For the author of the Derveni Papyrus, our primary evidence is of course the contents of his work, and they are not entirely homogeneous. In Figure 1 I give a conspectus of the total contents of the papyrus, with the column number, the section of the theogony discussed, and the author’s interpretation and any other texts quoted.

The concept of poetry as riddle is known before this text (Struck 2004:39–50), but I find particularly interesting the variant formulation of the principle in the pseudo-Platonic dialogue Alcibiades II 147D. After his interlocutor has expressed frustration with interpreting a poetic text, Socrates says that in fact the text is consistent: not only this poet but almost all other poets speak in riddles, since poetry is by nature enigmatic, and is not for any chance reader to understand (ἔστιν τε γὰρ φύσει ποιητικὴ ἡ σύμπασα αἰνιγματώδης καὶ οὐ τοῦ προστυχόντος ἀνδρὸς γνωρίσαι). And in addition to its being this way by nature, when poetry is in the hands of a resentful man who doesn’t wish to communicate with us but rather to conceal his wisdom as far as possible, the difficulty of understanding the thing that each poet actually means is stretched to an extreme (ἔτι τε πρὸς τῷ φύσει τοιαύτη εἶναι͵ ὅταν λάβηται ἀνδρὸς φθονεροῦ τε καὶ μὴ βουλομένου ἡμῖν ἐνδείκνυσθαι ἀλλ΄ ἀποκρύπτεσθαι ὅτι μάλιστα τὴν αὑτοῦ σοφίαν͵ ὑπερφυῶς δὴ τὸ χρῆμα ὡς δύσγνωστον φαίνεται͵ ὅτι ποτὲ νοοῦσιν ἕκαστος αὐτῶν). The text that he goes on to interpret is from an improbable source, the Margites: he uses the tactic of word-replacement (not found in the Derveni author) to alter the meaning of the line of the Margites to say not that Margites knew everything badly—since to know something badly would be unacceptable to Socrates—but that it was bad for him to know all the trivial things he did know. Certainly this broad statement of principle applied to an unlikely poetic text is not meant to be taken seriously—it seems rather to be a parody of such literary interpretations, like Tiresias’ linguistic reinterpretation of the birth of Dionysus in Euripides’ Bacchae 286–296, or like the Platonic Menexenus, which often seems to parody the institution of the funeral oration. And yet, though Socrates is less than generous in the motives he ascribes to his allegorical poet, he claims the same enigmatic license that the Derveni author does in column VII. {122|123}

And it is also an important characteristic of the Derveni interpreter, at least as I read him, that, for all his cavalier treatment of texts, nevertheless just like Didymus on Demosthenes or other prolific post-Alexandrian interpreters, he would be lost without them: the poem of Orpheus is cited again and again, even in the first six columns he cites Heraclitus, and I still believe it quite possible that in column XI and column XX he cites unnamed prose texts (see the appendix). I also cling to my opinion that, although he clearly uses the technical vocabulary of citation on several occasions for effect, not all citations have to be carefully introduced or their author specified; the principle of seeking parallels and ideas from texts is sufficiently clear that one need never be surprised by a citation around the next corner, and in these cases the scribe has often given us help.

And so, although one can have reservations about the methods and the originality of this writer, nevertheless his voracious reading, intimate familiarity with his text, zest in expounding it, and use of other texts to carry his ideas further, we can agree, all deserve admiration. {126|127}

Obbink does not give a list of examples for his claim that a scribe always fills the line at the end of a quotation fully, but I presume he means the like of VIII.3–4, XV.5–6, 10–11 (24, 26, and 19 letter–lines respectively preceding hexameter quotations) as well as the one he cited, XI.7–8 (11 letter–line preceding an unsourced prose quotation); especially lacking, however, is evidence for the claim that a line with 34 as opposed to 38 characters is not “full.” A glance at the photos in KPT shows that many lines are this short without any such external criterion, and a more important factor in line division is obviously that this scribe rarely divides syllables within words between one line and the next (XII.4–5 and XX.6–7 seem to be the only examples). In XX.10 he might have made the line slightly longer by adding πρὶν from the next sentence (as he does with short words e.g. VI.5 [new total 36 characters], XVI.8 [new total 40 characters], XVII.3 [new total 39 characters]), but at XXI.13 he avoids continuing a 34-character line with the available short initial words (ἦν μὲν γ[ὰρ) of a new sentence; at XX.10 there is in addition the factor of another awkward μέν without answering δέ (as at XIII.5–6), so that there are multiple reasons for beginning a new line. The “consistent graphic practice” does not operate here.

But Obbink goes on to cast doubt on the possibility of a quotation of any sort:

Apart from the difficulty of positing a scribe who is “consistent” at one time and “irrational” at others, the two paragraphoi considered anomalous (XIII.6, XXIII.7) are in fact also strikingly different from their surroundings, both being statements of interpretative method (cited above), quite possibly taken from another context (the first has μέν without answering δέ), but in any case explanatory interruptions in the process of interpretation.

Obbink concludes:

While my term “feeble prose paraphrase” was doubtless too harsh for lines 11–12, Obbink’s “rounding out” and “reiterating” seem too generous for them: while lines 1–10 never use the same form twice, every word in lines 11–12 is either a word that has appeared in 1–10 or a gloss on one (πρὶμ glosses πρότερον ἢ ἐπιτελέσαι,. ἐλπίζον̣[τε]ς + future infinitive glosses δ̣οκοῦντες + future infinitive and ἐλπίδος glosses γνώμης (i.e., the expectation of knowledge). Furthermore, these lines, although reproducing (in an anacoluthon) μὲν/ δ̣ὲ from 1–10, lack a sentence connective with what comes before.

Thus even apart from the paragraphos, there are good reasons for assuming a departure from the author’s voice in XX.1–10: the sole first-person verb, sophisticated sentence structure at odds with the commentary form in the section (and with the author’s practice even in cols. I–VI), and, between lines 10 and 11, a lack of explicit sentence-connection combined with total redundancy of contents. {131|132}