Homeric Scholia.” A New Companion to Homer, In Morris and Powell 1997:101-122.
Nickau, K. 1977. Untersuchungen zur textkritischen Methode des Zenodotos von Ephesos. Berlin / New York NY.
Parsons. E. A. 1952. The Alexandrian Library, Glory of the Hellenic World: Its Rise, Antiquities, and Destructions . New York.
Payne, M. 1991. “Alexander the Great: Myth, the Polis, and Afterward.” Myth and the Polis (ed. D. C. Pozzi and J. M. Wickersham) 164–181. Ithaca.
Pelliccia, H. N. 1997. “As Many Homers As You Please.” New York Review of Books 1997.20.11; 44.18:44-48.
Pfeiffer, R. 1968. History of Classical Scholarship from the Beginnings to the End of the Hellenistic Age. Oxford.
Porter, J. I. 1992. “Hermeneutic Lines and Circles: Aristarchus and Crates on the Exegesis Of Homer.” In Homer’s Ancient Readers: The Hermeneutics Of Greek Epic’s Earliest Exegetes (ed. R. Lamberton and J. J. Keaney) 67-114. Princeton NJ.
Rengakos, A. 1993. Der Homertext und die Hellenistischen Dichter. Hermes Einzelschriften 64. Stuttgart.
Richardson, N. J. 1994. “Aristotle and Hellenistic Scholarship.” In Montanari 1994:7-28.
Robert, L. 1940. “Bibliothèque de Nysa.” Hellenica 1:144–148.
Schenkeveld, D. M. 1994. “Scholarship and Grammar.” In Montanari 1994:263-298.
Thiel, H. van. 1996. ed. Homeri Ilias. Hildesheim.
Urios-Aparisi, E. 1993. “Anacreon: Love and Poetry (on 358 PMG, 13 Gentili).” Quaderni Urbinati di Cultura Classica 44:51-70.
Wachsmuth, C. 1860. De Cratete Mallota. Leipzig.
Weber, M. 1922 (4th ed. 1956). Wirtschaft und Gesellschaft . Tübingen.
West, M. L. 1966. ed. (with commentary). Hesiod: Theogony. Oxford.
West, M. L. 1978. ed. (with commentary). Hesiod: Works and Days. Oxford.
West, M. L. 1983. The Orphic Poems. Oxford.
Zetzel, J. E. G. 1983. “Re-creating the Canon: Augustan Poetry and the Alexandrian Past.” Critical Inquiry 10:83-105. Also in Canons (ed. R. von Hallberg 1984) 107-129. Chicago.


[ back ] 1. Pfeiffer 1968.
[ back ] 2. Pfeiffer 1968:117, 204, 206-207, 242, 269. Compare Nagy 1989, esp. p. 1. For a more detailed discussion, see Nagy 1990a:61-62, 85, and 402-403 (with special reference to the wording of Plato).
[ back ] 3. Pfeiffer 1968:206-207; compare Horace Odes I 1.35 and the comments of Pfeiffer (206).
[ back ] 4. See Pfeiffer 1968:207, tracing the modern usage back to David Ruhnken, who used the word ‘canon’ in his 1768 edition of Rutilius Lupus; for the ancient usage of the original Greek word kanōn (‘weaver’s rod; carpenter’s rule; model’), Pfeiffer draws attention to Eusebius (Historia Ecclesiastica 6.25.3), who speaks of Origen’s observance of the ecclesiastical kanōn by way of recognizing only four gospels.
[ back ] 5. Strabo 14.2.19 C657. Pfeiffer 1968:89.
[ back ] 6. Pfeiffer 1968:157-158, 259. According to Pfeiffer (158), a probable source for the philosophical distinctions that had led to the displacement of the term kritikoi by grammatikoi was Asclepiades of Myrlea (second-first century BCE), whose crucial role in the transmission of information about the Library of Pergamon I will discuss further below. On the term philologoi, see Suetonius De Grammaticis et Rhetoribus 10 (who attributes the formalization of the term to Eratosthenes), along with the comments by Nagy 1996b:3, 150.
[ back ] 7. Pfeiffer 1968:159, 238, 242 (with special reference to Sextus Empiricus Adversus mathematicos 1.79),
[ back ] 8. Sextus Empiricus Adversus mathematicos 1.79: τὸν μὲν κριτικὸν πάσης, φησί, δεῖ λογικῆς ἐπιστήμης ἔμπειρον εἶναι, τὸν δὲ γραμματικὸν ἁπλῶς γλωσσῶν ἐξηγητικῶν καὶ προσῳδίας ἀποδοτικὸν καὶ τῶν τοιούτων παραπλησίων εἰδήμονα. Compare Pfeiffer 1968:242n8. See also Sextus Empiricus 1.248, where he ascribes to Crates’ student Tauriscus this further elaboration: that grammatikē is subordinate to kritikē, which consists of three parts, namely, (1) logikon, (2) tribikon, and (3) historikon. He associates these parts with the study of (1) diction [lexis] and ‘grammatical variations’ [grammatikoi tropoi]; (2) distinctions in styles of speaking; (3) subject-matter that is as yet unsystematized [amethodos].
[ back ] 9. Dio Chrysostom Orations 53.1 (speech on Homer). See Pfeiffer (1968:67), who objects to this formulation; for a defense, see below.
[ back ] 10. Nagy 1990a:61n52 (also 83n3), following Zetzel 1983.
[ back ] 11. For the basic information on the Pinakes of Callimachus, see Pfeiffer 1968:127-134. Pfeiffer does not stress, however, the working distinction between the concept of a canon and the concept (not just the reality) of a complete collection. This distinction is the point of Zetzel’s argument, cited above. See also Nagy 1990a:83n3: “For the Alexandrian scholars, exclusion of an author from the canon does not preclude an active interest in that author, even as a model for imitation.”
[ back ] 12. Pfeiffer 1968:88-89.
[ back ] 13. See especially Richardson 1994; see also the articles, in the same volume (Montanari 1994), by Irigoin and Schenkeveld; also the comments of Montanari himself (29-31). Richardson (8) defends the formulation of Dio Chrysostom, quoted above, despite the objections of Pfeiffer 1968:67.
[ back ] 14. Richardson 1994:14-17. From here on in my inquiry, I prefer never to use the terms ‘grammar’ for grammatikē (tekhnē) and ‘grammarian’ for grammatikos, since they fail to do justice to the science of grammatikē as defined by Dionysius Thrax himself in the first sentence of his Ars grammatica (ed. Uhlig): γραμματική ἐστιν ἐμπειρία τῶν παρὰ ποιηταῖς τε καὶ συγγραφεῦσιν ὡς ἐπὶ τὸ πολὺ λεγομένων ‘grammatikē is the empirical knowledge of what is for the most part being said by poets and prose writers’ (5.1-2). See Pfeiffer 1968:268-269.
[ back ] 15. Richardson 1994:17.
[ back ] 16. Richardson 1994:17.
[ back ] 17. Richardson 1994:9, with reference to Pfeiffer 1968:71-72.
[ back ] 18. Aristotle F 427.5 (ed. Rose). For a survey of the various ancient references to an ekdosis of Homer by Aristotle, see Pfeiffer 1968:71-72.
[ back ] 19. Nagy 1996a:115-116.
[ back ] 20. Nagy 1996a:116-119, with extensive bibliography.
[ back ] 21. Nagy 1996a:119-122. For arguments showing that Zenodotus as “editor” of Homer did indeed make public his own version of the Homeric text, see also Rengakos 1993:12-14.
[ back ] 22. Nagy 1996a:135-152, 192-193. I argue that such “editorial” activities in the era of Aristotle seem to have involved not only Homer but even other classics, like Alcaeus and Sappho. I invite the reader to give a fair hearing to the argumentation in these pages before taking seriously the assertions of Heath 1997. I question many of Heath’s assumptions, as when he claims that Aristotle’s reports of prosodic solutions to Homeric problems “presuppose that the current [that is, in Aristotle’s time] oral realization of the text might have arisen from a false interpretation of an ambiguous written [his emphasis] text.” Here his non-technical use of the term “oral” in the vague sense of “performed” leads to a serious confusion. To say “the oral realization of the text” is already a contradiction in terms - if we are to use “oral” in the technical sense of composed-in-performance. Moreover, “false” interpretations can indeed happen in oral poetics: a composer-in-performance can “misunderstand” and thereby reshape the wording of the formulas that he uses. If a performer “mistakes,” say, δίδομεν for διδόμεν or vice versa, we cannot simply assume that his mistake is due to the wrong reading of a text. In an era like Aristotle’s, when texts were still being written without accent-markings, the “correction” of a Homeric accent could apply independently to rhapsodes as well as to readers of any transcript of Homeric poetry.
[ back ] 23. Nagy 1996a:121-122, 186, 200-206, where I examine the reports about an ekdosis or diorthōsis of Homer by Aristotle, as attested respectively in the Vita Marciana and in Plutarch Life of Alexander 8.2. Both these reports specify the Iliad, but Strabo (13.1.27 C595) speaks more generally of ‘a diorthōsis of Homer’s poetry’ (διόρθωσις τῆς Ὁμήρου ποιήσεως), adding that the traditional name for this diorthōsis is ‘the one from the narthēx’ (ἡ ἐκ τοῦ νάρθηκος λεγομένη). Clearly, the noun diorthōsis is used here to designate not the process of ‘correcting’ that results in the definitive text but the result itself, the text; Strabo (again, 13.1.27 C595) goes on to say that Alexander himself ‘put some markings’ (σημειωσαμένου τινά) into this text, and so too did the scholars Callisthenes and Anaxarchus. The implication, I suggest, is that the markings of Alexander continue the markings of Aristotle, who is earlier designated in this same context as the one who performs the diorthōsis. Also in this same context, Alexander is described as philologos and philanagnōstēs, a lover of language and reading out loud. Finally, Alexander is said to have deposited (καταθέντος) this Homeric text into the ‘Receptacle’, the so-called narthēx, which he considered the most prized of the treasures confiscated from Darius. By contrast with Strabo’s reference to ‘the diagnōsis of Homer, the one called ‘the one from the narthēx’, Plutarch, Life of Alexander 8.2, refers to ‘the Iliad from the narthēx’ (τὴν μὲν Ἰλιάδα . . . ἔλαβε μὲν Ἀριστοτέλους διορθώσαντος ἣν ἐκ τοῦ νάρθηκος καλοῦσιν). Heath (1997:242n24) objects to my not distinguishing between Homeric poetry and the Iliad as the object of Aristotle’s diorthōsis. The interchangeability is already attested in Strabo and Plutarch.
[ back ] 24. Here I am following the general outlines of Max Weber’s sociological models: see especially Weber [1922]. On wealth and prestige in terms of an “economy of prestige,” see in general Bourdieu 1972. See also Leppert and Lincoln 1989, esp. pp. 6-8. On power and prestige, compare Bloch 1977.
[ back ] 25. On aristocratic ideologies of power, wealth, and prestige in ancient Greek culture, see Nagy 1996c.
[ back ] 26. Nagy 1989b.
[ back ] 27. Strabo 17.1.8 C793-94. He does not explicitly equate the Museum with the space that serves as the setting for the storage of books. That equation, however, is inferred by most modern scholars (cf. nn. 33 and 72).
[ back ] 28. The manuscript tradition of Strabo has Σῶμα, but modern editors conventionally emend to Σῆμα. My arguments against the emendation are developed below.
[ back ] 29. Compare Vitruvius De Architectura 5.11.12.
[ back ] 30. A feature of traditional Greek hero-cults is the idea that a hero’s corpse is a talisman, and that possession of the body is therefore a key to power, wealth, and prestige: see Nagy 1990a:178.
[ back ] 31. For the specialized translation of komizein here as ‘bring home’, not just ‘bring’ see the next note. The component ‘home’ is “subjective,” reflecting the standpoint of the speaker.
[ back ] 32. Strabo 17.1.8 C793-794. On the ex-post-facto concept of Alexander’s tomb at Alexandria as his ‘home’ (oikos), see the Alexander Romance 3.24.4 (also 1.33.9) and the comments in Nagy 1990a:271-272, especially with reference to the traditional use of oikos to designate the tomb of a cult hero.
[ back ] 33. Compare Fraser 1972 I 325. Fraser’s own usage in his book often blurs the distinction between Library and Museum. On the synecdoche of canon as a model, see Nagy 1989:1.
[ back ] 34. Nagy 1996a:196-197; compare Canfora 1996:7-8.
[ back ] 35. As I have already noted, the manuscript tradition of Strabo has Σῶμα, but modern editors conventionally emend to Σῆμα. I submit that Σῶμα is the lectio difficilior. (I now retract my reading Σῆμα in Nagy 1990a:272n110.) In the Alexander Romance 3.34.5, it is made explicit that the taphos ‘tomb’ of Alexander was named the σῶμα Ἀλεξάνδρου ‘the Sōma of Alexander’. The traditional name of Alexander’s tomb works as a metonymy (specifically, a synecdoche), whereby the body (sōma) par excellence is the tomb (sēma). The phonetic parallelism of sōma / sēma serves to reinforce the metonymic device. On the Orphic identification of sōma and sēma, see Plato Cratylus 400c. Compare Payne 1991, esp. pp. 174-175.
[ back ] 36. Cicero borrows the Greek word sōma in referring to a ‘corpus’ of books (Letters to Atticus 2.1.4; Letter to Luceius ). Note too the expression τὸ σῶμα τῶν γραφῶν ‘the corpus of scriptures’, drawn into a parallel with τὸ σῶμα τὸ Μωυσέως ‘the body of Moses’ in Clement of Alexandria Stromateis 6.132.2-3. For these and related earlier examples (Philo Vita Contemplativa 78 and The Assumption of Moses 93), see Hoek 1989.
[ back ] 37. Compare Iliad XIII 196, telling how the Achaeans κόμισαν ‘brought’ (komizein) the corpse of Amphimakhos. Subjectively, they ‘brought it back to their side’, which is here by necessity the substitute for ‘brought it home’.
[ back ] 38. In the Platonic Hipparkhos (228b-c), it is said that one of the Peisistratids of Athens, Hipparkhos, ἐκόμισεν ‘brought home’ (from komizein) the epē ‘poetic words’ of Homer to Athens. As in the previous case of the Alexander narrative of Strabo, the component ‘home’ is here, too, “subjective,” reflecting the standpoint of the speaker. In this case, the speaker is Athenian, and so ‘bringing home’ is the same as ‘bringing to Athens’.
[ back ] 39. Nagy 1990a:58-61, 77n21, 188-189. When I say “production” I include composition-in-performance as well as composition-for-performance.
[ back ] 40. Nagy 1990a:82-84. Performance, not just writing, can be an aspect of preservation.
[ back ] 41. See Fraser (1972 I 316, 320, and esp. 325) on the Lyceum as a model for the Museum; also his comments (I 314) on the will of Theophrastus.
[ back ] 42. Strabo 13.1.54 C608.
[ back ] 43. Athenaeus 1.3a-b. Richardson (1994:11), following Pfeiffer (1968:7-8), considers this particular piece of “information” to be “suspect.” Compare Fraser 1972 II 473n100. On Neleus, see also Diogenes Laertius 5.62.
[ back ] 44. Compare the title of the book, Canfora 1990: The Vanished Library: A Wonder of the Ancient World.
[ back ] 45. The bibliography concerning this contradiction is vast: for a most useful survey of the whole problem, with extensive citations of secondary sources, see Barnes 1997. For an earlier discussion, see Blum 1991:52-64. Also Lindsay 1997.
[ back ] 46. Strabo 13.1.54 C609.
[ back ] 47. Strabo 12.3.16 C548. The pitfalls of translating grammatikos as ‘grammarian’, which is how modern handbooks conventionally describe Tyrannion, will become more apparent as the discussion proceeds.
[ back ] 48. Suda s.v. Τυραννίων. Pfeiffer 1968:266, 272-273. See also Lindsay 1997:297, esp. n. 35.
[ back ] 49. Pfeiffer 1968:273; compare 264 n. 3 on the testimony of Porphyry (Life of Plotinus 24). Strabo mentions Andronicus (14.2.13 C655); see Lindsay 1997:296 n. 32.
[ back ] 50. As we have seen from previously surveyed contexts, the verb komizein implies that whatever has been ‘brought’ to a place belongs there ex post facto, as it that place had always been ‘home’.
[ back ] 51. For more on the pinakes of Andronicus, see Barnes 1997:20-66; compare Blum 1991:55, 63, 93, esp. 194-196.
[ back ] 52. Athenaeus 5.214d-e.
[ back ] 53. FGrH 87 F 36. It is significant that this report comes from a Stoic source: see below.
[ back ] 54. Blum 1991:61-64.
[ back ] 55. Athenaeus 1.3a-b.
[ back ] 56. Nagy 1990a:157-162.
[ back ] 57. Nagy 1990a:157-162; also 168-169.
[ back ] 58. For background on this construct, see Nagy 1996a:69-71, 77-81.
[ back ] 59. “Plato” Hipparkhos 228c.
[ back ] 60. καὶ τὸν Πολυκράτεα τυχεῖν κατακείμενον ἐν ἀνδρεῶνι, παρεῖναι δέ οἱ Ἀνακρέοντα τὸν Τήιον ‘and he found Polycrates reclining in the men’s quarters, in the company of Anacreon of Teos’. On the explicitly sympotic features of the description given by Herodotus, see Urios-Aparisi 1993, esp. p. 54.
[ back ] 61. Nagy 1996a:70n34.
[ back ] 62. Athenaeus 1.3a-b.
[ back ] 63. Athenaeus 14.620b-c. Analysis of this most difficult Athenaeus passage in Nagy 1996a:157-161.
[ back ] 64. Analysis of this Plutarchean passage in Nagy 1996a:174-177.
[ back ] 65. Galen 17.1.607-608, Commentary on the Hippocratic Epidemiai Book 3.2.4.
[ back ] 66. Pfeiffer 1968:82.
[ back ] 67. Extensive commentary appears in Nagy 1996a:158-161; compare 81 on the parallelism (and mutual assimilation) of tragedy and epic as the two premier genres of the performing arts in Athens.
[ back ] 68. This correlation is explored at length in Nagy 1996a:161-186.
[ back ] 69. On the traditional role of the Muses as the divinities who preside over the politics of performance, see Nagy 1989b.
[ back ] 70. Compare Fraser 1972 I 334-335, II 479-80, 493-494.
[ back ] 71. For comparative perspectives on the term “virtual library,” see Jacob 1996, esp. p. 62.
[ back ] 72. Compare Canfora 1990:141: the bibliothēkē should be understood as comprised of “all the bookshelves located in the Museum precincts.”
[ back ] 73. Compare Jacob 1996:56-69.
[ back ] 74. Compare Nagy 1989:1-2.
[ back ] 75. Vitruvius Preface to De Architectura 7.4-7 (ed. Fensterbusch). For an illuminating discussion of Pergamene cultural influences on this preface, see Kuttner 1995, esp. pp. 164-65.
[ back ] 76. This passage of Vitruvius is quoted in full in Nagy 1996a:227-228.
[ back ] 77. The wording perlegeret emphasizes the idea of reading a text from beginning to end.
[ back ] 78. I infer from the wording that ‘the first ordo’ of the festival was the poetic competition (the narrative also mentions athletic competitions). My point is that there is a sequencing of the competitors also within the ordo.
[ back ] 79. Compare Jacob 1996 (80n79), who notes that the logic of the story points to a hiatus between the absolute memory of the Library and the amnesia of the public.
[ back ] 80. “Plato” Hipparkhos 228b-c. General commentary on the whole passage appears in Nagy 1996b:80-81.
[ back ] 81. Nagy 1996b:86, 91.
[ back ] 82. Nagy 1996b:90-91.
[ back ] 83. Nagy 1996b:69-75.
[ back ] 84. Nagy 1996b:69-75.
[ back ] 85. Nagy 1996b:67-75.
[ back ] 86. Cicero De Oratore 3.137. On this passage, see Boyd 1995b. As is evident from my own work, I agree with Boyd that the “Peisistratean Recension” is a construct, not a historical reality. I also agree that the realities of oral poetics are incompatible with the purported realities of the construct. Still, I suggest that the realities of oral poetics are indeed compatible with the construct as an applied metaphor; see Nagy 1996b:67-75, 93-106.
[ back ] 87. The relevant passages are quoted by Allen 1924:230-232. See also Porter 1992:67-68.
[ back ] 88. In this connection, see Greek Anthology 11.442, featuring “Peisistratos” as speaker: ὃς τὸν Ὅμηρον ἤθροισα, σποράδην τὸ πρὶν ἀειδόμενον ‘I who gathered together Homer, who was previously being sung here and there, scattered all over the place’. (This epigram is also quoted by the scholia to Dionysius Thrax, in the context of the narrative cited in the previous note.) See also Pausanias 7.26.13.
[ back ] 89. On parchment (the word itself is derived from “Pergamon”), see the definitive discussion of Pfeiffer 1968:236.
[ back ] 90. See in general Baratin and Jacob 1996.
[ back ] 91. Compare Hansen 1971:47, 316.
[ back ] 92. Parsons 1952:22. On the Attalids in general, see Allen 1983.
[ back ] 93. Pfeiffer 1968:235; Parsons 1952:22.
[ back ] 94. Parsons 1952:23.
[ back ] 95. Diogenes Laertius 60.1: τὰς εἰκόνας πόρρωθεν δεῖν θεωρεῖσθαι. I infer Platonic implications of Form and Copy in this jeu d’esprit. It raises questions of virtual reality about Pergamon, library and all.
[ back ] 96. Pfeiffer 1968:235.
[ back ] 97. Pfeiffer 1968:235; he also discusses here the dating of Eumenes II of Pergamon.
[ back ] 98. Parsons 1952:23. Compare Fraser 1972 II 673. Compare Pfeiffer 1968:235: “and then the Stoics came.”
[ back ] 99. Strabo 13.4.2 C624.
[ back ] 100. Parsons 1952:23.
[ back ] 101. For background on the ancient analogy/anomaly debate, see Schenkeveld 1994:286-287 n. 13.
[ back ] 102. For background on the research of Crates on Homer, see Pfeiffer 1968:239-241. Also Porter 1992.
[ back ] 103. Vita Romana 32.2 (ed. Wilamowitz). There is further testimony of great interest, concerning a three-verse alternative to the standard nine-verse beginning of the Iliad known also from the medieval manuscript tradition of Homer. It is cited on the authority of Aristoxenus. In another work, I hope to discuss in detail this variant as the Peripatetic scholar quotes it. It is important in this context to distinguish the two quoted hymnic prooemia from the beginning of the Iliad as we have seen it. See Pfeiffer (1968:239n7) for an important correction of a typographical error in the edition of Wilamowitz.
[ back ] 104. With respect to Nicanor, I infer this later Alexandrian scholar (era of Hadrian) is referring to Crates as his source for this particular verse.
[ back ] 105. On the morphology of the prooemium in archaic Greek poetics, see Nagy 1996a:62.
[ back ] 106. West 1966:150; compare Ford 1992:26-27. I suggest, further, that even the wording memnētai ‘he mentions’ in the Vita Romana implies that Crates did not fully legitimize this one-verse prooemium in his text proper.
[ back ] 107. Nagy 1996a:134, 138, 146-147, 182.
[ back ] 108. Ford 1992:26.
[ back ] 109. West 1966:150; West 1978:137. Pausanias 9.31.4 reports that he saw at Helicon an archaic text of the Hesiodic Works and Days engraved on a lead tablet. Since he also says that the Heliconians accept as authentically Hesiodic only the Works and Days—but without the prooemium—one may infer that the lead tablet featured no prooemium.
[ back ] 110. Pfeiffer 1968:241; Porter 1992:98; West 1966:50, 150; West 1978:65-66, 137. See also the Life of Dionysius Periegetes, 59-60, 72 in Kassel 1973, esp. p. 72.
[ back ] 111. Pfeiffer 1968:220. It looks as if Aristarchus based his athetesis partly on the fact that the prooemium was missing from a copy of the Works and Days that Praxiphanes found (see Pfeiffer 1968:220 n. 2).
[ back ] 112. Pfeiffer 1968:220; also West 1978:64-65, 364.
[ back ] 114. West 1966:50. He infers that the twenty-four-book division of the Iliad and Odyssey was already in place in the era of Aristophanes; he says (50n3) about this division: “it may have been pre-Alexandrian.” See now Nagy 1996a:182 n. 107.
[ back ] 115. For a general survey, see Fraser (1972 II 674) with further references to Pfeiffer (1968:239-242). Pfeiffer sums up (242): “Even this short survey of scanty evidence gives the impression that Crates was a serious scholar capable of displaying solid learning who did not disregard the results of previous research, even though it was the work of scholars who were his opponents in principle.” See also in general Wachsmuth 1860.
[ back ] 116. A most reliable treatment is that of Apthorp 1980. I disagree with Apthorp, however, when he infers that Homeric plus-verses are spurious or not authentic from the standpoint of oral poetics (see Nagy 1996a:138-152).
[ back ] 117. The editorial disagreements of Crates and Aristarchus extend to the textual tradition of Hesiod and other classics as well. For an illuminating example, see the discussion of the variant readings mentioned by Crates (and also Zenodotus) at Hesiod Theogony 5, as summarized by West (1966:153-154). I do not share, however, West’s assumptions about the methodology of Crates: for example, West (208) accuses Crates, without justification, of making up a verse (a variant of the medieval manuscript reading adopted by West for Theogony , 142).
[ back ] 118. See in general McNamee 1992. See also McNamee 1981. See also Fraser (1972 II 672n164) on Diogenes Laertius 3.66, who reports on the practice of charging fees from readers who wished to have access to texts featuring margins marked up with signs.
[ back ] 119. See Helck 1905:50. He argues that Aristarchus in general was disputing the decisions of Crates, not the other way around: a clear example is the statement of Aristonicus (by way of the scholia) concerning the diplē attached to Iliad XX 7; compare also Wachsmuth 1860:30. Helck (1905:66) focuses on another instance of a diplē marking an Aristarchean disagreement with Crates (this one is not discussed by Wachsmuth), at Iliad XXI 323, where the argumentation of Crates seems more persuasive than that of Aristarchus. See also the summation by Helck at 1905:76, where he argues in general that Aristarchus tended to go on record as disagreeing with Crates, not the other way around, and that instances of counterarguments generally involve Crates’ pupils, not Crates himself.
[ back ] 120. Plutarch De facie in orbe lunae 938d.
[ back ] 121. For akouein ‘hear’ in the sense of ‘have something read out loud’, see Nagy 1996a:33 n. 94 (with reference to Aelian De natura animalium 5.38).
[ back ] 122. For anagignōskein ‘read out loud’ in the technical sense of an editorial speech-act, see Nagy 1996a:149-150, 174-177, esp. 175-176 n. 83.
[ back ] 123. For example, XIV 246a is relegated to the apparatus criticus of the Oxford classical Text of Monro and Allen (1920). It is missing altogether from the Iliad edition of van Thiel (1996).
[ back ] 124. Wachsmuth 1860:44; Porter 1992:88-103, esp. 92 (on Crates’ interpretation of the shield of Achilles in Iliad XVIII 481-489 as an imago mundi).
[ back ] 125. Pelliccia 1997: “trivial” (45) and “banal” (46). Citing as his authority Stephanie West (Heubeck 1988 I 40) Pelliccia asserts (46): “The major problem for Nagy’s new theory is simply that the variant recordings that we know of from the papyri and the indirect sources mentioned earlier are for the most part too boring and insignificant to imply that they derived from a truly creative performance tradition.”
[ back ] 126. An effective argument for such an assessment is that of Porter 1992:95-107.
[ back ] 127. Compare the transformation (in Iliad XXI) of the heroic theme of man-against-river into the cosmic theme of fire-against-water (see the discussion in Nagy 1996b:145).
[ back ] 128. For example, see again West 1966:208.
[ back ] 129. See in general Porter 1992:85-103.
[ back ] 130. For a survey of such assumptions, see Nagy 1996a:128-129.
[ back ] 131. See Ludwich 1884-1885:92, 97, 109, 114; also Nagy 1996a:129 n. 99.
[ back ] 132. Strabo 14.5.16 C676
[ back ] 133. Porter 1992:70; Pfeiffer 1968:232, 245, 270. See also Fraser (1972 I 471 and IIb 682 n. 225) on the converging Alexandrian / Attalid connections of Apollodorus of Athens, originally a pupil of Aristarchus.
[ back ] 134. Pfeiffer 1968:245.
[ back ] 135. Pfeiffer 1968:245.
[ back ] 136. Helck 1905:30-31. On the Orphic traditions in their ultimate textual form, see West 1983.
[ back ] 137. Nagy 1990a:216n10.
[ back ] 138. Seneca Epistle 88.39. Thanks to Jed Wyrick, who writes (per litteras 2 15 1998): “I think that there is a clear contrast here between two different kinds of ‘absurd’ approaches to learning - the one involving Orpheus, and the other, Aristarchus.”
[ back ] 139. See Pfeiffer 1968:115, 178. See also Michael Apthorp 1995 on “a new and peculiarly American critical sign, the bullet, with which [the editor] dispatches … interlopers in the right-hand margin.”
[ back ] 140. See Fraser (1972 I 468, IIb 676) on the epigram of Herodicus of Babylon, quoted by Athenaeus 5.222a, where the Aristarkheioi ‘Aristarcheans’ are described as monosullaboi ‘monosyllabic’, interested only in such monosyllabic pronouns as σφιν, σφῷν, μιν, and νιν.
[ back ] 141. Scholia A to Iliad XIII 197.
[ back ] 142. Scholia A to Iliad XVII 719. Further details and discussion in Nagy 1996a:151; Porter 1992:83. See also Janko 1992:71.
[ back ] 143. Helck (1905:63 and 74-75) notes that Crates did not think that Homer was an Athenian, though he may have thought that Homer used sporadic Atticisms; compare Wachsmuth 1860:39. Again, I see a reactive Aristarchus: he wants to posit an Athenian Homer by extrapolating beyond what Crates and even Aristophanes of Byzantium are willing to extrapolate.
[ back ] 144. Janko 1992:32.
[ back ] 145. Compare Pfeiffer (1968:239-240), who also cites the Allegories of Pseudo-Heraclitus and the Life of Homer by Pseudo-Plutarch; for the latter, see now Keaney and Lamberton 1996.
[ back ] 146. See Jensen 1980:216-218.
[ back ] 147. Tzetzes On Comedy 20 (ed. Kaibel).
[ back ] 148. West 1983:249.
[ back ] 149. West 1983:221n141 and 250n43.
[ back ] 150. West 1983:249.
[ back ] 151. West 1983:248: he thinks that the compiler was one Theognetos, as mentioned in the Suda, who divided his work into 24 rolls (232-233). In Orphica F 157, the scepter that passes to Zeus (six kings in all, and he is the fifth) has 24 measures. West (1983:232) cites Orphica F 356: ‘straight, in six parts, of four and twenty measures’. This formula, ascribed variously to Orpheus, Musaeus, or the Pythia, refers to the hexameter, which contains 6 feet and 24 morae. These same three “authors,” West points out (232), are the various traditional “inventors” of the hexameter. Likewise in the Odysseus attributed to Alcidamas, Musaeus is the “inventor” of the hexameter; so also in Democritus DK 68 B 15. I would add, however, that the division of the Orphic theogony into 24 rhapsōidiai is not necessarily modeled on the Homeric scheme, as West (249) infers.
[ back ] 152. Davison 1955, esp. 21; also Janko 1992:32.
[ back ] 153. Nagy 1996b:95-96.
[ back ] 154. Herodotus 7.6.3.
[ back ] 155. Nagy 1996b:73; also Nagy 1990a:174.
[ back ] 156. Consider the discussion above of Aristarchus’ use of the diplē in disputing the authenticity of Homeric verses accepted by Crates.
[ back ] 157. This explanation reverses directions for the scenario posited by West (1983:249): “To counter Aristarchus’ arguments for an Athenian Homer, the theory was developed that the rhapsōidiai, ‘recitations,’ into which the Homeric poems were divided, represented episodes which Homer had recited and left behind him in different towns; they had then been united by Pisistratus with the help of certain poets, who re-created an approximation to Homer’s original conceptions, but interpolated passages of their own, which accounted for Attic elements.”
[ back ] 158. Asclepiades is cited as the authority for associating Orpheus of Croton with Peisistratos: Suda s.v. Orpheus Krotōniatēs = FGrH 697 F 9. See West 1983:249-250, on Orpheus of Kroton as a “homonym split” (see other examples given by West a p. 250 n. 41).
[ back ] 159. Josephus Contra Apionem 1.12-13.
[ back ] 160. See Nagy 1997, esp. 106-109.
[ back ] 161. Strabo 14.1.48 C650.
[ back ] 162. See Robert 1940. (For the reference, thanks to Christopher Jones.) Also Lindsay 1997:296-297. See esp. Robert (146) on external evidence for the Homeric scholarship of both Menecrates and Aristodemus, as well as for comments on Strabo’s information (14.1.45 C650) about a local ritual at Nysa that was connected to a verse in the Iliad, II 461, which was understood to refer to the place where the ritual was held.
[ back ] 163. Robert 1940 (145) discusses archaeological evidence (with ample bibliography) for the Library of Nysa, comparing the Library of Celsus at Ephesus.
[ back ] 164. Robert 1940:145.
[ back ] 165. Nagy 1996a:187-206.
[ back ] 166. See n. 23 above.
[ back ] 167. The Orphic poem to which Cicero is referring here is what West calls the “Orphic Rhapsody,” which was probably known as the Hieros Logos (West 1983:248).
[ back ] 168. Richardson 1994:9.
[ back ] 169. Richardson 1994:12.
[ back ] 170. Richardson 1994:13.
[ back ] 171. Richardson 1994:13.
[ back ] 172. Richardson 1994:14.
[ back ] 173. Richardson 1994:17.
[ back ] 174. Richardson 1994:19. On pp. 19-20 he refers to Nickau (1977), who argues that Zenodotus was indeed working within the Aristotelian tradition. Still, Richardson (1994:20) remains skeptical. On p. 23, he stresses that there is no mention of Aristotle in the scholia ascribed to Aristarchus, but then he goes on to show that Aristarchus was Aristotelian in many ways.
[ back ] 175. Suetonius De Grammaticis et Rhetoribus 2. Pfeiffer 1968:246.
[ back ] 176. Fraser 1972 II 486n179. Compare Parsons 1952:29n1.
[ back ] 177. Dio Chrysostom Oration 53.1 (on Homer).
[ back ] 178. I gratefully acknowledge the help and advice of Alexander Beecroft, Erwin Cook, Christopher Douglas, Casey Dué, José Gonzalez, Sarah Gordon, Albert Henrichs, Carolyn Higbie, Christopher Jones, Ann Kuttner, Leonard Muellner, John Morgan, Gloria Ferrari Pinney, Fred Porta, Timothy Power, Miriam Carlisle, Sarolta Takács, Richard Thomas, Dimitrios Yatromanolakis, and Jed Wyrick.