The Classical Quarterlyhttp://journals.cambridge.org/CAQ

Additional services for The Classical Quarterly:

Email alerts: Click hereSubscriptions: Click hereCommercial reprints: Click hereTerms of use : Click here

ΓΕΝΟΣ and ΕΙΔΟΣ in Aristotle's Biology

D. M. Balme

The Classical Quarterly / Volume 12 / Issue 01 / May 1962, pp 81 - 98DOI: 10.1017/S0009838800011642, Published online: 11 February 2009

Link to this article: http://journals.cambridge.org/abstract_S0009838800011642

How to cite this article:D. M. Balme (1962). ΓΕΝΟΣ and ΕΙΔΟΣ in Aristotle's Biology. The ClassicalQuarterly, 12, pp 81-98 doi:10.1017/S0009838800011642

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/CAQ, IP address: 66.77.17.54 on 30 Jan 2014

TENOE AND EIAOZ IN ARISTOTLE'SBIOLOGY

IT is not certain when or by whom yevos and ei8os were first technically dis-tinguished as genus and species. The distinction does not appear in Plato's extantwritings, whereas Aristotle seems to take it for granted in the Topics, which isusually regarded as among his earliest treatises. In his dialogues Plato seemsable to use yevos, et8os, 184a, fxipos interchangeably to denote any group ordivision in a diairesis, including the group that is to be divided.1 As Rittershowed, it is possible to see an increasing preference for yevos in the latedialogues, where diairesis is in full use, as compared with the Republic andPhaedrus. yevos, as a 'kind' composed of related members, would be a naturalchoice for a class-word; it is indeed ready to mean the genus which is divisibleinto related species. But the restriction of ei8os to mean species would be, inPlato, an unexpected development. Species is necessarily subordinate, implyinga genus above it; moreover it is the group-concept nearer to particulars andtherefore—for Plato—farther from certainty and reality. It seems improbablethat, if he had wished to confine one word to the lower status, he would havechosen ei8os with all its tremendous associations. If he did so distinguish yevosand eiSos, it can only have been in the unwritten practice of the Academy. Butthose who have taken this view have had to rely mainly on the AiaipeaeisApiaroreXovs 64—65,2 which is weak evidence. These two sections are givenonly in the Codex Marcianus version, not even in the version that Diog. Laert.says Aristotle attributed to Plato (D.L. 3. 80); § 64 contains the reference todyyeXoL which Rose pointed to as one of the marks of a Christian editor ;3 thestatements about yevos and ei8os in §§ 64-65 are commonplaces of Aristotleand his Greek commentators, and could have been extracted from thesesources. While it is likely that parts of the Aiaipeaeis ApicnoTeXovs descend fromAcademic origins,4 only the thinnest chain of probability could connect anyparticular statement or terminology with Plato himself. That Plato's use ofanimal examples in the Politicus came from an extended interest in biologicalclassification, has been inferred from the Epicrates fragment;5 but this men-tions only the separating of plants into yevr\, and not only is there no use of theword eiSos but there is no suggestion of hierarchical or systematic classification,which could have made rich material for a comedian.

Such classification may develop in two stages, which give rise respectively tothe relative and the absolute distinction between genus and species. The rela-tive use is seen at Arist. Pkys. 227bI I-14: el 8' eanv ard* a /cat yevrj apa /ecu ei8r]«mV, . . . el rj emaTr\\i.T) et8os fJiev imoX^ijiews, yevos 8e ruyv mioTmiwv . . . . H e r ethe classification provides a pattern as an aid to analysis, without implying thatthese concepts are permanently classified thus in relation to other concepts.

1 Ritter, Neue Untersuchungen, pp. 230 f. s Fr. 287 (K.) ; cf.j Alexis fr. 1. Usener,2 Hambruch, Logische Regeln d. plat. Schule, Preuss. Jahrb. 1884; Wilamowitz, Philol.

p. 7, also cites Arist. Met. iO39a24, but A. is Unters. iv. 283; Stenzel, Studien1, 1931, p. 82;using his own terminology there. Herter, Platans Akad.1, 1952, p. s i . Cf. Dies,

3 Arist. Pseudepigr. p . 678. Plat. Polit. (Bud6), p. xxvii, and the more4 Cf. Mutschmann, Divisions Arist., pp. cautious view of Jaeger, Diokles v. Karystos,

vii f. p . 178.

4599.1 G

82 D. M. BALME

T h e absolute use, on the other hand, does imply a permanent and more com-prehensive scheme; this enables analogies to be drawn between one set ofgroups and another, and therefore has more significance for science. Thus wesay that Dog is a species but Cat is a genus, referring to a scheme which givesthose groups a precise status in the animal kingdom. This use can be seen atArist. de Sensu 448*13-17 and G.A. 784b2i. If Plato had wanted to make thedistinction, either absolutely or relatively, he could have done so by restrictingei8os to two of the senses that he already used, namely (i) a grouping ofparticulars, and simultaneously (ii) a division of a larger group: both senses canbe seen, not simultaneously but separately, at Soph. 219 a and d—rexywv traadveiSrj Svo (two forms among the plurality of rixvai), and KTTJTIKTJS 8VO et8rj (twoforms into which KTrjriKrj is divisible). But he would hardly have wanted tomake it unless he were developing classification in some degree. I t is muchdebated whether he intended diairesis for this purpose or not. The fact that hedid not find it necessary to create a verbal distinction between genus and speciesis perhaps a straw on the side of those who hold that he did not use diairesisfor systematic classification. Moreover, if it is right to think that he was lat-terly more concerned with the inter-relations between Forms than with therelations between Forms and phenomena, then a hierarchy of genera andspecies might not be the best conceptual framework for him but might evencreate difficulties.1

These considerations probably do not apply to Speusippus, who not onlyabandoned the theory of Forms but held that definitions can only be made aspart of a comprehensive classification.2 Yet the positive evidence that hedistinguished yivos from etSos is hardly better than in Plato's case. Stenzel,already believing in Platonic classification, and following Lang, relies on frs.8 and i6(L. ) : 3

Fr. 8 (L . ) : Athen. 3. 105 b Hirevoimros 8e iv fi' 'Ofiolcov napavArfcnd (fyqaw eivaiTWV fiaXaKoarpaKcov Kapafiov, aoraKov, vvpi<f>rjV, apKrov, KapKivov, irdyovpov.

F r . l 6 ( L . ) : A t h e n . 7. 318 e ei8r/ 8' ecrrl VOXVTT68(OV iXeSaivr), TToXmroSivrj,

fioAfiiTivT), oafivXos, <Ls M.pi<rroTeAT]s iaropet xal ZTrevonriros.

With these compare fr. io (L . ) : Photius, Lexic. s.v. TTTJVIOV: t,q>ov o/ioiov KCOVOJTTL-. . . o n Se Koyviuiros el86s ecrri, Eir&ionnTos h> TU> /$' TCOV 'OfioiOT^Tiov <f>rj<rlv

OVTCOS' TTTjvloV, ip-TTlS, KCOVWlfl.

In addition, the list of titles attributed to him by Diog. Laert. includes Txe.plyevdv KOX ei'Scor. This is all that can be called direct evidence; to it must beadded the fair probability that Aristotle's criticisms of al yeypapixivai. 8iaipecreisand 01 8ixoTOfw€vres in P.A. 1. 2-3, where yivos and el8os seem to mean genusand species, were aimed at Speusippus. Yet Aristotle's evidence cannot bepressed over terminology, as is known from the way he reports other philo-sophers ; and in any case it would be possible to take yivos and e'Sos in thesepassages as meaning nothing more technical than 'kind' and 'form'. The same

1 Cf. Arist. Met. 1039*24, 1085*23; Cher- better images than the family tree suggestedniss, A.'s Crit. of Plato, p. 46, Riddle of the by A.'s yevrj in' aAArjAa.Early Academy, pp. 40, 54; but see Ross, 2 Anon, in Arist. An. Post. 584. 17; Cher-Plato's Theory of Ideas, p . 241, on depart- niss, A.'s Crit., p. 59.mental hierarchies. Cornford's divided circle 3 Stenzel, R.-E. 'Speusippos', 1640, 1653;(Plato's Th. of Knowledge, p. 271) and Skemp's Lang, De Speus. Acad. scriptis, p. 9.hachured map (Plato's Statesman, p . 74) are

rENOZ AND EIAOS IN ARISTOTLE'S BIOLOGY 83

applies to D.L.'s attribution: if there was such a title, it still need mean nomore than 'On kinds and forms' as Plato used the terms. It is doubtful, how-ever, whether Diogenes' evidence should be accepted, since there are neitherfragments nor testimonia for this book, and he attributes the same title toXenocrates with equally little to show for it.1 Photius' excerpt (fr. 10) does notattribute the word ethos to Speusippus, and his actual quotation fails to supporthis interpretation, since in^viov is not shown as a species of Kcbvcoifi but both areshown together with iynris in a list: comparison with the other fragments showsthat this means that they form a group of opoia. No generic or class name issuggested, and in fact the two classes named in frs. 8 and 16 (fiaXaKoarpaKa andiroXvnoSes) are the only ones that Lang could find. The word etSos occurs onlyin fr. 16, and here Athenaeus claims to be quoting both Speusippus andAristotle: but Aristotle does not speak of eZS-q noXvwohtuv but of ye\n\.z In theother fragments Speusippus regularly says that animals are ofioia, Trapa-rrX'qai.a,epjf>epfj. The only other class-word to be found is pepos, fr. 7 (also fromAthenaeus). The Epicrates fragment uses only yevos. Therefore the positiveevidence that Speusippus distinguished yevos and etSos in the technical sensecomes really to nothing. It may well be that he was attempting a comprehen-sive 'natural' classification by grouping together similar animals and plants,and by arranging the groups dichotomously in sets, and that he called thegroups indifferently by the same class-words that Plato used, including Siaipemsfor a complete set; but there is no evidence that he went farther and dis-tinguished genus from species, or that he even reached the point where thisdistinction becomes useful.

There seems therefore to be no satisfactory evidence that anybody other thanAristotle originated this verbal distinction. It appears unmistakably in manypassages of the Organon (especially Top. 4 and 6) and Metaphysics, and oc-casionally in other works.3 In the same treatises the words sometimes seem alsoto be used without this technical sense, and to be interchangeable.4 That thesame words should have sometimes a technical and sometimes a non-technicaluse, is not necessarily significant. But some passages, such as Met. iO58b2<5-iO59ai4, exhibit both uses in a single discussion, in such a way as to cause con-fusion if the whole is read together. Here we may perhaps see one part writtenbefore Aristotle began to use the distinction, and another part added after-wards.5

In the Topics ethos as species tends to be applied to the subject of discussion.6

As such, it is a group-concept whose members are formally indistinguishable forthe purposes of the discussion. In order to define it, its yevos must be ascertained.So yevos, although it too has unity of concept, is essentially divisible into dif-ferent eloW) ;7 ethos is treated as indivisible, and is essentially a member of a

' D.L. 4. 2. 13. Xenocrates would prob- 3 Cf. Bonitz, Index 15^12-40.ably have meant the non-technical sense, as * Cf. Met. iO58b26, 28 (Ross), iO7ia25,used in his fr. 9 (H.): KIVT/JOCIOS ci&rj, <f>opas 27; Cat. 8b27, 9 a i4 ; An. Post. 97b24, 34;ttSij, T?js «'? fv6v (j>opds etSr). Top. IOIai8, IO5aI 1, 20, bI4, io6ai I, 24, 27

1 H.A. 523b2g, 525ai3, 622ai5; P.A. io7bi9, ioga2, al.685bi3. At 9. 393 f (= Arist. fr. 271 (R.), 5 Ross ad iO58b28. Cf. Long. Vit. 465a2-6,I527ai2) Athenaeus says even more pre- H.A. 497b9-i2 (both discussed below),cisely ApujTOTeXrjs ^>rjal nepioTepwv jtev elvai 6 ctBos vocat rem denniendam, Waitz adlv yevos, ciSij St ncvre . . ., and names the I4ib28.same five that A. gives at H.A. 593ai5; but 7 wavi-os ytvovs ciSy wXeico Top. I23a3o;A. calls these TrepiarcpofiSwv yevrj (544bi), cf. I27a24; Bonitz, Arist. Stud. iv. 355.never «Sij.

84 D. M. BALME

yevos. In a full analysis a hierarchy of yevrj VTT' aXXr/Xa1 may be disclosed, whilein the other direction the eiSos may be found to be itself divisible into differentforms: then what was etSos in relation to the higher group becomes the yevos ofthe lower groups,2 and so on until a truly indivisible form is reached, and this isthen the drofiov or reXenratov elBos whose members are particular individualsdistinguished only by differences of vXr).3

As a method of analysis this contains nothing that cannot be found inPlato, except the distinction between yevos and eiSo?. But this very distinction,which might have served little purpose for Plato, becomes important for thosewho find reality in the roSe n. Genus and species then differ in epistemologicalstatus, and tend to be used not relatively but absolutely. The infima species isabstracted immediately from sense-perception, whereas all classes above it areabstracted from abstractions. In Aristotle's theory of substance the form that isactualized in the individual's matter is that of the aro/xov eZSos, while the suc-cessive ranks of genera are successively remote stages of potentiality which canonly exist when the infima species is actualized. If the male sperma fails toinform the female matter fully, the offspring may not resemble its parents butonly their genus, and this is a step on the way towards the monstrosity: theripas is £<SoV TI but no more.4 And by a similar argument Aristotle explainsthat the mule, which cannot resemble both horse and ass, resembles onlytheir common eyyvrara yevos and is unnatural and defective, napa <j>vcnv,ava.TT7jpos.s The stock example in the logical works of yevos—ethos is £<3ov—dvOpanros, and the real significance of the distinction between genus and specieslies not in its relative use at all levels, but in its absolute use at the level of theinfima species.6

One would have expected Aristotle to use this distinction above all inbiology, where it is most evident that the final differentia expresses the essence,7

so that the immediate object of science must be species rather than genus. Butthe surprising fact is that he makes least use of it in this field. It does not occurin the explanations of family resemblances and repara, nor of the mule, justquoted; there the distinction is between the universal (both t,wov and dvBpamos)and the particular (Socrates); he speaks of yevos and KO.86\OV as opposed toiSiov, KO.6' exaarov, and in the only paragraph where elSos occurs it is inter-changed with yevos.8 It is true that P.A. i apparently distinguishes yevos,hiaj>opd, ethos in the same way as the Topics does. But P.A. i is a book apart: itis not so much biology as a philosophical discussion of biology. Its usage ismarkedly different from that of the works in which Aristotle actually practisesbiology. There, in the strictly biological works, although it has been traditionalto interpret yevos and ethos as genus and species in many passages, examinationshows that there are only a few in which the technical sense is obligatory; ina few more it is possible but not demanded, while in others it even causes con-fusion (as D'Arcy Thompson pointed out) and in some it is impossible. Now inthe case of the logical works the reader has a clear choice on each occasionbetween the technical and the non-technical senses: both have been established

1 Top. 107*22, Cat. i b 2 i ; Bonitz, Ind. between Plato and A. {Zw Lehre v. Eidos bet' 5 i a 45- P- "• A., 1941, pp. 12, 21) depends upon

2 Phys. 227bi4. identifying eidos with aro\xov elSos.3 An. Post. g6bi5, 97a35-b6; Met. 1034=7. ' M'1- iO38aig-26; P.A. 644*3.* G.A. 4. 3 (cf. 7 6 9

b9 ) . 8 G.A. 747b33-748a7- (At 77ob i 7 , 772b25

5 Met. iO34ai; G.A. 2. 8. elSos means the formal cause opposed to6 The contrast that N. Hartmann draws VXT).)

rENOZ AND E1A0Z IN ARISTOTLE'S BIOLOGY 85in practice, and, where the technical sense is required, it is required because thecontext distinguishes certain concepts in the way in which genus and speciesare distinguished, and these examples can be pointed to in the text. In thebiology, however, there is no such framework of reference. The traditionalinterpretation assumed that Aristotle did actually classify animals into generaand species, but this assumption is not supported by the evidence.1 He acceptsas data the animal kinds {yevrj) that are presented to him by common par-lance, but does not try to group them further by finding similarities, as Speusip-pus appears to have done. Such kinds are: LTTITOS, KVCDV, KTX. ; a very few widergroups such as TO. aeXdxrj, TO. X6<f>ovpa; and certain general groupings (iieyioTayevrj)2 which comprise most—but not all—animals, namely opvis, lx@vs, evTOfxa,KTX. The extant treatises examine the common and peculiar characteristics ofthese yevrj, their functions, their reproduction, but not their taxonomy; more-over there is no classification scheme in the background, and all attempts toconstruct one for Aristotle have failed.3 Consequently the question is not somuch why Aristotle usually fails to distinguish the words yevos and etSos inbiology: it is rather, how does the distinction come to be there at all, in thefew cases where it does appear ?

In the zoological works excluding P.A. i4 the occurrences of the two wordsare, so far as I have observed:

H.A. 1-9P.A. 2-4Inc. A.G.A.Paw. Nat.4

yevos

207647

122

' 3

4 ' 3

{avyyevij et sim.)

("3)(2 )

(0(21)

(1)

(38)

ethos

5280

333

96

(o/xoctS1^ et sim.)

(5)(0 )

(0 )

(6)(0)

(11 )

Of the 413 instances of yevos, 275 are applied to a named type of animal (e.g.yevos IITTTCOV), 79 to animals in general (e.g. 7roAAa yevrj £<Lwv), 25 to other kindsof thing (e.g. yevos <f>VTa>v, oSovrwv), and the remaining 34 mean 'kind' in theabstract, including sometimes the sense of genus as opposed to species. That isto say, in 354 cases yevos denotes a kind of animal. On the other hand, ofthe 96 instances of elSos only 24 denote kinds of animal. Thus yevos is far thecommoner word for a kind of animal, though it is of course by no means thecase that Aristotle is concerned with genera rather than with species. He usesyevos indifferently for the type that is visible in nature and for all groupings ofsuch types. Bonitz is mistaken in applying the usage of the Topics (where•navros yevovs eiSrj nXelo}, I23a3o) to the biology, and inferring that yevos isnever used of the infima species.5 For not only is every visible type called ayevos, but if it has sub-types these are yevrj too. Thus dog is a yevos 658*29, andso are the breeds of dog, irXela) yevrj KVVWV 574ai6; similarly TO yevos TO TWV

' 579b5 a n d y^") heovriov 8vo 62gb33; compare f3oa>v j82by with 666big,

Thierkunde, on which see my paper, loc. cit.4 i.e. H.A. 1-9, P.A. 2-4, Inc. A., G.A., and

Pan. Nat. 453bi 1-458*32 and 464bi9-48ob3O(de Somno, Long. Vit., and all from Juv. to endofResp.). Mot. A. and H.A. 10 can be ignoredfor this purpose.

5 Arist. Stud. iv. 386; Index I5ib57.

1 For details see my paper 'A.'s Use ofDifferentiae in Zoology', contrib. to 2ndSymp. Arist. (Louvain, i960).

2 H.A. 49Ob7, 5O5b26. Not a technicalterm: cf. P.A. 683b26, H.A. 4gobi6 ylv<)licyaXa, 5 3 4 b l 3 yeVi? a i j - e / u e x c i . . . .

3 Full examination in Meyer, Arist.

86 D. M. BALME

5o6ag; Terrlycov 682ai8, 25 with 535b8; KapiScov 525a33 with 34;ibid.; noXimoStov 622ai4 with 525ai3; ofecov 511 a i4 with 15, 5O5bio; axaXr)<f>iov531 a3i with b i o ; fxvwv 6y6b3i, 771 "23 with 581 a 5 ; exwa>v 528a2, 683bi4 with53°a34> 68oai5- yevr) is the word for the different types of haaxmoboov 5O7ai6,

4, TTOp<j>vp&v 547a4, Kvihwv 5 4 8 ^ 4 , cnroyywv 548a32, tfivxajv 55Ob26,551 a9, <f>9et,pwv 557a4, deTwv 563b5, 6 l8 b l 8 , Upaicuiv 564a5, 62Oa22,

8eX</>lvci)v 5 6 6 b I 2 , KeOTpecov 5^9a '7> epwSiwv 6o9 b 22 , SpvoKoXanTihv 6 l4 b 7 ) KOT-

TVtjxov 6 1 7 a l I , KOXOUOV 617^18, KopvSdXojv 6 l7 b 2O, (f>aXayylcov 622 b 28 , apaxyioiv

ib id . , apaxyiojv yXa<j>vpa>v 6 2 3 a 2 5 , o<f>7]Kcov 627 b 23 , o<f>r]Ka>v rjfj,fpcov 628 a 2 , Oo&v

63oai2. Each of these subdivisions is clearly the type visible in nature, i.e. thefinal differentiation known to Aristotle, what would now be called the speciesor variety, and what would in the Organon and Metaphysics be called theO.TO}XOV etSos.

Of the 96 instances of etSos, only 13 are applied to named types of animal,and a further 11 to animals in general; 4 refer to other things {eiSos Tpi-x&v49ob28, 78ib34; (fxovijs 536bio; KoiXlas 674a23, cf. 5O7b23), and the remaining68 bear an abstract sense and occur largely in oblique cases (e.g. eZSei 8ia<f>epei.).Of these 68 abstract instances, 16 (all in G.A.) mean the formal cause, andmany others have a general sense of 'form', 'appearance', even 'nature'; butthere are some that seem to demand the sense of 'species', and these will be dis-cussed below.

Taking first the 13 references to named animal-types, we find that 5 of themrefer to types that are also called yevr], so that the technical distinction is pre-sumably excluded:

H.A. 532bI4 Terriytav elSr); cf. 556aI4 rerrlycov yevr), 535b8 Te-nLywv TI yevos.H.A. 557a24 ev eiSos ion <f>6eipa>v OaXaTTiwv; cf. 557a4 £<JTI' §£ yevos <j>9eipu>v.H.A. 6 l 7 b l 6 KOXOI&V £1877 Tpla; cf. 617^18 aAAo yevos KOXOI&V.P.A. 68oaI5 OVTCDV 8e TfAeiovajv yeva>v, oil yap ev etSos TWV eyivwv ITOLVTCOV earl',

cf. H.A. 53Oa34 yevt] nXeia) ex^vaiv, b4~7 aAAa Svo yevr/, dXXo yevos.G.A. 758b9 8ei Taj Kappas eiSos rtOevai OKWXTJKOS ; cf. 723b6 yevos TI OKOiXriKwv.

Three refer to sub-types, such as have been seen above to be most commonlycalled yevtj: e'187] y\m&v 592b7, alyiOaXaiv 592bl8, KI ACUV 6l7bl8.

In one case only do yevos and etSos appear to be distinguished: H.A. 593b8TO Totv OXKVOVCJV he yevos TrdpvSpov eorw. Tvyxdvei 8' avrwv ovra Svo eiSrj, Inview of the overwhelming evidence of the other cases, it seems clear that et8-qmust be non-technical here, and is used to mean 'forms', mainly for literaryvariety.

The remaining four cases refer to more abstract groupings than the visiblenatural forms:

H.A. 486a24 eiSrj wXelcx) ixOvwv Kal opvWcov. This occurs in a context where thedistinction between genus and species seems to be really intended. It is dis-cussed below.

P.A. 679bI5 eori 8e yevT) Kal eiSrj iroXXa, T&V 6crTpaKo8epfia>v. No further sug-gestion is made of a distinction between yevos and et8os among testaceans, andthe word etSos does not again occur in connexion with them except in the non-technical use with ixtvot, quoted above (68oai5). Cf. H.A. 528a2 TO T&Vyevos, 53Oa34 f. yevr] TrXelw exlvcov, dXXa Svo yevrj, aXXo yevos; 528a24~26

eW, TI yevos. The use of both words together here may show that they

rENOZ AND EIAOE IN ARISTOTLE'S BIOLOGY 87

have not completely lost their basic difference of meaning: i.e. yevos is thegenetic 'kind', elhos is the morphological 'form'.

P.A. 683b26 eori 8e yemj p,ev rer rapa ra /xe'yior' ai5raiv (sc. TWV ixaXaKoo~rpa.KU)v). . . TOVTOJV 8' eKaarov irXelo) eZSrj earl. T h e same applies to this passage as to679bi5 above. el8os is not again used in connexion with crustaceans, exceptonce in the abstract sense of 'shape': H.A. 525bio Zrepov yevos (uxpov fxevwairep 01 Kapxlvoi, TO 8' elhos o/xoiov rots doraKols. yevos is the word used other-wise for all groups of paXaKoarpaKa: cf. 525a34 yevr] irXeLco KaplStov KCLI KapKivoov. . . b2 TO fiiKpov yevos.

G.A. 7 : 9 a 7 aficjiOTepcov pieTex^ov TWV el8wv (sc. £>WOT6KWV Kal CLOTOKCOV). T h eovo-viviparous selachians and vipers have some of the characteristics of eachgroup. el8os bears a general sense of 'class'.

The above evidence shows that, wherever it is possible to test the denotationsof yevos and el8os by reference to actual animals, Aristotle's usage makes notaxonomic distinction between them, yevos means a 'kind' at any level fromthe most abstract group to the immediately visible type. elhos is far less com-monly used, and represents no group that yevos does not also represent. If theydiffer in meaning, the difference is not that of higher and lower rungs on thesame ladder. They belong to different ladders, and the original differencesometimes shows through, yevos being a statement about kinship and elhosa statement about shape or form.1

There remains, however, a number of passages in these biological works inwhich elhos is used with a more general reference in opposition to yevos, and insome of these the sense of 'species' appears obligatory. In the following pagesI offer an examination of all passages that appear to be at least capable ofbearing this sense (these include the 11 general references to animals, and someof the 68 abstract instances, referred to above).

H.A. I. 486 a l6 ravra 8e ra /j.ev etSet TWV piopiwv ecrrlv, olov

17 dv6pc!)TTOv pis Kal 6<j>8aXp.6s dvdpconov p\vl Kal 6<f>9a\p.a>,18 . . . TOV avrov he Tponov Kal

I g LTTTTOV Kal Tojv dXXojv Z,totov, ooa ra elhei ravra. \eyop\ev . . .21 . . . TO. he Tavra. p,ev eartv,

22 hia<f>epei he Kad' imepoxhv KaL eWeupiv, oawv TO yevos earl23 TavTOV. Xeycj 8e yevos olov opvida Kal l^Ovw TOVTWV yap24 eKaTepov e^ei 8ia<f>opav Kara TO yevos, Kal eo~nv elhr) nXelco

25 l)(9v(ov Kal opvlBwv. . . .b 17 • • • evict he TWV £a><ov18 ovre eihet. TO. fj.6pia r aura e%ei ovTe Ka6' vvepo)(r)v Kal eX-19 Xeiipiv, dXXa. Kar' dvaXoylav, olov irenovQev ocrrovv Trpos d-

20 Kavdav Kal . . . Trpos

21 TTTepov Xeirls' o yap ev opviOi TTTepov, TOVTO ev l)(dvi ecrrl Xeiris.488b3O raOra S' earl Tavra Kal eTepa

31 /card TOVS elprjpevovs Tponovs, 17 f a r ' el8os rj Ka6' vnepoxrjv rj32 Kar' dvaXoylav . . .

491 a I 8 rdy elpTjfjLevas 8ia<f>opds, eihei Kal imepoxfj Kal dvaXoylq. . . .

These three passages together with 49 7b 10-12 (see below) distinguish thelevels of comparison in a way that is found in P. A. 1, and occasionally in logical

1 As also perhaps in Plato's diairesis: cf. Dies, Plat. Polit., p. xxi.

yevei

88 D. M. BALME

usage, but not elsewhere in the biology. There are three levels: between in-dividuals, between species, and between genera. The language of these passagesis over-condensed and confusing, but from other references1 it is clear thatAristotle means the following:

(a.piBfj.<x> ev —the individual)

»<> , / ! —different individuals in one specieseioei ravra )

Sei 8iad>epovra \ ,.™- ., ; , 1 —different species in one genusevei ravra )

yevei Siatbepovra \ ,.&•, . , , , t —different generaavaAoyia ravra )(dvaXoyia Suu/tepovra —incomparables)

Since biology is not concerned with individuals, there are in effect only twostandards of comparison, namely between species (vnepoxq) and betweengenera (dvaXoyia). Hence the expression at 488b3O is too condensed: Tavrd Kalirepa does not apply simpliciter to each of the three levels mentioned, but KOT'ethos goes only with erepa, KO.6' \me.poyr\v goes with both ravra and erepa, andKOLT' dvaXoyiav goes only with ravrd. At 491 a i8 the statement is further con-densed and has become actually incorrect, for Si.a<f>opds cannot apply todvaXoyia at all.2 At 497bio the theory is correctly stated, but comes in the samesentence with an incompatible usage of yevos and eZ8os, as will be seen.

Now the terms dvaXoyia and xnrepoxq (often called TO fxaXXov Kal JJTTOV) areused throughout the biology, but they are not distinguished in this way. Forexample, both dvdXoyov and TO fidXXov Kal JJTTOV are applied together to thecomparisons between man and other animals at H.A. 588a25~28; dvaTrvor/ iscalled dvdXoyov to the use of fipdyxia at H.A. 58gbi8 (also at P.A. 1. 5. 645b6)although this comparison is made within the yevos Ixdvojv; within a species,male sperma is called dvdXoyov to female KaTapr/via, G.A. 727a3; OOTOVV andXovSpos are dvdXoyov at P.A. 653b36 (cf. H.A. 511 b7) but differ T& fiaXXov Kal T^TTOVat 655a32; OO-TOVV is dvdXoyov to aKavda according to the theory at 486big, yetthis comparison occurs within both the yevos Ixdvoyv and the yevos TeTpcwrdSoiv(LOTOKWV, H.A. 5i6bi4—22: dvaXoyia is applied to the differences of hardness inthe bones of vet,d t,woTOKa, H.A. 5i6b4, although this type of comparison isdistinguished from dvaXoyia and classed as vTrepoxrj at P.A. 1. 644bn andcorrespondingly at H.A. 1. 486a25 f. In practice dvaXoyia is most often used tocompare evai[i.a with dvaifia, a difference that is wider than the /xeyiora yevrj,while TO [idXXov Kal JJTTOV is used promiscuously.3 The reason for this is that thedistinction between genus and species is not drawn in practice, and hence thetheoretical distinction between dvaXoyia and vnepox^, which is made to dependupon it in 486ai6, etc., is not and cannot be applied.

It is equally inapplicable to the hierarchical type of classification that isoften envisaged in the logic. It could only apply where there are just the twolevels of division, a genus and its member-species, of which the distinction canbe used absolutely. Thus it could apply to the simple classification l<x>ov-

1 Top. iO3a8, An. Post. 98a2O, Met. 2 Cf. Muskens, De Vocis dvaXoytas Significa-ioi6b33, ioi8ai3, Phys. 228bi2, 242b4, P.A. Hone, etc. (Groningen, 1943), p. 43.639a28, 644ai6, 645b26. Cf. Bonitz, Arist. 3 Cf. Meyer, Arist. Thierkunde, pp. 335-44.Stud. iv. 386.

rENOZ AND EIAOS IN ARISTOTLE'S BIOLOGY 89

opveov-Kopai; (Top. io7a23) but not to any longer ladder of yeVq \m' aAAijAa,such as would result if intermediate groups were inserted in that classification.Still less could it apply to Aristotle's flexible or indeterminate grouping ofmany animals.

Further, it sets the genus-species distinction at the level of opvis Kal Ix8vs.This is a very simple degree of classification. It is used in the discussion at P. A.1. 2-4, but does not appear elsewhere in biology except in the two statementsabout (jLeyurra yhn\ that are discussed below (49,ob7 and 5O5b26). It is also rarein the logic (e.g. An. Post. 97a4, Top. 107*23); there, if an example of genus-species is wanted, the commonest is ^coov-dvOpamos; but in practice a moreelaborate hierarchy is used, of which the biological counterpart would be amore advanced degree of classification than this.

These passages in H.A., therefore, are strangely out of key. The concepts ofdvaXoyla and xmepoxq proved useful to Aristotle, and led to the modern con-cept of homology; but they would have been useless if tied artificially to genusand species, as they are tied here but nowhere else. Seen against Aristotle'spractice in biology, these statements are too doctrinaire. Yet if they are com-pared with his logic, with which they seem more akin, they are too primitive.This incongruity, coupled with their curious incorrectness, suggests that theyare not integral parts of the biology.

H.A. I. 490^7 yevr\ 8e fieyuTTa TWV l,d>(av, els a hirjpr\Tai TaAAaL,wa, Tab eoriv, ev p.ev opviuojv, ev o i^pvcov, aAAo oe

9 K-qrovs. ravra fiev ovv iravra evai/j,d eaTiv. aAAo Se yevos10 earl TO Taw oarpaKohep^wv, o KaXelrai oarpeov aAAo TO TWV11 fj,aAa.KO<rrpdi«x)v, dvwwp.ov ivi oVdfiaTi, ofof KapajSoi Kal12 yivr) Tiva KapKivwv Kal aoraKwv aAAo TO TCDV //.aAa/a'cov,13 ofov TevdiSes T€ Kal Tevdoi Kal arjnlai' <=Tepov TO T£>V ivrofMov.16 . . . TCUV Se Aoi/n&v t,woyv OVK£TI TO. y4vrj fieydXa' ov yap TTepii^i17 TToXKa etSrj ev etSoj, dAAa TO (lev ioTiv dnXovv avro OVK18 <£xov ^a<f>opav TO etSos, olov avdpconos, TOL 8' e^et /xev,19 dAA' dvwvvfui. TO. CLST] . . .31 TO5 8e yivovs TOV TWV TeTpanoScav t,wiov Kal ^COOTOKCOV eiSr)32 ftev ioTi TToXXd, dvc!)vvfi.a 8e- aAAd Ka9' tKaarov avrwv33 cu? elnelv, wcrnep avdpamos eipr/Tai, Xdcov, eXarf>os, ITTTTOS,34 KVWV Kal TaAAa TOVTOV TOV Tpovov, ivel icrTiv iv TL yevos Kal

491 aI e-nl TOUS Xo(f>ovpoLS KaXovpevois, olov ITTTTUI Kal ovco Kal opet2 . . . /cat Taty ev Evpia Ka\ovy.evais rjfii6voi.s,3 at KaAoCvrai rjp,iovoi 8t' o/iotoTrjTa, OVK ovaai dnXcis TO avTO etSos.

'The whole passage is very troublesome, and Aristotle seems to juggle withthe terms yevos and etSos,' Thompson.

This is one of the only two discussions,of jxeyiora yevr) (the other is at 5O5b26,see below). It comes towards the end of the introductory section of H.A. (1.1-6), in which Aristotle has explained that the treatise concerns the likenessesand differences between animals, and has analysed these Sta^opat generallyunder certain main headings, according to which the treatise will be arranged.Now finally he says that there are certain main groups1 into which animalsfall, implying that he will take the animals by these groups, as in fact he does.

1 liiytara: not 'highest' but 'largest': not a vertical but a horizontal division. So Thomp-son: 'very extensive'.

go D. M. BALME

But here lies a difficulty: not all animals fall into main groups; or, if they do,their groups are not all recognized (4gob 16-19). ^n l i n e s I 9 - 3 ° n e discusses theconnexions between four-footedness, viviparousness or oviparousness, hair orscales, and concludes that there are two further main groups, namely viviparousand oviparous quadrupeds, but that no more grouping can be done by com-binations of these factors. Then comes 1. 31 where he says that the group ofviviparous quadrupeds contains unnamed eiSrj.

He has therefore made two chief points: (1) that these fieyiora yevrj do notinclude all animals (for the two reasons given in 11. 16—19) J (2) that the yevosTerpaTTohwv £WOT6KWV includes etSrj dva>vvfj.a. The difficulty in the passage con-cerns the meanings of eiSos and dvwwfios.

It is first clear that elSos is not being used absolutely: ov yap vepUxei TTOXXOLe!8rj ev eiSos (16). Thompson suggests that yevos and eiSos are used as relativeterms here, but even this does not solve the difficulty, since ev eiSos representsyivr) fieydXa (16) while iroXXa ethr) is on the same level as yevrj KapKivcov (12).Hence yevos and eiSos are being used interchangeably, 'kind' and 'form'.

Thompson also notes that TrepUxei is technical, and compares de Caelo3i2ai2 and Phys. 2O7bi, where ethos irepiexei vXrjv. But perhaps a better com-parison lies in the logic where classes embrace classes: Top. I2ib25 irepiexet, TOa/nohoOkv yevos . . . TO AOITTOV nepiexei . . . SoKei yap, orav ev eiSos viro 8vo yevr\fj, TO eTepov vno TOV eTepov irepiexeodai . . . I4-Oa2 57 dpeTT] TTJV avfj.<j>a>vlav vepiexei. . . I44aI2 nepUxovaiv dXXrjXa . . . T&yadov rqv e£iv irepiexet . . . I44bI3 17Trepiexopevov fj.T]8e nepiexovTos . . . al. (Cf. P.A. i . 644ai4, b5.) This usage seemsto be Platonic: cf. Soph. 250 b, 253 d, Partn. 138 a, b, 145 b, 151 a, b. 8ia<f>opdis also technical here in the manner of the Topics rather than in the looser senseof 'difference' that is common in H.A. This technical use too may derive fromPlato: cf. Theaet. 208 c f., Polit. 285 b, and the Academic Definitiones 414 d, 416.

This passage, then, is concerned with an elementary degree of classification,and uses terminology familiar from the Topics and from Academic usage. But itdoes not consistently distinguish yevos and et8os. It becomes manageable ifthese words are translated not 'genus' and 'species' but 'kind' and 'form'.

dvwvvjjios. What, then, are the et'817 avdivvfia at line 32 ? In order to preservethe sense 'species' here, Schneider1 altered Scaliger's translation from 'nominecarent' to 'uno aliquo communi nomine carent', understanding evl 6v6/xaTi withdvcowfia, and compared Aristotle's statement that the oo-TpaKoSepfxa are calledSo-Tpea while the paXaKoorpaKa have no single name (11. 10-11 above). Intaking dviLwjxos to mean evl ovofiarc dv<x>vvfi,os, Schneider is surely on goodground: cf. 4goai3, b n (above), 623b5, P.A. 642bi5, de An. 4iga4. But I canfind no parallel to his translation 'lacking a common name', meaning that thingswhich have single names of their own lack a name shared jointly with eachother (unless H.A. 5O5b3O has this sense: but see below). All the other instancescited by Waitz ad Int. igb6 and by Bonitz, Index 6gb2-26, have the straight-forward meaning 'nameless', i.e. the object referred to has no ovofia (cf. Int.igb6 rj ovop,a -q TO dvwwyiov). The same is true of the following instances whichthey do not cite: 492ai5, 493=28, 494=3, 14, 5i5bio, 552b3i, P.A. 669b9,683b24. When Aristotle means that a common name is lacking he says, as at66gb9, dvwwpov TO KOIVOV; cf. H.A. 53ib23, Phys. 226=30, 32, Pol. 1275=30. AtAn. Post. 74a2' 81a TO p.r) eivai wvofiacrfj.evov TL iravra Tavra ev, the subject isirdvTa Tavra and wvofj.aafj.evov TL is complement with ev; cf. ibid. 8 dvwwfiov

1 Following Camus, and followed by Thompson; so Waitz ad 1 gb6.

rENOZ AND EIAQE IN ARISTOTLE'S BIOLOGY 91

em 8ta<f>opois etSet irpdyfiaaiv (sc. TO dvwTepov). At 49ObII it is not the KapafioiKal yevrj Kapicivwv KTX. that are dvd>vvp.ov evl ovofian, but the fxaXaKoarpaKOivyevos: the reason is that an adjective is not an OVOJJM even when used 'sub-stantivally', unless it has really become a noun; cf. 490s 13 (TO. TTTiXarra /cat TOSep/xoVrepa . . .) dvwvvpM. evl oVd/xart, 623b5 yevos ri TWV evrofxiov evl OVO/JMTIdvcowfiov . . . oaa Krjpiovoid, P.A. 642bI5 dvwvviioi, olov TO evaifiov /cat TO dvai^ove<f>' eKarepu) yap TOVTCOV ov /cetrat ev 6Vojna, and 669^10, de An. 417^32, 4 I^a27-At 491 ai TO. \6<f>ovpa can count as a named yevos (if this passage is genuine inspite of the contradiction at Met. 1034*1) because it has become a noun:L.S.J. quote it only in the neuter plural and give no instance of its use as anadjective; therefore it differs from the adjectives which Aristotle calls dvu)vvp.a(cf. 6oiai7 for adjectival /xaAa/cdorpa/cos and oarpaKoSepfios). It may be alsothat a descriptive phrase or periphrasis is not a name: cf. above examples, andMet. IO56a25 TO p.ryre dyadov /jL-qre /ca/cov dvwvvfxov, E.N. IIO7bI, 7, 30. Thiswould explain why the aladrfrrjpiov of touch is dvd>wfu>v at 48gai8: it can onlybe expressed by the phrase aa.pt; 7} dvdXoyov (unless Aristotle here refers to hismore accurate doctrine that the aladrj-rripiov is not the flesh but TI eVrds1).

If eiSr; dvd>wp.a means etSij that have no 6v6p.aTa, it cannot refer to Aewv,eXa<f>os, KTX. It must refer to groups that contain these types and are themselvescontained within TeTpaVoSa £«K>TOKCI. What, for example, is the name of thegroup that contains the lion? We have one ('Cat'), but Aristotle has not. Allthat he has at this level is TO. X6<f>ovpa, and the rest are nameless. eZ8rj dvd>wp.atherefore denotes the same missing groups as at 1. 19 above, i.e. groups inter-mediate between the main groups and the types visible in nature.

At 4gia3, TO avro elSos, the sense 'species' would be possible, but the con-text demands no more than 'form'. This type is elsewhere called only a yevos:577^24 {at ev Evpia rjfxlovoi. . . .) eart TO yevos ojnoiov fiev eTepov Se, 5^O°I evZvpLq ol KaXovp-evoi, rjfj.lovoi, eTepov yevos.

H.A. 2. 497^9 o~)(e86v yap oaa y' earl yevei eTepa TWV ^axov, teal10 TO. irXelaTa TWV p.epu>v e^ei eTepa TU> etSet, Kal TO. p,ev Kar'11 dvaXoylav d8iaxf>opa povov, TW yevei. S' eTepa, TO Se rai12 yevei fiev Tavra TU> etSei 8'

Thompson, ad loc.: 'In the opening sentences, which must be read togetherwith those of Book I, brevity leads to a certain appearance of confusion: we arereminded that a generic difference between two animals carries with it genericdifference between certain parts as well as specific difference between many others.'But the words which I have italicized are not in accord with 1. 486ai6f.(above), yevei. eTepa are, for example, opvis and ixOvs (486a2i). Their parts areto be compared otire ei'Sei ovre Ka9* VTrepo^v Kal eXXeuftiv, dAAa /car' dvaXoylav. . . o ydp ev opviOi iTTepov, TOVTO ev l)(dvi earl Xenls. Kara. p.ev ovv fjuopta a exovaweKaoTa T&V t,wwv, TOVTOV T6V Tpoirov eTepa eari /cat Tavrd (486bl8—23). T h eexpression eTepa TS> etSet, if used technically, is applicable only to yevei Tavrd(e.g. the difference between px/cpw pvyxos and jSpa u p'vyxos 486bio). Henceat 497bio et'Sei cannot mean 'species' but must either mean 'form' or beequivalent to yeVei (as at Cat. lbl6 TCOV eTepoyevwv . . . erepat TW etSet /cat atSta^opat, cf. Top. ioyhiQ). The rest of the sentence, from /cat TO. pev KaTJ

dvaXoyiav, summarizes exactly the doctrine of 1. 486ai6, etc. Therefore

1 During thinks not (Arist.'s P.A., Critical and Literary Commentaries, p. 130).

92 D. M. BALME

in 1. 12 requires the sense of 'species', which in 1. io is impossible. This sug-gests that the two parts of the sentence originated separately.

H.A. 2. 5O4bI3 TO TWV l)(6vo)v yevos . . . noXXds vepi€)(Ov ISeas.

Thompson translates ISeas 'forms', but comments: 't'Se'a, frequent in Aris-totle of a logical species, is not used of the species of animals or plants. It mayhere mean the sensible species.' It is true that ISea is not used of animal or plantspecies, but I cannot find it used of any other sort of species either. In theTopics it is used always, I think, sensu Platonico. Bonitz, Index s.v. ISea 338b34—48'logice i.q. species generis' cites various passages, of which none is from thelogical works and none supports his interpretation. They are as follows: E.N.ioo,6bi6, 25 ISea is meant sensu Platonico; de Caelo 274b2 ISeas TSSV dnXcovcrajfidraiv means 'kinds' (Stocks) or 'forms' (Guthrie); 'form' is also the mean-ing at 276b2, 285a25, Meteor. 38oai7, b3o; de Caelo 268a2i ISea as 'form' isdistinguished from vXq; P. A. 656a3 ISea <f>vra>v means 'configuration' (Ogle) or'formation' (Peck); Poet. i458a26 alvlyfiaros ISea 'the essence of a riddle';I449b8 and I458bi8 'form'. Now Bonitz, loc. cit., refers to Waitz ii. 406 andmay have made an error of recording, for there Waitz quotes four of the abovepassages to demonstrate precisely the reverse, namely that ISea can mean (a)ovala, TO TI -fy/ elvai, or (b) outward form ('externa species') in distinction fromthe sense of 'species generi subiecta' which Waitz proposes tentatively ('haudscio an') in only two passages. These two passages are de Caelo 285"25 and H.A.5O4bi4 (above). But at de Caelo 285 a25 ISeai is used of TO dvw, TO KOTO, TO Setjiov,KTX., i.e. the 'forms' of position (Stocks, Guthrie): cf. Phys. 2O5b32 TOTTOV eiSrjical Sia^opal, 2o8bi3 TOTTOV fiep-q Kal eiSrj, clearly non-technical, in the Platonicusage. There remains therefore only the present H.A. passage as a supportfor Bonitz's sub-heading. But the other instances of ISea in H.A. all meanexternal 'form': 6i5b8 ISeav fia/cpos Kal Xevros, cf. 530a3o, 577aio, 58oa28,592bio, 6i5b28, 6i6bi , 63Obi3. The mere use of the word nepiexetv here isnot enough to give to ISea a meaning which would be unique, and the passagetranslates perfectly well without that. It is a general introductory statement thatthe class of Fishes comprises many animal forms.

H.A. 2. 5O5b26 Tovrcp yap Siatfiepei TO. [x.eyicrra yevrj irpos rd Xonrd27 TU>V aXXauv t,wwv, Tip TO fiev evaifxa TO. 8' dvaifxa elvai,.28 earl Se Tavra dvOpanros r e Kal r d t,tpoTOKa TU>V TSTpaTroScov,

29 e n Se Kal Ta eboTOKa TG>V TSTpairoSiov Kal opvis Kal ixOvs

30 Kal KrJTos, Kal ei Tt dXXo dvuivvfiov eaTi Std TO fir/ eivai

31 yevos dXX' aTrXovv TO eiSos em TWV Ka9' eKaaTov, olov o<f>i.s

32 Kal KpoKoSeiXos.

29 e n Se Kal T<J UIOTOKO. TWV rerpavoSiav om. D a .31 otf>LS Kal KpoKoSetAos] e^ts Kal KopSvXos ci. Di t tmeyer .

The natural reading makes eiSos opposed to yevos in 1. 31; but it is not clearwhat eiSos denotes, nor what II. 30-31 mean from Kal ei TI to eKao-Tov.

What is dvwvvfj.ov ? Obviously not 0^1? Kal KpoKoSeiXos, but the missing yevosof which they are ecSy. In that case the meaning is: 'and anything else (i.e.any other [j.eyicn-ov yevos) that is nameless because of there being no yevos butonly the eiSos which is simple and is predicated of the individuals'. There is aslight oxymoron: the (group) is nameless because there is no group—i.e. thereis nothing but unrelated simple eiS-q.

TEN OS AND EIAOZ IN ARISTOTLE'S BIOLOGY 93

etveu (30) surely means 'exist'. If it were copulative, with subject suppliedfrom TI aAAo and with yivos as complement, then Aristotle would be sayingthat the n 0AA0 is not a yivos but is the etSos by itself, e.g. snake; but if he weresaying this he could not also say that it is dvwwfj.ov, since o ty is its ovofia.Camus, followed by Schneider, Wirnmer, B. St-Hilaire, Thompson, takes etvatas copulative1 and dvcowfxov 'lacking a common name'. So Thompson translates:'and all the others that come under no general designation by reason of theirnot forming genera, but groups of which simply the specific name is pre-dicable', and refers in his note to the relative use of yivos-eiSos, comparing hisinterpretation of 4gobi6 (on which see above). But this interpretation leads todifficulties:

1. There is the objection, explained above, to translating a.vwwii.05 'lackinga common name'.

2. Schneider himself objected that 5<f>i,s has in fact already been called a yivos(no doubt thinking of 4gob24, 5O5b5), and concluded that the word is corrupt.Dittmeyer conjectured ex^s, to which Thompson (followed by Tricot) replied:'As Meyer, Arist. Thierk. p. 155, says, the serpent as yivos contains many et'Sr;,but is also, as here, a single ethos in the great yivos, TerpdnoSa WOTOKO. <f>oXiSwTa.'That is, he appeals to the relative sense of yivos-eiSos. But this answer raisesa further difficulty:

3. I f o<j>i,s Kal KpoKoBeiXos a r e m e m b e r s of t h e yivos TeTpwnohwv CLOTOKWV,2

then how can Aristotle say Sid TO fiij elvai yivos ? For he has just named thererpdnoSa WOTOKO. among his list of fiiyioTa yivrj. Meyer answered: the yivosthat is missing and nameless is the intermediate group between 6<j>t,s andrerpdnoSa woTOKa. But this answer would equally apply to the majority ofanimals, since there are extremely few intermediate groupings to be found.Moreover, Aristotle is speaking here of piyiora yivrj, and the plain implicationof the sentence is that some animals are not contained within the fj.iyi.o~ra yivr/that he has just listed.

The traditional interpretations therefore fail to solve the problem, and inaddition they depend upon an unlikely sense of dvcuwfios. Dittmeyer's emenda-tion does not suffice, since e^is *s certainly a yivos within the yivos o<f>ewv3 (eventhough KopSvXos is admittedly in an ambiguous position).

If TI aAAo is a missing piyio-Tov yivos, as seems likely, then what is TO eiSo? ?And what is exemplified by olov 6<f>is Kal KpoKoSeiXos ? This is the crux of theproblem. eiSo? may be either technical (species) or non-technical (form,group), while TO. KO.8' eKaarov is used by Aristotle to mean either dVo/ia eiSry orindividual particulars. Since it is stated not to be a member of a yivos (Sid TO /AT)elvai yivos 30), eiSos cannot be used technically here, unless by a sort of meta-phor, but more probably means eiSoy £d>ov, form of animal. It refers, accordingto the meaning of KO.8' eKacrrov, either to a group of types or to a type of in-dividuals. Either could be exemplified by 6(f>is (assuming that Meyer is wrongin grouping 0 1? under TerpaVoSa WOTOKO.) since o<j>is is the name both of a typeand of a group of types, while not itself falling under a piyio-rov yivos. The factthat it is elsewhere called a yivos (Schneider's objection) does not matter, if thewords yivos and etSos are not being used technically. On the other hand,

1 Karsch, Kiilb, Tricot also take ftvai 509*8), but doubtful of o^is; cf. my papercopulative, but avwmnov as simply 'name- quoted above.less'. 3 5 H a i 4 - i 6 ; Meyer, p. 312.

2 This is true of KpoKoSeiXos (cf. 5o6a2o,

94 D. M. BALME

must be wrong on any interpretation.1 It is certainly a member ofTerpatroSa WOTOKCL, and is itself a yevos possessing yevrj.2 There is no sense inwhich it exemplifies TO ju.17 etvai yevos KTX., whether it be regarded as anindividual, a type, a group of types, or a member of a peyurrov yevos.

There is evidently an error here, and with it, I suggest, should be linkedanother difficulty in the passage, namely the position of dvdpwnos, 1. 28.Although his characteristics are taken as standard and compared with thoseof entire fieyiara yevrj, Aristotle nowhere else says that Man is himselfa neyidTov yevos; and if this term means not a 'top' group in a hierarchy buta 'most extensive' group in a horizontal division, he could hardly call Man one.In the only other discussion of fxeycara yhn\ (4gob7), dvOpwrros is twice men-tioned as an example of a type that belongs to no named group and to no yevosfj.eya, but is OLTTXOVV ethos and must be dealt with separately. The same is statedin a very similar passage at P.A. 1. 644b4: opvis Kal Ixdvs Kal ei TI dXXo earlvavcvvv/jiov fj.ev, TU> yevei 8' O/AOICOS TrepU)(ei ra ev airrat eiSrj- oaa Se firj roiavra,Ka6* eKaarov, olov wepl dvOpojTrov Kal el TI TOLOVTOV erepov eariv. Both passages(4gob7 and 644b4) distinguish two cases: (i) unnamed groups which comprisesub-groups, (ii) (named) groups which do not comprise sub-groups, e.g.Man. But 5O5b3O telescopes the statement and produces only the second of thetwo cases. If the text is genuine, it shows a threefold misunderstanding: (r) ithas confounded two distinct cases; (2) it wrongly puts dvdpayjros as a ix.eyt.aTovyevos; (3) the example KpoKoSeiXos is false (if not 6<f>is too). The whole passagemay simply be an interpolation by an inferior student: its removal, in fact,would not only not upset the argument but would even improve the sequence.

Alternatively, it must be corrupt. The right place for dvdpanros is after olov,whence 6<j>is Kal KpoKoSeiAos should be removed. Possibly dvdpamos in itspresent position has displaced an original TO WOTOKO. ; this loss would then havenecessitated the addition of the words en 8e Kal ra woroKa TU>V rerpaTroScov tomake up the missing fieyto-rov yevos; these words are not read in Da, whichmay therefore be preserving an older tradition.

If the passage is so emended, it still only states one of the two cases; but thisone is now correct, and agrees with 49Ob7. It would seem that in both passagesyevos and elBos are used non-technically (or confusedly), and it is worth notingthat the similar passage at P.A. 1. 644b4 occurs in a context where the technicaldistinction has not been successfully preserved: cf. 644ai7 oaa 8ia(f>epei. TG>V

yevwv Kad' virepoxrjv . . . ravra VTre^evKTai evl yevei. In all three passages thegenus is only to be found at the level of Bird and Fish, the peyiora yivrj,if at all.

H.A. 4 . 5 2 3 b I 2 TO rcov ivroficov (yevos) o TroXXa Trepi.elXr](/>e eiSyj t,a>o}v.

Note t,wu>v: not, therefore, 'species of the genus Insects', as every translatorsince Camus has it, but 'forms of animals'. In the context, yevos is used ofva r ious g r o u p s : 523 b 2 yevr\ -nXeloi (dvaCfuov ^a>aiv), . . . b l I TO TWV ico^Aitov yevos

. . . b ig yevos olov [ivppTjKes. The only insects named elsewhere as eioWj arecertain forms of Terriyes and <f>9etpes, on which see above (532bi4, 557a24).The distinction here is simply that between 'kind' and 'form'.

H.A. 4. 53Ib2I TO yevos (evrofuvv) iroXXa e)(ov eibrj ev avrq>.

61817: s c- ^VW VJ a s a t523 b i2 (above). Only evia of them are avyyeviKa, ibid. 22.1 'Num 3<f>is genuina lectio sit, incertus censeo', Dittmeyer.

haereo; . . . at KpoxoStiXos certe mutandum 2 Cf. 5o6a2O, 508*5, 5OO,b8, 558" 15, 609*1.

rENOZ AND EIAOS IN ARISTOTLE'S BIOLOGY 95

H.A. 5. 539a27 ev Se TW ixBvtjv yevei evta yiveTai ovre dppeva ovre OrjXea, TWyevei jiiev ovra eTepois TCOV lyBxxav ra avrd, ra> eiSei 8' erepa, evia Se Kal

181a.

Evidently the technical distinction. The same information is given withmore detail at 569/17, 57Oa3, G.A. 74ibi, 762b22, from which it appears thatthe fish here referred to as TW yevei p.ev TO. avrd, TCO el'Sei S' ere/xi is one of theKeen-pet?, while the -napmav iSia are the eyxeXvs. But the distinction betweengenus and species in KeaTpevs occurs only here; no use is made of it elsewhere,and the fish is quoted only as yivos TI Kearpiwv or evioi rS>v KeoTpewv. Thisstatement (from 1. 28 T<3 ye'vei p.ev to 30 181a) is the only one in this chapter thatis not paralleled in G.A. (apart from certain introductory matter and re-ferences). Its removal would not affect the sequence of thought. It could there-fore be a later addition.

G.A. 5 . 784b2I TO p,ev yap TCO yevei TO Se TCO el'Sei TCIVTOV ecrnv, rj juev irdxvq TCO

yeVei (aTjiiJs yap afjxf>co), 6 Se evpehs TCO et'Sei {arjifjis yap a[icf>co).

The technical usage of yeVos-eiSo? is invoked to justify calling grey hair bothevpebs and ndxvrj. But this has already been justified above by another logicalmetaphor: 784^16 coenrfp yap avTearpa.iip.evov TTJ irdxyrj 6 evpeos ecmv av fikvyap Trayfj ij dviovaa <XT//.IV, TTIX VIJ yiVerai, edv Se aairfj, evpeos. Sio S^ eTrnroXrjsearlv dp.(j>co . . ., i.e. the process can go in opposite directions but still producethe same effect. (Cf. Phys. 2O7a23 where TrpoaQecns is dvTecrrpap.fi.evrj to xadalpeois,which is explained by 2o6b3 f. TO Se Kara irpocrdecnv TO avro eaTi nws xal TO Karahialpeuw KTX., i.e. an equivalent result is obtained either by indefinite divisionor by indefinite addition.) The use of this device is satisfactory and pointedhere. But the further explanation in terms of genus and species imports aquite different metaphor, which has no biological significance here and,although it is correctly used, is otiose. It is the only appearance of the technicalusage of yeVos-eiSo? in G.A.1 Its removal would not disturb the sequence. It maywell be a later addition.

Long. Vit. 465 a2 eo-ri S' e^ovra3 -rr)v hia<j>opdv Tav-rrjv oXa r e 77730? oXa yevrj, Kal TLOV v<j>'

4 ev efSo? eTepa Trpos erepa. Xeyco Se KaTa. yevos p-ev Sia-

5 <f>epeiv olov dvOpconov Trpos hrnov (pM.KpofiitoTepov ydp TO

6 T(bv dvOpcoTTCov yevos r) TO TCOV LTTTTCOV), Kar ' e'So? Se dv-

7 dpcoTrov npos dvBpcoTTOv elal ydp Kal dvdptoTTOi oi fj.ev [ia-

8 Kpofiiot. ol Se /3pa)(vf3ioi eTepoi Ka9* eTepovs TOTTOVS Ste<7T<3-

9 Te?- Ta jxev ydp ev TOIS 6epp.ois TCOV edviov paKpoflicoTepa,

10 Ta S' ev Tots ^ivypois jSpa^u/3tcoTe/JO. Kal T&V T6V axnov11 Se Totrov OLKOVVTCOV 8ia<f>epovcnv ofjuolcos Tives Tavrqv npos12 dXXrjXovs TTJV 8t,a<f>opdv.

Ross, ad loc, discusses the difficulty, and concludes that yeVos and eiSo? areused without distinction in a2~4, but in the technical sense in 4-6. The latteris then unique in treating dvdpwiros as a genus whose species are the eOvrj1 (forRhet. I427b34 eiSrj dvdpdmuv is non-technical, 'sorts of men').

1 Interchangeable at 725b27 ev -tip yivu 564a5 yivos n\ interchangeable at 747b3O-T<3 axnCt TOIS 6fio€tSeatt 74-6a3O Tots ofnoyevdaiv 74^a 15.. . . a&m<j>6poi.s TU> ciS«; 746b3 UpaKes Sia- 2 Bonitz, Index 218*41 'non videtur con-<j>epovres <?S«, cf. H.A. 62Oa22 yevr/ UpaKwv, cinere'.

96 D. M. BALME

Theiler1 quotes this interpretation together with P.A. 645b23 f. to support hisview of de An. 4<52b2 el fii] d/xoeiSijs (17 ^fx7?)) "^orepov etSei 8ta<f>epovaa T} yevet;where he regards the specific (CL'SEI) differences as being those between racesof men. But the immediately following lines 4O2b3-8 show that Aristotle iscriticizing those who deal only with the human soul, and is asking whether thesoul is the same for all £<Sa (e.g. as distinct from <f>vra) or whether it differs asbetween avQpoynos, ITTITOS, KVWV. The yevos is £<pov, the 61817 are avdpamos,117770?, KTX. Nor does P.A. 645^3 f. support his view, for it distinguishes Birdas a yevos from M a n as an eiSos o Kara TOV KaOoXov Xoyov fj.~q8ep.lav e%ei Sia-<j>opav. It seems therefore that de An. 4O2b2 does not agree with the aboveinterpretation of Long. Vit. 465a4-6.

It may also be doubted whether the reference to e6vt] at 465ag is sufficient togive them the status of species upon which this whole passage would depend.Elsewhere Aristotle uses TO. eOvq mostly to denote foreigners and especiallythose foreigners who live in primitive conditions and not in a 770A1S (cf. G.A.775a33; Pol. i252b2o, i26ia28, i276a2g, i324bio; Meteor. 350a34, 3 5 i b n ;compare 'tribesmen' in popular English). The nearest parallel to the presentpassage is probably Pol. I327b23 f.: ra p.ev yap ev rots tftvxpols TOTTOLS edvr) . . .TO. 8e Trepl TT]V M.aCav . . . TO 8e TS>V ' EXXrpxov yevos . . . ra TWV 'EXXTJVOJV edvr]; butthis cannot mean 'species', for if it did it would have to imply generic-specificgroupings of the Greek yevos with its member-species and the barbarians withtheirs. The context, however, shows that no technical sense is intended:1329/20 TO pdvavcrov . . . ovb" aXXo yevos, . . ., a2J TO TCOV iepecov yevos . . ., a4IhirfpfjoOai Kara, yevr) TTJV TTOXIV, TO Te (xdxniov Kal TO yewpyovv.2 Moreover,465aio—12 adds (as a further example of avdpumov trpos avdpamov 8ia<f>epetv):Kal Ttuv TOV avrov 8e TOTTOV OIKOVVTWV . . ., and these presumably belong to oneedvos.

Lines 7-12 are therefore quoting examples of individuals, not of species.Hence if Kara yevos and KOT' eiSos bear their usual meaning 'in respect ofyevos, eZ8os' (as seen at H.A. 486a24, P.A. 645b22-25),3 then 465a4~7 fromXeyoi 8e Kara yevos to avdpaynov npos avdpoynov is wrong on two counts: (I) itdoes not fit the examples that follow; (2) it conflicts with Aristotle's theory ofdvdpamos. Once again, these are lines that could be removed without affectingthe argument, which would connect smoothly from 1. 4 77poy erepa to 1. 7 elalyap ..., and would be making the point that is developed at 466bi6 f. Lines 4-7may therefore be an addition by somebody who misunderstood the usage.

The technical distinction between genus and species appears obligatory,therefore, in only seven of the above passages, while in two more it would seemthe natural meaning were it not that this involves difficulties. These passagesare: H.A. 486ai6-b2i, 488b30~32, (? 4gob7-49ia4), 49iai8-ig, 497b9-i2,(? 5O5b2&-32), 539a28~3O, G.A. 784b2i-23, Long. Vit. 465a4~7. The last twoappear to be inferior interpolations, and are in any case insignificant. H.A.539a28~3O is also suspicious, and the distinction it suggests is never referred toagain. These three passages may be disregarded, but the others require explana-tion. They are really concerned with two points: the levels of comparison, andthe definition of fieyiara yevr]. They share certain features, in which they

1 Arist. u. d. Seele (Grumach series), (Ber- such misstatements might cause an unwarylin, 1959), p. 89. reader to take tear' efSos as 'within a

2 The pygmies are a yevos at H.A. 597a8. species'.3 But see above on 488b3O and 49iai8:

TENOS AND EIAOZ IN ARISTOTLE'S BIOLOGY 97

disagree with the rest of the biological works but agree with P.A. 1 and tosome extent with the logical works:

1. They draw an absolute technical distinction between yevos and eiSo?which is neither absolutely nor relatively used elsewhere in the biology.

2. They distinguish accordingly two levels of comparison (dvaXoyua betweenyevrj, vnepoxq between EI'STJ) which, though much used in practice, are not sodistinguished.

3. They apply the genus—species distinction only to the p,eyiara yevrj. (Anexception to this would be 53ga28 if genuine.) This is a simplification whichdoes not correspond to Aristotle's practice either in biology or in logic; else-where it is found only in P.A. 1 and in a few logical passages. When the H.A.passages discuss the possibility of intermediate groups, the verbal distinctionbetween yevos and eiSos becomes blurred (4gobi6, 5O5b3i). They state thestandards of comparison wrongly (488b3o, 491 ai 8), and confuse the technicaland non-technical uses (497b9). So the degree of classification is more elemen-tary, and the terminology less precise, than in the logical works.

4. The position of dvOpconos. Throughout the logical works Man is a stockexample of species, whose genus is t,wov; often he is accompanied by 177770s orKVWV, as though they were all species on a level together.1 But only here in thebiological works is this the case, namely at 486ai7-i9, 4gobi8, 33, and bypossible emendation at 5O5b3i.2 Elsewhere he is always a yevos;3 nor is hetreated as one with 177770? and KVCOV, but they are two among many members ofthe yevos TeTpanoSeuv £O)OT6KO>V while Man is a type on his own. This is anenhancing of his status; for in the logic and metaphysics, where there is atension between the superior substantiality of the concrete individual and thesuperior knowability of the abstract, and where we are often reminded thatgenus is dnXdis yvajpi^dtrepov, Man is but one (even if a special one) of thespecies of the genus t,u>ov. In the simpler outlook of the biology, where theactualized form is more important than general groups, Man's higher positionis perhaps characteristic of Aristotle's more developed view of <jn>xQ a n d ofnature's teleology.

The effect of these passages is to provide a bridge between the biology andthe logic. Yet their treatment of genus and species is less elaborate than thehierarchy envisaged in the logic, and may therefore be more primitive, whileon the other hand such biology as there is in the logic (e.g. the position of Man)looks more primitive than the view in the biological works. If that were all, theexplanation might be simply that Aristotle took his logic farther in the logicalworks, and his biology farther in the biological works, without any too carefulaccommodation of the two. But this explanation, which may yet be right, doesnot answer the difficulty that these quasi-logical passages in the biology are atvariance with the rest of the biological works: their genus-species doctrinewith its theory of comparisons is never carried out. What then is their pro-venance ? And what happened to the theory of classification ?

To attempt an answer in terms of Aristotle's development would require anassessment of many more factors than this. But any answer will have to take

1 Cf. Top. :O3ai4, io8a i5, I33b2, An. Pr. did not think that interbreeding was limited26a8, 28a32, Met. ioi6a27, iO38bi8, iO58a4, to specifically identical animals, cf. 746*29 f.al.; P.A. 1. 639a25. 3 Cf. H.A. 584b29, P.A. 655bi5, 656a7, 16,

2 At G.A. 73Ob35 eiSei means 'kind' in 676b33, G.A. 73ib35, 767b33, 777b5.general, not species as opposed to genus: A.

4599.1 H

g8 D. M. BALME

account of Theophrastus, whose practical use of yivos and etSos seems no moretechnical than Aristotle's.1 This suggests that Aristotle's Peripatos got nofarther in developing genus-species classification in biology than is to be seenin these passages. If that is so, it would seem that, contrary to what is oftensaid, Aristotle must have developed the technical distinction from logic and notfrom biology.2 He must have intended to apply it to biology, for it is incrediblethat he should have abandoned systematics in the very field where it hasproved most fruitful, and proved so for the very reasons that he himself pre-dicted in An. Post. B 13-14. In that case his biological work is incomplete. Thisis indeed self-evident: what is missing is a straight descriptive zoology, togetherwith the classification system that a descriptive zoology needs if only fororderliness. The H.A. does not primarily describe animals but the likenessesand differences between them; as such, it must be a preliminary study. If oneasks to what it is the preliminary, a likely answer will be that an analysis ofdifferentiae is necessary to any descriptive zoology, but it is necessary above allto systematics.

The passages that use the genus-species distinction all occur in introductorysections of H.A. The general introduction extends from the beginning to 491 "26and thus includes the first three (? four) passages. 497b9 comes in the introduc-tion to the external parts of animals, while 5O5b26 is in the introduction to theinternal parts of animals (and 539 a28, if it has to be considered at all, comes inthe chapter introducing the whole section on yeveaeis, books 5-7). Introduc-tions are naturally written last, and it may be that in preparing the treatise forschool use Aristotle—or an editor—wished to give it its proper systematicbasis, connecting it with the doctrines of logic. He may have written thesepassages then, but perhaps more probably he incorporated older notes: thismight account for their somewhat elementary and doctrinaire character.However that may be, they seem to represent an intention that was neverfulfilled.

Queen Mary College, London D. M. BALME

1 Cf. Schneider ad Theophr. H.P. 6. 1.2. Logique el mithode chez A., p . 72, A. philosophy2 So I venture to differ from le Blond, de la vie, p. 59, note 3.