BMCR 2006.10.29

The Derveni Papyrus. Edited with Introduction and Commentary. Studi e testi per il “Corpus dei papiri filosofici greci e latini”, vol. 13

, , , The Derveni papyrus. Studi e testi per il Corpus dei papiri filosofici greci e latini ; 13. Florence: Casa Editrice Leo S. Olschki, 2006. xiv, 307 pages, 30 pages of plates : illustrations ; 24 cm.. ISBN 8822255674. €35.00 (pb).

1 Responses

The Derveni papyrus is the most important text relating to early Greek literature, science, religion and philosophy to have come to light since the Renaissance. It was found amid the excavation of a grave near Thessaloniki in January 1962, when Petros Themelis saved it from destruction. It was written by one of the rationalists who reinterpreted traditional Greek religion in the light of early science, and is in my view a pivotal text for understanding the religious crisis of the late fifth century B.C.E.1 It will be as central to the study of Hellenism as the Dead Sea Scrolls have been for Biblical studies. Its publication has been awaited for 44 years; exactly the same interval elapsed between the discovery of the Dead Sea Scrolls in 1947 and their release in 1991. The book under review presents the work of four scholars at the University of Thessaloniki (including the late Stylianos Kapsomenos), whom the Archaeological Museum of Thessaloniki has successively entrusted with editing the papyrus. That it has appeared is a major triumph of diplomacy over despair; those responsible, including Theokritos Kouremenos and Franco Montanari, are to be most warmly congratulated.

However, this volume is effectively a report of the present state of this papyrus rather than an edition which takes all the evidence into account. Whether you will think this is for better or for worse will depend on whether you want to coax new knowledge from the papyrus, or simply to know what it says. From the latter perspective the book is a step backward. But from the former it is a gold-mine, since it is an essential step towards a definitive edition.

The book, written in excellent English, consists of an introduction, mostly by Kouremenos (K.); a diplomatic text with papyrological apparatus and facing reconstructed text, followed by an English translation, by Tsantsanoglou (T.) and Parassoglou (P.); a commentary, by K.; a list of abbreviations and bibliography; an index verborum et locorum by David Muratore; and thirty excellent plates by Makis Skiadaressis.

Ch. I of the Introduction (1-19), by T., discusses the circumstances of the find and the habits of the scribe.2 Then K. surveys the fragments of the Orphic poem, and the physical doctrines of the Derveni author, whom he rightly assigns to the school of Anaxagoras in the late fifth century. As to his aims in writing his treatise, K. reports various views without comment, but not my central contention that most of the surviving treatise is a digression, meant to prove that taking traditional rites and texts literally is an impediment to true faith. So far as I know, this is the only hypothesis thus far advanced that explains the point made in cols. 5 and 20.

As for who its author was, K. suggests that we can never know. He justifies his refusal even to report the current hypotheses by writing: ‘since the author could be identified only on the basis of evidence from the work, deciding between incompatible readings of a fragmentary work in this way would turn into reinforcing one’s own preferred reading by attaching to the work a suitable name. It is difficult to see how this vicious circle could be broken’ (59). He adds that the author may be someone of whom we have no record. This is of course possible. However, the editing of the text will in the first instance be based on its language and on the well-established principles of textual criticism, not on hypotheses of authorship. Whether the treatise is by an unknown author, by Socrates’ teacher Archelaus of Athens, or by the ‘atheist’ Diagoras of Melos can only be settled once the text is put on a firm editorial basis, but will remain a question of the highest importance.

Ch. I also gives a few details about the unrolling of the papyrus by Anton Fackelmann, whose report on its original state is included (4-5). Such was his brilliance that he succeeded in separating, flattening and mounting between glass the 266 fragments. One must also admire the skill with which 153 of them have been combined to form a text of 26 columns.

At the same time, it is astonishing to learn that 113 fragments, or 42% of the total, remain unplaced. Many are tiny, but some are as wide as half a column. A more detailed description of the method of reconstruction might have helped both to place these fragments and to justify the delay of 44 years in bringing out this papyrus.3 Experience with the Herculaneum papyri enables me to understand the outlines of this process. The end of the papyrus (the ‘midollo’ in Herculanean terminology) was recovered first, and was published in 1964 by Kapsomenos; this was cols. 22-26. The rest consisted of five stacks of fragments from the outer layers of the roll (‘scorze’); the sequence in which Fackelmann removed the pieces has evidently been jumbled or lost.

The editors do not openly acknowledge the effect of flattening carbonized papyri, which I know from four Herculaneum papyri that Fackelmann treated in the same way.4 When the pieces are flattened, their fibres are crushed and they may deteriorate rapidly. This is a natural process, for which only the original treatment is to blame. In addition, as was noted of the Herculanean rolls in the eighteenth century, carbonized papyri quickly fade. The secret of success is to take good photographs at once, to use every possible method of recording from the beginning, and to work rapidly. Old notes, drawings and photographs soon become indispensable sources; in this case there is at least one unpublished set of early photographs, by Spyros Tsavdaroglou (6 n. 15).

However, this edition aims to present the text as it can now be seen. In col. 5 the editors exploit old photographs by Charalambos Makaronas, and sometimes they can read more on the original than was previously reported, but they nearly always see less. Thus at col. 26,10 they dot two letters in the word αποδειξαι where I read both of them beyond doubt by enhancing an old published photograph. If the continuing deterioration of the papyrus went unrecognized, this must greatly have hindered its publication: every time something needed to be checked, less could be read than the time before.

The editorial conventions are never explained. The proper use of a sublinear dot is to indicate that the letter can be read as a different one. However, here a dot has four different uses, as intensive study of the literature on the papyrus and of the available images of many dotted letters has shown me. Such study is needed in each case to clarify which usage is involved. In this edition, a dotted letter may mean:

1. The traces can be read as more than one letter.

2. The letter is damaged, but is still unmistakable. Its identity is often asserted in the papyrological apparatus (e.g. at col. 4,2, where the traces are ‘compatible only with kappa’).

3. No traces at all have been seen, presumably because the papyrus is stripped or the ink has flaked off; the dots are space-holders. However, if the editors fill the spaces with letters they print these letters with dots (e.g. the first letter of col. 5,9).

4. The letter was once read but is now damaged or destroyed. This seems to be what is meant when letters appear with dots in the diplomatic transcription on the left-hand pages as well as in the text itself, but the traces are not described in the papyrological apparatus, e.g., the epsilon in col. 2,2. In the extracts that T. communicated to A. Bernabé the letters in this category are consistently printed without dots.5 Unfortunately, the evidence that the letters were once securely read often remains unpublished.

By using a simple but bizarre expedient, P. and T. have contrived not to acknowledge that scholars other than themselves have toiled to reconstruct this text. They include no apparatus criticus! Crucial supplements to many lacunae have been made by numerous other scholars, but the editors do not even reveal which ones are owed to Professor Kapsomenos.6 The editors’ openness to supplements other than their own seems to have ceased at least a decade ago, but scholarship on the papyrus has moved on since then. Following the appearance in 1997 of the volume edited by A. Laks and G. Most,7 several scholars have brought out important work on the text. G. Betegh contributed a major interpretative book, together with a text and translation of his own, based on some of his own readings of the papyrus made during the one week when it happened to be on display in Thessaloniki.8 Working entirely from published materials, I contributed an article on the authorship of the papyrus, a translation with a basic commentary, and an ‘interim text’ with far better punctuation and facing translation.9 F. Jourdan also wrote a book which (by permission) included my text, to which she appended a commentary in which she systematically dissented from it.10 None of these contributions has been taken into account in the constitution of the text.

We have all been working blind for all these years. Now that we have eyes to see, what can we learn from the editors’ photographs and their very careful transcriptions?

The situation varies. Most of the text of cols. 6-26 has already been published correctly, apart from the marginal paragraphi, where reporting had been very incomplete. There is new text and/or supplements (many of which are excellent) at the bottoms of cols. 9, 11, 13-14, 18-20, 22 and 24-5; this usually amounts to one line or so. In places further supplementation is easy, e.g. πλείονα at col. 20,13 and τὰ ἀγαθὰ at col. 24,14. Where the editors report a letter without a dot, their work turns out to have been completely reliable. Hence all previous attempts, including mine, to suggest better readings of letters that had been reported as secure prove to be mistaken; we will have to learn to live with the ‘disputable riddles’ and new verbal adjective εὐκρίνητος‘easily clarified’ in col. 7,6-9.

However, the editors refuse to admit any emendations, even where the papyrus is blatantly ungrammatical or nonsensical. They allow that letters have fallen out at cols. 5,8 and 21,3 and a word at 25,8, but that is all. T. believes that the 60-75 years between the composition of the treatise and the copying of the roll were ‘too short a time for the text to have been considerably corrupted in the process of transmission’ (10). However, miscopying can occur in a second: our copies of Philodemus’ works, made during their author’s own lifetime, teem with errors. Here the need for emendation is often patent. At col. 9,9, a plural antecedent governs a singular participle. The editors have not made sense of col. 15,7-8: with better word-division and punctuation, this runs ‘(Orpheus) means that his rule has existed from the point when (Mind) became king. But his rule is explained: Mind, by thrusting . . .’. At col. 16,10 αὐτὸν lacks the article. At col. 21,13 ‘he was named to have been born’ is ridiculous: emend to ‘he was thought to have been born’. At col. 24,4 the editors print the non-existent ερησθαι rather than Ricciardelli Apicella’s easy correction erh=sqai.11 At col. 24,6 an accusative and infinitive lack a governing verb; at col. 25,1-2 a superlative is needed instead of a comparative, and καὶ has dropped out before κατὰ; and at col. 26,11 a dative agrees with an accusative. My list is very incomplete.

The edition retains a number of supplements that have been improved upon by recent scholarship, e.g., λύσιν at col. 7,4, where Betegh and I both proposed θέσιν. A number of new ones make no sense either. For example, at col. 7,5 κεἰ deprives the sentence of a main verb; at col. 8,9-10, by not supplying a main verb, the editors make a verb in a subordinate clause govern an infinitive in the main clause.

With regard to cols. 1-5, T. has often changed his text of a decade ago.12 He has modified some supplements and moved several fragments. Cols. 2-3 are almost completely different, with very unconvincing supplements. In col. 4 the famous fragment of Heraclitus is offered in a new and unattractive reconstruction. In col. 5,1 T. does not accept my reading of the old photograph, which shows that the ‘terrors in Hades’ are discussed. His new text appears to rule out my idea that col. 5,5-6 contains the blasphemous statement ‘we will enter the prophetic shrine to ask, regarding what is prophesied, whether it is permissible to disbelieve in the terrors in Hades’, since his new readings contradict the supplement ἀπιστεῖν τὰ ἐν. But the photograph proves that the letters at the end of line 5 come from a different layer, as they are at a different height from the rest of the line; multiple layers are common in Herculaneum papyri, and T. acknowledges this phenomenon both elsewhere and here (70). Moreover, enhancement of the plate still seems to show a nu at the start of line 6. It will be exciting to see what more can be made of these columns once advanced techniques of reading and reconstruction are applied to them, not to mention to the 113 unplaced fragments.

The English translation accurately renders the Greek. If there is much that makes little sense, this is often caused by the editors’ choice of readings. Here is col. 12,3-10 in their version:

Olympus and time are the same. Those who think that Olympus and heaven are the same are mistaken, because they do not realize that heaven cannot be long rather than wide, while if someone were to call time long, he would not be mistaken. Wherever he wanted to say ‘heaven’ he added ‘wide’, but wherever (he wanted to say) ‘Olympus’ he did the opposite, he never (added) ‘wide’ but ‘long’.

The logic and sequence seem crazy because a mistaken supplement has been adopted. Here is my rendering:

‘Olympus’ is the same thing as ‘time’. Those who suppose that ‘Olympus’ is the same thing as ‘sky’ are quite mistaken, as they do not comprehend that it is impossible for ‘sky’ to be ‘longer’ rather than ‘broader’. But if someone termed time ‘long’, he would not be at all mistaken. Wherever (Orpheus) intended the meaning ‘sky’, he added the epithet ‘broad’, but wherever he meant ‘time’ he did the opposite, never adding ‘broad’ but ‘long’.13

For an intelligible translation readers must still look elsewhere. Here is their translation of col. 17,1-11:

it (sc. air) existed before it was named; then it was named. For air both existed before the present ἐόντα were set together and will always exist. For it did not come to be but existed. And why it was called air has been made clear earlier in this book. But after it had been named Zeus it was thought that it was born, as if it did not exist before. He also said that it will be “last”, after it was named Zeus and this continues being its name until the present ἐόντα were set together in the same state in which they were floating as former ἐόντα. And it is made clear that the ἐόντα became such because of it and, having come to be, [are again] in it.

The problems here are largely caused by misidentifications of the subject. My own version runs:

(Zeus) existed before he was named; then he was named. For Air was pre-existent even before those things which now exist were put together, and he will always exist; for he did not come to be, but existed. Why (Zeus) was called ‘Air’ has been revealed earlier. But he was thought to have been ‘born’ because he was named ‘Zeus’, as if he had not existed before. (Orpheus) said that (Air) will be ‘last’ because he was named ‘Zeus’ and this will continue to be his name so long as the things which now exist are put together in the same element (i.e. air) in which they had been suspended when they were pre-existent. (Orpheus) reveals that the things that exist became such as they are on account of (Air), and, having come to be, [are all] in (Air).

I believe that this will be found more intelligible.

K.’s commentary is philosophical rather than philological; as such, it is well done. It addresses a number of issues (and readings) raised in recent scholarship, but passes over many philological questions. It has a special partiality for the work of F. Jourdan and is respectful towards that of G. Betegh; the textual proposals of the present writer are either rejected or (mostly) ignored. K.’s reports of T.’s suggestions and readings, and of proposals made privately to T. by other scholars at unspecified dates, are valuable.

In conclusion, P. and T. have chosen to benefit neither from the scholarship of the past decade nor from recent advances in reconstructing and reading carbonized papyri. They have been slow and timid, both in editing this text and in interpreting it, but at least they have brought it out. The flaws in their work cannot negate the fact that it will be indispensable for the definitive edition, translation and commentary on the Derveni papyrus which is still lacking.

[For a response to this review by Theokritos Kouremenos, George Parssoglou, and Kyriakos Tsantsanoglou, please see BMCR 2006.11.02.]


1. See my article ‘Socrates the Freethinker’ in S. Ahbel-Rappe and R. Kamtekar (edd.), A Companion to Socrates (Oxford: Blackwell), 2006, 48-62.

2. The dialect is not only a mixture of Attic and Ionic (11-14), but also contains a few Doric forms, which need explanation, as I noted in ‘The Physicist as Hierophant: Aristophanes, Socrates and the Authorship of the Derveni Papyrus’, ZPE 118 (1997), 61-94, at 62-3.

3. P. and T. promised to bring it out by mid-1984: see Gnomon 54 (1982), 855-6. Contrast the Strasbourg papyrus of Empedocles, where work began in 1990 and the text appeared in 1999: see A. Martin and O. Primavesi, L’Empédocle de Strasbourg (Berlin and New York: De Gruyter), ix.

4. For details see my Philodemus, On Poems Book 1 (Oxford: Clarendon Press), 2000, 48-9, 52, 55-6.

5. ‘La Théogonie orphique du papyrus de Derveni’, Kernos 15 (2002), 1-38; id., Poetae epici graeci II. Orphicorum et Orphicis similium testimonia et fragmenta (Munich and Leipzig: Saur), 2004. Yet at col. 5,11 letters that survive only in an old photograph are undotted.

6. See my apparatus criticus in ‘The Derveni Papyrus: an Interim Text’, ZPE 141 (2002), 1-62, where I sorted all this out so far as one can from published sources.

7. Studies on the Derveni Papyrus (Oxford: Clarendon Press), 1997.

8. The Derveni Papyrus: Cosmology, Theology and Interpretation (Cambridge: Cambridge University Press), 2004.

9. ‘The Physicist as Hierophant’, cited above, n. 2; ‘The Derveni Papyrus (Diagoras of Melos, Apopyrgizontes logoi ?): a New Translation’, Classical Philology 96 (2001) 1-32; and ‘The Derveni Papyrus: an Interim Text’, cited above, n. 6.

10. Le Papyrus de Derveni, Paris: Les Belles Lettres, 2003.

11. This emendation is correctly attributed in G. Betegh (cited above, n. 8), 50. I mistakenly took credit for it in ‘The Derveni Papyrus: an Interim Text’ (cited above, n. 6), 48.

12. ‘The First Columns of the Derveni Papyrus and their Religious Significance’, 93-128 in Laks and Most (cited above, n. 7).

13. It is gratifying to learn that Peter Parsons independently proposed δὲ χρόνον instead of δ’ Ὄλυμπον at col. 12,9. Following Jourdan and Betegh, K. opposes this supplement (191), but I cannot make head or tail of the logic of the passage as he explains it.