Reconstructing (again) the Opening of the Derveni Papyrus

1 R ECONSTRUCTING ( AGAIN ) THE O PENING OF THE D ERVENI P APYRUS 1 The Derveni papyrus is arguably the most extraordinary new text on Greek religion to emerge since the Renaissance. This carbonized book-roll, dating I believe from as early as 350 BC, was found in the remains of a funeral pyre at Derveni in Greece in January 1962; it is thus the oldest surviving European book. Its final columns were brought out by S. G. Kapsomenos in 1964. However, K. Tsantsanoglou first published its opening columns in 1997, announcing that there were 26 columns rather than 22 as has previously been thought,2 and the editio princeps of the whole papyrus was produced by T. Kouremenos, G. M. Parássoglou and K. Tsantsanoglou in October 2006.3 This latter edition included a set of photographs of all the fragments, a benefit of inestimable value. It also revealed the surprising fact that a total of 113 pieces, some of them as large as a third of a column in width, were still unplaced. While waiting for the papyrus to be fully published, I had produced an interim text and a translation, based on the information that was then in the public domain.4 The photographs in the editio princeps have now made it possible to apply to the reconstruction of this roll techniques which have been developed for putting together the dismembered volumina from Herculaneum. In the course of completing a review of the editors’ publication,5 I constructed a physical model of the entire volumen from the published images in order to test their results and to see if further progress could be made. This article presents the results. Our understanding of these columns is at an early stage, but progress can still be made. There is much at stake in attempting to solve the puzzle of this papyrus. Nobody would ever have predicted that we would find an analysis of an Orphic poem in terms of pre-socratic molecular physics, such as occupies columns 7–26 of the papyrus. The Derveni author’s style puts him without any question within the fifth century BC. To see the methods of etymology and allegory applied to the interpretation of such an important religious text was utterly unexpected. We knew of some such speculations about the names of the gods from Plato’s Cratylus, but the author seems to me to be far more daring and provocative than one could have imagined. His arguments culminate in column 20, when he attacks people who get initiated into the mysteries (he means the Orphic mysteries). They are gullible and waste their money, he argues, because they accept the priest’s explanation and do not enquire further into what they have heard. He can enlighten them further about the meaning of the text. He then goes on to explain how it is perfectly proper and pious for Orpheus to have said that Zeus raped his mother, provided that one understands his words correctly, that is allegorically. 1 I presented the new texts of cols. 6–7 in the Midwestern Colloquium on Ancient Religons at the University of Michigan in March 2007, cols. 2–3 at the XXVth International Congress of Papyrologists in Ann Arbor in July 2007, cols. 1–3 at the Fifth A. G. Leventis Conference at the University of Edinburgh in November 2007, and cols. 4–5 at University College London in January 2008. I am most grateful to audiences on those occasions for suggestions and ideas, and especially to Jan Bremmer, Fritz Graf, Sarah Iles Johnston, and Richard Rawles. I also thank Chad M. Schroeder for help in constructing the model and preparing the digital images that I used in these presentations, and the University of Michigan for the sabbatical leave which has given me time to write up my results. I am especially grateful to the Editors of ZPE, and in particular Jürgen Hammerstaedt, for improving this article in numerous ways (including the Latin of the apparatus). The remaining errors are mine. 2 K. Tsantsanoglou, The first columns of the Derveni Papyrus and their Religious Significance, in A. Laks and G. W. Most (edd.), Studies on the Derveni Papyrus, Oxford 1997, pp. 93–128. 3 The Derveni Papyrus. Edited with Introduction and Commentary. Studi e testi per il “Corpus dei papiri filosofici greci e latini”, vol. 13, Florence, 2006 (henceforward abbreviated to Ts.3). 4 The Derveni Papyrus (Diagoras of Melos, Apopyrgizontes logoi?): a New Translation, CP 96 (2001), pp. 1–32; The Derveni Papyrus: an interim text, ZPE 141 (2002), pp. 1–62. I argued that its author is the ‘atheist’ Diagoras of Melos, whom the Athenians condemned to death for impiety. Meanwhile Maria Broggiato has shown that John Lydus may have ascribed to Diagoras of Melos an etymological interpretation of the role of Amaltheia in the birth of Zeus (Giovanni Lido, Sui mesi 4.71: un nuovo frammento di Diagora di Melo?, Seminari Romani 5 (2002), pp. 231–7); the interpretation of Amaltheia as a synonym for ‘power’, via émalaki!t¤a, is just the kind of sophistry that could have been used to justify his overthrow of Cronus in the papyrus. 5 BMCR 2006.10.29, with a response by the editors in BMCR 2006.11.02 and my reply in BMCR 2006.11.20. 2 The author holds the remarkable belief that God is Mind, and is also identical with the physical element Air. He offers equations between a number of different gods, stating that ‘“Earth” (Ge), “Mother” (Mētēr), “Rhea” and “Hera” are the same’ (col. 22), and indeed that ‘“Heavenly Aphrodite”, “Zeus”, “Persuasion” and “Harmony” are conventional names for the same god’ (col. 21): gods of different sexes are equated, and Zeus is part of the equation. It is easy and I believe correct to infer from this that the author was a monotheist, one of those logioi who according to Democritus ‘stretch their arms towards the sky, which we Greeks now call “Air”, and say that all things are Zeus, he knows everything, gives and takes away everything, and is king of everything’.6 It seems that, when the author refers to multiple gods, he is reporting the opinions of others. According to the editors’ reconstruction of the opening columns, he does believe in daimones, which he thinks are souls of the dead. He aims to remove from religion any beliefs which seem impious, including the belief in bloodthirsty deities like the Erinyes (col. 6): he thinks that it is progress to replace them with vengeful souls. In col. 5 as I supplemented it, he also mocks those who refuse to believe literally in the terrors of Hades, ironically offering to enter a shrine to consult an oracle about them. He offers to ask an impertinent question, namely whether it is right to disbelieve in the terrors of Hades, rather like how Chaerephon asked Apollo whether anyone was wiser than Socrates. As I argued in 2001,7 the opening columns, combined with the criticism of the Orpheotelestai in col. 20, make it clear that the author was discussing the Orphic poem not as an end in itself, but as an illustration of another point. I suggested that his aim was to show that, although religious texts and religious rituals must not be rejected entirely, they both need interpretation and could not be taken literally. To take them literally is a danger to one’s faith in the divine itself. In my view, this author’s extraordinary approach explained why, in the last third of the fifth century BC, many Athenians who held traditional religious beliefs became so suspicious of the presocratic physicists that they exiled Anaxagoras and sentenced Socrates to death.8 Accordingly, a successful reconstruction of the opening columns will be vital to confirming or refuting this interpretation, or the opposite view that the author was an Orphic priest and religious practitioner. The topics of these columns seem to vary so wildly that one might well have doubted that the reconstruction was correct. However, by making a scale model of the roll and testing every possible reordering of the fragments, I successfully verified the editors’ physical reconstruction of columns 4–7. Indeed, I have been able to place some further pieces in columns 6 and 7, which mostly confirm their text. My model also shows that there were in fact at least 28 columns.9 Since the first two columns are in a desperate condition, I shall retain the editors’ numeration here. 1. New joins in Columns 6–7 First, a tiny triangular piece, fr. I 70, belongs in col. 6, between the lower parts of frr. G 14 and G 2.10 The fibres are clear and match perfectly, and so do the letters in lines 8–9, which confirm the earlier supplements; the new piece completes the M in mãgoi!. It also yields a new reading in line 10. I omit lines 1–7 (the inserted letters are in bold): 6 Democritus fr. B 30 D.–K.: t«n log¤vn ényr≈pvn Ùl¤goi énate¤nante! tå! xe›ra! §ntaËya, o nËn “±°ra” kal°omen ofl ÜEllhne!, "pãnta, , “ZeË!” muye›tai ka‹ pãny' oto! o‰de ka‹ dido› ka‹ éfaire›tai ka‹ ba!ileÁ! oto! t«n pãntvn". 7 Art. cit. (n. 4). 8 The link between Anaxagoras and Socrates is of course the fact that Anaxagoras’ pupil Archelaus of Athens was Socrates’ teacher. The reality of this link is confirmed by the report of Ion of Chios, a contemporary witness who had no reason to invent or conceal the facts, that Socrates voyaged to Samos in Archelaus’ company (D.L. 2. 23). 9 I would need copious photographs to demonstrate this. For all illustrations the reader will need to have on hand the new edition of 2006. 10 For col. 6 see the new edition, Plate 6, and for fr. I 70 see Plate 30. Reconstructing (again) the Opening of the Derveni Papyrus 3 Col. 6 7 10 11 xoå! p`oioË!i. énãriyma` [ka]‹` poluÒmfala tå pÒpana yÊou!in, ˜ti ka‹ afl cuxa[‹ én]ã`riymo¤` efi!i. mÊ!tai EÈmen¤!i proyÊou!i k[atå tå] aÈtå mãgoi!: EÈmen¤de! går [ cuxa¤ e`fi`!in. œn ßnek`[en ı m°llvn fl]e`rå yeo›! yÊein ÅÙÄ_y`´[r]n¤`y`[e]ion prÒteron [yÊei . . . . . . .]a`i!`p`o`t`e`[. .]tai ‘They sacrifice cakes which are countless and many-humped, because the souls too are countless. Initiates make a first sacrifice to the Eumenides in the same way as magoi do; for the Eumenides are souls. For these reasons a person who intends to make offerings to the gods first sacrifices a bird . . .’ The editors read line 3 of fr. I 70 (= col. 6,10) as ].ra[, commenting that the damaged letter is ‘very likely’ to be I. The letters ra rule out their supplement tÚn m°llonta. It was not hard to find the right supplement to replace it. I tried ı m°llvn m]∞`ra yeo›! yÊein, but the correct phrase for ‘offer thigh-bones to the gods’ is m∞ra ka¤ein. Instead, digital enhancement of the published photographs suggests that the alleged iota may be made up of the tips of the square capital E that the scribe uses, and that one should therefore read ı m°llvn fl]e`rå yeo›! yÊein. There are parallels in Plato and Lysias.11 Secondly, I have been able to place within col. 7 two narrow vertical strips of papyrus, which were published as frr. I 7 and I 55 in the new edition.12 These strips fit perfectly at different heights in the place where they must have formed a sovrapposto, since detached, to F 3a from the layer below, which is part of col. 6. (The straight right edge of F 3a is not a kollesis, as I at first thought, but must be the result of a razor-cut made by Anton Fackelmann when he opened the roll in July 1962.) The fibres match well on either side in both pieces, so far as one can tell from the published photographs. The new joins confirm the previous editors’ supplements, save that Orpheus’ poem is quoted in direct speech rather than in oratio obliqua, i.e. as yÊra! §p¤ye!ye rather than yÊra! §piy°!yai. My text includes one minor change that the spacing requires (kayareÊ]o`nta! rather than ègneÊ]o`nta! in line 10), now that the photographs reveal what the papyrus actually looks like; it also reverts to the readings of the editors where they have proved to be correct (the newly joined fragments are in bold): Col. 7 2 5 10 11 . . Ï]mnon` [Íg]i`∞ ka‹ yem[i]t`å l°go[nta: flerologe›]to går t]∞`i poÆ!ei, k`a‹ efipe›n oÈx oÂon t' [∑n tØn t«n Ù]nomãtvn y°]!ig ka‹ t[å] Ér=hy°nta. ¶!ti d¢ m`[antikØ ≤] pÒh!i! ka‹ ényr≈[poi!] afini`[gm]at≈dh!. [ı d]¢` [ÉOrfeÁ]! aÈt[Ú]! §]r¤!t' afin[¤gma]ta oÈ`k ≥`yele l°gein, §n` [afin]¤gma!`[i]n d¢ me]g``ãla`. fler[olog]e›`tai m¢n oÔg ka‹ é`pÚ [to]Ë pr≈tou ka‹] m°xri oË` [tele]uta`¤`ou =Æmato!, …`[!] dhlo`[›] ka‹ §n t«i eÈk]rinÆtv[i ¶pei. “y]Ê`ra!” går “§p¤ye![y]e`” ı [ke]l`eÊ!a! to›`[! »!‹]n aÈt[oÁ! oÎ ti nomo]y`e`te›m fh[!in to]›`! pollo›!, éllå didã!kein toÁ! tØ]n ékoØn [kayareÊ]o`nta!, kat[å hymn that tells of wholesome and permissible things. For he was speaking allegorically with his poetic composition, and it was impossible to state the application of his words and what was meant. His composition is prophetic and riddling for people. But Orpheus himself did not wish to utter disputable riddles, but important things in riddles. In fact he is speaking allegorically from his very first word right through to his last, as he reveals even in his easily-explained verse: for the one who bids them ‘shut the doors’ on their ears is saying that he is certainly not making laws for the many, but instructing those who are pure in hearing . . .’ 11 Lys. 26.8, m°llonto! êrjein flerå yË!ai; Pl. Leg. 10. 910c, yÊ!a! flerå yeo›!. 12 For col. 7 see Plate 7, and for frr. I 7 and I 55 see Plates 29 and 30 respectively. 4 Fr. I 7,1 (= col. 7,6) is read by the editors as ]e.[, with the damaged letter ‘part of a midline curve meeting an upright, possibly f’; this must be N. Lines 2 and 3 read ]po[ and ]dh[ respectively. The editors read the damaged letter in fr. I 7,4 (= col. 7,9) as the tip of a high horizontal; this must be E. In fr. I 55 (= col. 7,8–11), much of the ink is invisible to the eye or indeed completely lost. The L is clear in line 1 (= col. 7,8), there is a diagonal in line 2, and a S in line 3. In addition, I have adopted M. L. West’s proposal m`[antikØ13 in place of the j`[enÆ ti! of the editors: they report traces of Z or J, but since West had earlier read a`[ÈtoË14 it is possible that there were or still are traces of an entire upright rather than the top and bottom of two horizontals. 2. The Reconstruction of Columns 1–3 The editors’ reconstruction of the first three columns is profoundly mistaken. To explain this, I must briefly recapitulate some bibliological aspects as they can be deduced from Herculaneum papyri and from Fackelmann’s report.15 Like the rolls from Herculaneum, the Derveni papyrus was difficult to unroll. Yet Fackelmann opened it successfully, using a combination of papyrus-juice and static electricity to separate the layers and put them between glass. This extraordinary achievement was made possible by the fact that at Herculaneum the rolls were contorted by superheated steam before they were carbonized and then crushed by debris from the volcano, whereas the Derveni papyrus was simply carbonized. Most of its layers came apart easily, as is evident from the extraordinarily good preservation of the pieces, at least when they were first photographed. The Derveni papyrus could not be unrolled continuously. In this it resembles other carbonized papyri. The last few columns still formed a scroll.16 However, its outer parts were broken into hemicylindrical stacks of fragments;17 there were some small multi-layered slivers of papyrus too. The single layers of each hemicylindrical piece have the important property that they have the same shape throughout. They are effectively a stack of fragments, that ought to get larger as you approach the exterior of the roll, as the circumferences broaden. In Herculaneum papyri they do get larger, but that is not the case with the Derveni papyrus. Its pieces diminish again as you reach the outermost layers, which, as a result of incomplete carbonization, had lost larger parts of their margins when Fackelmann separated the layers.18 But at least they still have much the same outline as one goes from one layer to the next. Except in cases where these hemicylindrical pieces have been broken into smaller fragments, two such pieces form a circumference. The pattern repeats with each successive circumference, until you reach either the middle or the exterior. For instance, in col. 20 the editors call the alternating fragments B and D;19 there are four hemicylindrical pieces, making two circumferences. The editors noticed these recurrences, and exploited them to make a reconstruction of the last 22 columns that is for the most part physically correct. Thus in the middle of the roll, as is seen in the editors’ Plates 7–22, alternating pieces 13 In P. Struck, Birth of the Symbol: Ancient Readers at the Limits of their Texts (Princeton, 2004), 33 n. 4, who hit upon the same reading independently. Unfortunately Struck’s text of the rest of the papyrus, on which his discussion of this item relies, was outdated soon after his otherwise excellent book appeared. 14 The Orphic Poems (Oxford, 1983), 78. 15 In Ts.3 pp. 4–5, with notes by Tsantsanoglou. 16 Scholars of the Herculaneum papyri call this a midollo or ‘marrow’. Fackelmann called it ‘piece I’ (now frr. A 1–9 and D 12–14). 17 The Neapolitans called these scorze, ‘bark’ like that on a tree. Fackelmann records that, when he came to work on the papyrus, these outer pieces comprised ‘piece II’, a detached stack of 24 layers (now frr. B 1–12 and C 1–12), and ‘piece III’, a stack of 33 layers, the outside of which was completely burned and stuck together. Tsantsanoglou notes that this corresponds to frr. D 1–11, E 1–13, and F 1–9, but frr. F 1–15 can also be referred to it. 18 Cf. especially Fackelmann loc. cit. 5, on piece III: ‘die äußere Seite war total verkohlt, die Lagen ineinander verschmolzen . . . der Rand war zu 2/3 ebenfalls total verkohlt, während am übrigen Teil des Randes die Lagen schwach zu erkennen waren . . . Die Lagen klebten aneinander. Das Ablösen ergab 33 Lagen, die zum Schluß durch die starke Verkohlung immer kleiner wurden.’ 19 See their Plate 20. These are Fackelmann’s pieces II and III respectively. Reconstructing (again) the Opening of the Derveni Papyrus 5 from both hemicylinders20 build up 23 complete circumferences. Both of them find an immediate continuation, with no layer lost, in the inner continuous scroll,21 which suggests that they were both broken off from the latter during the excavation.22 Another stack of fragments was the outermost portion of one of the two hemicylinders.23 However, there are serious obstacles to the reconstruction of the opening columns. Many of the unplaced fragments are tiny, and many are very hard to read. The high number of letters per line (36–38) makes filling the lacunae difficult; the ratio of lost to surviving text is very high, and the great width of the circumferences means that it is hard to find matches between the fibres, since the lacunae are broad. Worse still, outer layers of the hemicylindrical pieces were broken into small fragments; the fragmentation is greater in F-fragments 1–15, which belonged to one half-cylinder, than in the G-fragments, which belonged to the other. Most of the large unplaced pieces come from these two series.24 Thus there are three constraints on any reconstruction. It must preserve the correct ordering of the layers within their stacks, it must give reasonable lengths for the kollemata, or sheets of papyrus, and it must give reasonable widths for the distance from the left margin of one column to that of the next. The principles outlined above prove that the editors’ reconstruction of columns 1–3 cannot be right. In the opening columns there was just under one column per circumference, as cols. 4–7 make clear. The alternating fragments from different hemicylinders come from the F series and the G series. The reader needs to bear in mind that a single hemicylinder may break into several fragments. However, in reconstructing col. 2 the editors have juxtaposed three G-layers one after the other.25 This is physically impossible, since it contradicts the principle of hemicylinders: fragment G 7 cannot belong between G 8 and G 15+G 6, because only a piece from the F series can stand here to complete the circumference, whereas these pieces derive from a piece of the same shape and are therefore from different layers of the same half-cylinder. In the right half of col. 2, the editors placed fragments G 15, G 6 and G 5 next to each other to form a single hemicylinder. Both the letters and the fibres confirm that G 15 joins with G 6, to form two thirds of the hemicylinder; I will call the resultant composite G 15+G 6. However, the narrow fragment G 6 in the centre of the hemicylinder has the same shape as the left side of G 5, and cannot therefore join it; this is confirmed by the fibres, which do not match. Instead, G 6 must belong in a layer that lay above or below G 5. Now G 5 has a vertical crack running down it that corresponds to the location of the right edge of G 6. Digital enhancement of the photograph shows that the right edge of G 6 is in fact a kollesis, or join between two sheets of papyrus, where parts of two layers are visible. The editors missed this kollesis, no doubt because it is so close to the edge of the fragment.26 The extra thickness where there were two overlapping sheets of papyrus at the kollesis caused the vertical crack that runs down G 5. It also caused another phenomenon in G 5, namely that part of a different, overlying layer, called G 5a by the editors, has remained stuck to the upper right portion of G 5. They rightly interpreted this as a layer from one circumference further in,27 and therefore placed this fragment one circumference later, in col. 3.28 This is correct, but fragment G 5a needed to take fragments G 15+G 6 with it: these pieces too belong in col. 3. Fragment G 6 fits well in shape beside G 5a, although one cannot confirm this from the fibres, because the kollesis comes exactly in between and so no match in fibres is expected or possible. When 20 These are Fackelmann’s pieces II and III respectively. The Italians call these hemicylinders sezioni. 21 I.e. Fackelmann’s piece I. 22 I owe this suggestion to J. Hammerstaedt. 23 Fackelmann’s piece IV. This was broken into two smaller stacks. It was the outer part of the hemicylinder of his piece II. It joins the layers of piece II, with no layer lost in between, at cols. 6–7 (Ts.3 Plates 6–7). Fackelmann’s mention of a photograph permits one to identify piece IV with frr. G 1–21. 24 See the new edition, Plate 27. 25 See their Plate 2. 26 The first kollesis that they observed is in col. 4, near the left edge of fr. F 15 (see their Plate 4). 27 The Neapolitans call this a sovrapposto. 28 See their Plate 3. 6 fragment G 15+G 6 is placed in col. 3, the fibres of the left edge of fragment G 15 match the right edge of fragments F 9+F 8, which the editors put in the left half of column 3. Moving fragment G 15+G 6 to col. 3 has the valuable effect of bringing the discussion of the Erinyes closer to col. 4, where Heraclitus refers to them, rather than leaving them in col. 2, where they are further removed from that discussion. It also implies that fragment G 11 cannot be from col. 3 where the editors had put it. Fragment G 11 is from the left half of a G-layer. My reconstruction established that it belongs to the previous layer, i.e. in column 2, where it joins the left edge of G 5. The fibres confirm this join. Does this reconstruction pass the twin tests of width of kollema and width of column? The distance from the left margin of the new col. 3 to that of col. 4 is 11.7 cm., which is average for this part of the roll. However, the length of the kollema which begins at the right edge of fr. G 6 half-way across the new col. 3 and ends near the left edge of fr. F 15 near the right margin of col. 4 works out at 12.85 cm., which is shorter than the average in the roll of 15.8 cm.; the shortest kollema previously known is reported to be 13.7 cm. wide (col. 26),29 and the second-shortest is 14.4 (col. 25). But the kollemata early in the roll seem to be shorter than the rest: cols. 7–25 are all written on sheets 16.0–17.0 cm. wide, whereas the first two kollemata are 14.0 and 14.6 cm. wide respectively.30 I do not think this finding disproves my proposed order of fragments: all the other arrangements of fragments that are possible (and I have tested them all) give readings that diverge from what is expected much more than this does. I shall return to the text of col. 3 below. Col. 2, according to my reconstruction, is in a sad state. It was 10.65 cm. wide, which is close to the average of 11.1. A kollesis ought to have fallen about one-third of the way across the column, but no such kollesis can be identified in any of the surviving pieces. This coincides with where the F-fragment ought to be. Since the F-layers were breaking up into small fragments, I believe that the left half of col. 2 disintegrated because of the kollesis (or else stuck to the overlying layer) and will be very hard to reconstruct, if it is not destroyed entirely. All that can be recognized of the left part of col. 2 is the margin, and little sense can be extracted from the remaining text. Col. 1, however, is much better preserved. There are two G-fragments, G 7 and G 17+ G 8. The latter contains the ends of col. 1, the intercolumnium, and a few letters from the left edge of col. 2. The former, G 7, is displaced from the old col. 2, where the editors had put it between other G-fragments, impossibly, as we have seen. If G 7 is moved to col. 1, it accords well with the right edge of col. 1 (= fr. G 8 col. i), since both G-fragments are concerned with signification (!hma¤nein) and both also have the distributive adjective ßka!to! in them. This led me to search for unplaced F-fragments that might occupy the middle of the column, where the F-layer must have broken into pieces. I found that F 10 and F 18 are both concerned with signs (!hme›a), and both also mention prayer (eÈxÆ). F 18 also mentions ßka!to!. F 14 mentions prayer also, and like F 18 mentions fire. Hence I propose that these pieces form the middle of col. 1. The high level of repetition in the style of this treatise makes such a tightly bound verbal nexus plausible in principle. I fitted the pieces together according to the shapes and the sense that emerged. Although it is hard to be certain of the fibres because of the distances between the pieces, the right edge of G 7 appears to match the left edge of F 14. The distance from the left margin of col. 1 to that of col. 2 was 11.9 cm., which is wide but paralleled in col. 5. The widest column, col. 20, was 12.2 cm. The kollesis would have fallen in the lacuna to the left of G 7, affecting only a few letters close to the left margin. This rearrangement of fragments resulted in the following new text of cols. 1–3. (In the headings, the hemicylinders are delimited by upright bars ||.)31 29 Ts.3 p. 6. I could not measure this, since the editors do not illustrate the agraphon at the end of the roll (one would like to be sure that it has no writing!). 30 These are my calculations, based on the model. The editors think the first two kollemata are wider (14.6 and 14.9 cm. respectively), but the model and experiments with Adobe Photoshop show that this is inaccurate. 31 Note that I have printed without dots all those letters where the editors put dotted letters on the left-hand as well as the right-hand page, which I took to mean that the letters are damaged but nonetheless certain (see BMCR 2006.10.29); the editors’ Reconstructing (again) the Opening of the Derveni Papyrus 7 Col. 1 (G 7 || + F 10 + F 14 + F 18 + H 45 || + G 17 + G 8 col. i) 1 5 10 12 . . . . . . . . . . . . . . . . . . . . (.)]keu`[. . . . . . . . . . . . . . . . . .(.)]h`id[. . . . . . . . . . . . . .] §p‹ ta[. . . . . . . . . . . . .(.) . . . .]daram`[. . . . . . . . . .(.)]k`o›! ka‹ ka[. . .]in[. . . . . . . . .(.)].ntana`[. . . . . . . . . . .]ai tå !hme[›a é]n' ßka!ton`[ mer]¤di neim[. . . . . . . . . d]Êo` eÈx∞! [. . .].a`!` . . . .] §p°yhke[n À]!`per fu!ik[Ú! .]e`x.` [. . ., m]h`d¢` §å`n` kat]å` tå !hmai[nÒ]mena eÈxa`[›! flere›a] y`e`«n •kã]!tvn kãv[!in] énhmm°[na . . . . . .].e`!.` .(.)u` . . . .]m`v! §p' [eÈ]x∞! . . . . .l[. . . . . . . .(.) mi]ã`n`eie . . . . . . . .(.) p]urÒ!, Ïdato! de`[. . . . . . . .]. .[.(.) !hm]e›a . . . . .(.) §!t]i`n ßka!ta !hme›a` [. . . . . . . . . . . . . . . .(.) . . . . . . . .(.)]l`u! ka‹ tîll' ˜!a` [. . . . . . . . . . . . . . . . . incertum quot versus desint ‘. . . towards (?) the . . . for (plural noun missing) and they burn (?) . . . allocate the signs to each (object missing) in turn (1–2 words missing) two (noun missing) of prayer (word damaged). He added these things (?) like a natural scientist (word damaged), and not even if, according to (?) the things that are signified, with prayers (word missing) for each of the gods they burn sacrificial victims (?) that are ignited . . . but all the same (?) in the case of a prayer . . . might pollute . . . of fire, but (subject missing) of water . . . signs reply in BMCR 2006.11.02 did not contradict my interpretation, though one may heartily wish that they had explained their practice. Such letters are shown underlined (e.g. d) in the apparatus criticus. Thus the sigla there are as follows: Bernabé Poetae Epici Graeci. Testimonia et Fragmenta. Pars II: Orphicorum et Orphicis Similium Testimonia et Fragmenta fasc. 1 (Munich and Leipzig, 2004); cf. La Théogonie orphique du papyrus de Derveni, Kernos 15 (2002) pp. 1–38 Burkert W. Burkert, Orpheus und die Vorsokratiker. Bemerkungen zum Derveni-Papyrus und zur pythagoreischen Zahlenlehre, A&A 14 (1968) pp. 93–114, esp. p. 93 n., and new conjectures communicated to me privately ed. Der orphische Papyrus von Derveni, ZPE 47 (1982), after p. 300 (edition of columns 3–26) Ferrari F. Ferrari, Note al testo delle colonne II–VII del papiro di Derveni, ZPE 162 (2007) pp. 203–11 Kouremenos Th. Kouremenos in Ts.3 Laks et Most A. Laks and G. W. Most (edd.), Studies on the Derveni Papyrus, Oxford 1997 Livrea E. Livrea, Eraclito nel papiro di Derveni, ZPE 164 (2008) Par.1 G. M. Parássoglou and K. Tsantsanoglou, Heraclitus in the Derveni Papyrus, in A. Brancacci et al. (ed.), Aristoxenica, Menandrea, Fragmenta Philosophica, Studi e Testi per il Corpus dei papiri filosofici greci e latini 3, Florence 1988, pp. 125–33 (col. 4) Par.2 G. M. Parássoglou and K. Tsantsanoglou, Heraclitus 1T, Corpus dei Papiri filosofici I.1**, Florence 1992, pp. 221–6 (col. 4) Ts.1 readings and conjectures by K. Tsantsanoglou in Laks and Most 1997, pp. 9–22, with his edition of columns 1–7 (ibid. pp. 93–128) Ts.2 readings and conjectures of K. Tsantsanoglou in Bernabé 2002, 2004 (vid. sup.) Ts.3 T. Kouremenos, G. M. Parássoglou and K. Tsantsanoglou, The Derveni Papyrus, Florence 2006 * the present editor a` littera dubia quae aliter legi potest a littera fracta quae tamen secundum edd. pr., ut videtur, non dubia est [a] littera ab editore suppleta {a} littera ab editore deleta littera ab editore inserta $a¸ littera e fonte gemino ab editore suppleta _a´ littera a librario deleta ÅaÄ littera a librario supra versum addita [.] littera deperdita [] una vel nulla littera deperdita [.(.)] una littera vel duae litterae deperditae reliquiae totidem litterarum incertarum. ..... 8 . . . all (plural subject missing) are signs . . . much (?) (singular noun missing) and the other (plural noun missing), as many as . . .’ imagines contuli et frr. coniunxi (nisi quod F 18 + H 45 et G 17 + G 8 coniunxerat Ts.3); fibrae frr. G7 et F14 bene congruunt 1 u` vel f`, c` oÈ]k eÈ` vel kef`[ Ts.3: oÈ]k eÈ[x- * 2 h` vel m`, n` potius quam p` EÈme]n`¤d[i Kouremenos p.143 3 dara Ts.3 m`[Òno! Ts.3 p. 143: n`[hfal¤oi! Ferrari m` vel n`, p` k` vel x` kã[ou!]in * 4 ka‹ * ]. vestigia incerta p]ã`nta Ts.3 a` vel l` na`[Òn Ts.3: na`[!mÒn Ferrari fort. k]a‹ vel e‰n]ai vel -t]ai !hme›[a Ts.3 n` vel i`[ nihil in fine versus deesse censeo 5 mer]¤di *: EÈmen]¤di Ts.3 et prap]¤di Ferrari, longiora ne›m`[ai Ts.3, sed m monstrat imago d]Êo` *: ta]Ëy`' vel êne]uy`' Ts.3 o` vel y` a` potius quam l`, d` !` vestigium incertum 6 taËt' potius quam pur‹ * §p°yhke[n Ts.3 (e` potius h`; p` vel g`) À]!`per *: oÂ]a`per vel kay]ã`per Ts.3 !` potius quam a` fu!ik[Ò! Ts.3 e` vel j` x`. vestigia dubia m]h`d¢` §å`n` * h` vel n` e` vestigium valde dubium a` potius quam d` imago: d` Ts.3 n` vel h` 7 kat]å` tå * (a` vel !` imago ut videtur): ] tå Ts.3 !hmai[nÒ]mena *: !hmai[nÒmena Ts.3 eÈxa`[›! flere›a * a` vel l`, sed fort. ad stratum alium pertinet y` vestigium incertum e` vel t`, p`, g`, z`, j` imago ut videtur y`e`«n *: ÉEri]n`Ê`vn Ts.3 (n` vel i`; u` potius quam t`) 8 •kã]!tvn *: mÊ]!`t«n Ferrari: ]!tvn Ts.3 kãv[!in *: kat[ Ts.3, haud recte énhmm°[na *: énhmm°[n- Ts.3 vestigia incerta in fine versus: nihil leg. Ts.3 9 éll' ˜]m`v! * mv!e Ts.3: pro m` fort. potius !` imago, ut mihi videtur eÈ]x∞! Ts.3 . . . . . vestigia valde dubia mi]ã`n`eie *: kerd]ã`n`eie vel jhr]ã`n`eie vel f]ã`n`eie possis a` vel d`, l` n` vel p`, h` ut videtur fragmentum abscissum litt. ! praebens temere sub finem versus collocatum esse iudicavit Ts.3 10 p]urÚ! Ts.3 in asyndeto: distinxi e` vel i`[ (i praef. Ts.3) ]. .[ vestigia in alto !hm]e›a potius quam mant]e¤a * 11 §!t]i`n * i` potius quam h`, p` 12 po]l`Á! Ts.3: fi]l`Á! vel éx]l`Á! * l` vel a`, d` a` vel l`, d` The new col. 1 is about divination from signs, apparently using lots. énã with the accusative has its distributive function in line 4.32 In line 5 the editors restored EÈmen]¤di ne`›m`[ai, but the Eumenides are always a plurality. Instead I suggest (§n) mer]¤di, ‘in turn’. Sortition is a known form of divination both in Greece and the Near East; the number dÊo ‘two’ appears next to a mention of prayer. Another possibility is that the topic is the division of meat at a sacrifice, where the officiant allocates more to a given portion to keep the portions equal. Line 5 is short. The scribe’s practice elsewhere suggests that it is the end of a sentence, possibly indeed of a quotation. In accord with this, line 6 may well have said that someone ‘added this like a natural scientist’, unless §p°yhken refers to putting a substance on a sacrificial fire. We may compare how in col. 4 the author says that Heraclitus ‘was speaking like a muyÒlogo!’ when the Ephesian says that the Erinys will keep the Sun within his bounds. Each time the author is concerned to note that someone who might have been expected to speak in religious terms, i.e. a religious expert, is speaking like a physicist, or vice versa.33 After something about things that are signified and prayer, some people burn an object or objects that have been lit. Prayer and pollution seem to be mentioned. Pollution would be a perfectly appropriate topic, since one of the purposes of divination is to find out what is causing pollution so that it can be remedied. At the end of the column fire is contrasted with water; the editors had already suggested that F 18 referred to pyromancy and hydromancy. Divination from fire and from water are rarely attested directly in Greece until the Roman period. However, Philochorus and Apollonius Sophista34 tell us that the yuo!kÒoi, a kind of mantis known from Homer onwards,35 divined the future from offerings made by means of fire. A scholiast on Il. 24. 221 calls them libanomãntei!, ‘diviners from incense’. Later the main form of divination from fire was lychnomancy, which meant gazing into the flame of a lamp, as in Lucian’s Lychnopolis in Vera Historia 2. Both lychnomancy and lecanomancy, i.e. divination by gazing into a bowl of water, are well attested in the magical papyri from Egypt of the Roman period. It is not clear to me whether the author is speaking of Persian or Greek practices, just as the identity of the mãgoi in col. 6 remains hotly debated; are they the Persian caste of priests or the itinerant Greek magicians? We are very ill informed about Persian divination. Herodotus36 and Diogenes Laertius37 credit the magi with divination, but the only methods speci32 LSJ 9 s.v. énã, C. II. 2. 33 A philosopher can charge with atheism either mythographers and poets, as Isocrates does in his Busiris, or scientific materialists, as Plato does in Leg. 10, 886a–b, where he gives émay¤a alongside ≤donÆ as the causes of such beliefs, exactly as the Derveni author does (col. 5, 8–10). This does not of course make Plato any less of a physikos than our author is. 34 Philochorus 328 fr. 178b Jacoby; Ap. Soph. p. 88,33 Bekker. 35 Il. 24. 221, mãntie! . . . yuo!kÒoi. 36 7. 37, 7. 43. 37 1. 7. Reconstructing (again) the Opening of the Derveni Papyrus 9 fied are divination from dreams and from eclipses. However, Strabo38 states that ‘the Persians’ practised necromancy and hydromancy. Lecanomancy was already known to the Hittites and to Joseph in Genesis 44. Writing in the sixth century AD, Agathias alleges that the Persian magi practised pyromancy.39 Heraclitus, who is soon quoted in col. 4, was undoubtedly influenced by Persian religion, above all in the role he gives to fire, but the author quoted him for a different reason – because he was a physicist who resembled a religious expert, or vice versa. In my reconstruction, col. 2 is still very fragmentary, since I have been unable to find any F-fragments that can be proved to belong in the missing F-layer. In addition, the photograph of much of fragment G11 is very hard to decipher, even though the fibres prove that it belongs to the left of fr. G 5a. Accordingly, I will print only lines 7–9 here, and cannot supply an apparatus criticus either. Col. 2 (G8 col. ii || + || + G11 + G5 + H7) 7 9 ka`[. . . . . . . . . . . . . . . . . . . . . . .(.)]i` p`ã`nt`a! x`rØ [ ] ê` d ikoi` o` ye›Òn ti [ . . . . . . . . . . . . . . . . . . . . . . . . . . .(.) ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .(.)]l`ou d`[¤]khi [ ‘. . . all (masculine plural noun missing) must . . . unjust persons (word damaged) something divine . . . but not (?) with justice . . .’ The editors saw in line 8 a reference to a kind of bird (Ùrn¤yeiÒn ti), but the bird belongs in col. 6, not here. References to wrongdoers and justice lead well into the appearance of the Erinyes in the next column. A reference to justice seems more apt to the context than had the editors’ Ïmnou! èrm]o!to[Á]! t∞`i` m`ou!`[i] k∞i. Col. 3 (F9 + F8 || + G15 + G6 + G5a || + F7 col. i) 1 5 10 . . . . . . . . . . . . . . . . . . . .(.)]vn: [. . . . . . . . . . . . . . . . . . . yu]h`l`a¤`, …`!.[. . . . . . . .(.)] ÉE`rin[u. . . . . . . . . . . . .(.) . . . . .(.)]v`g g¤neta[i . . . . .(.)] tim«!in` [.(.)]i`r`h`l`[. . . .(.)]r`h toÁ! ≥]d`h §j≈lea! [. . . .(.) x]oa‹ !tagÒ!in` ÉE`rinÊv`[n. ofl] d`¢ d]a¤mone!, o„ katå [toÁ! m]ã`gou! timå! [t]h`ro`Ë`!i [t«n] [ ye«n, Íphr°tai d[¤kh! . .]. •kã!toi`! or[. . . . . . .]i efi!in, ˜pv!per a[. . . . . .(.)].o!to[.]!to[. . . . . . . . .]noi: afit¤hn [d' ¶]xou!i[. . . . . . . . . . .(.)].! t[oio]uto`[. . . . . ., o·ou! g`[. .]e`[. . . . . . . . . . . . . . . . .(.)]n`e[. . . . . . . . . . . .]u!t[. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . incertum quot versus desint ‘. . . divinations from burnt-offerings (?), as . . . Eriny(e)s . . . becomes . . . libations in drops to the Erinyes honour (word missing) those people who are already annihilated. But the daimones, who according to the magoi observe the honours of the gods, are servants of justice . . . for every (plural noun missing) . . ., just as . . . (plural participle missing). But they are responsible . . . such persons . . . as . . . later (?) . . .’ imagines contuli et frr. coniunxi 1 n imago: i Ts.1, Ts.3 spat. vac. ii litt. ut videtur: incertum est an verba in fine versus desint 2 yu]h`l`a¤` * h` vel p`, n`, k`, i`i` l` vel a` ai`v`! Ts.3 i` pes lineae vert. fort. v` pars dext. .[ vestigia incerta e` vel k`, x` ÉE`rin[u- Ts.1: num k`rin[-? 3 da¤m]vg Ts.3 g¤neta[i ed. n` vel h` i` vel n`, h` r` vel b`, p`, g` h` vel e`, n, p`, g`, i`[` (linea vert. sin.) l` potius quam a`, d` r` vel b` x]r`Ø Ferrari: sed fort. etiam é]r`Ø` vel blã]b`h 4 toÁ! ≥]d`h *: ≤ d¢ d¤]k`h Ferrari d` vel l`, k`, a`, x` §j≈lea! Ts.3 afl * x]oa‹ Ts.3 n`: potius g`, p`, r` e` vel t`, z`, j`, g` ÉE`rinÊv`[n ed., sed litt. v paene certam praebet imago (v` potius quam !`) ofl Battezzato et Ts.3 5 d]a¤mone! ed. o„ ed.: ofl West ap. Laks et Most p. 83 katå Ts.3: kãtv ed. toÁ! m]ã`gou! * (a` vestigium medium lin. vert. praebet imago; g imago): ] t`oÁ! Ts.3 t]h`ro`Ë`!i Ts.3 t«n vel tå! * 6 d[¤kh! * d[ Ts.3 potius quam z`[, !`[: k[ ed. ap. fr. B 5 •kã!toi`! * (i` vestigium medium lineae vert. in spatio angusto praebet imago): 38 16. 3. 39. 39 Hist. 2. 25. 10 •kã!to[i]! Ts.3 ır[k- * (litt. r certam praebet imago): Ùr[n¤yeion Ts.3 7 ]. fort. l` vel !`, a`, d`, k`, x` ut videtur èrm]o!to[Á]! Ts.1: fort. Ù!to[›]! vel pol]l`o!to[›`]! vel -o! to[Á]! potius quam n]Ò!to[.]! 8 d' Burkert: t' Ts.3 ¶]xou!i Ts.3: ‡!]xou!`[i ed. fort. to]Á`!: ].! Ts.3 t[oio]uto`[- *: t[. .]uto`[ Ts.3 o` vel y` 9 g` vel p` 10 fort. to]Á! t[ vel m]u!t[- vel Í!t[er- The new col. 3 may open with a reference to yuhla¤, which Hesychius defines as ‘divination by means of burnt offerings’.40 Next we have a mention of the Erinyes and ‘people who are wiped out’ (§j≈lei!). This is the normal term for people who are destroyed so utterly that even their descendants are annihilated, often because they are under a curse. Thus Antiphon speaks of ‘swearing the greatest oath, putting yourself under the curse that you and your family and your household will perish utterly’.41 People who die before their time or under a curse can be expected to cause problems after their deaths. The new reconstruction seems to say that ‘libations in drops of the Erinyes honour those who are already wiped out’, but xoa‹ ÉE`rinÊv`[n] must be an objective genitive, meaning ‘libations to the Erinyes’; cf. xoa‹ kekmhkÒtvn, ‘libations to the dead’.42 The opinion that libations offered to the Erinyes placate the souls of those who have perished §j≈lei! is, I believe, an interpretation by the author. The traditional view was that the Erinyes can act on behalf of the dead to carry out their anger; thus in Aeschylus’ Eumenides it is not Clytaemestra’s ghost that takes vengeance for her killing, but rather the Erinyes at her behest. The Derveni author is explaining that people placate the Erinyes with libations when, in his view, they are really mitigating the wrath of the angry dead. This is definitely an anti-traditional view, i.e. a learned interpretation, for the following reason: as S. I. Johnston remarked,43 in traditional belief those of the untimely dead who were men would not be expected to become female ghosts. Thus it is an interpretation of the same allegorical kind that we know from the rest of the papyrus, and not evidence for popular belief. Indeed, it implies that the Erinyes as such do not exist. The view that the Eumenides are the souls of the dead was embraced by intellectuals like Plato, who believed that the souls of the angry dead could take revenge on their oppressors directly and without the intervention of deities like the Erinyes: compare Leg. 927A–B, in a discussion of the proper treatment of orphans, where Plato says ‘the souls of the dead have some power, even though they are dead, by which they are concerned with human affairs . . . Let people fear . . . the souls of the dead, who have it in their nature to care in particular for their own offspring, and to be kindly (eÈmene›!) towards those who honour them and hostile to those who dishonour them’.44 Plato’s use of eÈmene›! of course evokes the Eumenides without acknowledging their existence. Somewhat similarly, an unidentified Pythagorean writer reported by Alexander Polyhistor thought that the air was full of souls, which people considered to be daimones and heroes and deemed responsible for divination of all kinds; but that writer also held that the Erinyes exist, since they fetter the souls of the impure in unbreakable bonds.45 The Derveni author continues that ‘the daimones, who according to . . . observe the honours of the gods, are servants’. tima‹ ye«n and Íphr°tai ye«n are familiar expressions.46 He attributes this opinion 40 y 837, tå! diå yumãtvn mante¤a!. 41 De caede Herodis 11.9. 42 Pseudo-Aristotle, De mundo 400b22. 43 The Restless Dead: Encounters between the Living and the Dead in Ancient Greece, Berkeley and Los Angeles 1999, p. 274. 44 afl t«n teleuth!ãntvn cuxa‹ dÊnamin ¶xou!¤n tina teleutÆ!a!ai, √ t«n kat' ényr≈pou! pragmãtvn §pimeloËn- tai: . . . fobe¤!yvn . . . tå! t«n kekmhkÒtvn cuxã!, aÂ! §!tin §n tª fÊ!ei t«n aÍt«n §kgÒnvn kÆde!yai diaferÒntv! ka‹ tim«!¤n te aÈtoÁ! eÈmene›! e‰nai ka‹ étimãzou!in dusmene›!. 45 D.L. 8. 31–2 = Alex. Polyh. F 93 Jacoby (58 B 1a D.–K.): ka‹ êge!yai tå! m¢n kayarå! (sc. cuxã!) §p‹ tÚn Ïci!ton (add. Cobet), tå! d¢ ékayãrtou! mÆte §ke¤nai! pelãzein mÆte éllÆlai!, de›!yai d' §n érrÆktoi! de!mo›! ÍpÚ ÑErinÊvn. e‰na¤ te pãnta tÚn é°ra cux«n ¶mplevn: ka‹ taÊta! da¤monã! te ka‹ ¥rva! nom¤ze!yai ka‹ ÍpÚ toÊtvn p°mpe!yai ényr≈poi! toÊ! te Ùne¤rou! ka‹ tå !hme›a nÒ!ou te ka‹ Ígie¤a!, ka‹ oÈ mÒnon ényr≈poi!, éllå ka‹ probãtoi! ka‹ to›! êlloi! ktÆne!in, e‡! te toÊtou! g¤gne!yai toÊ! te kayarmoÁ! ka‹ épotropia!moÁ! mantikÆn te pç!an ka‹ klhdÒna! ka‹ tå ˜moia. 46 Cf. A. PV 954, ye«n Íphr°tou; Soph. Ant. 845, timã! ge tå! ye«n pat«n; Xen. Mem. 4. 3. 14, toÁ! Íphr°ta! d¢ t«n ye«n eÍrÆ!ei! éfane›! ˆnta!; D.S. 8. 30. 1, throËnta tå! prÚ! toÁ! yeoÁ! timã!; D.S. 1. 73. 3, tã! te yu!¤a! . . . !unteloË!i ka‹ toÁ! Íphr°ta! tr°fou!i . . .: oÎte går tå! t«n ye«n timå! ’onto de›n éllãttein; Plut. De def. or. 417A–B, ≤me›! d¢ Reconstructing (again) the Opening of the Derveni Papyrus 11 to someone, using katå plus the accusative. With the old arrangement of fragments, the editors read t`ou! as part of the adjective §jair°]t`ou! timå!, but their photograph shows G, not T, probably with the base of a diagonal before it. Hence I have supplied katå [toÁ! m]ã`gou!. Some mãgoi are then the source of this piece of theology, which suits both the genuine Persian magoi, who had an elaborate demonology, and Hellenic magicians, who certainly regarded chthonic daimones as servants of the gods (they were often thought to be led by Hecate, for example). There was of course confusion, since Dionysius of Halicarnassus says that the Erinyes are servants both of the gods and of the daimones.47 I doubt whether Persian mãgoi are meant anywhere in the papyrus; but in either case, the Derveni author disputed their interpretation, as we know from col. 6. There he says that daimones who cause trouble are vengeful souls (da¤mone! §mpo[dΔn ˆnte!] c[uxa‹ timv]r`o¤), and that the Eumenides are souls (EÈmen¤de! går cuxa¤ e`fi`!in). If he thinks the Erinyes are daimones, as seems inevitable from Greek usage of the latter term, it follows from his argument that they are the same as the souls of the angry dead and would not properly exist. He does not deny that the incantations of the mãgoi are effective; what he denies is that they placate supernatural beings other than the souls of the dead. His opposition to the view of the mãgoi helps to confirm that he is arguing against them throughout his treatise. To return to col. 3, of whom or what are the daimones, who are presumably the Erinyes, said to be servants? The clue may lie in the adjacent columns, which mention justice; indeed, col. 4 probably quotes Heraclitus’ claim that the Erinyes are ‘allies of Justice’ (d¤kh! §p¤kouroi). The following lines, which described the responsibilities of the daimones, are damaged. However, a word beginning in ır- could well have referred to ˜rkoi or oaths, which would be fully appropriate in the context of the Erinyes and of people who are obliterated for committing perjury; for it is hard to supply a form of ırçn so near to efi!in in line 7. Depending on how we restore it, col. 3 breaks off with a reference to an initiate or to mysteries. This too would not be out of place, because S. I. Johnston has shown that initiands made offerings to malevolent spirits, and specifically to the Erinyes, at the start of the Eleusinian rites, and the Derveni author himself says in col. 6 that the initiates make a preliminary sacrifice to the Eumenides in the same way as do the magoi (mÊ!tai EÈmen¤!i proyÊou!i k[atå tå] aÈtå mãgoi!). She has also explained why initiation and the afterlife are so closely related: it was the initiator’s job to procure for the initiate a special relationship with the underworld powers, which the spirits there might not be eager to permit.48 Hence the Orpheotelestai and the initiates, like the mãgoi, are pertinent to the Derveni author’s argument: in my view he is attacking all of them. 3. New proposals regarding columns 4–5 The sequel to the new col. 3 is the famous col. 4 on Heraclitus. Its connection with col. 3 remains unclear, probably because the latter is so severely damaged in its lower parts. In the light of the new edition I would suggest the following text of col. 4, with three new supplements (in bold face): Col. 4 1 t]oË ei`[. . . . . . . . . . . . . . . . . . . . . (.) per‹ t«n y]e`«n [ ı ke¤m[ena] met`a`y`[e‹! . . . . . . . . . . . . .(.) §]kdoËnai mçll[on μ] !`¤netai [. .].[. . . . . . . . . . . .] t`å t∞! tÊxh! p`ã`[yh mÆte mante¤a! tinå! éyeiã!tou! e‰nai l°gonta! μ teletå! ka‹ Ùrgia!moÁ! émeloum°nou! ÍpÚ ye«n ékoÊvmen mÆt' aÔ pãlin tÚn yeÚn §n toÊtoi! éna!tr°fe!yai ka‹ pare›nai ka‹ !umpragmateÊe!yai dojãzvmen, éll' oÂ! d¤kaiÒn §!ti taËta leitourgo›! ye«n énatiy°nte! À!per Íphr°tai! ka‹ grammateË!i da¤mona! nom¤zvmen §pi!kÒpou! {ye«n} (del. Paton) fler«n ka‹ mu!thr¤vn Ùrgia!tã!. 47 D.H. AR 8. 28. 4, tå! går dØ parå ye«n te ka‹ daimÒnvn §pipempom°na! to›! énÒ!ia ka‹ deinå diaprajam°noi! ÉErinÊa! §«, Íf' œn ôkizÒmenoi (correxi: afikizÒmenoi codd.) cuxã! te ka‹ !≈mata kakoÁ! m¢n diantloË!i b¤ou!, ofiktrå! d' Ípom°nou!i teleutã!. 48 Op. cit. pp. 130–6. 12 5 10 14 oÈk e‡`[a la]mmãnei`n`. îr' oÈ tã`[jin ¶xei diå t«]n`de kÒ!mo!; katå [taÈt]å` ÑHrãkl`e`ito!, ma[rturÒmeno!] tå koinã, kat[a!tr°]fei tå ‡d`[i]a, ˜!per ‡kela` [muyo]lÒgvi l°gvn [¶fh: “¥li`$o!¸ [meyÒ]d`ou katå fÊ!in én`yrv$p¸[h]$˝ou¸ eÔro! podÒ! [§!ti, toÁ`[! oÎrou]! oÈx Íperbãllvn: efi k[atã ti oÎ]rou! •[vutoË §kbÆ!eta]i`, ÉErinÊe$!¸ min §jeurÆ!ou`$!i, d¤kh! §p¤kouroi, ˜pv! mØ . . . . . . Íper]batÚm po∞i k[. . . . . . . . . . . ..” . . . . . . . . . . . . . . . . . . . .]a`i` yÊo`u`[!i. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .]adikh![. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . é]mÆnita k[. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .]!`. i`p`a`i!e`[. . . . . . . . . . incertum quot versus desint ‘. . . regarding the gods he who alters established making public , rather than causes harm. prevented from undergoing the affects49 of fortune. Does not the world have its order by means of these? In the same way Heraclitus, bearing witness to shared ,50 overturns those which are individual. Speaking like a mythologist (?), he said: “the sun, in accord with the nature of method,51 is in breadth the size of a human foot, and does not wax beyond its size; for if it exceeds its own boundaries at all, the Erinyes, allies of justice, will discover it, (singular subject missing) may make excessive .” . . . they sacrifice . . . (in)justice . . . things lacking in divine wrath . . . ’ imagines contuli frr. coniunxerunt edd. 1 t]oË Par.1 ei`[ Ts.3 (i` vel k`, m`): •a`[utoË Lebedev ap. Par.2 (a` vel l`, d`) per‹ t«n *: oÈk ≥yelen ı patØr t«n Livrea y]e«n Ts.3 (e` vel z` praebet imago) 2 ke¤m[ena Ts.1: ke¤m[eno! Par.1, sed lacunam iii litt. statuit Ts.1 p. 96 met`a`y`[e‹! * et »fele› diå tÚ Hammerstaedt: met`a`y`[e‹! efi! kÒ!mon NoË! Livrea: met`a`y`[e- Ts.3: m`et`a`y`[°meno! »f°leian y°lei Ts.1 p. 107: m`et`å` y`[e«n Par.1 t` vel u` a` vel l` y` potius quam o` fu!ikÚ! coniecerim §]kdoËnai Ts.3: ].d`oËnai Par.1 (d` vel a`) 3 mçll[on Par.1 μ Ts.1: ì Ts.3: num ˘? !`¤netai Ts.3 (litt. ! paene certam nec g nec e esse dicit Ts.1 p. 96) ].[ vestigia ii pedum lin. vert. hab. imago ut videtur toÁ! ényr≈pou! vel tÚg kÒ!mon Ts.1 p. 107 t` vel u` pars extrema dextra et fort. pes (sed atramentum non esse crediderunt Par.1) t`å Ts.3: t]å Par.1 p`ã`[yh * (p` vel g` imago; a` vel l`, x`): gå`[r Ts.3: g[år Par.1 4 oÈk e‡`[a Par.1 (i` vel k`, h`): oÈk <ín> e‡`[h Lebedev ap. Ts.1 p. 108 la]mmãnei`n` Ts.3 (i` caput; n` vestigium medium fort. lin. vert.) îr' oÈ Par.1: p]ar' o Lebedev ap. Par.2: g]år o[to! Par.1 tã`[jin ¶xei Ts.1: tã`[!!etai olim * a` vel l`, x` diå Par.1: §k Ts.1: ı Burkert t«]n`de Lebedev ap. Par.2: tÒ]n`de Par.1: tÆ]n`de Par.1 n` linea vert. in fine notam interrogationis Par.1 5 katå Ts.1 (a` vel l` imago) taÈt]å Ts.3 (a` vel d` imago): taËt]a` Par.1 ÑHrãkle`ito! Ts.3 (l` vel a` praebet imago; e` vel z`, j`) ma[rturÒmeno! Ts.3: me`[ta!keuãzvn Par.1 (litt. e` potius quam a` legi dixit Ts.1 p. 97): me`[gãla nom¤zvn olim * tå Ts.3 (t` vel g` imago) 6 kat[a!tr°]f`ei Ts.3: kat`[agrã]f`ei brevius esse monuit Ts.1 p. 97: kat`[agg°ll]ei Par.1 ‡d`[i]a Ts.3, qui diaeresin praesto esse notaverunt (d` vestigium sin. inf.): ‡d`[i]a Par.1 ‡kela` Ts.3 (a` vel l`, x` imago): ‡ke[la Par.1: fikel`[o› Ts.1: fik°[lv! Par.1 muyo]lÒgvi Ts.1: fu!io]lÒgvi Par.1: flero]lÒgvi Sider ap. Laks–Most p. 135: yeo]lÒgvi brevius esse dixerunt Par.1 ¶fh Par.1, qui verbum non fuisse fh!i monstrant, nullo vestigio apicis supra litt. primam eminente: œde Ts.1 7 sqq. Heracliti Ephesii frr. B 3 + B 94 D.–K. (agn. West), sed etiam v. 10 ad verba Heracliti pertinere opinor 7 ¥li`$o! Par.1 i` pars inf. lin. vert. meyÒ]d`ou *: . . .]d`ou Ts.3: •vu]t`oË Ts.1: •au]t`oË vel ynh]t`oË Par.1: !kÊ]f`ou Livrea, haud recte d` potius quam t`, g`, e`, z`, j` fÊ!in Par.1: fÊ!in Ts.3 ényrv$p¸[h]$˝ou Ts.3: én`yrv$pe¤ou Par.1, ut testis Heracliti (sc. Aët. II.21) eÔro! Ts.3 (e` vestigia trium bracchiorum praebet imago) §!ti Par.1 8 toÁ`[! oÎrou]!` Ts.1 (u` vel t`): tÚ m`[°geyo]!` Ts.3, brevius (m` vel k`, i`, e`; !` in frustulo parum a suo loco remoto): toÁ`[! ˜rou]!` Par.1, qui dixerunt oÎrou]!` longius esse Íperbãllvn Ts.3 (e` partem dext. praebet imago) efi k[atã ti oÎ]rou! * (]ro:u! pap., cum atramenti macula supra lin.): efik[Òta! oÎ]rou! Ts.3: efi gã`[r ti eÎ]rou! Par.1 (a` pes sin.): .i`.[. . . . ˜]r`ou! ed. •[vutoË Ts.1: •[autoË vel §[kbÆ!etai Par.1: e[Îrou! Ts.3 9 §kbÆ!eta] i` * (i` vel h`, n`): §]k`[bÆ!eta]i` Par.1, sed vestigium caput litterae k` esse non potest (tantum f` vel c` fuerit) et macula esse videtur: Í]p`[erbale]›` Par.1: •oË: efi d¢ m]Æ` Ts.3 ÉErinÊe$! ed.: -e$! Ts.3 min imago et testes Heracliti (Plut. Mor. 370d, 604a): nin perperam ed., Ts.3 §jeurÆ!ou`$!i Par.1 u` pars extrema sin. D¤kh! §p¤kouroi Par.1 ex Heraclito (etiam 12 contuleris), in lacuna litt. fere viii quam statui: §pikourÆ!ou!i Sider, ZPE 69 (1987) 225–8: taËta doke› e‰nai Burkert: ka‹ fulãjou!i Par.1 49 This is my attempt to render pãyh here, which needs to mean both ‘changes in the heavenly bodies’ and ‘mental sensations, experiences’. English has no word with exactly this semantic range. 50 I.e. the pãyh we all experience while waking, as opposed to individual visions in dreams. The neuter antecedent is understood from line 3: cf. D.L. 9. 7 = fr. A 1 D.–K., e‡rhke d¢ ka‹ per‹ t«n §n kÒ!mvi !uni!tam°nvn pãntvn pay«n, ˜ti te ı ¥liÒ! §!ti tÚ m°geyo! oÂo! fa¤netai. 51 The ‘method’ (a new and no doubt controversial supplement) is Heraclitus’ epistemology, which is what unites the ideas in this fragment with the sentence that introduces it: see especially fr. A 19, quoted in n. 54 below. Reconstructing (again) the Opening of the Derveni Papyrus 13 10 ˜pv! mØ *: ˜pv! d¢ mhd¢n Ts.3: går D¤khi §ãn ti Sider loc. cit.: êtopa, μm mÆ ti! Burkert: oÏtv d' ¶fh ·na Lebedev ap. Par.2 k[a‹ Lebedev ap. Par.2 podÚ! oÎrvi Burkert: épo!b°!ou!i vel kata!b°!ou!i Sider: é!af∞ tÚn lÒgon Lebedev ap. Par.2 11 ¥lion §mm°nonta ékoÊhi Burkert ]a`i` Ts.3: k]a`‹` *: ]a` Par.1: pÒpan]a` Lebedev ap. Par.2 a` vel l`, d` yÊo`u`[!i Lebedev ap. Par.2 o` vel y` u` vestigium sin. sup. 12 d¤kh! Ts.1: D- Ts.1 p. 97 §p¤kouroi * 13 é]mÆnita *: mhn‹ ta[kt«i potius quam du!]mÆnita vel Ùju]mÆnita Ts.1 p. 97 14 !` vel t` .: vestigium incertum i` pars sup. lineae vert. p` vel t`, z`, e`, j` fort. a` vel l`, d` (apex) i! Ts.3 e` potius quam g`, p` §]![t]‹ p`a`›! leg. et suppl. Ts. per litt. ap. Ferrari, qui Heraclit. fr. B 52 agnoverit (quod plane incertum esse censet Livrea) The editors had already restored y]e`«n in line 1; their new supplement §]kdoËnai ‘divulge, make public’ in line 2 seems certain, according to their own reports. Hence I suggest that col. 4 began with the claim that a religious innovator is right to publish his views, since doing so confers benefit rather than harm. The reference to someone or something that prevented the world from being subject to the vicissitudes of destiny, which the author thinks give it its order, will make sense if the author is denying that the views of the physicists could cause harm. His point will have been that the pãyh that we observe in the heavens continue unchanged whatever we consider their causes to be – gods, da¤mone!, d¤kh or tÊxh: in other words, free thinking cannot cause eclipses or other phenomena that were popularly thought to signify divine wrath, as we see from Aristophanes’ Clouds or the conduct of the Athenian forces laying siege to Syracuse in 413 BC. He supports this point by citing Heraclitus, who asserted the claims of koinå pãyh at the expense of ‡dia pãyh such as we experience in dreams.52 The contrast between junÒn, i.e. koinÒn, and ‡dion was central to Heraclitus’ thought: for he held that reason is shared and divine, whereas sensation is individual and unreliable (êpi!ta): this latter term supplies the missing link in sense with col. 5, which is on épi!t¤h.53 The author probably likened Heraclitus to a religious expert, since the latter’s expression brings together physics with the Erinyes; if so, the author regards him as essentially a fu!ikÒ!.54 The editors did not manage fully to restore the fragment of Heraclitus itself. My suggestion meyÒ]d`ou replaces the previous proposal •vu]t`oË. The editors report that, although the letter appeared to be T, it is actually the right half of D; it sounds as if there is no doubt about it. They have allowed only three letterwidths before it, but my model shows that there is certainly room for four. ‘Method’ illustrates perfectly the point that the Derveni author makes in introducing the extract. I think the fragment continued for one additional line, since the content of line 10 seems still to relate to it. The editors place fr. H 8 below fr. G 4 in the middle of the column. This is plausible enough but may never be proved for certain, because the H-series are to be placed so far down in the columns that confirmation from the horizontal fibres of adjacent fragments will always be lacking. If, however, fr. H 8 is rightly placed, I suggest that the text broke off during a discussion of how ordinary people try to placate the Erinyes by sacrificing and avoiding unjust actions that might provoke divine anger. In col. 5 the author 52 Cf. B 89, ı ÑHrãkleito! fh!i to›! §grhgorÒ!in ßna ka‹ koinÚn kÒ!mon e‰nai, t«n d¢ koimvm°nvn ßka!ton efi! ‡dion épo!tr°fe!yai; also fr. A 16, cited in the next n. 53 The connections between these terms and ideas emerge a long and important testimonium to Heraclitus’ epistemology, which I need to quote in extenso, namely S.E. Adv. math. VII 127–34 = fr. A 16: tÚn d¢ lÒgon kritØn t∞! élhye¤a! épofa¤netai . . . tÚn koinÚn ka‹ ye›on. t¤! d' §!t‹n oto!, !untÒmv! Ípodeikt°on: ér°!kei går t«i fu!ik«i tÚ peri°xon ≤mç! logikÒn te ¯n ka‹ fren∞re! . . . (129) toËton oÔn tÚn ye›on lÒgon kay' ÑHrãkleiton di' énapno∞! !pã!ante! noero‹ ginÒmeya, ka‹ §n m¢n Ïpnoi! lhya›oi, katå d¢ ¶ger!in pãlin ¶mfrone!: §n går to›! Ïpnoi! mu!ãntvn t«n afi!yhtik«n pÒrvn xvr¤zetai t∞! prÚ! tÚ peri°xon !umfu˝a! ı §n ≤m›n noË!, mÒnh! t∞! katå énapnoØn pro!fÊ!ev! !vizom°nh! . . ., xvri!ye¤! te épobãllei ∂n prÒteron e‰xe mnhmonikØn dÊnamin: (130) §n d¢ §grhgÒr!ei pãlin diå t«n afi!yhtik«n pÒrvn . . . prokÊca! ka‹ t«i peri°xonti !umbalΔn logikØn §ndÊetai dÊnamin . . . (131) toËton dØ tÚn koinÚn lÒgon ka‹ ye›on ka‹ o katå metoxØn ginÒmeya logiko¤, kritÆrion élhye¤a! fh!‹n ı ÑHrãkleito!: ˜yen tÚ m¢n koin∞i pç!i fainÒmenon, toËt' e‰nai pi!tÒn (t«i koin«i går ka‹ ye¤vi lÒgvi lambãnetai), tÚ d° tini mÒnvi pro!p›pton êpi!ton Ípãrxein diå tØn §nant¤an afit¤an . . . (133) diå toÊtvn går =ht«! para!tÆ!a! ˜ti katå metoxØn toË ye¤ou lÒgou pãnta prãttom°n te ka‹ nooËmen Ùl¤ga pro!dielyΔn §pif°rei “diÚ de› ßpe!yai t«i koin«i: junÚ! går ı koinÒ!. “toË lÒgou d' §Ònto! junoË z≈ou!in ofl pollo‹ …! fid¤an ¶xonte! frÒnh!in” (fr. B 2). ≤ d' ¶!tin oÈk êllo ti éll' §jÆgh!i! toË trÒpou t∞! toË pantÚ! dioikÆ!ev!. diÚ kay' ˜ ti ín aÈtoË t∞! mnÆmh! koinvnÆ!vmen, élhyeÊomen, ì d¢ ín fidiã!vmen, ceudÒmeya. (134) nËn går =htÒtata ka‹ §n toÊtoi! tÚn koinÚn lÒgon kritÆrion épofa¤netai, ka‹ tå m¢n koin∞i fh!i fainÒmena pi!tå …! ín t«i koin«i krinÒmena lÒgvi, tå d¢ kat' fid¤an •kã!tvi ceud∞. 54 The alternative would be ‘physicist’ (fu!io]lÒgvi), but since Heraclitus believes that fire is divine and permeates everything, a notion very like that of the Derveni author, I think it most likely that the latter regarded him as a physicist. In the light of the model my former supplement flero]lÒgvi turns out to be too short. 14 goes on to challenge their excessively literal belief in the terrors of Hades, which is of course the place where the Erinyes were thought to punish the souls of sinners after their deaths. New suggestions are in bold type: Col. 5 1 5 10 14 . . . . ÜAi]do`u dei`n`[å . . . . . . . . . . . . . . . . . . . . . . . xrh[!th]riazom[e. . . . . . . . . . . . . . . . . ]doig`e[. .(.) xrh![t]hriãzon[tai . . . . . . . . . . .].[.] . . . . .[. . . .]i aÈto›! pãrimen [efi! tÚ ma]nte›on §per[v]t`Æ!`[onte!, t«m manteuom°n[vn ßn]eken, efi y°mi[! épi!t]∞`!`a`[i] í`n` ÜAidou deinã. t¤ é[pi!]toË!i; oÈ gin≈!`[konte! §]n`Ê`pnia o`Èd¢ t«n êllvm pr`[a]gmãtvn ßka!t[a], d`iå po¤vn ín par`adeigmãtvm p`[i]!teÊoien; ÍpÚ t`[∞!] è`martŤÄh!` k`a‹ [t]∞! êllh! ≤don[∞]! nenikhm°n[oi, oÈ] ma`ny[ãno]u!in oÈd¢] p`i!teÊou!i. ép`[i]!t¤h d¢ kéma`[y¤h tÚ aÈtÒ: μg går mØ ma]n`yãnv!i mh`[d]¢ gin≈[!]k`v![i, oÈk ¶!tin ˜pv! pi!teÊ!ou]!`i`n ka‹ ır[«nte! §nÊpnia . . . . . . . . . . . . . . . t]Øn épi!t¤[hn . . . . . . . . . . . . . . . . . . . . . . . . . . .]fa¤netai` [. . . . . . . . . . . . . . . . . . . . . incertum quot versus desint ‘. . . the terrors of Hades . . . ask an oracle . . . they ask an oracle . . . for them we will enter the prophetic shrine to enquire, with regard to the things that are prophesied, whether it is right if one were to disbelieve in the terrors of Hades. Why do they disbelieve? Since they do not understand dreams or any of the other things, what sort of proofs would induce them to believe? Since they are overcome by error and pleasure as well, they do not learn or believe. Disbelief and ignorance are the same thing. For if they do not learn or comprehend, it cannot be that they will believe even if they see dreams . . . disbelief . . . appears . . . ’ imagines contuli frr. coniunxerunt edd. 1 tå ÜAi]do`u *: tå §n ÜAi]do`u olim * ex imagine vetere: . . . . . .]h` Ts.1 (pes dext. lineae vert.) o` vestigium valde dubium dei`n`[å * ex imagine (i`n` lineae vert. pes et vestigium incertum): de.[ Ts.1 2 xrh[!th] riazÒm[enoi (cf. 4) vel -Òm[eya * (cf. pãrimen ap. 5): xrh[!th]riazom[ ed. litt. ]do`i`g`e`[ fort. alium ad stratum pertinent d` Ts.3, sed litt. certam praebet imago litt. o`i` et e` certae habendae g` vel r` potius quam k` 3–12 Orph. T 473 Bernabé 3 xrh!`[t] hriãzon[tai * (litt. r alterae pedem sub lin. praebet imago): xrh!`[t]hr`iãzon[tai Ts.2: xrh![t]hriãzon[tai Ts.3 ].[.]. . . . . . 3 3 3 .[ Ts. (papyrus abrasa est) ka]‹ potius quam !Á]n` *: -meno]i Ts. : ¶`p`[ei!]i Burkert: a`È`[to]‹ Ferrari 4 u Ts. post aÈto›! dist. Laks et Most pãrimen` [efi! Ts.1, sed n certa habenda ex imagine vetere: parime. .[ ed. tÚ Ts.1: ti bene Ferrari ]n ed.: ] n Ts.3 §per[v]t`Æ!`[onte! Ts.1 (t` potius quam g`; litt. ! certa habenda): -!`[ou!i Burkert 5 manteuom°n[vn ed.: manteuom°nv`n` Ts.1, haud recte y°mi[! *: yemi[tÚn Kouremenos épi!t]∞`!`a`[i *: épi!te›n tå olim *: tÚn p]e`d`ç`[n §n Ferrari yemi[. . .]. . h`d`a`[.] Ts.3 (h` vel n`; d` potius quam !`; a` vel l`) 6 í`n` * (a` vel l`; r` potius quam n`, g`, ut videtur): §`n` Ts.1: î`r`' Ts.3, cui obloquitur Ferrari post deinã dist. * é[pi!]toË!i, ed.: ép`i!toË!i Ts.1, haud recte gin≈!`[konte! Ts.1 (! vestigium incertum) §]n`Ê`pnia Ts.1 (n`u` vestigia incerta): tå §nÊ]p`nia ed. 7 oÈd¢ Ts.1 (o` vel y` imago) pr`[a]gmãtvn ed. (r` vel k`, p`): pr`agmãtvn Ts.1, haud recte ßka!t[a *: ßka!t[on Ts.3 dist. Ts.1 d`iå po¤vn Ts.1 (d` vestigium inf.) 8 par`adeigmãtvm Ts.3 (r` vel n`, h`; a` vel d` imago) p` vel g` dist. Ts.1: hypostigmen post p`[i]!teÊoien posuerit Burkert t`[∞! * (lineola in alto ut videtur): [te går Ts.3, longius: t`[∞! te Burkert: t[∞! te Ts.2 a` vel l`, d` è`martŤÄh!` * (litt. i supra t scripta est): è`mart<¤>h!` Ts.1 !` vestigia 9 k`a‹ Ts.3, sed litt. k` vestigium minimum est t]∞! Ts.1 ≤don[∞]! Ts.3: -! Ts.2: -!` ed. nenikhm°n[oi Ts.3 dist. Ts.1 oÈ] m`a`n`y`[ãno] u`!in Ts.1 (litt. m et ny et u certae habendae; a` vel l`, d`) 10 oÈd¢ Ts.1 p`i!teÊou!i Ts.1 (litt. p` paene certam habuit ex imagine vetere) dist. Ts.1 ép`[i]!t¤h * ex imagine vetere (p` linea vert. incerta): é[pi]!t¤h Ts.1 kéma`[y¤h Ts.1 (a` vel l`) tÚ aÈtÒ *: taÈtÒn Ts.1 μg går Ts.1 11 mØ ma]n`yãnv!i Ts.1 n` linea vert. mh`[d]¢ Ts.1 h` pars sup. lineae vert. gin≈[!]k`v![i * (k paene certum habendum): -k`v![in Ts.1: -k`v`![i ed. oÈk ¶!tin ˜pv! Ts.1 12 pi!teÊ!ou]!`i`n Ts.1 (!` vel e`; i` pes lineae vert.) §nÊpnia * 14 i` vestigia lineae vert. Despite my initial doubts, the model that I constructed supports the editors’ belief that the traces visible in lines 2–5 of fr. G 10 belong to the same layer as those in lines 6–9 of the same fragment, since there is no location available in the columns on either side where they might belong. Hence I suggest efi y°mi[! Reconstructing (again) the Opening of the Derveni Papyrus 15 épi!t]∞`!`a`[i] | í`n` ÜAidou deinã in lines 5–6. There is room for only a single narrow letter at the end of line 5, because from the very left edge of the following fr. G 3 the surface is preserved with no ink on it: this is the intercolumnium after col. 5. At the start of line 6 the editors had read e`n` but now read a`r`, which they interpret as îra with an elision, but this yields impossible syntax: hence I suggest í`n`. 4. Summary and Conclusion In the light of the analysis above, my current understanding of these columns is as follows. Col. 1 is about how signs from divination are interpreted: the Derveni author quotes a writer on divination who talks ‘like a natural scientist’. By this he intends to show that even religious authorities offer interpretations, just as he himself wishes to do. Col. 2 begins a discussion of the punishment of injustice, which popular belief ascribed to the Erinyes. Col. 3 suggests that those who pour libations to the Erinyes are trying to placate the souls of those who have been killed. The mãgoi consider the da¤mone!, i.e. the Erinyes, to be servants of the gods, whereas in the author’s view they are the souls of people who have been killed; he returns to this in col. 6, when he says that the Erinyes are actually the souls of the dead. Cols. 4–5 supply a bold justification for the author’s right to reinterpret matters of religion. In col. 4 he argues that those who change established views about the gods do not do harm, since private opinions cannot alter celestial phenomena: these are controlled by tÊxh. Likewise Heraclitus, he says, insisted that our shared sensations about phenomena such as the size of the Sun are credible, whereas private sensations like those in dreams are not. But when Heraclitus said that the Erinyes control the size of the Sun, he was speaking like a religious storyteller; perhaps the author implies that Heraclitus did not mean that the Erinyes actually exist as such. Col. 5 argues that private beliefs which we cannot all verify from observation, such as the existence of terrors in Hades, are incredible: the author offers to prove that they are false by asking an oracle whether it would be right to disbelieve in them. People disbelieve because such beliefs are incredible, even though they may dream about them; therefore phenomena like dreams need interpretation. In col. 6 the author returns to his interpretation of the Erinyes and da¤mone! as the souls of the dead, arguing that the rites used by the mãgoi and the initiates support his interpretation. In col. 7 he embarks on an interpretation of a similarly incredible text – the Orphic poem which has the gods violate the most fundamental taboos about sex and violence among kindred. I need not repeat here my argument that the author need not have been a priest or diviner, but only a layperson like Chaerephon, to enter the oracular shrine and ask his question; nor that this question, which may have seemed innocuous to the writer, would have seemed just as provocative as Chaerephon’s was to the Athenian jurors who sat in judgement upon Socrates. There is nothing in the new edition that undermines this interpretation, and much in the preceding columns, insofar as the published photographs permit them to be reconstructed, that seems to me to support it. The author of the Derveni papyrus was a fu!ikÒ! and, from a traditional viewpoint, a blasphemer against the gods. If the Athenians did sentence him to death for impiety, as they sentenced Diagoras of Melos, this would certainly have been in accord with the attitudes that they are documented to have held in the closing decades of the fifth century BC. In conclusion, these remarks can mark only yet another stage in our effort to understand the opening columns of the Derveni papyrus. However, I think I have shown on papyrological grounds that the editors’ reconstruction of these columns is mistaken, and that something more both correct and more coherent can be put in its place.