Critical Notes on Plato's "Politeia", IAuthor(s): S. R. SlingsReviewed work(s):Source: Mnemosyne, Fourth Series, Vol. 41, Fasc.

3/4 (1988), pp. 276-298Published by: BRILLStable URL: http://www.jstor.org/stable/4431734 .Accessed: 03/12/2011 23:37Your use of the JSTOR
archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jspJSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact support@jstor.org. BRILL is collaborating with JSTOR to digitize, preserve and extend access to Mnemosyne.http://www.jstor.org

Mnemosyne, Vol. XLI, Fase. 3-4 (1988) CRITICAL NOTES ON PLATO'S POLITEIA, I BY S. R. SLINGS With the publication of G. J. Boter's excellent study of the MSS of the Politeia, work on the text ofthat dialogue can now for the first time be based on a trustworthy examinatio and eliminatio. The results are not startlingly new: it had of course been known for a long time that A and F are primary witnesses, and such a status had been suspected for D from Schanz onwards (Boter, 19f.), though it was denied by Chambry. Therefore, one should not expect a new edi- tion, based on ADF (plus papyri and indirect tradition), to diverge widely from those of Burnet (based on ADF and M) and Chambry (based on AF), In fact, given Burnet's impressive qualities as a tex- tual critic, which are only partly counterbalanced by the inade- quacy of his collations and his cavalier treatment of part of the indirect tradition (cf. my remarks Mnem. IV 40 (1987), 28), a new edition of the Politeia can only be justified if it is based on an exhaustive examination and evaluation of papyri, indirect tradition and medieval MSS. Boter's work has now paved the way for such an undertaking. The following notes are intended as a companion to such an edi- tion. Of course, they do not constitute a commentary: I have gone into matters of interpretation only where this was absolutely necessary to reach a decision as to what Plato wrote. Besides, I have restricted myself to those places where I think I have something new to say?as a further restriction, I have refrained from saying things that may be new but are not, in my opinion, of importance for anyone but those interested in the history of the text or in the niceties of Platonic orthography and morphology or other minutiae. As a consequence of these criteria, the reader will look in vain for treatment of some familiar cruces, and he must not expect treatment of every passus where I think Burnet and/or Chambry are wrong.

CRITICAL NOTES ON PLATO'S POLITEIA, I 277 There are two points where I part company with Boter. I am not convinced that DF go back to a common source which is a gemellus of A (Boter, 99-104), and I believe that authentic ancient readings may be found (through contamination) in secondary MSS to a larger extent than he is inclined to assume (242-294). In Book One, the second point is of academic interest only. In all places where a secondary MS is obviously right against ADF, their readings are conjectures of a rather trivial character: 337a7 ?p???????? ?c: ?p???????? AF: ?p?????a?? D b4 ? p???????t? ?c: ?p???????t? ADF 342a4 ??p?????s?? ?c: ??p??????s?? ADF b2 a?t?? ?: a?t?? AF: a?t?? D b5 a?t? ???: a? t? ADF 352e9 a?.,.fa??e? ?c: a?.,.fa?e? ADF Stob.: a? seel. Adam (cf. below ad 353d7) 354b3 ??? ??? ^Sc: ??'???a? A: ??????1 F*: ????a? F2m: ??? ???a? DF4sl In the following places, one or more secondary MSS offer a reading, not found in ADF, that may have some claim to authen- ticity, but which it would not be prudent to put in the text: 332e5 ?a? ?? t?(?) ????a?e?? ADF: ?a? ????a?e?? ??? 334cl e??a? ADF: om. Se 335dl 1 ????? ADF: i.m. T: om. ?? 343c3 ?a? ADa: tea? F*: te ?a? ? (te s.l.) ScD2F* 352b8 ??de? ADF Stob.: ??de ?? (adopted by Burnet) As for the first point, since on Boter's own showing A, D and F are more or less heavily contaminated (99-104), their stemmatic relationship cannot be used as an argument in choosing between their variants. But Boter is inclined (104) to prefer the reading of AD against F in divergencies of word order and in variations involving particles and other small words: F is indeed very careless in both aspects. But there certainly are cases where F inserts a small but necessary word not found in AD: 345e2 a????? AD: ?? a????? F Eus. 353d9 ????? AD: ?? ????? F Stob. And if F is here obviously right against AD, it may also be right in the following places, where I am inclined to prefer its reading to that of AD (it should be noted that later hands in F have no primary value):

278 CRITICAL NOTES ON PLATO S POLITEIA, I 329e3 se ?a(?)d??? AD: ?ad??? se F Stob, (se ?ad??? se F2) e5 pa? a????a fas?? AD: fas? pa?a????a F (-??a or -???) Stob. : fas? ta pa?a????a Proci, (in Tim. I 42 Diehl) 339d5 d? ADF?: de F* (adopted by Burnet and Chambry) 345d5 d? AD: d? F 348a 1 ????sa? ?: ????sa? * *? D (two letters erased, reading a.c. uncertain but not ????sa? ???): ????sa? ??? F (not reported by Burnet or Chambry) 350a6 de AD: d? F Stob, (adopted by Burnet) 353b4 ?st?? AD: est? t? F Stob, (adopted by Burnet) Although in these cases the AD reading cannot be excluded, I feel that in all of them F's reading is better (in some cases decidedly so). It is clear that Boter's canon cannot be rigidly applied; there- fore it is sensible to apply it only in those cases where there is little or no difference in value between the readings of AD and F. Of course this does not mean that I have refuted Boter's hypo- thesis of the relationship between DF and A; I intend to deal with it elsewhere. Whenever in the following notes my report of MSS readings dif- fers from those of Burnet and Chambry, they have been checked against the microfilms of the Free University Plato MSS Collection, both by Dr Boter (for ADF) and myself (for ADF and secondary MSS). In addition, most of the ADF readings have been checked against the originals by Dr Boter, to whom the distinction of various hands is due entirely. I wish to thank Dr Boter for his comments on an earlier draft of this paper. SIGLA A Parisinus gr. 1807, s. ix (for later hands, cf. Boter, 106-112. I do not distinguish here between APC and A2). D Marcianus gr. 185 (coll. 576), s. xii (for later hands, cf. Boter, 119-121). F Vindobonensis suppl. gr. 39, s. xiii-xiv (for later hands, cf. Boter, 128-131). ? Marcianus app. cl. IV, 1 (coll. 542), ca. 950. M Caesenas gr. D 28,4 (Malatestianus), s. xv. ? Laurentianus conv. soppr. 42, s. xii-xiii. ?? go back to a common exemplar which is a gemellus of ?. ??? derive from a MS copied from A after A3 and before A4 had introduced variants and corrections. Sc Scorialensis y.1.13, s. xiii-xiv. Derives from ? for Book One.

CRITICAL NOTES ON PLATO'S POLITEIA, I 279 ? Marcianus gr. 187 (coll. 742), ca. 1450. An indirect copy of ? for Book One, written and heavily corrected by Bessarion. a Laurentianus gr. 80,7, s. xv. A heavily contaminated MS, deriving, via various other preserved MSS, from ?. Par Parisinus gr. 1810, s. xiv. Derives from D as corrected by D2 and D3. ? Laurentianus gr. 80,19, s. xiv-xv. An indirect copy of Par, corrected (some- times from its exemplar) by a very intelligent scribe (?c, cf. Boter, 222 f. and 245-260). q Monacensis gr. 237, s. xv. A copy of ?, often quoted for readings that should be attributed to ?c. ? Laurentianus gr. 85,7, s. xv. A copy of F. Some later hands add a number of good corrections. 327c 10 ?????? ?? d' ??? ?t? ???e?peta? t? ?? pe?s??e? ??a? ...; ???e?peta? ADF: ??. e? ?e?peta? ?3 Against Adam's excellent defense of e? ?e?peta? (?e?pes?a? said of a third alternative, cf. Grg. 510c7: a better man can be no friend of the tyrant, nor can a worse one. ?e?peta? d? e?e???? ????? ????? ????? f????, to wit, the ???????) may be brought forward Phlb. 18d4-6 t? d' a?t? ??? t?? ?????...???e?peta? ? ??? ? F????e t? t? p??? ?p?? a? ta?t' est??; where the quotation preceded by t? is exactly parallel. The slightly abnormal use of et? (''isn't there still the possibility left out by you") may have caused the correction to e? ?e?peta? (so Jowett-Campbell), but it is equally possible that a non-assimilative spelling (???e?peta?) was misunderstood. BIBLIOGRAPHY J. Adam, The Republic of Plato, Cambridge 1902. D. J. Allan, Plato's Republic, Book I, London 19552. G. J. Boter, The textual tradition of Plato's Republic, diss. Free University Amsterdam 1986 (a new edition with a different pagination is to appear: Leiden 1989). J. Burnet, Platonis Opera, IV, Oxonii 1902. E. Chambry, Piaton, Oeuvres compl?tes, VI, Paris 1932. CG. Cobet, Variae Lectiones, Lugduni Batavorum 18732. F. M. Cornford, The Republic of Plato, London 1941. J. L. V. Hartman, Notae Criticae ad Platonis de Rep?blica libros, Hagae Comitum 1896. ?. Jowett-L. Campbell, The Republic of Plato, Oxford 1894. H. Richards, Platonica, London 1911. C. E. C. Schneider, Platonis Opera Graece, I-III, Lipsiae 1830-1833. P. Shorey, Plato, The Republic, London-Cambridge, Mass., 1930-1935. G. Stallbaum, Platonis di?logos selectos..., Ill, Gothae-Erfordiae 1829. T. G. Tucker, The Proem to the ideal commonwealth of Plato, London 1900. U. v. Wilamowitz-Moellendorff, Platon, II, Berlin 1919.

280 CRITICAL NOTES ON PLATO'S POLITEIA, I It must further be noted that the authority of A3 in the Politela is uncertain; apart from some scholia and repetitions of phrases in the margin (with s????te? possibly a variant for d?as????? te? at 329a3), the only other variant reading which we owe to this hand in the Politeia is 329e3 s??] se (falso). In the Nomoi, A3 is very busy in recording variants and supplying lacunae; there he had an ancient MS at his disposal. His behaviour in the Politeia suggests the contrary. Therefore, despite the ??., e? ?e?peta? may very well be a conjecture made by A3. Cf. Boter, 110 and my data on A3 in the Clitophon, Mnem. IV 40 (1987), 36 f. 329e7 ?a? ?????s? ??? t?, ?? ???t?? ?e ds?? ???? ta? ???t?? ?e ds?? ADF: ???t?? ds?? ?e Proci. In Ale. 12,10 (p. 5) Westerink: ???t?? ds?? Se Stob. IV 50a,31 (= V 1035,5 Wachsmuth-Hense) ???t?? ?e is juxtaposed in Plato here: Cra. 424c8; as a variant for ???t?? Phdr. 230d5 (cf. De Vries ad loe); Phd. 61c9; cf. also Sis. 388a5; Ale.II 139e9. Prt. 312d8 the reading ?? ???t?? ?e ??a??? ?e (?; ?? ???t?? ??a??? ?e TW) is a conflation of the TW reading with ?? ???t?? ?e ??a???; similarly Euthphr. 12c4 ?? ???t?? ??a ?e BDTW Stob. ML ?? ???t?? ?e ??a C ?? ???t?? ?e ??a ?e Ven. 184 Stob. S. The juxtaposition is rare throughout in Classical Greek (Denniston, 405), but there is no good reason for doubting it (statistics in J. Blomqvist, Greek particles in Hellenistic prose (Lund 1969), 30). Here, ???t?? ?e may have been used to avoid ds?? ?e, which com- monly means "only in as much as" (cf. however ad 331b5 for another explanation); at Cra. I.e., the particles were probably juxta- posed for euphonic reasons. There is no good reason for assuming ?e to be interpolated: in Hellenistic prose ???t??...?e almost disap- pears and ???t?? ?e becomes the norm, but ?? ???t?? ?e is rare. The Atticists revive ?? ???t??...?e (Blomqvist, 29; 32). I would rather assume that the rareness of ???t?? ?e in Plato brought about the variants ???t?? and ???t??...?e (the latter had been conjectured by H. Hoefer, De particulis Platonicis capita selecta (Bonnae 1882), 39; Blaydes ad Ar. Thesm. 709; J. L. V. Hartman, Mnem. II 42 (1914), 221. The conjecture is ascribed to Cobet in Mnem. 1884 by J. L. V. Hartman, Notae Criticae, 11 f. I have not found it either there or in any of Cobet's works, though there is a general rejection of ???t?? ?e in Mnem. II 11 (1883), 138 in a note on Hdt. II 93,2

CRITICAL NOTES ON PLATO'S POLITEIA, I 281 and 98,2. But Hartmann wording seems to refer to a more specific treatment). Other authors borrow this dictum, but usually rephrased in such a manner that it loses its value for the text of Plato (e.g. Choric. 157,24 Foerster-Richtsteig ???e? ?e? t?...?? ??? ds??pe? ??eta?). Arethas is fond of it; six times out of seven he gives a ?a? before or (once) after 8s??, which must be his idea of elegant Greek, and should not be used for Plato. The fact that he has ???t?? ?e ?a? ds?? three times (Ser. min. I 117, 10 f.; 335,2 f.; II 57,28 f. Westerink) as opposed to one ???t?? ?a? ?s?? (? 292,17) indicates that his text was identical to that of ADF. 330c5-6 ta?t?? te d? ?a? ?? ????at?s?? e??? t? ????ata sp??d????s?? ?? ????? ea?t??, xat ?at? t?? ??e?a? ??pe? ?? ?????. ta ????ata F Stob. (IV 31d,118 = V 775,4 W.-H.): pe?? ta ????ata AD / ?a? ?at? AD Stob.: ?a? ?? ?at? F Both sp??d??? t? and sp??d??? pe?? t? are found in Plato; contrast Phd. 114e3f. t??.,.pe?? t? ?a????e?? (se. ?d??a?) ?sp??dase and 64d2f. ?sp??da???a? pe?? ta? ?d??a? ?a??????a?. But the construction with the simple accusative (cf. Sph. 251c3; 259c3) is definitely lectio dif- ficilior; it is far easier to explain how pe?? was inserted than how it came to be dropped. If the preposition is omitted, as I think it should be, there is a slight ambiguity: ta ????ata may also be taken as internal object of ????at?s??e??? (cf. X. Cyr. Ill 3,5). This is not a valid counter- argument: in fact, it is probable that ta ????ata, given its position between ????at?s??e??? and sp??d????s??, should be taken ap? ?????? with both. The negative in F was conjectured independently by H. Wolf (cf. Philol. 10 (1855), 347) and endorsed by J. C. Vollgraff (Mnem. II 44 (1916), 4; his statement that it is in Stob, is wrong); it may be attractive at first sight, but it leaves te unexplained; besides, the entire sentence is meant to explain (??? c3) the statement ?? d? ?t?s?? e??? d?p??? ? ?? ????? asp????ta? a?t?. I see absolutely no reason for Vollgraff s omission of the article in his proposal <?? > ?at? ??e?a?; it may be sheer carelessness. 330e5 ?p???a? d' ??? ?a? de??at?? ?est?? ????eta? ?a? a?a?????eta? ?d? ?a? s??pe? e? t??? t? ?d????e?.

282 CRITICAL NOTES ON PLATO'S POLITEIA, I ?d????e? AF Stob. (IV 31d,118 = V 775,20 YV.-H.): ?d???se? A2 (s s.l.) D Justin. (Coh. Gr. 26 = 111 90 Otto) The perfect has a very close parallel in Phd. 113d7f. ?a?a????e??? t?? te ad????? t?? d?d??te? d??a? (cf. supra d8-el) ap?????ta? e?' t?? t? ?d????e?; cf. Men. 92b5. I have found no similar use of the aorist in Plato. "Le parfait...insiste souvent sur la responsabilit? de l'agent" (P. Chantraine, Histoire du parfait grec (Paris 1927), 177, quoting Lys. 30,24 t?? ???tt? t?? p???? a?a?? pep????e? ? p?e?? ?d????e?;). Choric. 100,9 f. Foerster-Richtsteig ??de? e?s??e? de???? ??de ????s??? ?'st?efe? a?t?? t?? ????? (cf. 330e 1-2), ?? t??a ?e??p??e? shows that he is alluding to a text with the perfect, not the aorist. 331b5 ??e? d? ?a? a???? ??e?a? p?????, ???a ?e e? ???' e??? ??? e????st?? e???e ?e??? a? e?? t??t?...p???t?? ???s???tat?? e??a? ?ee? ADF: e? ?e Stob. (IV31d,118 = V 776,17 W.-?.): ?e seel. Stallbaum The combination ???a ?e is found five times in (some of) the MSS of Plato. At 543c4, Phd. 86e4 and Hp.Ma. 287b4 ???' a?e is demonstrably better; the remaining two instances are this place and Phdr. 262a2 (???a ?e d? BD Hermias ???a d? TW ??? a ??? Galenus). Juxtaposition of ???a ?e is rare, not only in classical Greek (where its occurrence has been doubted altogether, cf. Den- niston, 23) but also in later times (contrarily to what is suggested by Neil, Comm. on Ar. Eq., Appendix I, 193): no examples are recorded in Preisigke-Kiessling, only two in the New Testament, probably none in Polybius (XII 4,10 ? e is Casaubonus' correction), none in Epictetus and Plotinus; from the Atticists no instances are recorded in W. Schmid's Atticismus ? very many of ???a...?e ? whereas Arr. Cyn. 15,1 is textually uncertain; Neil mentions Pausanias without giving any more information. However, in classical Greek there are some instances of ???a ?e that merit closer investigation. Arist. EE. 1216b20 ?? ??? ???a ?e is a case in itself; Oec. 1343b25 ?e is not in the MSS; Archestr. 177,2 Suppl. Hell. Wilamowitz' t? for ?e is certainly right. But there is no good reason to doubt Gorg. Pal. 10 and 14 (both times answer- ing a hypothetical objection beginning with ???a).

CRITICAL NOTES ON PLATO'S POLITEIA, I 283 Now, when we are confronted with a rare phenomenon estab- lished certainly for only one classical author, should we accept it at only one place in another author for whom it is not unanimously attested (Phdr. 262a2 a??? ?e d? is absolutely unique and almost certainly wrong, cf. Denniston, 242)? In principle, no. But there are two considerations which may lead one to accept ???a ?e here. First, since ???a ?e is also uncommon in later Greek, there is no convincing case for its being interpolated (if we follow Stallbaum, who incidentally in the second edition of his commentary recanted his seclusion of ?e and adopted Stobaeus' ???a e? ?e) or transposed (Stob.' reading is lectio facilior). Secondly, it may be more than a coincidence that the author for whom ???a ?e is attested is a Sicilian, whereas the speaker here comes from Syracuse (maybe this also accounts for 329e7 ???t?? ?e; Adam defends both ???t?? ?e and ???a ?e on the ground that Cephalus was not an Athenian?he did not know about the passage from Gorgias but his intuition may be right). Cf. Olymp. In Grg. 37,3-5 Westerink fa??? ??? dt? epe?d? G????a? ? ?????, ?? ?p' e?e???? p??f??e? t?? ???e?? (se. ?e????????a, ????s??) e???????? ??sa?? ?e??t???? ?a? ??; Dodds ad Grg. 450b9. Though neither of these arguments is incontrovertible, together they give sufficient reason for retaining the MSS reading. For what it is worth, I note that when two forms of e?? are juxta- posed, Plato nowhere has a particle in between: Phlb. 63cIf. e? ???' e???; Lg. 705b4 f. ???' e??? e?; Lg. 647b4 f. e? p??? e?, cf. 738e5; Sph. 229b9 e? ?f' ???, cf. Lg. 758b7; La. 182b2 ?p? ?e e??? e??. But cf. Epin. 976el f. ??a ?a? ?? e?pe?? p??? ??a?. 333e6-7 ??' ??? ?a? ??s?? ?st?? de???? f????as?a?, ?a? ?a? e?? ??t?? de???tat?? ??p???sa?; f????as?a? ?a? ?a?e??, ??t?? ADF / ?a?e??] ?? pa?e?? ?c / ??p???sa? ADaF: ?a? ??p??? sa? D2 See the notes of Stallbaum, Schneider, Tucker and Adam (and his App. II, I 62 f.). ?? pa?e?? (cf. Boter, 256) and ? a? ??p???sa? are clearly conjectures. Of the other remedies proposed, only Schneider's ??p???sa? and ?a???, considered (and rejected) by Tucker (cf. O. Apelt, BPhW 20 (1903), 340 f.; A. Prandtl, Analecta critica ad Platonis de rep?blica libros (M?nchen 1904), 5; ?. Vretska,

284 CRITICAL NOTES ON PLATO'S POLITEIA, I WS 66 (1953), 76 f.) are at all probable. I prefer ??p???sa? for the following reasons: (a) Corruption of ??p???sa? to ??p???sa? is easy, given the word order, which easily leads to f????as? a? ?a? ?a?e?? being taken together (of course, the MSS punctuation is no valid argument); corruption of ?a??? to ?a?e?? is less obvious. (b) ?a??? as a circumstantial participle, instead of ?a????? with a participle, is definitely the rarer construction in Attic prose. In Plato's authentic works, it is only found Prt. 321el, that is to say in Protagoras' myth with its numerous poeticisms and Ionicisms. It has been argued (especially by Tucker and Vretska) that with ?a???, the emphasis falls on ?? p???sa?, as it should. I am not quite certain that it should (one might equally argue that the emphasis should fall on ?a?e?? as this facilitates the transition pat??a? - ?a?e?? ??p???sa? - ????a?) and given the word order, the notion of secretiveness is stressed anyway, whether ?a?e?? or ?a??? is read; I cannot agree with Vretska that ?a??? is "der gleichsam unbemerkt eingef?hrte neue Begriff (77). 335a8-9 ?e?e?e?? d? ???? p??s?e??a? t?? d??a??? ? ?? t? p??t?? ????????, ?????te? d??a??? e??a? t?? ??? f???? e? p??e??, t?? d' ?????? ?a???, ??? p??? t??t?? ?de ???e?? [?t? est?? d??a???] t?? ??? f???? a?a??? d? ta e? p??e??, t?? d' ?????? ?a??? ??t? ???pte??; The words I have bracketed are found in AD but omitted in Fa (added in the margin by F2). As far as I know, nobody has ever taken offense at them, although Schneider and Chambry duly record their omission in Fa. It is possible, but no more than that, that they were omitted by sheer accident in F's exemplar (from A. C. Clark, The Descent of Manuscripts (Oxford 1918), 415, one gets the impression that lacunas of 12 to 17 letters are relatively frequent in F, but since his data are based on Burnet and Schneider, they are practically worthless outside the Politeia?those reported from Hp.Ma. may be due to the well-known injury of F's direct exemplar, which is irrelevant for the Politeia). It is far more likely, however, that they were interpolated in AD in order to ease the construction. Though there are in F a fair amount of arbitrary changes made with this aim (cf. 332e3 ?? t??? p???e?] ?? t? t? p??tte?? F; 344c2

CRITICAL NOTES ON PLATO'S POLITEIA, I 285 ?d?????ta] pep?????ta F; 347a3-4 infra; Boter, 134; 137 f.), interpolation (except of small words) is more typical of A and D than of F. Therefore, I have little doubt that dt? est?? d??a??? should be removed; to my mind, the sentence gains with their deletion. Com- pare, in a similar context, 334el0. 337a6-7 ?a? ta?t' ??? ??d? te ?a? a?t??? p????e???, dt? s? ?p?????as?a? ??? ??? ??e??s???, e????e?s??? d? ?a? p??ta ?????? p??? s??? ? anoxptvo?o e? t?? t? s* ???ta? a6 ??e??s??? AD: ??e??se?? FSc / a7 p???s??? seel. Cobet (Variae Lectiones^, 526 G)/?p???????? ?c: ?p???????? AF: ?p?????a?? D/???t?? AD: ????t? Fa: e???t? F2: ???t?? Goodwin (??? 690). In a6, ??e?? se?? may very well be right; for the switch from indicative to optative in indirect discourse, cf. K.-G. II 366; Good- win (MT ? 670). However, F's numerous confusions of e?, ? and ot do not encourage adopting its reading here. Cobet's seclusion of p???s??? (cf. van Herwerden, Mnem. II 11 (1883), 333; J. L. V. Hartman, Not?t criticae, 25; J. J. Hartman, De emblematis in Platonis textu obuiis (Leiden 1898), 113), for which he quotes Thphr. Char. 25,5 ?a? pa? ?????? ? ???es?a?, is more convin- cing than most of his deletions; interpolations facilitating the con- struction of a sentence (cf. above) are indeed found occasionally in all MSS (e.g. 553b4). The expression p??ta p??e??, adduced by Adam to defend the text, does not seem relevant to me. But in the expressions ??d?? ???? ?, t? ???? ?, which Cobet cites as parallels, we do find elliptical and nonelliptical constructions side by side (cf. K.-G. II 285); thus, there is no general rule by which to justify the deletion here. In choosing between ???t?? and ????t? (this had been conjectured by H. Richards, CR 7 (1893), 14 = Platonica, 87; neither he nor Burnet could have known that this is the reading of F), we must ask ourselves whether the sentence would be good Greek in direct speech: p??ta ?????? p???se? ? ?p?????e?ta? e? t?? t? a?t?? ???t??. ? have seen no parallels for such a construction: if the e?-clause con- tains a present indicative and the main clause a future indicative, the e?-clause never refers to the future (e.g. Hdt. VII 15,3 e? ?? ?e?? ?st?? ? ?p?p??p??.,.?p?pt?seta? ?a? s?? t??t? t??t? d?e????). I cannot make head or tail of Allan's suggestion that the indicative is used

286 CRITICAL NOTES ON PLATO'S POLITEIA, I because "e? is not genuinely conditional, but temporal, and tem- poral clauses are liable to such changes, which make for liveliness." So much for the exigencies of grammar. On the other side of the balance we may note that (a) "the indicative may perhaps be allowed in loose conversational style" (Adam); (b) the change from e???t? to ???t?? in some ancient MSS is harder to explain than vice versa; (c) F has a tendency to regularize the moods, cf. 466a 1-6 where in indirect speech F has optatives throughout, A only from a4 e??e? onwards, while D sits on the fence, as usual (but cf. above on ??e??se?? for an opposite case). The aspectual difference between ???t?? and e???t? does not help to reach a decision: in general clauses of this type both present and aorist forms are found (cf. Grg. 447e7f. ? p?????es?a? ?t? a? t?? se ???t??; Men. 70b7 ?p?????es?a? ??? t?? t? e??ta?). Hence Goodwin's conjecture is superfluous. On balance, I prefer ???t??, mainly because I think it more likely for Plato to make a minor grammatical error (in what is, after all, a complicated oratio obliqua structure) than for a later copyist to change a regular form deliberately into an irregular one. There is a third possibility, namely that Plato wrote ???...???t??, that ??? was subsequently corrupted to e? and ???t?? corrected in one of the ancestors of F. But corruption of ??? to e? is very doubtful hypothesis. 340e3 ?p???p??s?? ?a? ep?st???? ? ??a?t???? ??a?t??e?, ?? ?? ??? est? d?????????. ?p???p??s?? AaD Stob. (? 8,38 = II 163,14 W.-H.): ?p?- ? e?p??s?? ????5 (or ?2?) Theoretically, one expects the present and the imperfect tenses to be used for the process of failing, and the aorist when the end of the process has been reached (ta ????ata ?e ?p??e?pe? "I was get- ting short of money"; t? ????ata ?e ?p???pe? "I have (had) run out of money"). At the two passages in Plato where there are no variants in the primary witnesses, this distinction is not altogether confirmed: R. 574dl dta? d? t? pat??? ?a? ??t??? ?p??e?p?? t?? t???? t??, p??? d? ?d? s??e??e?????? ?? a?t?? ?? t? t?? ?d???? s?????, ?? p??t?? ??? ????a? t???? ?f??eta? t????? ?t?(?p??e?p? (?) AF ?p??e?pe? D ?p???p?(?) ??) the most logical interpretation is that the young

CRITICAL NOTES ON PLATO'S POLITEIA, I 287 tyrannical man has already squandered his parents' resources, though "begin to fail" (Cornford) is not entirely impossible. Pit. 274d4 epe?d? t? ??? ?? ?e?? ... t?? ep??e?e?a? ?p???pe? a????p??? fits the theory (Phlb. 52dl0; Prt. 310el; 334e6 the verb is used in a dif- ferent construction). Since there are three passages where the primary MSS are divided (h.l.; 568el; 573e3), a look at usage outside Plato may be useful. An examination of all occurrences in Hdt., Th., Lys., D. (14 oc- currences in all: discounting D. 20,23; prooem. 55,1 where the verb has a slightly different meaning) shows the following results: when only the interpretation as a more or less gradual process is possible, the present is used (only D. 14,30 bis); conversely, when it is clear that completion is meant, the aorist is normal, but the present can be used for speciali reasons (Lys. 30,19, where there is an emphasis on the process leading up to exhaustion, cf. ??a??s? es?a?). In ?dl three passages in Plato where there are variants, the con- text indicates that the exhaustion has been completed; in none is there any special reason for concentrating on the process, but only at our place is such a concentration definitely excluded (and I do not find it plausible for 568el). There is no sign of a conscious preference for either the present or the aorist in our MSS. At our place as at 573e3, AaD have the aorist, A2F the present; at 568el F has the aorist, AD the present. If we were to follow A (or Aa) everywhere, we would have to accept the aorist here and at 573e3, and the present at 568el. How- ever, if we accept the aorist at 573e3, I feel we should read it a for- tiori at 568el. Therefore, I submit that the aorist should be read at all three places. Allan notes "?p??e?p??s?? ?a? ep?st???? and ?? ? ??? ?st? d????????? convey the same notion. For this reason it is desirable to read ?p??e?p??s??". I am not sure that I understand this, but if the implication is that ?p??e?p??s?? ep?st???? is a state, not a process (or rather, event), the argument is contradicted by the use of ?p??e?p? everywhere else. 341e7 t??t?? ??? dp?? ??p?????? ta ???f????ta, ?p? t??t? pa?es?e??s?? ? t???? ?p? t??t? Fa: ?p? t??t?(?) ADF3

288 CRITICAL NOTES ON PLATO S POLITEIA, I The accusative is regular in final expressions after pa?as?e????, cf. 422d7 ?p? t??t? pa?es?e?as?????; Grg. 509el; 510a3. No similar use of ?p? t??t?? is found. For the same reason t??t? (AD) is to be preferred to t??t?(?) (F Stob.) at 353a5. 344el ? s?????? ??e? ep??e??e?? p????a d?????es?a?, ???9 ?? ???? d? a?????, ?? ?? d?a???e??? ??ast?? ???? ??s?te?est?t?? ???? ????; ???' ?? ADF3: d??? Fa Adam (CR 16 (1902), 217) claims that ???' ?? is "simpler, livelier, and in every way better than d???". There is a close parallel in form and content: La. 185a3-5 ? pe?? s?????? ??es?e ???? ???d??e?e?? ... ???' ?? pe?? t??t?? t?? ?t??at?? d t?? ??et???? ????st?? d? t?????e?; Against this, W. L. Lorimer (CR 45 (1931), 213) wishes to read d??? with Burnet, but puts commas after p????a and d? a?????, comparing Cri. 45c5 f. ??de d??a??? ??? d??e?? ep??e??e?? p????a, sa?t?? p??d???a?, e??? s????a?. As an ancillary argument, Lorimer adduces a number of transla- tions in which words like * entire' creep up even if ???' ?? not d??? is translated (e.g. Shorey: "and not the entire conduct of life"). These are, I think, due to the fact that it is hard to bring out in a translation that ???? d?a????? (whether or not with d??? added) is a foil for the really important words ?? a? d? a???e??? ?t?. In fact, it might be argued that d??? would perhaps be more appropriate than d???. Another, less trifling argument against d??? could be the absence of the article; though the construction is occasionally found in Plato (e.g. R. 577e2 pe?? d??? ... ?????; cf. Gildersleeve, II 303), it seems rather poetic or at any rate high-flown language to use here. But in this case palaeographical considerations appear to swing the balance. It is not easy to see how d??? could have been corrupted to ???' ??, whereas it is rather easy to explain the other way round: if ????? was misunderstood as ?????, which makes no sense, d??? would be an easy correction. 345c3 t?? ?? a????? p?????a ????t? ?itou de?? ?'ste??? a?????? f????a?, ???a p?a??e?? ??e? a?t?? t? p???ata ?a?' ds?? p????? ?st??, ?? p??? t? t?? p????t?? ???t?st?? ???p??ta ???', ?spe? da? t????a t??? ?a? ??????ta ?st??ses?a?, p??? t?? e????a?, ? a? p??? t? ?p?d?s?a?, ?spe? ????at?st?? ???' ?? p?????a.

CRITICAL NOTES ON PLATO'S POLITEIA, I 289 p?a??e?? A Eus. (PE XII 44,2 = II 134,1 Mras): p???a??e?? D: ??. p???a??e?? A1 i.m.: pa???e? Fa: pa???e?? et s.l. p???a??e?? F3 Most commentators retain p???a??e??, on the assumption that a shepherd qua shepherd cannot be considered by Thrasymachus to fatten his sheep ("how Thrasymachus errs is explained in ?? p??? t? ?t?" Adam). But Socrates' point is a double one: not only has Thrasymachus lost sight of t?? ?? a????? p?????a (as in fact he has, cf. 343b 1-4), but also his description of the ruler qua ruler (cf. 343b5) seeking only his own interest, leads to the assumption that the shepherd qua shepherd fattens his sheep (instead of preserving their good shape by a normal quantity of food), because it is in his own interest to do so. Note pa???e? at 343b2. p???a??e?? may be a corruption of p?a??e??, but I rather think it is a conscious correction made by someone who had problems with p?a??e?? similar to those of modern commentators. It is just possible that the first explanation in Hsch. p 2237 p?a??e?' s?te?e? ?? pa??e? derives from a commentary on this passage, in which the difficulty was circumvented (rather like Tucker's explanation "the art of the p?????, which might be expressed by p???a??e??, is at least equally expressed by p?a??e?? ta p???ata"). F's pa???e? is either a simple misreading of a gloss p????e! or the remnant of a longer note (cf. Boter, 134); in either case it is p?a??e?? not p???a??e?? which is explained. Cobet (Mnem. II 9 (1881), 356) saw that in Galen Plac. IX 5,8 (= II 564,35-566,1 de Lacy) d?af????ta? d? ?a??s?? ? ??? ?? ????at?st??, ? d' ?? ?atada?s?ta? (cf. c5-dl), ?atas?e?????s?? a?t? p???a, the last three words prove that Galen read p?a??e??. 346b4 ??d?? t? ??????, ??? t?? ???e???? ????? ?????ta? d?a t? ???f??e?? a?t?? p?e?? ?? t?? ?a??tt??, ??e?a t??t?? ?a?e?? ?????? a?t?? ?at?????; ???f??e?? Fa?: ???f???? ADF3 With the exception of Allan, commentators have gone remarkably astray in explaining this passage. The only way of keeping ???f???? is given by Burnet: "The construction surely is d?a t? p?e?? ?? t?? ?a??tt?? * thanks to his being at sea', ???f???? a?t?? 'when it is good for him to be so' (accusative absolute)" (CR 18 (1904), 204; cf. 19 (1905), 296 and n. 5). But first, the word

290 CRITICAL NOTES ON PLATO'S POLITEIA, I order is clumsy, because any Greek reader would connect d?? t? with ???fe??? rather than with p?e??. Secondly, ???fe??? a?t?? as an accusative absolute wrongly suggests that becoming healthy is the reason for being at sea (Burnet's 'when' only obscures the point further: are there moments for a captain when it is not useful for him to be at sea?). Consequently, even though ???fe??? may be considered lectio dif- ficilior, it is illogical and should not be accepted. As in the entire passage t? ???fe??? is a highly frequent phrase, it is not hard to account for the corruption of the original ???f??e?? (so rightly Tucker). 347a3-4 ?? d? ??e?a ?? e???e ??s??? de?? ?p???e?? t??? ??????s?? ??e??se?? ???e?? ?? FaM^ Eus. (PE XII 9,3 = II 98,16 Mras): ?? A: ?? DF* / de?? ADF3: de? Fa Eus. ?? is obviously right, ?? a correction of ?? (perhaps a misreading of ?). de?? is at any rate lectio difficilior: the infinitive, however, is not to be explained as indicating that Socrates' previous words (cf. 345e5 f.) are being repeated?which they are not?, but rather as a type of anacoluthon familiar after ?spe? ?'?e??? and the like, where the sentence proceeds in the accusative and infinitive.. Cf. Phlb. 20d7 f. t?de ?e ??? ?? ???a? pe?? a?t?? ??a??a??tat?? e??a? ???e??; L. Reinhard, Die Anakoluthe bei Platon (Berlin 1920), 11 f.; 24; K.-G. II 581. Reinhard points out that F normally replaces the infinitive by the indicative in such cases, e.g. R. 485a4 f. d t?????...????????, t?? f?s?? a?t?? p??t?? de?? ?ata?a?e?? (de?? AD, de? F, wrongly adopted by Burnet). ?, to which she ascribes the same tendency, reads de?? at our place. 347d6 ?ste p?? a? ? ?????s??? t? ?fe?e?s? a? ?????? e???t? ?p' ????? ? ????? ?fe??? p????ata e?e??. t? secl. ?. Richards (CR 7 (1893), 15; Platonica, 87) In Plato, a?????a? is nowhere else followed by t? + infinitive (there are some twenty instances of the simple infinitive), and the construction may well be unique in classical Greek. However, this does not seem to me a good reason for deleting it. In the first place, one gets the distinct impression that in the fifth

CRITICAL NOTES ON PLATO'S POLITEIA, I 291 and fourth centuries, t? + infinitive and the single infinitive are more or less interchangeable; cf. K.-G. II 44 f. Secondly, Plato may very well have written the article here to prevent construing ? ?????s??? (used absolutely) with ?fe?e?s?a? ("to learn", cf. 619a5). 348b6 ("If we conduct the discussion this way, ... , if we do it that way...") ? p??? ??? ???, ef?. ? p?te??? ??? s?? ?? d' ??? a??s?e?; ? ??t??, ?'f? p?t???? Fa: ?p?t???? ADF3 F's reading is unobjectionable; ?p?t???? looks, however, more idiomatic: "Whichever way you want". JowettCampbell (II 214) compare Euthd. 271a6 (??? ?? ... ?? ??e? ... d?e?????; ...) ? ?p?te??? ?a? ???t???, ? ???t??; where few scholars would not accept Her- mann's correction p?te???; Pit 261e4 ("Are we going to call this ??e?a??t??f?a or ?????t?? f????") ? ?p?te??? a? ?? t?? ????? ????a????, the main clause (?????s??e?) is easily supplied from the preceding ???????? e?. Neither of them is a good parallel. Even in comedy, the phrase ?t? ????e? as an answer (absent from Plato) always depends on a predicate (e.g. Ar. Lys. 98 ?pe??s??a? t? ??????; ? dt? ????e? ?e s?). But the really suspicious part of the AD reading is ???: in Plato, after indefinite relatives, ??? is restricted to 'elliptical* use (Den- niston, 422). In other words, whereas we might find G????a? pe?? ?t????? (?t?? ???) ???e? we shall never find G????a? pe?? dt?? ??? ????e? ???e?. We must therefore conclude that ?p?te??? is a corruption, prob- ably due to the later Greek habit of using forms of ?st?? etc. in direct questions (cf. Euthd. I.e. ; Lys. 212c4 p?te??? Hirschig ?p?te??? BT). 349b7 ? d??a??? t?? d??a??? d??e?t? s?? a? ????e?? p???? e?e??; ??da??? ... ? t? d?; t?? d??a?a? p???e??; ? ??de (ta?)t?? [d??a?a?], ?'f?. ta?t?? scripsi: t?? d??a? a? ADF Stob. (III 9,58 = 111 382,5 W.-H.): t?? d??a?a? p???e?? Stallbaum: t?? p???e?? t?? d??a?a? Adam: t?? W?amowitz W?amowitz is certainly right in rejecting ??d? t?? d??a?a? (? 377: "Er konnte sagen 'auch als keine gerechte Handlung', allenfalls 'auch als keine Handlung', aber nicht 'auch (nicht) als eine

292 CRITICAL NOTES ON PLATO'S POLITEIA, I gerechte', denn dann lag der Unterschied in dem Adjektiv"). Stallbaum's conjecture (inspired by Ficinus' translation Neque iusta etiam actione) does not explain the error; Adam's does, but is very unelegant. Wilamowitz himself must have been somnolent: I do not see how ??d? t?? ef? can be at all possible (the article can be used as a pronoun with ??te /??te but not with ??d?). Read ??de ta?t??, ef?. The first syllable of some forms of ??t?? is occasionally dropped, thus yielding an article. The ancient scribes, who had better Greek than Wilamowitz, felt that a noun or adjective was called for, and came up with the senseless d??a?a?. Had they interpolated p???e?? instead, nobody would have taken offence. A similar process happened Cra. 385cl0 ?a? t??t? ??a t? t?? a?????? ????? ???eta?; "That part of a true statement is also pro- nounced" (meaning d???a as the smallest part of a ?????). The true text is preserved by C (Tubingensis). In an ancestor of BW, t??t? was corrupted to t?, which of course needed a noun, and d???a was easily found: BW read t? d???a and ? the conflation t??t? d???a. Other cases of corruption from demonstrative to article are: Grg. 492cl t??t? Hoenebeek Hissink t?? BTWF; Epin. 981a8 t??t? ?e 02KC recte t??t? te AO t?te ?? (O4 i.m., the famous "book of the patriarch"); R. 340b7 t??t? AD recte t?? Fa (corr. F3); Ti. 60d5 ta?ta FCY-family recte t? A. Another solution is given by G. E. Vasmanolis, Platon 16 (1964), 184: ??de t?sde, ?? ??a. I cannot take ?? ??a seriously, but t?sde deserves consideration (-de may have fallen out, and d??a?a? sup- plied). It is, however, far less idiomatic than ta?t??. 349c7-9 t? d? d? ? ?d????; ??a ????? t?? d??a??? p?e??e?te?? ?a? t?? d??a?a? p???e??; ? p?? ??? ??? ef? ... ? ?????? ?a? ? d???? ?????p?? te ?a? p???e?? ? ?d???? p?e??e?t?se? ?a? ??????set<?? ?? ap??t?? p?e?st?? a?t?? ?????; ad???? ADF: ? d???? ?e Stob. (Ill 9,58 = III 382,13 W.-H.)/ ?a? ??????seta? ? ????? om.Stob./ ?? ap??t?? AD: d? a? p??t?? F/ p?e?st?? a? t?? AD: a?t?? p?e?st?? F Though ?????? ... ?e is not abnormal in Plato (cf. 353e7; 456c 11; Smp. 200d8; Prm. 141a8 f.), the fondness for ?e of later ancient MSS and secondary witnesses should count as an argument against adopting it here.

CRITICAL NOTES ON PLATO'S POLITEIA, I 293 The clause omitted by Stobaeus contains a syntactic unicum that makes it rather suspect. Plato avoids the use of ?? with the subjunc- tive as a final conjunction (whether in adverbial or object clauses). Indeed the only certain instance of this construction is Io 537b4 (?? a?) in a quotation from Homer. Phlb. 55c7 and Lg. 779b7 the MSS have e??. The papers of final ?? with the optative are slightly better (Ti. 92a4, without a?; Phd. 82e6 and Smp. 187d5 with a?). Plato is not the only Attic author to shun this usage, cf. H. A. Short ap. Gildersleeve, AJPh 4 (1883), 419 n. 2 (no certain example in the orators apart from Antiphon). The reading of F is hardly an improvement. It could be inter- preted as "he will compete (in order to find out) who ... ", but a?t?? is slightly awkward, and to my mind, a Greek would have preferred ???eta? to a? ... ????? in such a case. Combining AD's ?? with F's a? does not contribute to better Attic syntax either. On the other hand, though the clause could be missed, it is far from being an obvious, or even plausible, interpolation?indeed it is hard to see why anybody should have wanted to interpolate it at all. Integrating a previously conceded point into a second question which thereby covers two points is standard Platonic procedure (ap??t?? or p??t?? refers both to the ?d???? ?????p?? and p????? of this question and to the d??a??? and d??a?a p????? of the last one). Besides, in this and the following excerpts Stobaeus leaves out clauses that are not vital to the argument (e.g. Ill p. 385,3 om. 351b2 ?a? ? b3 d????sa?????; b4 ?a? ? b7 s??p?). Therefore, it would be unwise to use Stobaeus' omission of the entire clause to get rid of a syntactic anomaly. How, then, do we proceed with the anomaly? One way out would be writing <e>?? a? p??t??. But this presupposes a double corruption: ?'?? to ?? and, in AD, a? to a-: surely, it is better to explain F's d? a? as originating from AD ?? a-: the initial error was ap??t?? to a?pa?t??, after which the unfamiliar ?? a? was corrected to ?? a?. Perhaps ?? can be retained if one realizes that its choice (over ?p??, cf. X. HG VII 2,14) may have been motivated by the superlative p?e?st??: in the instances of final ?? ( + a?) c. opt. in Plato, either a comparative or a superlative is present (cf. Gildersleeve, I.e., on Ti. 92a4). This consideration may also justify

294 CRITICAL NOTES ON PLATO'S POLITEIA, I Apelt's change of e?? to ??, Phlb. I.e.; Lg. I.e. e?? is probably cor- rect (though the word order is not: perhaps e?? ??? a? ???). 351b2 p???? fa??? ?? ?d???? e??a? ?a? a???? p??e?? ep??e??e?? d?????s?a? ad???? |?a? ?ataded????s?a?|, p????? d? ?a? ?f' ?a?t?? e?e?? d????sa?????. Cobet brackets ?a? ?ataded????s? a? with the following comment: "Grammaticus aliquis in margine explicuit quid esset e?e?? d????sa?????. Recte ille quidem, sed non erat id Platoni obtruden- dum" (Variae Lectiones^, 257). I cannot accept the MSS text, but I am not sure that his remedy is correct. 1. The scholiasts normally explain the construction e?? c. part.aor. with aorist, not perfect, forms, e.g. Schol. E. Med. 33 (?t???sa? ??e?) ?tt????, a?t? t?? ?t??ase. Cf. further Schol. S. Ant. 1068; OC 1141; E. Hipp. 932; Or. 451; Tro. 1123; Ar. Av. 852 (Rutherford). Only twice have I found a perfect: Schol. E. Hipp. 932, which goes on to compare Hes. Op. 42 (????a?te? ??? e???s? ?t?) ...a?t? t?? ?e???fas? (cf. W. J. Aerts, Periphrastica (Amsterdam 1965), 129 f.); S. Ant. 31 f. in a paraphrasis of ?????a?t' e?e?? the Schol. gives ?e???????a?. The replacement of a verbum simplex by a compound is of course standard scholiast practice, but at the same time it is also Platonic, cf. ??. 240a2 ded???????a? ? a4 ? ataded????????. 2. If Cobet wants to imply that e?e?? d????sa????? is a periphrastic construction, he is wrong, as is shown cogently by Aerts, 157, who compares Grg. 456a8 (? ??t?????) ?p?sa? t?? d????e?? s???a???sa ?f' a?t?? ??e?; Hdt. ? 28 t??? ?????? p??ta? ?p' ???t?? e??e ?atast?e???e???: the expression ?f' ?a?t?? e?e?? (p??e?s?a?) prevents us from taking e? e?? d????sa????? as a periphrasis of ded????s?a? (though cf. Epist. 334c6 f. ?? d?????s?a? S??e??a? ?p' ?????p??? desp? ta??). In reality, d????sa????? is to be regarded as a participle of circumstance (note the word order at all three places) and p????? is object both of ?'?e?? and of d????sa?????. Indeed, there are only three instances of periphrastic e?? with aorist participle in Plato (Aerts, 156-158). An examination of these instances shows that Plato uses the construction only when he has no other option: at Lg. 793b8 pe???a???a?ta ??e? and Cra. 404c 1 ??as?e?? e?e?? there was no corresponding perfect form (?e????fa,

CRITICAL NOTES ON PLATO'S POLITEIA, I 295 ??as?a?) in Plato's time, as Aerts points out. Phdr. 257cl f. p??a? ? a???sa? ??? Plato problably could not have used the perfect te?a??a?a because, like most fourth-century writers, he uses the lat- ter only as a so-called intensive perfect; in other words, te?a??a?a always denotes a state in the present and does not combine easily, if at all, with p??a?. I may refer to my analysis of the perfect of state verbs in Demosthenes in: Ophelos: Zes studies voor D. M. Schenkeveld (Amsterdam 1988), 72-76, esp. 75 f. Returning now to our textual problem, we must conclude that Cobet's explanation of ?ataded????s?a? as an explanatory gloss (adapted later to the context by means of ?a?, as often) is not with- out its problems. Yet what sense can be elicited from ?a? ?ataded????s?a?? After ?p??e??e?? d?????s? a? ad???? the words must mean 'and has (already) enslaved them'. This is odd in itself, and besides, it makes no sense when the text continues 'and it also holds many (cities) in subjection, having enslaved them'. The words p????? d? ?a? indicate that something new is added to the preceding and that the addition is valid for many cities?if ?a? ?ataded??- ??s?a? is retained, the addition is not new and the fact has already been stated for al cities. Most editors have countered Cobet's deletion with the hypothesis that ?a? ?ataded????s?a? refers to the city's succeeding in enslaving others. But as the lasting result of this success is already covered by ?f' ?a?t?? ??e?? the point would be valid only if Plato had written ?a? ?atad???? sas?a?. In fact, in a recent discussion of this problem with some of my Amsterdam colleagues, two of them (Professor C. J. Ruijgh and Dr A. Rijksbaron) independently suggested correc- ting the text to ?a? ?ata[de]d????<sa)s?a? (sas?a? becoming s?a? by haplography and the reduplication inserted afterwards to correct the nonsensical ?atad????s?a?). This yields a smooth text: the unjust city endeavours to enslave others unjustly and succeeds in doing so. Still, I have several objections. 1. After fa??? a? an aorist infinitive would normally refer to a past action, whereas here the context requires a simultaneous one: with ?atad????sas?a? the clause would probably mean 'the city is trying to enslave others and did enslave them', which is out of the question.

296 CRITICAL NOTES ON PLATO S POLITEIA, I 2. The notion of success is superfluous at this stage of the argu- ment (though it becomes important from 351c8 onwards). What Socrates suggests in this sentence amounts to the following: (a) Let us assume the existence of an unjust city (p????...e??a?). (b) Such a city would try to enslave other cities unjustly (? a? ???a?... ad????). (c) In addition, such a city would keep many other cities in servitude, which it had enslaved in the past (p????? d? ?a? (!) ... d????sa?????). The addition of 'and would be successful at it' (?a? ?atad????- sas?a?) in (b) brings in a notion which has nothing to do with the city's being unjust. In fact, Socrates' point further on is that success depends on the city being just, not unjust (cf. b7-9; c8 ff.). Thus, there are grave objections to the transmitted text but also to the two corrections discussed here. If I were forced to make a decision, I would follow Cobet, but it seems more prudent to use the obelos. It may be noted in passing that the parallels quoted by Aerts disprove F's ?f' ?a?t??. 351cl-2 e? ??? ef? ?? s? ??t? ?? e?e? ?e?e? ? d??a??s??? s?f?a?, ?et? d??a??s????* e? d' ?? ??? e?e???, ?et' ad???a?. ??e? ADF Stob. (Ill 9,60 = 111 385,7 W.-H.): ?st?? ? / s?f?a ADF: s?f?a? xc Stob. I cannot think of a satisfactory reading in this place. The transmitted reading cannot be right. Commentators compare various explanatory asyndeta (cf. 337e2), but here there is no asyndeton since ?et? d??a??s???? is the apodosis. Burnet makes ? d??a??s??? s?f?a a parenthesis, a practice of which he is far too fond in general, and which is clearly a remedy of despair here: if anything, it strengthens the suspicion. Of the variants, N's ?st?? (presumably a correction made c?rrente calamo by Bessarion) does not explain ??e?. s?f?a? is either a further corruption or a desperate attempt to make at least some sense. The remedies that have been proposed can be found in Adam and Tucker. Perhaps the most obvious one is J. L. V. Hartman's deletion of ? d??a??s??? s?f?a (Notae criticae, 41), but this cannot be right: Socrates had said that justice is a?et? as well as s?f?a (cf.

CRITICAL NOTES ON PLATO'S POLITEIA, I 297 350d4); but a?et? only confounds the issue here. Thrasymachus says in fact: it is wise and therefore, in your view, just for cities to be imperialistic?in my view it is wise and therefore unjust, since injustice is e??????a (348d2). Therefore, s?f?a is in fact very much needed here (the same argument goes for Tucker's deletion of s?f?a alone). Baiter's ??e? <e?>, endorsed by Stallbaum and J. B. Mayor (CR 10 (1896), 111) is puerile; Adam's statement that in ? ? is written over an erasure is more than dubious, and anyway the erasure is certainly not "large enough to have contained e? ?". The only conjecture which is at all acceptable is Tucker's dele- tion of ??e?. However, it is not easy to see how it came to be inter- polated, as the clause is clear enough without any verb. Besides, the entire sentence would become too nominal to my (subjective) taste. In the next line, Stobaeus' omission of ??e??? looks attractive until one realizes that, if it is adopted, the logical supplement is ??t? ??e???, whereas in fact the two verbs refer to two different statements (cf. above). 353d7 t? ?p??e?e?s?a? ?a? a??e?? ?a? ????e?es?a? ?a? ta t??a?ta p??ta es?' dt?? a???? ? ????? d??a??? a? a?t? ? p?d???e? ?a? fa??e? ?d?a e?e???? e??a?; ? ??de?? a????. e?e???? ADF Stob. (Ill 9,60 = III 389,12 W.-H.): e?e???? Par2q2a Commentators (especially Schneider and Adam) appear unduly embarrassed by e?e????; even Burnet calls it 'illogical' (CR 19 (1905), 296 n. 5). Madvig, Adversaria critica ad scriptores Graecos et Latinos, I (Copenhagen 1871), 416 went so far as to delete it altogether, which leaves ?d?a strangely unspecified. His paraphrasis shows what has gone wrong: "cui [rei.] ea tribuamus et (cuius) propria dicamus". But in the Greek, the clause is only formally a relative clause: if it were a read one, we should have had a?t?? (or t??t??) for e?e????. But in fact es?' dt?? a???? ? ????? equals ??' ?? ????? t?? ?????, just as es?* dst?? (in ques- tions) equals ??' ??de?? (cf. Thrasymachus' answer). See K.-G. II 403. Hence, changes like ?a? fa??e?] ? fa?e? (Richards, Platonica, 88; fa?e? is in fact read by AaF and two of the three MSS of Stob.) and ?a? <??> fa??e? (G. E. Vasmanolis, Platon 16 (1964), 298 CRITICAL NOTES ON PLATO'S POLITEIA, I 188; R. Renehan, Studies in Greek texts (G?ttingen 1976), 126; cf. d9 where AD omit ?? as against F Stob.) are superfluous.* Amsterdam, Free University *) The bulk of these notes was written during my stay as a Junior Fellow at the Center for Hellenic Studies, Washington, D.C. I wish to thank the Trustees for Harvard University and the Netherlands Organization for the Advancement of Pure Research (Z.W.O.) for making this sojourn possible.