BMCR 2005.01.27

The Derveni Papyrus. Cosmology, Theology and Interpretation

, The Derveni papyrus : cosmology, theology, and interpretation. Derveni papyrus.. Cambridge: Cambridge University Press, 2004. xii, 441 pages ; 23 cm. ISBN 0521801087. $110.00.

The Derveni papyrus, containing a treatise by a follower of Anaxagoras probably written in the 420s B.C.E., is the most important new piece of evidence about Greek philosophy and religion to come to light since the Renaissance. It is also the hardest to understand, and all work on it is inevitably work in progress. This is the first book-length study of this text since 1997, when its crucial opening columns, plus an updated translation of the whole, were published.1 Betegh (henceforth ‘B.’) has made a major contribution to understanding both the thought of the Derveni author (henceforth ‘D.’) and the Orphic poem which D. interprets; his reconstruction of the Orphic theogony and of D.’s physical system should command wide assent. B. rightly concludes that D. ‘is trying to make Orpheus’ teaching up to date by providing it with an allegorical interpretation involving the conceptual and explanatory frameworks of late Presocratic speculation’ (372).

However, we can be sure that B. has not written the last word about this extraordinary text, for he also argues that Anaxagoras’ theory became ‘a major impetus both for the Platonic Socrates in pursuing a more thorough and radical philosophical analysis of the teleological structure of the world, and for an intellectually active Orphic initiator to reinforce some forms of traditional religiosity by adopting a more up to date and appealing theoretical framework to prop up his priestly lore and practice’ (283). B. develops an approach that was pioneered by Richard Seaford and variously modified by André Laks, Glenn Most and Dirk Obbink. B.’s major premises, (a) that D. was an Orphic priest rather than a sophist or Presocratic physikos, and (b) that the discovery of the papyrus in the ashes of a funeral pyre relates somehow to his doctrines about fire, motivate his conclusions as to D.’s ultimate beliefs and the overall purpose of his work. These rely on speculations about D.’s attitude to Heraclitus, an amazingly obscure Orphic gold leaf, and ancient beliefs about death by lightning. I suspect that few will find them convincing, for reasons that will appear below. B. lapses into petitio principii when he claims that ‘the technique of allegorical identifications does not have the same theological import in the Derveni papyrus as in other allegories’ (204). Many writers used allegory to defend poets like Homer or Orpheus from charges of impiety precisely by arguing that they had privileged knowledge of the divine; indeed, B. himself compares D. with the Stoics (205). Finally, his approach does not adequately explain the opening columns or the apparent excursus in col. 20 which mocks incurious initiates into the mysteries.

B. begins with a text and translation of the papyrus. This ‘makes no claim to be a critical edition’ (1); he refers instead to my own provisional text,2 which appeared when he was about to submit his manuscript to the press. The continuing lack of a definitive edition of this text 43 years after its discovery is most regretable. B. contributes some new knowledge, since he is able to report K. Tsantsanoglou’s readings in three columns, with parts of three new lines in col. 25, plus a few observations of his own. His text and apparatus do not take account of all available sources (including the few published photographs) and unfortunately abound in misprints in accents, dots, brackets and even letters that suggest imperfect revision.3 Its publication in this form, which is sure to be cited by incautious scholars, risks perpetuating the confusion as to what the papyrus says. Many suggestions are misattributed or simply omitted from the apparatus, even if they are printed in the text, e.g. at col. 17.2. All this is a pity, since B. offers some attractive textual supplements,4 and his choice of readings, albeit less courageous than I could wish, is generally sensible. There are some errors in his translation.5 The exigencies of production no doubt explain why the rest of the book takes no account of the new text, nor indeed, save in occasional footnotes, of my translation with lengthy introduction and notes which appeared in 2001.6

Ch. 1 well describes the finding of the papyrus, its script (but with little on its orthography or dialect), the editorial situation and attempts at identifying D.; here I miss a reference to my earlier elimination of several possible candidates, including Stesimbrotus.7 B. denies the relevance of Martial’s allusions to the use of papyrus to kindle funeral pyres (what better kindling did the ancients have?); instead he rehearses the familiar argument that the discovery of this roll in a funeral pyre suggests that it refers to funerary ritual (68), as if its contents resembled the ritual instructions in the Egyptian Book of the Dead or the Orphic gold leaves, which they do not. This argument shapes his understanding of its first six columns.

These crucial columns, discussed in Ch. 2, prove that the purpose of the text was not the commentary on the Orphic poem, which begins only in col. 7. B. suggests that ‘the width of the columns suggests that the discussion of hexameter verses was the main purpose of the entire text’ (92), but this width was standard for prose of whatever kind until about 260 B.C.E.8 These columns seem to jump wildy between different topics. B.’s solution to this puzzle is that ‘the ritual activities dealt with in cols. 2 and 6 constitute the immediate topic of this part of the text, and the rest is invoked in order to explain these cult activities’ (75). He does not consider the alternative that D.’s focus throughout his treatise is on the need for interpretation in both rituals and holy texts; to take them literally is to risk losing one’s faith in the divine.9 B. does not discuss the significance of the ‘terrors of Hades’ in col. 5, with which the Orphics threatened sinners (Plato, Rep. II 365a); these were a cause célèbre among religious radicals and sceptics, and Protagoras wrote a work about them (Diogenes Laertius IX 55). Nor can B. entertain the supplementation of col. 5.4-6 which the run of this passage demands, namely ‘for them we will pass into the prophetic shrine to ask again, regarding what is prophesied, whether it is right [to disbelieve in] the terrors of Hades’.10 B. leaves the lacuna unfilled and takes the apparent statement that D. enters the shrine as a proof that he was a seer. To my supplement B. objects that ‘it would be quite paradoxical to inquire of an oracle whether there is divine sanction not to believe’ (90 n. 48). Indeed; but in my view D. set out to shock traditional believers, just as some of his contemporaries outraged the Athenians by their freedom of religious thought (recall the case of Chaerephon’s consultation of the Delphic oracle).11

This reading influences B.’s approach at several points. He rejects the obvious interpretation that cols. 2 and 6 mention the sacrifice of a bird, on the ground that animal sacrifice is unlikely in a context of Orphic ritual, preferring to think that the Erinyes are compared to birds (77-8). In col. 6 he assumes that the magoi and initiates are presented as knowing that daimones are avenging souls (85), whereas D. is interpreting rituals of whose meaning their practitioners are unaware, as he complains in col. 20. B. also suggests that we philologists must not wax ironical about D.’s methods of argument (93), as if one can study religious thought without remarking its frequent departures from rationality. B.’s repeated and valid parallels with Plato, who attacked the Orpheotelestai and magoi while being an arch-heretic in terms of traditional religion, point in an opposing and far more profitable direction.

Chs. 3 and 4 reconstruct and interpret the Orphic poem, using first internal and then comparative evidence. The procedure followed, as throughout, is very methodical and involves considerable repetition, as B. acknowledges, but valuable insights abound. He rightly takes αἰδοῖον as ‘penis’ (113), persuading me to change my mind about Burkert’s interpretation of col. 14.1 (155).12 He rightly thinks the ‘first-born king’ is Ouranos (118), with the primeval couple as Night and Aither. The handling of the Orphic cosmogonies is clear and sensible.

Ch. 5, entitled ‘the Cosmic God’, uses internal evidence to investigate D.’s theology. I had concluded that he was a radical monotheist who identified all the traditional gods with each other, with Air, and with Mind (Nous). B. has reached largely the same conclusion, arguing for instance that, when Zeus rapes his mother, D. reinterprets the verses because otherwise the perpetrator and victim would be one and the same (188). However, B. blurs the identification of all the gods by suggesting that e.g. ‘Moira/pneuma/phronesis is not the same as Zeus/air/Nous, but an aspect or function of it’ (202). I do not see how this helps. B. rightly concludes that D. was not a monist, since Mind is set against everything else (186) and ‘air dominates the other things not because the air is in the other things, but because all the other things are in air’ (270); in this he follows André Laks.13 But he thinks D. also allegorizes as aspects of God verbs like ‘have sex’ and ‘mate’ (191), whereas I still hold that a short phrase is misplaced in col. 21. B.’s rendering of ἀρκέσαι in col. 11 is ingenious (pp 210-12), and his understanding of ἀκούειν in col. 8 is preferable to that of Laks and Most and of myself, even though his supplement ἀμφισβη]τεῖ gives the wrong sense. B. claims that D. denies that Achelous is water, supplementing οὐ in col. 23.12, but D. could easily be asserting that the rule of Air extends over water too; on Achelous in the papyrus see now G. D’Alessio in Journal of Hellenic Studies 124 (2004), 17-37. But B.’s main conclusion is new and compelling: that D. is transforming into Presocratic terms (those of the ‘One/Many’ problem) Orpheus’ tale that Zeus, last-born of the gods, swallowed the others to become first and alone (219-23). This finally explains why advanced thinkers like D. and Plato were so attracted to radical interpretations of Orphism.

Ch. 6, ‘Cosmology’, intelligently explicates D.’s physical system. B.’s interpretation that God ‘removed’ ( ἐξήλλαξ[εν) fiery matter in order to create the sun (231-5) is attractive, but he denies that D. could have placed the sun in the middle of the universe, even while acknowledging that this was a Pythagorean belief that predated Aristotle (240). As for D.’s identification of Olympus with time, B. rejects my supplement ‘time’ in col. 12.9 (249-50), but does not notice that my translation takes λέγειν as ‘mean’ rather than ‘say’, viz. ‘wherever Orpheus intended the meaning “sky” he added the epithet “broad”, but wherever he meant “time” he did the opposite, never adding “broad” but (sometimes) “long”‘; D. has in mind the standard formula μακρὸν Ὄλυμπον, which must have appeared in a lost verse of Orpheus’ poem. At col. 17.6-9, B. rejects his own attractive emendation of συνεσταθῆι into a subjunctive in favour of taking this aorist in a future sense (258); I do not see how this can work, but his translation of εἶδος as ‘form’ is appealing.

In Chs. 7-8 B. compares what we have learned about D.’s thought with that of Anaxagoras, Diogenes of Apollonia and Socrates’ alleged teacher Archelaus of Athens. The two latter are the closest analogues, and B. does well to remind us of Archelaus’ importance (321-4). As several scholars have argued, D. belongs among those disciples of Anaxagoras who sought to endow the divine Mind with a teleological role in the universe, the move that Socrates desiderated (Plato, Phaedo 97b-98b). B. is not completely clear on this point (315). He has not understood my own position on the authorship of the treatise. In 1997 I thought that D. was a disciple of Diogenes who held the same tenets as he, but I did not claim that D. is Diogenes himself (306); rather, I argued that D. is Diagoras of Melos.14 Subsequently I was convinced by Laks’ proof that Diogenes was a monist whereas D. is a pluralist; B. usefully points out other divergences between the two. It follows that D. either adapted Anaxagoras’ ideas in his own way or followed another Anaxagorean, perhaps even Archelaus. However, it is captious to deny that Diogenes’ fragment A 1 implies a ‘cosmogonic vortex’ (319); if his vortex created the earth, then it was cosmogonic enough for Aristophanes to mock it ( Clouds 830). D. has a vortex too: Burkert’s supplement τὴν δ[ίνην] at col. 18.1 is better than my τὴν δ[ὲ γῆν], which B. prefers (377-8), because the syntax shows that δέ cannot be postponed so late in its clause.

After the excellent contributions of Ch. 8, Ch. 9 is the weakest in the book. B. tries to find in D.’s quotation of Heraclitus in col. 4 a link between his theory of elements and a doctrine of salvation (329-30), without weighing the alternative that the link actually lies in the need for interpretation. To fill in the details he turns to external sources, notably the unbelievably obscure Orphic gold Tablet C from Thurii, as well as to ancient beliefs about death by thunderbolt and the immortalizing power of fire. This is fascinating material, but, as B. is aware, a Heraclitean belief in fire as the supreme divine power is scarcely compatible with D.’s equation of God, Mind and Air. To reconcile the two, B. suggests that D. accuses Heraclitus of contradicting himself. If the Erinyes are airy souls, then ‘unwittingly even Heraclitus agrees that the sun is controlled by the divine air and its airy helpers’ (346). Although B. translates κατ[αστρέ]φει τὰ ἴδ[ι]α at col. 4.6 as ‘overturns what is private’, he understands it to mean that Heraclitus confutes his own doctrine, overlooking David Sider’s demonstration that D. is using Heraclitus’ distinction between ‘shared’, waking perceptions of the kosmos and ‘individual’ ones such as we experience in dreams.15 In my version the sentence runs ‘[deeming] the common (sensations) [important], Heraclitus overturns those which are particular’. B. does not explain what ‘the common’ means here; this undermines his theory that D.’s ‘ultimate message is the proper explication of the relationship of air and fire’, in that ‘the fire of the pyre, or that of the thunderbolt, destroys only the corpse, but the (superior) soul is airy, thus fire has no power over it’ (347). Nor is this confirmed by the utterly obscure transition between cols. 19 and 20 (347-8).

Ch. 10 sums up, making a few additional points. B.’s treatment of the opening columns is crucial to his understanding of the rest. Deeming the treatise a polemic by one religious practitioner against others, he shows that such polemics were common among the doctors (350-9). He proves that religious experts ( exegetai) like Philochorus wrote works of a quite different character from the Derveni treatise (359-60); this is why I think D. was actually a natural scientist ( physikos). B. rightly perceives an ‘inner tension in the Derveni text’, which stems from D.’s ‘epistemological attitude’ and places him ‘outside the limits of his profession’ of priest and closer to Heraclitus (362-3). He well compares D.’s methods with the exegesis of oracles, but claims that this differs from allegorical interpretation in which the credit of holy texts is saved by claiming that they have a deeper meaning (368). He ends by comparing Empedocles, but he could equally well have adduced Aristophanes’ portrait of Socrates as a mystical hierophant in the Clouds.16

In the Appendix, B. attacks my theory, without fully presenting it, that D. is none other than the notorious ‘atheist’ Diagoras of Melos, whom the Athenians sentenced to death in 415/14 for mocking the Eleusinian mysteries and discouraging people from getting initiated. He can point to several errors in my original article of 1997, but does not engage with my revisions and further arguments of 2001.17 B. thinks Diagoras’ defamation of the Mysteries prompted Aristophanes’ joke about him at Clouds 830; however, it is his beliefs about Zeus and his playing on names that are mocked there, since Epicurus makes it plain that Diagoras used etymology to explain the names of the gods, listing him in the same company as Prodicus, Critias and Antisthenes (Philodemus, De piet. 518-41 Obbink). B. trusts the testimony of writers like Epicurus that Diagoras was an outright atheist, when the term atheos could also mean an advocate of a new religion. One must not underestimate the capacity of fanatics to propagate misrepresentation, even against other believers, when their core values are at stake (one need only consider the recent American election). Following Socrates’ execution, radical theologians like Plato and Epicurus needed to distance themselves from predecessors who had sought to explain the workings of the world through the actions of a material principle like Air, even if they also divinized it as Spirit. When Epicurus and others say that Diagoras believed in no god at all this does not invalidate the rest of their reports. Indeed, B. is obliged to concede that the contemporary public could have regarded D. as ‘impious’ and called him ‘atheist’ (380), acknowledging the significance of the papyrus in explaining the terrifying popular reaction against science and philosophy in late fifth-century Athens. Not to see this would be not to live in our own times either.


1. André Laks and Glenn Most (edd.), Studies on the Derveni Papyrus, Oxford: Oxford University Press.

2. ‘The Derveni Papyrus: an Interim Text’, Zeitschrift für Papyrologie und Epigraphik 141 (2002), 1-62.

3. The most unfortunate are ἐφόντων for ἐόντων (col. 13.11) and the report that the papyrus has ]. ουντ[ rather than ]. υοντ[ (col. 20.13).

4. Excluding changes to dotted letters (which I do not know how to reproduce in this review), the following new readings deserve consideration: col. 3.6, δεχόμ[ενοι; col. 7.3, ἅμα; col. 7.4, θέ]σιγ (proposed independently by myself); col. 9.1, ἀλ]κὴν; col. 9.7, εξήλλαξ[ε; col. 11.9, ἐπὶ τούτωι (also proposed by Tsantsanoglou, unless he obtained it from B.). The following are mistaken: col. 8.9, ἀμφισβη]τεῖ, which cannot mean ‘maintain’; col. 9.5, ἅτε, because the syntax will not work; col. 9.7, ὅς]ον γε, which is too wide; col. 15.1, κἂ[μ, since D. uses the form κἤμ, cf. col. 24.5; and the negative in col. 23.12, for the reason given below.

5. παραδείγματα does not mean ‘warnings’ (col. 5.8), nor αἰνίζομαι‘give a riddling meaning’ rather than ‘hint’ (col. 9.10), nor ἀνύσασθαι‘hear’ (col. 11.10, where B. translates my supplement without printing it), nor αἰθέρα ἔχθορε‘sprang from the aither’ rather than ‘sprang into the aither’ (col. 13.4). The aorist μέχρι . . . συνεστάθη cannot mean ‘until . . . have got set together’ as if it were a subjunctive in an indefinite temporal clause referring to the future (col. 17.8).

6. ‘The Derveni Papyrus (Diagoras of Melos, Apopyrgizontes Logoi ?): a new translation’, Classical Philology 96 (2001), 1-32.

7. ‘The Physicist as Hierophant: Aristophanes, Socrates and the Authorship of the Derveni Papyrus’, Zeitschrift für Papyrologie und Epigraphik 118 (1997), 61-94.

8. Janko 2002 (above, n. 2), 2 n. 5.

9. See my article cited above (n. 6).

10. So my translation of 2001; in that of 2002 I mistakenly took τῶμ μαντευομένων as neuter.

11. See my lecture ‘God, Science and Socrates’, Bulletin of the Institute of Classical Studies 46 (2002-3), 1-18, with new supplements for the end of col. 20.

12. Hence I accept his correction of my translation at 113 n. 49.

13. ‘Between religion and philosophy: the function of allegory in the Derveni papyrus’, Phronesis 42 (1997), 121-42, at 130.

14. Art. cit. (above, n. 7).

15. ‘Heraclitus in the Derveni Papyrus’, 129-48 in Laks and Most 1997 (above, n. 1), at 134-5, citing frr. A 16 and B 89.

16. Two minor errors: for ‘Eritrea’ read ‘Eretria’ (167, 275); Metrodorus of Lampsacus is cited in Philodemus’ On Poems, not his On Piety (203). I forbear to list a number of misprints.

17. Artt. cit. (above, nn. 6-7).