1 Other authors quote only a couple of short sentences (see Diels-Kranz).
2 See Lore and Science in Ancient Pythagoreanism (Cambridge, MA, 1972), p. 379 n. 46.
3 Bowen, A. C., ‘The Foundations of Early Pythagorean Harmonic Science: Archytas, Fragment I’, Anc. Philos. 2 (1982), 79–104, p. 84f.
4 ‘The Authenticity of Archytas fr. I’, CQ 35 (1985), 344–8, p. 346.
5 von Wilamowitz-Moellendorff, U., Geschichte der griechischen Sprache (Berlin, 1928), 18 n. I = ‘Sein Schriften 3 (Berlin, 1969), 461–95, p. 473 n. 1: ‘Sein Dialekt ist ganz verwüstet: da, steht διαγνὼμεναι, ϕρονειν, ἂμμεσ, ζατειν, um nur die gröbsten Fehler aufzugreifen’. διαγν ὼμεναι and ϕρονειν will be discussed presently; on ἂμμεσ (Vors. 47 B 1; vol. 1, p. 434,3) see n. 15. ζατειν is surely the ζατειν of Vors. 47 B 3 (vol. 1, p. 437,5f.), probably a double slip γ into Attic; Archytas is likely to have written ζατειν (For ν as the infinitive of verbs in -ω cf. ϕιλοσοϕν in the letter of Lysis, p. 158,57 Städele [see n. 8]. Such infinitives are very common * in the MS. tradition of Hellenistic and later Pythagorean texts; however that of Philolaos and Archytas offers only -εῖν).
6 Uguzzoni, A., ‘Note sulla lingua dei Pitagorici: Filolao ed Archita’, Quad. Ist. Glottologia Univ. Bologna, 7 (1962–1963), 53–71, esp. 61–3.
7 Cardini, M. Timpanaro, Pitagorici: Testimonianze e frammenti 2 (Firenze, 1962), p. 360.
8 See Stādele, A., Die Briefe des Pythagoras und der Pythagoreer (Meisenheim am Glan, 1980), p. 208 n. 12.
9 See Kieckers, A. Thumb-E., Handbuch der griechischen Dialekte i (Heidelberg, 1932 2), pp. 74—7 on Die Dorische Betonung.
10 It should be noted that -γνὼμεναι is likely to be an artificial form taken from Homer (Il. 21.266). The situation in Lesbian is not very clear: one would expect *γνων on the analogy of πρóσταν, but γνναι seems to be the only attested form; see Blümel, W., Die aiolischen Dialekte (Göttingen, 1982), pp. 208–16 (the doubts raised by Blümel 209 n. 262 on the reading γ[νω]ναι in IG 12.2, 526 d 21/22 are now solved by the photograph published by Heisserer, A. J., Alexander the Great and the Greeks [Norman, 1980], p. 46; I owe this reference to R. Hodot). No aorist infinitive of γιγὼσκω seems to be attested in the Lesbian poets.
11 Wackernagel, J., ‘Akzentstudien III’, Nachr. Kön. Ges. fViss. Gött., Philol.-hist. Klasse, 1914, 97–130, p. 102 n. 1 = Kleine Schriften 2 (Göttingen, n.d.), 1154–1187, p. 1159 n. 1. See also Ahrens, H. L., De dialecto Dorica (Gottingae, 1843), pp. 316 and 322.
12 Obviously Archytas must have written ἦμεν, the Doris severior form, which is rightly printed by Diels-Kranz.
13 I have also collated two MSS. not considered by Hoche, Vat. gr. 186 and Neap. Ill C 1: both of them have διαγὼμεναι, ἂμμιν and ἔμμεναι.
14 See Städele (op. cit., n. 8), p. 207 and n. 11, with previous bibliography.
15 Obviously not all the learned people who quoted or copied Doric texts will have been as daring as Nicomachus was. Apparently most scribes confined themselves to alterations involving minimal changes of the originals (e.g. ἔμμεσ instead of μσ): see e.g. the apparatus of the Letter of Lysis, p. 158,64 Städele. There is also one example of ἔμμεσ in a passage of Archytas, fr. 1 (Vors. 1, p. 434,3) where Porphyry is the only source: obviously Nicomachus has nothing to do with this alteration, for which an early scribe of Porphyry is likely to be responsible.
16 See Ahrens (op. cit., n. 11), p. 316.
17 See Wackernagel (n. 11).
18 ”εμμεναι is normal in Aretaeus, the Cappadocian physician writing in Ionic in the age of Nero (Kudlien, F., Abh. Akad. Wiss. Lit. Mainz, Geistes- und Sozialwiss. Klasse, 1963, nr. 11). στμεναι is in most MSS. at Hdt. 1.17.
19 De Archytae Tarentini fragmentis mathematicis, in Mélanges Graux (Paris, 1884), pp. 573–84.
20 On the -ν infinitives of verbs in -ω see n. 5. It is unlikely that Archytas used θεαρω, as θεαρω comes from the Ionic an d Attic intellectual vocabulary. In the letter of Lysis (p. 156,28 Städele) we find θεωρματα On the dialect forms θεᾱρóσ etc. (in the religious technical sense of the word), which are only superficially doricized, see Harlow, R. B., Eine Dialektanalyse der koischen Asylieurkunden (Dunedin, Ne w Zealand, 1972), p. 5f.
21 It seems to me that the present infinitives of ϕρονω regarded as Aeolic by us would have caused some difficulties to Nicomachus. ϕρονμεναι would have been acceptable as a Homeric form (Chantraine, P., Grammaire Homérique 1 [Paris, 1948], pp. 305f.; cf. e.g. Il. 15.310 ϕρονμεναι), but -μεναι would have corresponded to no -ναι in Attic. On the other hand, the correct Lesbian form for ϕρονειν, namely, ϕρνην, is no t Homeric, whereas all the other Aeolic forms interpolated by Nicomachus we have come across (-γνὼμεναι, ἂμμιν, ἒμμεναι, δμεναι) are attested in Homer. As I said, ϕιλειν, ϕρονειν etc. are in the MS. tradition of the Aeolic poems of Theocritus; see Gow, A. S. F., Theocritus (Cambridge, 1950), ii.500 and n. 2; they are corrected into -εην by Gow but not by Gallavotti, C., Theocritus quique feruntur Bucolici Graeci (Romae, 1946).
22 Given these premisses, one might wonder whether Nicomachus did not deliberately falsify Archytas' thought by writing περ γρ…ἒχει in dialect (ὂντοσ for ντοσ might be a scribe's mistake, and MS. Vat. gr. 186 has πρ γρστα instead of πρωτιστα). Yet, if so, ἒχει strangely careless, as Archytas has τ μαθηματα at lines 8–9; see also Huffman (n. 4), 346. In any case I do not want to commit myself on this point: videant doctiores. -It is curious that ταν αναστροϕαν ἒχει is metrically even more striking is the fragment of Androcydes in Ionic (probably a late forgery) quoted by Nicomachus immediately before (p. 6,11–16 Hoche), which ends with a series of pure trochaics, πρσ λγων μαθσιασ συνεργíην ἒχουσιν Did Nicomachus in writing his own sentence employ the rhythm with which the preceding quotation ended?
23 This article is a revised version of part of a paper on Doric literary prose read to seminars in Oxford and Cambridge in May 1987.1 am very grateful to Dr Colin Austin, Professor Hugh Lloyd-Jones, and Professor M. D. Reeve for inviting me, and to all those present for their comments. I am also indebted to Mr Nigel G. Wilson for reading a previous draft of this paper, and especially to Professor Anna Morpurgo Davies for corrections and suggestions concerning dialect problems.