Unlike most ancient Greek texts we have, the Aetia has not been transmitted to us intact in manuscript form. Therefore, each of the texts printed on this site will have gone through a complex and layered process of reconstruction, and a number of decisions about what to print. Each text represents some combination of the following:

(1) Before the discovery of papyri, the Aetia was known only by a handful of book fragments: these could be lines quoted in other ancient authors like Athenaeus or Strabo, or in sources like the scholia to Homer or Pindar, or in ancient lexica like the Etymologicum Magnum, Etymologicum Genuinum, or the Suda. The earliest modern collections of these fragments date from the seventeenth century (see Pontani 2011), and, as more papyrus fragments of the Aetia are discovered, hitherto unsuspected book fragments may also emerge.

Example 1. When the papyrus of the prologue Against the Telchines (P. Oxy. 2079) was discovered, about six letters were missing from the opening lines. The editors conjectured that [πολλάκ]ι might be the missing word from line 1, but this awaited confirmation until 1999, when F. Pontani actually found a scholium on Od. 2.50 with this explanation: ὡς πολλάκις Τελχῖνες, clearly echoing the first line of Callimachus’ poem.

Example 2. The second line of the prologue: νήιδες οἳ Μούσης οὐκ ἐγένοντο φίλοι was quoted in several ancient sources including Choeroboscus, Hephaestion, and Dionysius Thrax. Hence it is printed without brackets although the papyrus does not have the line complete. (Many editors will use half brackets to indicate these lines found in other sources.)

(2) Papyri of Callimachus began to come to light in the late nineteenth century, most notably the fragments discovered by Grenfell and Hunt and published in the Oxyrhynchus Papyri. They are identified in various ways: (a) the identification may be obvious from the content itself,  or from marginal comments, as with the Victory of Berenice; (b) the content may be known from other sources, as with Acontius and Cydippe, or (c) book fragments sometimes coincide with a line or more of a papyrus find, and thus guarantee the assignment to poet and text, as with example 2 above.

(3) The discovery of the Diegeseis, a papyrus roll containing incipits and brief paraphrases of the contents of each individual aition, has been of enormous advantage in establishing the order of text, especially for the very lacunose book IV. Other discoveries of papyrus scholia with lemmata or commentaries with glosses have helped in the restoration of individual lines.

Example 3. Line 11 of the prologue Against the Telchines was restored by Pfeiffer as ending: αἱ κ[ατὰ λεπτόν, and this reading stood for a generation. But the line end as part of a lemma occurs on the London scholium, and rereading of that scholium originally cast doubt on Pfeiffer’s restoration. In 1997 W. Luppe reread the scholium as αἱ γ’ ἁπαλ(αί), and this has been accepted by Harder, though earlier editors print αἱ κατὰ λεπτόν.

(4) Because the poem was so often imitated in antiquity, especially its opening, allusions to or translations of lines or parts of lines are also useful in establishing a text. The best example of this is Catullus’ rendition of the Lock of Berenice into Latin, which is always printed with Callimachus’ text and is often of help in restoring the Greek of the original. This must be done with caution, however, since Catullus’ poetic agendas did not necessarily coincide with Callimachus’ (see Bing 1997).

(5) Occasionally an ancient source has summarized the plot of an aition: for example, Aristaenetus has provided a summary of Acontius and Cydippe that has helped with the ordering of the many of fragments, and his summary of Phrygius and Pieria provides information about the plot that the exiguous remains cannot.

(6) Finally, editorial supplements and conjectures, often based on (3)-(5) have played an essential role in fleshing out fragmentary lines and reconstructing the whole. In this process scholars depend on Callimachus’ metrical tendencies with respect to elegiac couplets and his linguistic preferences. The tendency of modern editors is to be conservative—not to restore unless the lines of restoration are clear—though even this cautious practice does not guarantee that every supplement is correct.

Example 4. Line 11 of fr. 80 (Phrygius and Pieria) has been restored as: ἔνν]επες ὀφ[θαλμο]ῖς ἔμπαλι κ[λιν]ομέν[ο]ι[ς, on the basis of Aristaenetus 1.15.44: τὸ πρόσωπον ἐξ αἰδοῦς ἀποκλίνασα. G.-B. D’Alessio (2007) prints the supplement, though most editors do not.

Evaluating supplements

In evaluating editorial supplements readers should be guided by the following rule of thumb: if the syntax of the line or lines is clear and uncontroversial, then the grammatical shape of the restoration can be trusted as well as the parameters of the restoration even if there is editorial disagreement about the exact word or words.

Example 5. All editors adopt Wilamowitz’s line 29 of the prologue: τῷ πιθόμη]ν, “I obeyed him.” The justification for the restoration is that Apollo is speaking at the end of line 28: ἐλάσεις, but Callimachus is clearly speaking in line 29: ἀείδομεν. The missing phrase, which cannot contain more than 7-8 letters and must fit the metrical shape: ‒ ⏑ ⏑ ‒ or ‒ ‒ ‒, has to be a transition that indicates Callimachus’ assent to Apollo’s advice.

Example 6. All editors adopt Housman’s conjecture of ἀ[ηδονίδες] in line 16 of the prologue. The reason is that the rest of the line: δ’ ὧδε μελιχρ[ό]τεραι makes it clear that the thought is complete, that only one word can be missing, and that it must be feminine plural. Since the missing word falls at the end of the first hemistich of the pentameter its metrical shape can only be: ⏑ [‒ ⏑ ⏑ ‒] or ⏑ [‒ ‒ ‒]. Shorter words are not possible: although Callimachus will postpone δέ until after the article + noun unit (as in line 12: ἡ μεγάλη δ’), if the article αἱ is restored in line 16, it can only be long; but the visible α[ must be the final syllable of the preceding metron and thus necessarily short. Housman’s conjecture is bolstered by the fact that the diminutive ἀηδονίδες does occur in Callimachus’ Hymn 5.94, and ἀηδόνες in Ep. 2.5 Pf., where it means “poems.”

Example 7. Line 5 of the prologue ends: ἔπος δ’ ἐπὶ τυτθὸν ελ[, while the following line opens: παῖς ἅτε (guaranteed because it was quoted in an ancient source). Thus ελ[ belong to the verb that is missing from the phrase, and παῖς ἅτε requires it be Callimachus speaking in the first person. Either a verb beginning with ελ[ or beginning with λ[ + a temporal augment may be restored. There are no other possibilities. The three most plausible conjectures are: ἑλ[ίσσω, ἐλ[αύνω, and ἔλ[εξα. Each is dependent on a different set of arguments: the first (ἑλ[ίσσω) means “roll out” or “roll around” and was originally proposed with ἔπος δ’ ἐπὶ τυτθὸν to mean “roll out my writing (ἔπος) in a small compass,” i.e., write a short poem, where the image of rolling would allude to a papyrus roll as the material of composition. It is supported by Posidippus’ autobiographical elegy (118.17 A-B: βιβλίον ἑλίσσων), where he is unrolling a papyrus in performance. Harder accepts the restoration but takes the sense to be “turn over in one’s mind.” The second (ἐλ[αύνω) was Friedländer’s conjecture and seems to be supported by AP 14.121.11, where Metrodorus is clearly imitating the Aetia prologue. The meaning would be “guide” or “steer.” The third is a conjecture of Benjamin Acosta-Hughes and myself (2001); it depends on the fact that Callimachus uses ἔπος in his poetry very frequently (17 times) and almost always with a verb of speaking. We argued that the meaning would then be “I told my story bit by bit, like a child.” We were also attracted to a past tense because, given Callimachus’ advanced age, the critique of the Telchines would seem to leveled against something already done (e.g., line 4: ἤνυσα). There is no clear way to decide which of these is correct, but the sense of the passage is not in doubt, namely, Callimachus is chastised for writing 'small' poems.

But when the grammatical contours are not clear, then any restoration is untrustworthy and modern editors tend not to print.

Example 8. Although not much is missing from the opening of line 23 (approximately 10 letters having a metrical shape of ‒ ⏔ ‒), it is unclear if a verb is missing to control θρέψαι (line 24) and/or an adjective to modify ἀοιδέ or a string of particles.