1. The Manuscript: [1] Posidippus on Papyrus

Susan Stephens, Stanford University
Dirk Obbink, University of Oxford
Posidippus is no stranger to papyrus. The mentions of Posidippus in finds of Hellenistic epigrams exceed that of all other epigrammatists, the majority of whom are known from Roman papyri that reflect the ordering of Meleager’s Garland. Apart from P.Mil.Vogl. VIII 309, [2] there are five poems of Posidippus that have been found on papyrus, four of which are otherwise unknown:
i. P.Petrie II 49a (= P.Lond.Lit. 60 = AB 114 = SH 961), the so-called Epithalamium for Arsinoe. Now only the foot of a column of a third-century BCE papyrus roll that derives from papyrus cartonnage discovered in the Fayum (Gurob). On back of the roll, the title σύμμεικτα ἐπιγράμματα is written (and again the other way up), together with the name Ποσειδίππου. [3]
ii. P.Louvre 7172 (P.Firmin-Didot), containing AB 115-116 (= 11–12 GP = Page, GLP 104a–b). Written before 161 BCE, opisthographic papyrus roll, consisting of poems written by the orphaned sons of a Macedonian mercenary in the Memphite Sarapeum. [4] The roll includes excerpts from comedy and tragedy as well. The name Ποσειδίππου and the title ἐπιγράμματα are given as heading to the texts of two otherwise unknown epigrams—one (AB 115 = 11 GP) on the dedication of the lighthouse by Sostratus of Cnidos:
Ποσιδίππου ἐπιγράμματα
Ἑλλήνων σωτῆρα, Φάρου σκοπόν,
ὦ ἄνα Προτεῦ, | Σώστρατος ἔστησεν Δεξιφάνους
Κνίδιος· οὐ γὰρ ἐν Αἰγύπτωι σκοπαὶ οὔρεος οἷ᾿ ἐπὶ
νήσων κτλ.
The other epigram (AB 116 = 12 GP) refers to the shrine of Arsinoe-Aphrodite at Cape Zephyrium by Callicrates of Samos. AB 119 of the Milan roll concerns the same dedication by the same Callicrates; the text of this epigram is preserved by Athen. VII 318d (ἔγραψε δὲ καὶ Ποσείδιππος). [5]
iii.P.Berol. 14283 (AB 118 = SH 705), a waxed wooden tablet from the first century CE. This is a long elegiac poem on the poet’s old age, his poetry and his statue set up in the square in Pella. Posidippus’ name together with the place-name are mentioned in the poem.
iv.P.Freib. 4 (AB 65 = 18 GP = SH 973). This is a papyrus sheet of the first century BCE with a text of the Lysippus epigram, preceded by another by Theodorides (first century BCE), likewise on a statue. No author is named, but the epigram on Lysippus (also known to Himerius Or. XLVIII 14 with a variant reading and no author’s name) is given by AP XVI (Planudean Appendix) 119 with the ascription Ποσειδίππου.
v.P.Tebt. I 3 (AB 117 = 24 GP, from a first-century BCE papyrus roll). Names of poets survive from what appears to have been a collection, but they are all broken: – – – ]ιππου might be Posidippus. There are only ends of several lines, but the poet seems to praise a friend for his taste or skill in poetry.

What We Can Deduce From the Milan Roll

In addition to these—all of them informally written texts [6] —we now have P.Mil.Vogl. VIII 309, a papyrus roll recovered from a section of mummy cartonnage forming a chest piece or pectoral. [7] It now measures 19.6 x 152.8 cm. and contains 16 columns and 606 lines of text. The text has numerous errors, many of which appear to have been left uncorrected, though there are a fair number of corrected errors, either made by the original scribe, by a reader, or a second hand. There is no sign of a diorthotes; but the practice of Ptolemaic copying suggests somewhat greater tolerance for orthographic variation and the kind of errors that a reader can correct for himself.
We can identify four stages in the life of this papyrus:
  1. The original manuscript written along the fibers of a roll, containing epigrams, and written in a hand that likely belongs to the later part of the third century (ca. 230–200 BCE). The hand is that of a professional scribe, so this is not a privately made copy.
  2. Subsequently, the original protocollon—the sheet attached at the opening of the roll to strengthen it—was replaced, either because the opening was somehow damaged or for another reason.
  3. A text containing mythological material of some kind was written on the back of the roll. [8] This text can be dated by comparanda to the early second century and the hand is documentary in style (which makes the dating somewhat more secure). It is in four columns extending 72.5 cm. It is written upside down with respect to the text of the epigrams and covers an area running from cols. 6–13. From the fact that this second text only occupies the center of the roll, we can infer that the two sides were in simultaneous use. If not, we should expect that the roll would have simply been cut down to the necessary size for the new text.
  4. Finally, at some point around or after 176 BCE, the papyrus was sent to a recycling center and formed into the pectoral from which it was recovered. The papyrus itself had a useful life, we may infer, of at least 30 years (giving a low date for the initial writing) to 176, very possibly much longer—50 or more years.
The habit of recycling paper, whether it contained a literary text or a document, seems to have been a common practice during the Ptolemaic period. The Lille Callimachus (third century BCE), for example, came from a mummy mask and chest piece, found in Magdola (1901-1902). [9] The Milan cartonnage almost certainly came from one of the necropoleis in the region of the Fayum—an area to the south and west of Alexandria in which in the third century a large number of Ptolemaic veterans were settled. This would provide a plausible context in which such a roll might have been read, though nothing about where it might have been written. Unfortunately, there is no consistent relationship between a Greek literary text found in this way and the documents discovered with it. [10] It may have belonged to whomever the documentary texts belonged; it may just as easily have found its way to the recycling center via an independent route. There is at least one other cartonnage find that includes material from Alexandria as well as locally written documents, so the Milan epigrams might have been a text disseminated from the center. But equally they could have been copied locally. In the Tebtunis collection in Berkeley (P.Tebt. I 1–2), for example, there is a lyric anthology copied next to and in the same hand as a royal decree of Euergetes II (118 BCE). This is the hand of a professional scribe. There is also the example of the brothers living in the Serapeum at Memphis in 160 BCE who copied, along with other literary extracts, two dedicatory poems of Posidippus (P.Firmin-Didot). They also included on the same roll a bread account and a private letter. [11] Their hand, however, is not professional. Still, these two examples serve to indicate a context in which an anthology of epigrams might have been copied, read, and later recycled.
For evidence of anthologies of epigram in this period, two other papyri are relevant:
  1. The first is an Oxyrhynchus papyrus (P.Oxy. 3724) of the late first century CE. It contains a list of seventy-five incipits written in a documentary hand, which should indicate that it was a privately made copy. Twenty-seven of the incipits have been identified as belonging to Philodemus and there are good reasons to think that all of them belonged to that author. If so, we have an incipit list organized to produce a selected edition of epigrams of Philodemus. It should be noted that these incipits are occasionally reversed, some crossed out, with occasional ticks in the margin—all of which suggests a process of reorganizing.
  2. The second is an unpublished Vienna papyrus (P.Vindob. G 40611), also from cartonnage and found with documents datable to the last third of the third century BCE. [12] It consists of five fragments from a roll measuring 17 cm x 70 cm and contains incipits from over 200 epigrams. Only one of the incipits has so far been identified—and not to everybody’s satisfaction—as from Asclepiades (AP XII 4b). It bears the title τὰ ἐπιζητούμενα τῶν ἐπιγραμμάτων ἐν τῆι α´ (= 1st) βύβλῶι. Each incipit is accompanied by a number, e.g., 4, 8, 10, that is, presumably, the numbers of lines of the epigram to which the incipit belongs. At the bottom of col. 1 is the notation (εἰσι) κ´ στίχοι πη´, that is there are 20 lines, 88 verses. Col. 2 has the same notation as col. 1, col. 3 has (εἰσι) κδ´ στίχοι ρ´ (= 24 lines, 100). The total for book one is given as 83 incipits and 344 lines. Other identifiable numbers indicate that the Vienna epigrams like the Milan will have had mainly four and six line epigrams, with a few longer poems, though one poem of 21 in the Vienna cannot have been an epigram. In both we have mainly four and six line epigrams, with a few longer poems. These numbers are typical.
The Milan papyrus has two features relevant for understanding it as a literary manuscript: the repair of the opening of the roll and the stichometrics found throughout. Specifically:
  1. The opening of the roll as it stands seems to have been damaged or cut in some way and its original cover sheet replaced with another, which is rather narrower and attached not by overlapping the left edge (which is the usual practice) but by being fastened underneath. The editors assume that the roll is complete and that we have the original opening. They infer this from the fact that they read the category heading [λιθι]κ̣ά̣—in the top margin. They also point to the fact that the stichometric number for this section is correct for the actual lines still extant, thus guaranteeing that the section is complete. We cannot read the letters they claim to see, which in any case would be rather high in the margin, not in line with the top of the subsequent columns. The anomaly in the placement of this heading (if it is there) in combination with the recut left margin raises the question of whether we in fact have the original beginning of the roll. William Johnson has argued that line 1 of col. i is in fact an addition in the upper margin of col. i, written by a second hand (after the original beginning of the roll was detached) in order to make col. i begin with a complete poem. [13] According to him it is written noticeably above the level of the first lines of the subsequent columns in the roll. An examination of the papyrus, however, reveals that the graphic typology and ink of the hand that wrote line 1 of col. i differs little, if at all, from the text of the rest of the roll: its differences are certainly not sufficient to establish the presence of another writer at work. In addition, the level of line 1 of col. i does not appear to be significantly different from that of the first lines of the subsequent columns: the disparity is caused by the necessity of straightening out in Photoshop the angle at which fr. 1 of col. i (containing line 1) appears in the volume of plates accompanying P.Mil.Vogl. VIII 309 in order to effect a comparison with the level of the first line of the subsequent columns.
    It would be quite easy in principle to cut off one or more of the opening columns (perhaps another full section) and attach a new protocollon. But if the scribe numbered the sections as he copied, then this could only happen if the beginning of the category—λιθικά—coincided with the top of a column (if not the title). Otherwise the stichometrics, which appear to have been written by the original hand, would need to have been added after the fact. Either scenario is possible, but there is a greater likelihood that the scribe would have totaled as he went. The lack of title at the opening cannot be used as evidence of anything in particular. Literary texts in antiquity regularly bore end-titles not initial titles, though occasionally we have evidence of titles added at a roll’s beginning or on the outside of the protocollon at right angles to the text. [14] Should we not expect, if we have the original beginning, some kind of dedicatory poem? We initially thought that the opening poems, now fragmentary, of the Lithika would have contained the necessary dedicatory information. But there are reasons to doubt this, given what we can see of the poems. [15]
  2. There are two separate stichometric systems on this papyrus. The first is a total of lines (M = 40) entered at the foot of col. 1. If the scribe wished to have a total of the number of lines copied (and from the presence of this number, presumably he did), all he needed to do was multiply this number by the total number of columns in the completed text to arrive at a sufficiently accurate count. There is, however, a second set of numbers used in conjunction with a point placed at every tenth line of each section. (The points begin with a new section and are not cumulative for the whole papyrus.) The stichometrics occur at the end of each section and provide a total for that section. The numbers are not cumulative, which is the normal practice, so their purpose is not to provide a simple count of the whole, but must perform some other function. (In fact adding up these individual numbers to reach a total for the whole would be quite tedious.) Why section totals? We can think of three possibilities:
    • (a)  for balance
    • (b)  for control
    • (c)  for extraction from another longer roll (or rolls).
With respect to balance: if you were constructing a roll of some 600+ lines in total, jotting down the number of lines in a section would allow you, as you progressed, to think about balancing long against short sections without having to look over the whole section. The Milan text has section lengths of respectively 126, 80, 38, 116, 50, 98, 26, 32, and 32 before it breaks off. While this suggests a certain amount of variation of section size, the sizes are not varied exactly, nor would this degree of variation be particularly difficult to achieve. For this reason we are inclined to discount this explanation.
The use of such numbers as a control is more likely—the totals close the sections and would serve to guarantee completeness in comparison with an exemplar also so numbered or for a corrector or a future copyist. It is worth considering these numbers as a method of control in a particular environment. If the Milan papyrus was a collection produced from an incipit list like that of the Vienna papyrus we can imagine the scribe who produced a new roll based on such an incipit list would first keep his running totals (hence the dots at every tenth line), then set out the final count of each section to indicate to himself or his employer that all of the poems marked for extraction had been found.
There is, of course, a fourth possibility—that the numbers were already in the exemplar and were simply copied in by the scribe as if it were an integral part of the text. But this would only push back the explanation a generation; it does not explain their original function. While the incipit lists we have from Oxyrhynchus and Vienna are presumably locally produced for a private collection, it is worth considering whether an author or editor organizing an epigram collection would have proceeded much differently. Do we imagine him to be producing an autograph and laboriously copying out his poems in the order he wanted them? Or is it more likely that he indicated his own preferences via such an incipit list and gave it to his scribe to copy?


[ back ] 1. Our references to the Epigrams (and Elegy in the case of the poem on Old Age) of Posidippus are to the numeration of the editio minor of Austin and Bastianini (= AB).
[ back ] 2. Ed. pr.; commentary, analysis of readings, plates: BG.
[ back ] 3. Plate in Lasserre 1959:224, 227; Fraser 1972:1.607 and 2.858n403 reports autopsy of the papyrus by T. C. Skeat that failed to find any trace of names of other epigrammatists in addition to Posidippus’ on this papyrus as suggested by Lasserre and reported e.g. by Gow-Page in HE comm. (introd. to Posidippus). Presumably the title was copied from the collection from which the writer copied the poem itself.
[ back ] 4. For the Egyptian context: Thompson 1987:105–121; Thompson 1988:261. The title and author, together with the text of the epigrams, contain numerous copying errors and orthographical idiosyncracies characteristic of the level of competence of the writer.
[ back ] 5. Bing 1998:21–43 argues that Posidippus’ epigram records the dedication not of the lighthouse per se but of the statue of Zeus Soter which stood atop it. The new AB 110.1 εἴαροϲ ἡ Ζεφ[υρ- may derive from such a context, and the dedications in AB 37–38 may refer to this same temple: cf. 37.8 ναοπόλο[υ.
[ back ] 6. That is to say, they consist of copies of publicly circulating epigrams made by private individuals, and are not to be confused with informally composed and recorded poems: cf. Wissmann 2002:215–230.
[ back ] 7. From the Fayum, according to the antiquities dealer who acquired and sold the papyrus. For illustrations of the cartonnage pectoral and accompanying head-piece (both subsequently destroyed in the process of extricating the papyrus layers), see Bastianini and Gallazzi 1993.
[ back ] 8. To be published by C. Gallazzi.
[ back ] 9. Meillier 1976:261-26, 345–346; Parsons and Kassel 1977; Livrea 1979; SH 254–269.
[ back ] 10. See van Minnen 1998 for a list of papyrus texts and the context of their finds.
[ back ] 11. Lewis 1986:69–87; Thompson 1987:113, 116; Thompson 1988:261; Gutzwiller 1998:22–23. See also Acosta-Hughes, Nagy, Obbink, and Sider (this volume).
[ back ] 12. It is described in Harrauer 1981; CPR XVIII p. 1; Cameron 1993; Gutzwiller 1998, 23–24. The papyrus, to be published by B. Krämer and P. J. Parsons, is illustrated and discussed in Gööck 2002:18, who argues for a connection between the verse-counts marked by numbers in the margin before the epigram-incipits of the Vienna papyrus and Callimachus’ cataloguing of poetry in his Pinakes.
[ back ] 13. Johnson in Gutzwiller 2004 (forthcoming).
[ back ] 14. On the placement of titles in general, see Albino 1962–1963:219–34; Cockle 1987:219–222; Fredouille and Deléani 1997; Hengel 1984; Luppe 1977:89–99; Nachmanson 1941; Oliver 1951:232–61; Schmalzriedt 1970. On titles and agrapha at the beginnings of rolls, see Bastianini 1996:25–27.
[ back ] 15. See Hunter (this volume), who argues that the Lithika may have commenced from Zeus.