Grammatical Observations on Euripides' Bacchae (Amsterdam Studies in Greek Philology, Vol 1) 9050630413, 9789050630412

Gieben, Amsterdam, 1991. 217 S., goldgeprgtes Ln. Amsterdam Studies in Greek Philology; Band 1/Herausgegeben von Irene J

134 98 20MB

English Pages [228] Year 1991

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
GRAMMATICAL OBSERVATIONS ON EURIPIDES' BACCHAE
Preface
Table of contents
Grammatical observations on the Bacchae
Appendix 1. Differences with Murray's text
Appendix 2. The speeches of Pentheus and Teiresias, 215-262 and 266-327
Appendix 3. The constructions of oὐ μή
Appendix 4. Forms in -σω in questions: future indicative or aorist subjunctive?
Bibliography
Index locorum
Index graecus
Index rerum
Recommend Papers

Grammatical Observations on Euripides' Bacchae (Amsterdam Studies in Greek Philology, Vol 1)
 9050630413, 9789050630412

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

GRAMMATICAL OBSERVATIONS ON EURIPIDES' BACCHAE

AMSTERDAM STUDIES IN GREEK PHILOLOGY EDITORS

IRENE J.F. DE JONG & ALBERT RIJKSBARON VOLUME ONE

ALBERT RIJKSBARON

GRAMMATICAL OBSERVATIONS ON EURIPIDES' BACCHAE

J.C. GIEBEN, PUBLISHER AMSTERDAM 1991

CIP-DATA KONINKLUKE BIBLIOTHEEK, DEN HAAG Rijksbaron, A. Grammatical observations on Euripides' Bacchae / A. Rijksbaron. - Amsterdam : Gieben ISBN 90-5063-041-3 SISO 804 UDC 801.5:87 Subject heading: Euripides "Bacchae". Cover design: Ester Wouthuysen

Voor Caroline

Preface 'No other play of Euripides has been so much discussed as the Bacchae; very few have been the subject of such exact and careful study on the linguistic side'. Thus opens the preface to the first edition of Dodds' commentary. One might subscribe to these words nowadays even more readily than at their original date of publication (1944), if only because Dodds himself has added considerably to our understanding of the play. And yet, as I hope to show on the following pages, the linguistic side is perhaps due for a reappraisal. I am not so much thinking here of applying the latest insights of general and Greek linguistics (although naturally these may very well have a bearing on the interpretation of the text), but rather of making use of the impressive grammatical apparatus which is at the disposal of classical philologists, but whose value is not always fully acknowledged. I have selected for discussion notably (a) passages where scholarly opinions vary, (b) passages where one individual commentary seems wrong, and (c) a number of passages which are not treated by any commentary - unjustly, to my mind. I have included some observations that are not strictly grammatical. Two problems of a more technical nature are discussed in an appendix. Jargon has been avoided as much as possible, but perhaps I should note that as a general rule I use the neutral term 'state of affairs' instead of 'action' for the lexical meaning of a given verb. The book originates from two seminars devoted to grammatical problems in the Bacchae. I wish to thank the students who participated in these seminars for their critical cooperation. Special thanks are due to Sjoerd van den Berg, Amber Chabot and Emilie van Opstall, who contributed papers on the semantics of the middle, the constructions offi and~. and the uses of TlKTE"L, respectively. Amber Chabot and Emilie van Opstall also provided invaluable assistance in checking the references. I am indebted, moreover, to the members of the 'Amsterdamse Hellenistenclub', to Dr J. Ysebaert, to Hotze Mulder, and above all to Marietje van Erp Taalman Kip and Kees Ruijgh, who were

vii

true collegae, in the sense of 'co-readers', if I may be allowed this Gelegenheitsetymologie. Finally, the book could not have been produced without the editorial and word-processing skills of Noor Kegel.

Amsterdam, January 1991

viii

Table of contents Preface

vii

Grammatical observations on the Bacchae Appendix 1. Differences with Murray's text Appendix 2. The speeches of Pentheus and Teiresias, 215-262 and 266-327 Appendix 3. The constructions of ou

µ11

1

163 165 167

Appendix 4. Forms in -0w in questions: future indicative or aorist subjunctive?

175

Bibliography

191

Index locorum

199

Index graecus

212

Index rerum

215

ix

A note on texts etc. References are to the plays and fragments of Euripides, unless indicated otherwise. Abbreviations are in accordance with LSJ; note, however, that Heracleidae = Held., Phoenissae = Pho. and Troades = Tro.

The following text-editions have been used (only authors referred to regularly are included): Homer: Pindar: Aeschylus: Sophocles: Euripides:

Herodotus: Aristophanes: Thucydides: Xenophon:

X

D.B. Monro & T.W. Allen (OCT) B. Snell & H. Maehler (Teubner) D. Page (OCT) A.C. Pearson (OCT) for Cy., Ale., Med., Held., Hipp., Andr., Hee., Supp., El., HF, Tro., IT, Ion: J. Diggle's edition (OCT); for Hel., Pho., Or., Ba., IA, Rh.: G. Murray's edition (OCT) C. Hude (OCT) V. Coulon & H. van Daele (Bude) H. Stuart Jones & J.E. Powell (OCT) E.C. Marchant (OCT)

2 TllCTfl R.:* 'Present historique, frequent avec les verbes exprimant la descendance, tels que TLICTE"LV, yevvav'. L.: 'Present avec valeur de passe proche, precise ici par TTo8'. Cf. 11, 42, 44'. V. 2: 'The historic present is used with or, as in this case, without expressive force (praesens annalisticum): cf. Schw. II, 272. R. and L. fail to distinguish this use from the perfective present at 42 TlKTE"L '.

I do not think that these explanations are satisfactory (Lacroix' 'valeur de passe proche' does not exist at all); more in general this use ofTlKTw and related verbs in drama would seem to be ill-understood. I will therefore discuss it in some detail. One of the main functions of the historic present is to signal that the state of affairs concerned is of special importance in a given context or, more precisely, in a given narrative context, i.e. in a report of past events. Often the historic present marks a turning-point in a series of events; it 'hebt entscheidende und neue Momente der Ereignisse heraus', in S-D's words(: 271). Now verbs denoting birth, marriage and death, too, may have this function: in fact, they show a marked tendency to occur in the present. This can readily be understood, since such verbs refer to events that by themselves are important in a human being's life. Herodotus has many illustrative examples of this use; see e.g. 1.111.1 (icaTa 6alµova TlKTn, sc. the child that will be replaced by the future Cyrus), 5.41.1 (: Cleomenes' birth, a much longed-for event). 1 An example from Euripides is Pho. 55 (TlKTw). Apart from having this 'highlighting' function in a full-fledged narrative, i.e. in alternation with imperfects and aorists, the historic present is also used, again within the overall framework of a narrative, in rapid succession with other historic presents, i.e. without being interchanged with imperfects and aorists. In this way e.g. the main events of the life of some person who is not particularly important for the main story may be summarized. Good examples are Hel. 6-8 (yaµe-t, nKTEL), Hdt. 6.71



For the abbreviations see the bibliography. 1. The aorist, the most obvious competitor of the historic present, is used when the birth involved has no special importance in that particular context (e.g. 1.84.3, in a relative clause) or when the event of the birth has already been mentioned (e.g. 5.41.3, where ws hEKEV refers back to TlKTouaav in the previous sentence, in itself an important event). Cf. also below, p. 4.

(with verbs expressing 'life-events'), Th. 1.136.1 - 138.2 (the end of Themistokles, which does not, of course, belong to the main story),2 and X. An. 1. 1.10 (a short digression about a certain Aristippos, who will play no further role). It is this use for which the term praesens annalisticum (or tabulare) should be reserved. Obviously, it is much less expressive than that discussed in the previous paragraph, since these historic presents are not followed by a story illustrating their significance. Although, then, the historic presents discussed above differ contextually, they have one important feature in common: both types occur in narrative discourse. This is not so in the case of the presents to be discussed presently, which are sometimes called 'perfective' presents. T(KTE"L at Ba. 2 is a case in point. It is discussed by K-G, along with a number of similar verbs, on p. 137 of vol. I. Some other examples given by them are: S. OT 118: (OL. ov8' liyyE>.IY:; TLS' ... / KaTE'i&- ... ; /)

Kp. OT 437:

Ion 1560:

8v(]crKOVO'L yap, TTXT)V EtS' TLS', KT)..

ws

(TE. T)µE'iS' TOLot8' l4>vµEv, µEv crol 6oicE'i, µwpOL, yoVEoot. 8', oC. cr' l4>ooav, lµcf>povEs-.l OL. TTOLOLO'L; µE'ivov. TlS' BE: µ' !iccf>l'EL 13po'rwv; ft& nKTfl cr' !~ 'ATT6>.>.JAVOS' TTaTp6s-, 8l&xn. 8' cis- lBwKEV, ou cf>luacr( crE,

ws

These verb forms are presented by K-G (: 1, 135) under the general heading: 'Das Prasens bezeichnet Handlungen, die zwar der Vergangenheit angehoren, aber in ihren Wirkungen noch im Augenblicke des Sprechens fortdauem'. The latter modifier ('im Augenblicke des Sprechens') is essential: the presents involved occur typically in expository discourse, e.g.prologues (: Ba. 2) and epilogues (: Ion 1560), and in dialogue(: S. OT 118 and 437), and denote, in fact, some state of affairs that was carried out in the past but is still relevant at the moment of speech.3 By this feature these presents differ from the narrative historic presents discussed above, and it is clear, I think, that they can hardly be treated on a par. To mention only one important difference: the use of TT0TE with a 'perfective' present, as at Ba. 2, is wholly excluded in the 2. Some good remarks on this passage can be found in Eriksson (1943: 9-10). Incidentally, the whole Einleitung of Eriksson's monograph gives a useful summary of the various approaches of the historic present in Greek. 3. At IA 1164 we find a similar use of TlKTw in an expository monologue.

2

case of a real historic present. 4 On the other hand, on a more general semantic level the 'highlighting' historic present of narrative discourse and the perfective present of expository discourse and dialogue are, I think, related: in both cases the present gives prominence to states of affairs that belong to the past, either because they are important for subsequent historical events(: historic present sensu stricto), or because they are important for a certain person in the particular situation in which he finds himself. Note especially, in this connection, the use, at OT 437, of lKc/>vn by Oidipous, as opposed to the neutral lc/>uaav of Teiresias. T(KTn at Ba. 2 would seem to be no exception to this: the fact that Semele is his mother is particularly important for Dionysos in the context of his arrival at Thebes.5 Summing up: although T(KTfl at Ba. 2 is semantically related to real historic presents, i.e. presents in narrative discourse, it has a semantic value of its own: it indicates that a past state of affairs is still relevant at the moment of utterance; accordingly, it should not be called a historic present. 6 I think, finally, that all other presents in the prologue of the Bacchae exhibit the same semantic features, viz. T(&JicrL (11), T(KTfl (42), 8l8waL (44).7 4. Some other examples of presents with 1TOTE are given by Stahl p. 92 (who calls them 'historic', wrongly): El. 416: CliiVT' daaKo6aas- 1rat8' 8v lKa~CEL 1TOTl,HF 252:@ yiis- >.oxe6µa8', oOs- "Apns- atrdpEL 1TOTt. LSJ S.V.1TOTE mention S. OT 715: Kal Tov µtv ... ~lvm troTl I ... cf,ove6ouaw. These presents are no more historic presents than those modified by mi>.aL, which indicate that a slate of affairs which srarted in the past is still going on at the moment of speech; K-G I, 134. 5. Observe that the translation 'is mother/father or, which is often suggested for TlKTEL, is blocked here by the presence of 1TOTE. 6. As for the term 'perfective present' for such presents, this is an appropriate term, since the perfect typically expresses 'present relevance'. Compare, for instance, 8V!jaKouaL at S. OT 118, quoted above, with TE8vdaL at S. Ant. 1173. This variation raises the question of which factors determine the choice of one rather than the other form. In some cases metrical factors may be involved. Thus, the perfect TlToKa could not be used in stichic parts; in fact, it is avoided altogether by the tragedians (but not by Aristophanes). In other cases we probably have to allow for stylistic variation, see also next note. 7. The same use of 8(8waL occurs at El. 34, and at S. OT 1173 (not a historic present (Kamerbeek)). -Apart from yevvtii, TlKTw etc., K-G discuss some other verbs and verb classes under the heading 'Handlungen die ... in ihren Wirkungen noch im Augenblicke des Sprechens fortdauem', notably perception verbs like ala6dvoµaL, 1ruv8dvoµm, dKo6w, certain verbs of movement, e.g. -f!Kw, otxoµaL, and the verbs dBLKlii, cf,e6yw, VLKlii, KpaTlii, l'iTTliiµaL. It must be asked, of course, whether the

3

To conclude I add a few observations on the uses of lnKTE" and he-KE". The aorist would seem to have a neutral meaning, and is mostly found in what may be called a genealogical use: it simply states, or inquires after, somebody's descent, e.g. Ba. 988, IA 198, Ion 340, 344, 989. "EnKTE", on the other hand, often occurs in situations of dramatic tension, where it would seem to describe the bearing as going on in the past, thereby probably emphasizing the emotional involvement of the speaker in the childbirth concerned, e.g. El. 1183-4: 8La trupbs lµo>.ov a TQMlLVa µaTpl TQ.8 d µ' hLKTE" ICOUpav, 1212,cf.alsoHF2801: lyw L>.tii µtv TEKva· TTWS' yap OU LAW / liTLKTOV, aµ6x0JJcra;, Med. 930: (Iason) Tl 8fjTa Al.av To1a8' lmaTEVE"LS TEKVOLS; (Medea) lTLKTov auTous, 1247 ... µT)8' d.vaµVT)a8ijs TlKvwv, I ... ws lTLKTE"S (Medea, apostrophizing her hand when she is about to kill her children). 8 In other contexts it has rather an iterative meaning (e.g. IT 274). Finally, the imperfect, too, is used 'genealogically' (cf. above), e.g.: Pho. 289, He/. 1645 (with TTOTE"), HF 3 (with TTOTE"). Cf. also Hom. Od. 16.118: µoiivov AalpTT)v 'ApKE"laLos ulov lTLKTE", µoiivov 8' aliT' '08oofia TTaTTIP TEKE"V. Possibly the distinction between aor. and impf. is that 'whereas the aor. states the naked fact of the bearing (or begetting) the impf. dwells also on the continuing result: I bore them (and am their mother.)' (Barrett on Hipp. 419-21).9 Cf., however, also K-G 1, 144, who suggest that metrical factors may be involved. Observe, in this connection, that Hdt. does not use hLKTE" in this way. 1 ,

/

'relevance' value postulated for TlKTEL is present here as well. Generally speaking I think it is, although in many cases the use of the present appears to have become idiomaticized or even fixed. A clear case in point is t\Kw, which virtually has a lexical, and thus not a contextually detennined, perfective meaning. A kind of idiomaticization appears to have been operative in the case of the many present participles with 'perfective' meaning, e.g. l') TlKTouaa, b 1rpo8L8ovs 'traitor', b d,ro· ~6pCf 1rup( D.: 'd.. may be either passive in sense, 'lightning-borne' (from d.aTpalTl14>POS), or active, 'lightning-carrying' (from d.aTpa1TT14>6pos)'. R.: 'd.aTpa'ITT14>opos (Bruhn) meilleur que d.aTpa1TT14>6pos ... (Wecklein, Dalmeyda)'. So too Lacroix. V.2: 'The translation 'carrying lightning' is suggested by fr. 312 d.aTpa1TTl4>J>E"t, but 'carried by lightning' (R., L.) seems to be more probable on account of S. OR 199-2001rupq>6pwv d.aTpa1rdv'.

'AaTpa'ITTl4>P0S belongs to the large group of adjectives that consist of a substantival first member and a verbal second member; these adjectives may have either active or passive meaning and are accentuated as paroxytonon (: active) or as proparoxytonon (: passive), a difference which is lost when the final syllable has a long vowel or a diphthong, as in most inflected forms. Cf. Herodianus, Gramm. Graeci 1, 234, who mentions e.g. AL8ol36AQS (throwing stones) as against A1.86!3o>..os (hit by stones), Schwyzer (1939: 429); Kuhner-Blass 1, 329; 526, and especially Vendryes (1945: § 249). On the basis of the adjectives in -4>opos given by Kretschmer (1944) our conclusion should be that we are dealing here with the dative of active d.aTpa1TT14>6pos: the great majority of these adjectives have active meaning (in fact, Kretschmer only has d.aTpa1TT14>6pos), the only exceptions being, if we can trust Kretschmer, 1rap8a>..1'4>opos in S. Fr. 11 Radt (= 10 Nauck2), and probably A. Ag. 1150, although we read there 8rn4>6pous, which presents, of course, the same problem as d.crTpamr q>6pl¼). 10 Statistically, then, d.aTpa1TT14>6P£!l has a greater chance of having active meaning. On the other hand, however, the sense '(fire) carried by lightning' would seem to be more in accordance with our (and the Greeks') views of the functioning of lightning, and this is borne out, indeed, by a number of passages dealing with lightning. Thus we find (in the nominative and accusative, where possible differences in accent are not blurred): Pi. N. 10.71: ZEvs 8' hr' ·18Q. 1ruP4>6pov 1T~f 1/JOM>fVTa KEpauv6v A. Th. 444: TOV 1rup4>6pov . . . KEpauv6v

S. OC 1658: 1ruP4>6pos ... KEpauv6sll 10. For the same reason S. OT 199-200, adduced by Verdenius, is inconclusive. 11. That the 'active' accentuation is justified appears from cases like Pi. Dith. II (= fr. 70 b) 15: b .. . KEpauvos dµ1rvtwv I 'lriip, A. Pr. 1084: l>,LKES .. . anpo,rijs

5

It is clear, then, that lightning is viewed as the carrier of fire; consequently, the fire in Ba. 3 should be taken as having been carried by lightning.

9 d8dvaTov ••. Ol3pLV D.: 'in apposition to the sentence'. R.: 'apposition a 4>Mya'. V.2: 'Not an apposition to 4>Mya (R.) or to the whole sentence (D.), but rather to the internal object of Ciiiaav which has been left unexpressed'. In my opinion Dodds' explanation is the right one: when action nouns like tJl3pLS occur in apposition to a sentence (or, as here, to an embedded sentence, viz. the participle 12 dependent on opw) they summarize the whole of that sentence, i.e. the predicate and its arguments (subject/object) and modifiers, adding some further qualification to the action expressed by that sentence.1 3 I fail to see what advantage is gained by Verdenius' approach (who follows Wilamowitz on HF 59), which connects the apposition with a postulated internal object only. Verdenius, and Wilamowitz, do not tell us why they view the appositional accusatives in this way; possibly, to expiain the very fact that the apposition appears in the accusative. 14 As to this latter phenomenon, this can perhaps best be accounted for by analyzing the apposition as a kind of extended internal object, especially in those cases where the apposition involves an action noun. 15 A clear case in point is: Ca1rupOL, Ion 212: KEpauvbv dµcj,(,rupov, Tro. 80: ,rilp KEpavvLov. A further parallel may be found in Ar. Av. 1749: w b.Lbs clµl3p0Tov lyxos / ,rupcj,6pov. Verdenius' reference to E. Fr. 312 is misleading, since these lines are from Ar. Pax 722, where the subject is the Beetle, not ,rOp or the like. 12. I.e. the whole of lpdma Tucj,6µ£va ... cj,Mya; or, alternatively, if TE is maintained (for which cf. Verdenius ad loc.), both lpdma Tucj,6µ£va and Cwcrav cj,Mya. 13. S-D 617 correctly speak of 'Satzapposition'. See also K-G 1, 284-5, whose analyses, however, are questionable, especially their paraphrases of the appositions by consecutive clauses (wcrn ... Elvm), as they do e.g. in the case of Hom II. 24.735 (see below). Astyanax' being thrown from the wall does not result in a grievous death, it is a grievous death (i.e. way of dying). 14. Verdenius' reference to K-G 1, 284-5 wrongly suggests that they support his analysis. 15. A similar line of reasoning is followed by Barrett on Hipp. 752-57, although he is too much inclined to extend the use of the internal accusative at the expense of appositions. He notably tends to neglect the fact that internal accusatives in principle should be semantically related to the verb. See also Moorhouse 45f. and (1988: 214),

6

... fi TIS' 'Axmwv XflpS' EAWV a.no trvpyou, X.uypov .E8pov makes explicit the nature of Astyanax' death, just as it would in a sentence like 8avij X.uypov .6ya in a rather pregnant sense, since it is not, properly speaking, the result of Hera's action which should be qualified as vj3pLS', but rather that action itself (viz. her causing 2.eus to appear to Semele in his true form). Dodds draws a parallel with El. 1261 µfjvLv; compare also Ion 506, TTLKf>WV yaµwv Oj3pLv: the exposition of the child as food for the birds and the wild beasts is qualified as an outrage.1 8 although he, in turn, underestimates, I think, the similarities between apposition and internal accusative. 16. Indeed, the internal object interpretation is perhaps preferable, since dv6cnov 8fov is not an apposition in the strict sense. Note, too, the semantic similarity between da(&iµev and 8fov, a common feature of internal accusative constructions. S-D (617) cite HF 323 as a case of 'Beriihrung mit dem Akkusativ des lnhalts'. 17. That Verdenius' approach is wrong appears clearly from a sentence like Or. ll05: 'E>.lvr,v KTciVtilµEv, Mevl>,EI\I >.(11rriv mKpciv, where taking >.fi1TT]v mKpciv in apposition to an implied 'killing', to be taken from KTcivwµev, causes the apposition to lose all meaning. It is not the killing as such which is a MlTTJ mKpci for Menelaos, but the killing of Helen. 18. Verdenius calls O~pLV abstractum pro concreto, similarly Owen on Jon 506, needlessly. Cf. Chantraine (1933:112): 'Suffixe anime, !'element -i- s'est prete a former des noms d'action. Ces noms d'action, en particulier µijvLS' et O~pLS', expriment des notions traditionnelles et religieuses. Ils ont conserve neue leur valeur "active"'.

10 cll3aTOV

Verdenius2 rightly draws attention to what he, not quite correctly, calls the -TEOS' sense of dl3aTov: 'which may or can not be entered'. 19 Two factors seem to be involved in this 'axiological' meaning, as it perhaps may be called, of verbal adjectives in -Tos. 2 First, in many languages adjectival suffixes which express 'can be done', often have a secondary meaning 'deserves (not) to be done', 'must (not) be done', especially in adjectives with a moral meaning. Cf. Dutch: verwerpelijk ('which should (not: can) be rejected'), verachtelijk ('contemptible'), belachelijk ('laughable, ridiculous'), wenselijk ('to be wished for'), also onbespreekbaar ('which may not be discussed'), etc., English: despicable, unspeakable, see Quirk et al. (1972: 1006). 21 A similar meaning can be postulated for Greek evaluating adjectives in -TOS like: d mcrTos, du6~A~Tos, dpp~Tos, lnaLVET6s, 8avµacrT6s, KaTayEAacrT6s, µaKapLcrT6s (for the accent cf. Kilhner-Blass 1, 538 f.) Second (cf. also fn. 23), the verbal adjectives in -TEOS' are used exclusively as predicative constituents with ElµL, or, without ElµL, in nominal sentences, not, then, attributively, 22 or predicatively with other

°

19. Not quite correctly, because the adjectives in -Tlos, whether used personally or impersonally, typically occur in direct speech and have a concrete deontic meaning, i.e. they express that, in the situation at hand, some action has to be carried out. Adjectives in -TOS, on the other hand, occur in all types of speech and assign 'a property to the referent of the noun. For cl~aTos 'which may not be entered' cf. also Pl. La. 183 b, cl~aTov lEpov. 20. Verdenius notes that K-G and S-D do not recognize this use; add Chantraine (1933: 306 ff.). It is, however, recognized by Kuhner-Blass (2, 288f.) and by Schwyzer (1939: 50lf.). Compare also Stahl (762): 'Wenn die Mtlglichkeit eine moralische ist, so entsteht der Sinn des Wertseins' (762). Strangely enough, Stahl (761) holds that the adjectives in -TOS occur only as substantivized adjectives or as predicative elements. Possibly, his remark is meant to apply to the Homeric poems only, but even there attributive adjectives in -Tos- are found, cf. JI. 5.113 6LCi crrpEirToto XLTwvos, 21.31 l,rl crTpEirTotcrL XLTWO'L. For the range of meanings of adjectives in -TOS' see also Barrett on Hipp. 677-9. 21. The new Quirk et al. (1985: 1555 on -able) is less informative on this point. Ultimately, the axiological meaning can be considered a development of the 'can'meaning, inasmuch as it expresses a 'one only' possibility. Thus, verwerpelijk in een verwerpelijk voorstel ('a proposal which should be rejected') can be paraphrased by ('this proposal is such that) we can do only one thing: reject it'. 22. Cf. Schwyzer (1939: 810f.). Unlike Latin Greek did not develop an attributive use of the gerundival suffix.

8

verbs than e-lµL. Possibly some verbal adjectives in these missing functions of adjectives in -Tlos. 23

-TOS,

at least, fulfil

11 T£8T1C1L Above (ad 2) I argued that this should be taken as a past present with present relevance: 'is the maker of' .24 Observe, however, that whereas, by our knowledge of the world, TLKTEL unquestionably has past reference, this is not so with TLEhim. In principle, this could be an actual present: 'is busy making'.

1 7 'Aa(av TE ,rdaav, f\ ,rap' ciAµupclv 4>.a. D.: 'in the restricted sense of western Asia Minor, as the context shows'. R.: 'ici, la cote d'Asie Mineure'. This is correct, but depends on taking the relative clause as a restrictive relative clause. To bring this out the comma after mioav should be omitted. ii etc. has as its antecedent 'AaLav only, not TTdaav, which stands predicatively, i.e. 'I visited (that part of) Asia which lies along the sea completely'.

20 ls T,iv8E ,rpwTov ~X8ov 'EU,ivwv ,r6XLv, 21 KdKEt xopEvaas Kal KaTaaT,iaas lµds 22 TEMTds, tv' dT1v lµcf>all'l'js 8a(µwv ~poTots. Thus the text and order of LP, as well as of the Berlin papyrus (Brashear 1975; ::; IT2 in Kopff's app.). At 21 Wilamowitz conjectured TaKEL, which has been adopted by e.g. Murray. D.: 'with this text we must, I think, translate "I came to this city of Greeks first when (i.e. only when) I had set Asia dancing" (-) The common rendering of 20, "I came to this city first in Greece", (a) leaves the following participles as a weak tail dangling from the sentence, (b) makes 23 a repetition of 20 (-) but I now incline to prefer ... Pierson's remedy of putting 20 after 22'. 23. Observe also that an adjective d~aTloS' does not exist (the rule being that compound adjectives in -Tlos-, unlike those in -TOS', may only be formed from compound verbs, not, then, from other adjectives). But even if it did, it could not have been construed as a predicative adjective with Tl8rJcrt. 24. Whereas this use is quite common with TlKTEL and the other verbs of fn. 7, it is exceptional in the case of Tl&r]crt. In fact, it is passed over in silence by grammars and commentators alike. A parallel in Eur. is IT 34.

9

R. and L. keep the text and the order of LP, but make a fresh start at l. 21, taking KO.KE"L xope-foas etc. as a modifier of dvo>.w>..v~a; in 23 they read ye- (: Triclinius) instead of Bl. This approach was already taken by Schone. V. 2: 'It is misleading to say that this line(: 21) "introduit une idee nouvelle" (R.): it introduces the specification of what the coming of Dionysus both to Asia25 (16 hre->..8wv) and to Greece (20 n>..8ov) meant'. Verdenius rightly rejects all readings involving the replacement of Kci.Kd by Ta.Ke-t, as well as taking KO.KE"L xope-vcras as introducing a new idea, to be connected with dvo>.w>..v~a. 26 However, as far as n>..aov, at least, is concerned, his own 'specification' solution is impossible. First, the states of affairs referred to by xope-vcras and KaTaCJTTJcras, being aorist participles, precede that ofn>..8ov, and, second, Verdenius' proposal would mean that Dionysos' arrival in Greece has already had effect, which it has not. The traditional text and order are rightly defended by Kannicht (1962: 97); cf. also Erbse (1984: 92) and Oranje, p. 34, fn. 85. Although the participles refer to states of affairs that are chronologically prior to that of the main verb n>..8ov, they are postponed because their implication, viz. that Dionysos in Thebes, too, will act as he did in the other countries, meaningfully leads over to line 23ff. 27 The transposition of line 20 after 22 ruins this semantic link.

25-6

~>.os·

I l,rd ...

So most recent editors. Gregoire and Lacroix, however, while translating by 'car', have a comma after l3l>..os. This is incorrect; l,rd here= ycip, not only semantically but also syntactically, i.e. it functions as a near-coordinator; cf. Rijksbaron (1976: 79ff.). Punctuation should reflect this status. The l1rd-clause explains why Thebes was the first Greek city visited by Dionysos. 25. Rather 'the going over, the traversing' of Asia, cf. LSJ s.v. l1mµL (8), III. 26. Observe that Ka( in KCIKE1 is not assigned any function at all by Roux, neither in her commentary nor in her translation. 27. Some other examples of aorist participles which follow upon the verb, although they are chronologically prior to that verb, are HdL 4.91.1: a-r1'XTJv l.lyovTa Tei&, and S. OT 1047-9: lanv ns ... /

!Sans KciTm& Tov ~onjp' ... / dT' oliv hr' dypwv dTE Kdv8ci8' daLBwv;.

Cf. also Hdt. 1.14. l; Hel. 15, 47, 597-98; IA 15.

10

3 3 (,rapdico,roL)

♦pEvci>v

R.: 'genitiflocatif'. V.2: "'genitive of limitation" ("within the limits of', "restricted to the field of')'. K-G 1,401 give ,r. cf>. as an example of 'ablativischer Genetiv', together with l: °Mv8Epos, Ka8ap6s TLVOS' etc., so probably as 'bereft of'. K-G's analysis (which is also that of Wecklein) is, I think, the right one. ITapd.Kotrs should be taken as an adjective with passive meaning, and the genitive is used as in A. Ag. 1252: ~ KapTa . . . ,rapEK6m1s XPT'lcrµwv lµwv, i.e., in fact, as an ablatival genitive, so ,rapd.Kotrot cf>pEvwv = 'knocked off their senses'. Cf. also A. Ag. 479 cf>pEvwv KEKoµµlvos 'lit. "knocked from his wits" (genitive of separation)' (Denniston-Page; cf. also Fraenkel's note ad loc.). As for the 'genitive of limitation', Verdenius supports this with a reference to S-O 132; however, as appears from the relevant passage (section y), this is an idiomatic use, since it is confined to the genitive with lxw, ~Kw and related verbs plus adverb.2s

41-2 EEµl~11s TE µ11Tpi>s icT~ R. (in her translation): 'Je dois defendre la memoire de Semele, ma mere, en me manifestant aux yeux des mortels comme le dieu qu'elle a donne pour fils a Zeus'. Similarly Gregoire, Lacroix and Kirk, i.e. they make ci.1ro>.o'Y1icraa8a( µ' depend on 8Ei, on a par, by TE, with tr6Mv TI)V8' E"Kµa8Eiv. So too Wecklein ad locum (Dodds and Verdenius2 are silent on this point). This would seem doubtful Greek; especially the wholly unprepared transition from n;v tr6Mv to µE is awkward. Rather, TE indicates that the clause tr6Mv TI)V8' EKµa8Eiv ... «TEMOTOV ofoav is followed by a second predicate having tr6Mv for its subject, viz. either ci.1ro>..o'Y1icracr8m, on a par with e-Kµa8Eiv, or (µE) cf>avlvTa, on a par with a.TEMOTOV oliaav. Since the former is clearly excluded, the subject of ci.1ro>.o'Y1icracr8at being µE, the second option is preferable. There are, then, two things the city should become thoroughly acquainted with, first that it has, as yet, 28. The parallels given by Verdenius are of doubtful value. At Med. 649, olKTp6TaTov dxiwv, dxlwv is probably just a partitive genitive, like ,ra8lwv at 658 (&w6TaTa ,ra8lwv). As to cj,pevwv at A. Pr. 907 (au8d8TJS' cj,pevciiv; other mss. have au8d8TJ cj,povcilv), this is perhaps best taken, with K-G I, 371 Anm. 19, as a kind of objective genitive.

11

not been initiated, and second, that Dionysos has appeared at Thebes as a god to defend his mother. 29 'AiroXoy,iaaa8m is, then, an infinitive of purpose; for such infinitives after verbs of coming etc. cf. K-G 2, 16-7, Stahl 608-10, S-D 362-3, Moorhouse 237f., e.g. Th. 1.128.3: dq>LKvnTaL ES 'EUT)0'1TOVTOV • . • Ta ,rpos l3aaL>.fo ,rp.a.Lm, "bola portier'".

ev

These lines pose two problems, one of punctuation, and connected therewith, of the interpretation of lKTOTTOS', and one concerning Tls. On lKTOTTOS' see below. Roux is alone among the commentators in observing that T(s cannot just be a normal Tls-question, with Tls = who? For the latter is normally used in a context where someone is present, whose identity, however, has as yet not been established, cf. e.g. Ar. Ra. 38: Tls TT)V 8upav lmhaeEv;. The interpretation suggested by Roux is a possible way out, although strictly speaking we have not been told that the inhabitants of Thebes are present. But in the situation at hand their presence can perhaps be considered presupposed. Another solution was proposed- long ago by Bruhn (see Murray's app. crit. and Kopff's 'appendix altera'), viz. to take Tls, or rather Tls or TLS, as an indefinite pronoun. Although at first sight this seems rather improbable, there are, in fact, a number of examples of independent n(s) (i.e. 'someone' or 'something') in initial position. See Ktihner-Blass 1, 345, and cf. also Wackernagel (1953: 1220, original date of publication 1908): 'Im Griechischen ist dies (viz. indefinite pronouns becoming orthotonic AR) unstreitig der Fall', and Moorhouse 159. Unfortunately, the only certain instances come from Plato and Aristotle and these do not involve uses where ns can be confused with the interrogative pronoun, as here. Other cases cited in this connection are A. Ch. 654, Pi. P. 10.4 and S. Tr. 865, OT 1471. What is needed, it seems, to settle this question is an unequivocal positive answer, like >Jyns following >Jyw TL; (at S. OT

13

1474),32 after Tl (Tl) 11µL; which we find at S. Tr. 865 and OT 1471, as well. Since such cases would seem to be lacking, it is perhaps best not to regard cases like Ba. 68 as illustrations ofTls (ns) ('someone') in initial position. But the whole subject deserves a more thorough study. 6 9 l KTO'll'OS' l aT(I) D.: 'Those who are in the street (664>) are to make way for the procession (lKTOlToS fon>); all, including those indoors (µe-M8poLs) are "to hush their tongues to reverence"(-)' R.: 'comme dans Phaethon fr. 773 N2, v. 68-69,33 E"KT6TILol TE" 66µwv aTiadpe-TE", iii he- >.aot: "sortez de vos demeures et accourez! o peuples, venez!" (-). Les bacchantes adressent ici la meme invitation aux Thebains'. L.: '(les Thebains) sont (ou se placent) deja sur le pas de leur porte et lKTolTs lOTw les exhorte a se tenir al 'ecart (-)'. Dodds' analysis is impossible, since it means applying lKTOlToS foTw only to Tls b64> Tls b64>, and ignoring Tls µe-M8pOLs; 34 Lacroix' analysis reads far too much into the text. Roux' analysis, then, (which is also adopted by Verdenius3) is the most plausible one. It can be further supported by 62 KTVlTE"LTE" . . . ws bpq. KaBµov TI6>.i.s (Dionysos wants the Thebans to be witnesses of the Chorus' performance); also, if lKTOlToS lcrTw were to mean 'let him be withdrawn in his house' (so Elmsley, Gregoire and others) the ensuing clause becomes rather pointless. The same objection can be raised against Dodds' approach.

70 Cf. for aT6µa £-ll4>11µov lfoaLo-6a8w also A. Ag. 1247: e-rxt,,,µov ... Kolµ11crov crT6µa. In both cases e-rxt,,,µov is used predicatively.

71 ald D.: 'with voµLcr8lVTa: cf. 201' 'Do my words make sense?' 'They do'. 33. = line 111 in Diggle's edition, see also Diggle's note ad loc. Cf. also Call. Lav. Pall. (Hymn. 5) 1, where l~m, means 'come out'. 34. Dodds' note suggests that he reads Tls b&il Tls b&i\; Tls µEM8pms; but this is not in Murray's text, which follows Elmsley's Tls b&il TLS' b&i\; Tls; µEM8polS' lKTOTTOS' foTW. 32.

14

R.: 'Bien qu' ald puisse grammaticalement se rapporter avµvfiaw, ii est preferable de le rattacher avoµLaelVTa (Dodds)'. I do not think Dodds' reference to 201 is valid. There, Teiresias is vehemently defending the ancestral customs, i.e. specifically those that are connected with the other, older, gods. Since at our line, however, it is the Chorus who are speaking, they must be referring exclusively to the Dionysiac rites that had been established shortly before in Asia, cf. 21: KaTa in 96. Verdenius refers to K-G 2, 36. While 'se faire' is excluded one might posit an indirect-reflexive meaning, on the analogy of l3v>.£6w ('be member of a council') : ~ov>.efoµaL ('be a member of a council to serve one's interests'), thus: 'be a member of the thiasos for one's own benefit'. But this leaves iJ,ux6.11 unexplained. 37. This includes ritual treatments like K6TTToµaL, KdpoµaL. 38. Lacroix wrongly suggests that 'middle' is just 'middle' and fails to distinguish between the direct-reflexive (so Roux) and the indirect-reflexive (so Dalmeyda and Gregoire) variants. Notice that in Dalmeyda's and Gregoire's lranslations 'l'ame' is the direct object of 'se faire'. It is only in Roux' analysis that iJ,ux6.11 functions as an 'accusatif de relation'.

16

562 to support his view, but as a matter of fact this reference only permits talcing µT)p6V from µT)JXi>: 'Wenn zwei miteinander verbundene Verben (as here 1Ca~{nJ1as and auve-pd8n - AR) ein gemeinsames Objekt haben, so pflegt im Griechischen, selbst dann wenn die Rektion derselben verschieden ist, das Objekt nur einmal gesetzt und der Rektion des zunachst stehenden Verbs, haufiger des ersteren als des letzteren, unterworfen zu werden. Am haufigsten tritt dieser Fall bei der Verbindung eines Partizips mit dem Verbum finitum ein (-)'.39 As for KpuirT6v, this should be connected with vw; it is precisely the fact that the acc. µTlp6v is not expressed which makes it possible for KpuirT6v to be connected unambiguously with VLV.

102-3 c!ypav 8Ti/p0Tp64>ov Musgrave: 8rJpoTp64>oL (LacP: 8upao4>6pot. l) D.: "'their beast-brood spoil", i.e. the snakes they have caught: the adjective is hardly more than a poetic variant for 8llpu:i8Tl '. R.: 'l'epithete, ici simple synonyme de 8llpLw8TlS' ... '. L.: reads 8llpoTp64>oL, on which see below. V.3: 'I do not believe ... that the word is "hardly more than a poetic variant for 8llpLw8T)": the parallel Phoen. 820 8llpoTp64>ou 8paKOVTOS shows that the meaning is "feeding on beasts"'. Again (cf. I. 3), we are confronted with a compound adjective consisting of a nominal first and a verbal second member. As it is accentuated in all texts which have this reading, 8llp0Tp64>ov should mean 'feeding the beasts'. Since this is said of snakes, this is rather improbable: snakes eat, rather than feed, other beasts. It is for this reason, probably, that Dodds and Roux have taken recourse to equating 8T)pOTp64>ov with 8llpLw8Tl, needlessly (incidentally, the meaning of Dodds' 'beast-brood' is not clear to me). I agree with Verdenius that 'feeding on beasts' is far better but then, of course, 8rJp6Tpoct,ov should be read, with 'passive' accentuation. In fact, LSJ give 8llp6Tpoct,os for the parallel passage Pho.

820. 40 39. 'Objekt' is used somewhat loosely, for although the non-accusative constituent, mostly a dative, is an obligatory constituent with the participle, it is not an object; cf. our line and also e.g. Th. 1.20.2: Tl\i 'hmcipxl\l 1TfpLTux6vTES' ... dirlKTELvav (sc. auT6v), 1.144.2: TOUTOLS' diroKpLvciµEvoL diroirlµ-i,wµEv (sc. auTous). 40. Perhaps to be connected with intransitive, rather than with strictly passive Tpllj>oµaL.

17

As for 8T)poTp6ct,oL, retained by Lacroix, this is hardly meaningful; it is not clear in what way the Maenads are 'beast-feeders' .41 However, if 8T)p6Tpocf>oL is read the meaning is fully acceptable; again, 'feeding on beasts', viz. during the wµoq,ay(a (cf. Dodds XVI-XX and ll. 734ff. below). On the other hand, Dodds and others are probably right in thinking that dypav is more in need of a (qualifying) adjective than µaLvd8E"s (dypav alone = 8pdicovTas is rather stretched); 8T)p6Tpo4>ov should, then, be preferred.

106-14 106 aTEcf>avo0a8E", 107 pp-6E"TE" pp-6E"TE", 109 ic:aTapaic:xLo0a8E", 112 aTEcf,E"TE", 114 6aLo0a8' Inceptive present imperatives, cf. note on 173-98. 113- 4 dµcf>l Bt vdp&r)icas flj:!pLaTcls / 6aLo0a8 "· D.: 'It seems to mean "be reverent in your handling of the violent wands"'. R.: 'a.µq,t a ete explique de deux fa~ons: a) "rends-toi sainte a l'egard des thyrses" (-); b) "rends-toi sainte par le fait de tenir le thyrse"'. R. herself proposes a third explanation: 'Le chreur invite ici Thebes a rendre saints les narthex tout autour, i.e. ales entourer de la guirlande de lierre qui sanctifiera et transformera les batons profanes en thyrses doues d'une violence magique et irresistible'. L. (following Gregoire): 'portez pieusement le narthex'. V.3: 'ooLoo means "to make holy", "purify"; accordingly, "make yourselves holy by handling ... "'; he compares X. Cyr. 5.5.44 dµq,l &1wov E"lxE-v and LSJ s.v. dµq,l CI 6. The interpretation hinges on oaLoiia8E and dµq,L Formed from active oaL6u1, OOLoiia8E must have a middle meaning. For this reason Dodds' 'be reverent' and Lacroix' 'portez pieusement (or: religieusement)' drop out ('portez' is very strange indeed). Verdenius' 'make yourselves holy' is a possible interpretation; in that case bO"Loiiµm has to be viewed as a middle which belongs to the class of habitual ritual verbs, cf. ancf>avoiia8E at 106 and fn. 37. However, his 'by handling' for dµq,(, further specified as = 'engaged on', cannot be supported by referring to X. Cyr. 5.5.44 dµq,l BiiTTvov ElXEV or to LSJ. As appears clearly from 41. In an additional note Dodds suggests 'beast-keepers', but this has a homely flavour which is slightly out of place.

18

LSJ, aµcf>t has this meaning only in the idiomatic combinations with lxw and dµL. See also K-G 1,491. Ifit is to be construed with vdp611Kas its meaning should be: 'with respect to' (or local 'round, at', but this hardly makes sense). Roux' analysis, in which ocrLo0a8e- is taken as an indirect-reflexive middle and aµcf>l adverbially or in tmesi ('make them pure all around for your own benefit'), avoids all these problems and should, therefore, be preferred. 42

13Sff. T)Bus, lv c5pe-aaw lhav / .. ./ 1rl01J

1rE86aE,

KTA..

Thus Kopff, following LP, except for ope-aaLv, which is Schone's correction of LP's oprnLv. Verdenius 3, in part following Kamerbeek (1953: 192-3), has a very detailed discussion of this much-debated passage; he defends, successfully, I think, the reading given above, in which a new sentence starts at 135, Dionysos is the subject of,;Bvs, stands predicatively, meaning 'welcome' (viz. to me, the Chorus), 43 and c.lTav etc. refer to Dionysos' partaking in the atrapayµ6s and the wµocf>ayta. The latter point, which has given rise to many doubts, is conclusively proven by the iconographical evidence assembled by van Straten (1976: 66). 44 All other proposals involve far greater difficulties than those presented by the text given here. If the above is correct I think we should read at 140, with Verdenius: o 8' Uapxos, Bp6µLos ('and he is the leader, Bromios', implying that the other members of the thiasos followed suit).

,;svs

139 atµa TpayoKTovov, wµo4>dyov xdpw D.: 'Blood of the goat that is slain, joy of the living flesh devoured'; R.: 'litteralement "le sang tueur de bouc", c'est dire qui tue le bouc au moment ou ii s'ecoule de la blessure';

a

42. For adverbial d.µ1/)( cf. especially Hipp. 770, with Barrett's note (although B. translates, rather misleadingly, 'about her'). 43. 'fi8us, lhav ... looks like a personal variant of 'fi8oµllll\l µoL lhav .... Cf. Pl. La. 187 c: l-µol µEv ... Kal ME~TJO'(opos, e,,poTpoq>os, the passive accentuation Tpay6KTovov yields 'blood killed by a goat', which is nonsense. The first of LSJ's translations given above (which boils down to taking Tpay6KTovov as a kind of perfect participle) can therefore be discarded, I think. Observe, incidentally, that Dodds, who reads TpayoKT6vov, has virtually the same interpretation. 2. So it should be TpayoKT6vov, which, as Roux at least is perfectly aware, in principle must have the meaning 'which kills goats'. Verdenius' criticism of Roux is not justified, and his own solution is arbitrary, and has, moreover, a definitely passive tinge. Nortetheless, there remains the question how alµa can be said to kill. The clue to the right answer is provided by Groeneboom's note on A.

Eum. 282-6: Ka8apµo'Ls xoLpOKT6voLs, and by LSJ s.v. XOLPOKT6vos. Groeneboom remarks (fn. 6 on p. 135): 'Ka8apµo'Ls XOLPOKT6VOLS = reiniging, die in het dooden van een xo'Lpos bestaat' ('a purification which consists in the killing of a xo'Lpos'), and LSJ note, quoting this line: XOLPOKT6vos 'purification by the sacrifice of swine'. In other words, the action noun Ka8apµ6s 45 is viewed as the 'performer' of the killing of the swine. Applying this analysis to our line entails taking alµa as an action noun, i.e. not as 'blood' but as 'bloodshed'. Compare LSJ s.v. alµa II: 'bloodshed, murder', e.g. Or. 284-5: elpyaaTaL 8' e-µol / µT)TpQov alµa. 46 All in all, then, my interpretation of aype-vwv alµa 45. Chantraine (1933: 144f.). 46. While not being derivable from a verbal stem, atµa apparently behaves like the regular nouns in -µa. Observe, in this connection, that the meaning 'result of some state of affairs' is by no means the only meaning of this suffix. Some other action nouns in -µa are Kci8apµa (IT 1316), cnrcipa-yµa (Antir. 826), llpE-yµa, 8(w-yµa (A.

20

TpayotcT6vov runs: 'hunting after/thirsting for the bloodshed which consists in the killing of goats' .47 By the same token wµocf>ayov xapLv

must be interpreted as: 'joy which consists in eating raw (flesh)'.

144-6 l:vp(as- 8' C:,s- >.a.pctvov ica1rv6s-. 6 BaicxEus- 8', dvtxwv 'll'VpaeaS811 cf>).6ya 'll'EUICaS-, 145 dvexwv Wilamowitz : lxwv LJ>48 Thus the text as printed by Roux. This means that at 144 she retains LP's tcmrv6s- (so too Verdenius3), not improbably. With tca1rv6s-, pEi. should be supplied from the preceding line, as proposed by Kamerbeek (1953: 193). Murray, Dodds, Lacroix and Kopff, however, adopt Wilamowitz' full conjecture tca,rvov b Ba.tcXEUS' dvlxwv. Dodds and Lacroix construe :2;uptas- . . . tca1rv6v as modifying 4>>..6ya, apparently connecting ws- with tca1rv6v, but this hardly makes sense, unless one manipulates the Greek, as Dodds does: 'lifting high the blazing flame of the pine torch, sweet as the smoke of Syrian frankincense'; 'sweet', however, is not in the Greek text, and should therefore be omitted. But then the result is rather odd: ' ... the blazing flame of the pine torch, as the smoke of ... '. The same oddity attaches to Lacroix' translation: 'Baccheus tient la torche flamboyante de pin, pareille a une fumee d'encens syrien'; the torch as such is not similar to smoke.

14 5

o

8atcXE'6S'

D.: 'The Bacchic One (i.e. the god temporarily incarnate in the celebrant)'. R.: 'le bacchant par excellence, le dieu lui-meme ... '. Eum. 139, Andr. 993), examples taken from Chantraine (1933: 184-5). At S. OT 101, where atµa is the subject of XELµci(ov, it probably= 'bloodshed' as well. 47. This is also Sandys' interpretation ('hunting after a goat-killing slaughter'), which apparently was not taken into account by later editors. - The parallels adduced by Dodds can be explained in the same way: Or. 833 atµa µaTpoKT6vov 'the bloodshed that consists in the killing of the mother', similarly A. Fr. 327 arµaTOS" xoLpoKT6vou, and others, e.g. HF 1201 atµa TTaL8ocj,6vov. Cf. also Cy. 127 ~opq dv8pwTTOKTOVI\I, IT 384 8uo(aLS" ~poTOKT6voLS", A. Pr. 530-1: 8o(vaLS" ... ~oucj>6voLs.

48.

According to Roux' app. crit. L has TTiip ... / 1Tupow811, but this is wrong. Its text runs: TTiip (or TTup) EK vcip81]Kos cilooEL / ow811 cj>>..oya. See Zuntz

(1965: 119-20).

21

Roux' opinion is, I think, the right one. The choice of this name (and not e.g. of BMya (-). Or, sur les nombreuses representations figurees que nous possedons, ces deux accessoires sont toujours distincts: le thyrse n'est jamais utilise comme porte-torche. Le complement EK vap&TJKOS marque l'origine, le principe de l 'action exprimee par le verbe dlcrcrn '. This verb is taken transitively by Roux, but she connects it with ,r;\avdTas, not with cf>Mya. Like Dodds she takes 8p6µqi Kal xopotatv as instrumental datives with Epe-8lCwv: 'il les dechaine par la course et les chreurs de danse'. LACROIX (who has the same text as Dodds) translates: 'Baccheus tient la torche ... attachee au narthex ', i.e. he seems to take EK vcip0TJKOS as if it is construed with a verb like Kplµaµat EK; he has no note on EK vcip&TJKoS. Lacroix takes dlcrcrn intransitively: 'bondit'. 49 ~p6µq> Kal xopots (L. does not accept Murray's xopotaw) are taken as a locative modifier with Epe-8lCwv: 'dans la course et dans la danse il invective ... '. 49. The comma after .6-ya in Lacroix' text is clearly a mistake. Lacroix alleges that Roux construes dlaaEL with .6-ya, but this is incorrect.

22

VERDENIUS3 follows Roux in rejecting the connection between TTfVKT) and vap8l't. He talces e-K vap&flKOS' as instrumental, comparing S. OC 848-9: o~Kovv trOT' E"K TOUTot.v ye- µT) aic1'trTpoLv ln / b&w TTOPTlO'IJS. As for d(aan, he notes: 'Probably used absolutely: "causes movement". The connection with lK vap&flKOS' shows that the sense cannot be intransitive'. V. takes 8p6µCj) Kal xopotaLv as modal datives; he further thinks that 'the object' (of lpe-8(Cwv, presumably) 'may be supplied from 8p6µCj) Kal xopo1aLv ... and the predicate tr>.avaTas has resultative force'. ORANJE, p. 190: 'In 147 lK vap&flKOS' is instrumental, d(aafl is intransitive. Dashing off with his narthex the god springs into movement and leads his bacchae in a dancing run (8p6µci> Kal xopotaw) over the hills'. This is a rather bewildering variety of opinions, some of which, however, are more bewildering than others. Roux and Verdenius are certainly right in disconnecting TTfVKas from lK vap&f1KOS'; iconographical evidence makes it clear that torch and narthex were two different objects. Cf. especially Van Straten (1976: 58). But what, then, about E"K vap8T]Kos? According to Roux, l K va.p&flKOS' 'marque l'origine, le principe de l'action exprimee par le verbe d(oofl', while Verdenius and Oranje prefer a strictly instrumental value. Actually, both semantic features may be relevant: Dionysos dashes on (d(oafl being intransitive, see below) 'starting from the narthex', i.e. using his narthex to push himself off. It must be said, however, that if one looks for parallel uses of E"K, the number of occurrences of E"K with a concrete entity50 turns out to be very small indeed. Apart from S. OC 848, adduced.by Verdenius, one may compare S. Ph. 91-2: ou yap l~ €VOS tro8os / ~µas ... XflPOOfTQL (lit. 'starting from one foot', i.e. 'when he has the use of only one foot', Jebb), Tr. 875: ~E~11KfV ... l~ aKLvr,TOu tro86s ('she has gone ... without stirring a foot', Moorhouse 110),51 Hee. 573: E"K xe-pwv I uMoLs l~Mov. 52 One factor, how50. Nouns occurring as instrumental modifiers typically refer to concrete entities. For a succinct discussion of the relationship between instrumental and manner adverbials see Vester (1982: 37-47). 51. Moorhouse doubts the instrumental value of these expressions and views them rather as referring to 'the base of an action, its underlying condition or manner of performance', as in EK 86Aou (S. El. 279), EK u1r00Tpo4riis (S. El. 725), EK j3(as (S. Ph. 563). Note, however, that in none of these instances the noun refers to a concrete 23

ever,. that may have favoured the choice of (instrumental) €K + genitive at Ba. 147 is that the dative of a concrete noun with cUcrcrn could easily have been taken as a sociative (or: comitative) dative, as in Hom. II. 8.88 cf>a.cryav'\l dlaawv, cf. also 5.81, 10.456, 11.484, 17.460: hmoLs d(aawv, 18.506: TOLO"LV (: O'KTJTTTpoicrL) l'jLaaov. Compare also the comitative use of vap8JiKL at Ba. 251: vap&r,KL f3rucxe-ooVT '. As for dlcrcre-L, we have seen above that this is taken transitively by Dodds and Roux; they differ in that Dodds connects it with 1/)Mya, and Roux with TT>..avaTas. Somewhat confusingly she renders in her translation by 'donne le branle leurs errances', and in her commentary by 'lance les errantes', while explaining TTAavaTas some lines further as 'proleptique'. Be that as it may, I do not think there is a good case for the transitive analysis. There are only two parallels, Or. 1427ff. and S. Ai. 40, whereas the number of intransitive instances of dlcrcrw is rather overwhelming, both in Euripides and in the other tragedians, as well as in Homer (see also above). Moreover, there is definitely no need here to take it in this way, the intransitive use being fully acceptable. Verdenius' objection that, by being connected with (instrumental) EK vap0r)KOS', it cannot be intransitive, is unwarranted. Instrumental modifiers show, to be sure, a preference for occurring with transitive verbs, but they are by no means excluded with intransitive ones. A nice example is S. OC 848, quoted by Verdenius himself. Cf. also S. OC 146 (dMoTploLs 'oµµacrLv elpTTov) and Moorhouse 88-9, § 16. As for the meaning of dlcrcrn here, it combines, I think, as suggested by Oranje, 'springing into movement' (viz. from his position on the ground, I. 135) with 'rushing along'. Both actions are carried out by means of the narthex. This leaves us with 6p6µCt> Kal xopoiaLv TTM.VaTas e-pe-8l(wv laxais T' dvaTTaAAwv. Although 6. Kal X• might perhaps be taken as instrumental datives with e-pe-8l(wv, I prefer to take them, with Verdenius and Oranje, as modal, to be construed with dlcrcrn. Verbs of (rapid) movement are often accompanied by such modal datives, e.g. El. 474: foTTe-v6e- 6p6µ(\l, IA 427, Ion 1556, Or. 726, below 727, 748. dp6µ(\l Kal xopoicrLv possibly form a hendiadys, cf. Oranje: 'in a dancing

a

entity, a fact not considered by Moorhouse. Roux inclines to view the narthex as a more or less animate entity, pointing to v~pLa~s- in 113, because this makes the use of tK more viable. But this seems rather arbitrary. 52. tK XEp.Aw. Observe that rather more often dn6 is found with concrete nouns in instrument-like constructions, cf. K-G 1,458 and LSJ s.v. drr6 III 3.

24

run'. 53 Next, if a.(aan is intransitive, TTAaVJ.wv only, viz. as a direct object. It can hardly be proleptic (Roux) or have resultative force (Verdenius), if there is no real object. Verdenius, it is true, supplies an object from 8p6µ'¼) Kal xopoLO"LV, referring to 97 avvEpd8n. But this is impossible. For one thing, for this analysis to be possible there has to be a finite verb besides the participle; this cannot be d(aon, since he takes it absolutely. For another, even if d(aaEL were transitive, it cannot 'borrow' its object from 8p6µ1¼) Kal xopo'LaLV, for these words, whether they are instrumental or modal, form an optional constituent; and it is only when the constituent involved is obligatory both with the participle and with the finite verb that it can be omitted with one of these verb forms. Cf. also on 97, and

fn. 39. All in all, then, I suggest the following translation by way of a conclusion (11. 145-149): 'And Bakcheus, while lifting high the fiery flame of the torch, springs into movement and rushes along with his narthex, in a dancing run, provoking the stragglers and by his cries urging them upwards'. 54

150 Tpv4>Ep611 irMKaµov ds- al8lpa ~£TrTWV Tf add. Wilamowitz Since LP's Tpvcf>Epov TTMKaµov does not yield proper ionics, we must either read TptJ£p6v , which is adopted by e.g. Murray and Roux, or follow Earle in reading 136aTpvxov (thus e.g. Kopff), taking TTMKaµov as a gloss on ~ooTpvxov. Dodds has no preference. The addition of Tf brings p(TTTWV etc. in the same syntactic position as the participles tpe-8(Cwv and a.vami>.Xwv; thereby phrTwv, like these, would express an attendant circumstance of merely d(aaEL: 'he rushes along, provoking ... and urging ... and throwing ... '. 55 Without Tf, p( TTTWV etc. express an attendant circumstance of the whole of a.laaEL ... t pE8lCwv ... T' a.va53. This does not exclude, of course, that ultimately Dionysos' dancing run is the means by which the tpe8(CeLv is brought about. 54. dva- in dvairdllwv probably expresses an upward movement, cf. 165 els ISpos els lSpos. Cf. also Oranje's 'over the hills'. 55. Willink (1966: 36), too, rejects TE (followed by Verdenius), but his syntactic analysis lacks precision.

25

mi>.>.wv: 'he rushes along, provoking ... and urging ... , while throwing •.• '.56 Since this seems more natural, I prefer Earle's ~pvxov.

lSlff. ToLd8" l1rL~piµ.£L· / '3 tTE" ~dicxaL, ic:T~. Commentators hesitate as to the precise scope of TOLd8e-. Some are inclined to let the BaKXE"VS stop after the first instance of [3dKXaL (cf. Dodds' note), others after the second instance of [3dKXaL (e.g. Oranje), whereas others, again, think that he speaks up to ds lipos in 1. 164 (thus e.g. Dodds and Roux). For a number of arguments against the first two positions see Dodds and Verdenius3. Another argument would seem to be that the content of TOLd8e-, if this covers only the first or the first two lines, is surprisingly brief and monotonous. In the other cases of TOLd8e- with verbs of saying in Euripides this word covers a minimum of 3 lines (El. 804ff.) and a maximum of 27 1/2 lines (IT 364ff.; incidentally, all other cases of TOLd8e- occur in spoken passages). Also, the imperative µl>,trE"TE" cannot be spoken by the Chorus itself, since they are already celebrating Dionysos, cf. 65f.: (8odCw ... ) BdKXLOV e-vaCoµlva. 154 X~L8ci LP : x>.i.8Q. Musurus, x>.i.8d Victorius Since the proponents of the reading x>J.8q (Roux, Lacroix, Verdenius3) fail to produce a convincing explanation of the dative, 57 it seems best to read x>.i.8d, as an apposition to BdKXaL ('pride (or: splendour) of Tmolus', not 'O pride ofTmolus' (Dodds)). Alternatively one might consider reading x>.i.Bav (Seyffert, Willink), to be taken as an accusative 'of end 56. Verdenius, who does not accept TE, nevertheless keeps 1r>.6Kaµov, and suggests that the last element of the first ionic might be taken as anceps. But it is perhaps more simple, if one wants to keep the transmitted text, to take 1r>.6Ka- as being resolved and -µov els al- as a syncopated ionic. 57. The only instrumental datives found with µ0.1rw are x6poLs and similar words, not, then, words referring to musical instruments (i.e., if x>.i.81' can denote such an instrument, for which, again, no parallels are given). Roux thinks that x>.i.84 refers to the music itself, but such a use of x>.i.81' is, once more, unparalleled. (Pi. 0. 10, 99 x>.i.&ooa µohmi is not a real parallel, for we cannot infer, from the fact that µobd can be characterized as 'delicate, soft, rich', that x>.i.81' can mean 'delicate, soft, rich music').

26

of motion' (Moorhouse 45; cf. also De Boel (1988: 52-3)). 58 Cf. 139-40: llµEVOS' ES' 6pe-a pi,yta.

157 dyallcSµoai. The middle has indirect-reflexive meaning: 'for your own benefit'.

159 lv ... ~oats

A good analysis of the instrumental use of EV may be found in Moorhouse 106-7. This use is not a Hellenistic development, as Dodds suggests, but is already fully present in Sophocles, cf. especially Tr. 887, where EV Toµq cn&ipou (if we follow the codd. in giving these words to the Nurse) answers a question with TTWS'.

165 ds- c5pos-

If the speech of the Bakcheus ends here, a full stop should be printed after ~pos (so Roux). See also note on next line.

166 clpa. "Apa draws our attention to the effect of the Bakcheus' words; 'expressing a lively feeling of interest' (Verdenius3, following Denniston) is not specific enough. This means that, with ,;Boµlva etc., we have left the direct speech of the Bakcheus. 170 TCS' lv 1r6~a1.a1.; Cf. note on 68-9. Unlike such forms of address in modem language ('Is anybody there?', 'II y a quelqu 'un?', 'Is er iemand?') the Greek formula apparently presupposes that someone is present (not unnaturally, since Teiresias is in front of a palace), and immediately calls for this person's attention.

173 [Tw TLS', dadyyEUE" For the transition 3rd ps. imperative ➔ 2nd ps. imper. see K-G 1, 85-6. The use of the 2nd ps. probably indicates that, following the general order with TLS', one of the attendants comes forward, who can, thus, be addressed directly. 58. '(i)t matters little whether it is interpreted as an internal accusative ... , or as an accusativus loci ... ' (Willink). Actually, however, it cannot possibly be an internal accusative. For a bare accusative to be used as an internal accusative (which, incidentally, is extremely rare), the noun must be semantically related to the verb, which x>..L8ci obviously is not. Cf. also n. on 247.

27

173-4 dadyyfllE TnpEa(as- lSTL / C11Td vLv The front position (prolepsis) of Tnpmlas probably gives prominence to the name; so too Lacroix and Roux. The effect may be as described by Roux: 'Le personage se nomme pour se presenter'. Comparable cases are IT 1323-4: ... EKp6VTLO'OV / 8uuyµos lSO"TLS TOUS etvovs &ripdanaL, 59 Pl. Men. 81 e: l8lN.il µnd aoO (T1TELV l>..rn8', are different, inasmuch as yap explains the form of address. Thus, at Rh. 608 Odysseus, with cf>9lyµaTos yap 11a86µT1v / TOO ao0 auvft9ll yfipvv, explains why he calls his addressee BfolTotv' 'A8dva; cf. also Hee. 1114f.: cf>l>..rnT', -Oa86µT1V yap, 'Ayaµe-µvov, al8E"v I cfxi>VllS dKouaas, E"laopq.s miaxoµe-v; Similarly, in S. OC 891: cf>l>..TaT', lyvwv yap TO lTpocr cf>wVT1µ0..TaTE" ('I can call you "dearest friend" because I have recognized your voice'). Note also that in these cases there are explicit indications that the speaker cannot see his addressee: either the latter is physically separated from the speaker (Rh. 608), or the speaker is blind (Hee. 1114f., S. OC 891). 63 In the present instance, however, there are no such indications. Observe, incidentally, that Dodds' 'recognized' is not in the Greek text. Moreover, 'I heard and recognized your voice' is rather strange, since Teiresias has used his own name (173). So ws tia96µTlv possibly explains just Kadmos' (implied) quick arrival, as proposed by Tyrrell: 'ws, since, to account for his coming out unsummoned. He had heard the voice of Tiresias and therefore had not to await the summons of the porter'. (But see below). Syntactically, thewsclause, like the yap-clauses discussed above, strictly speaking has no main clause. However, Bl in l. 180 possibly marks the inception of what at least semantically is the 'main clause'. Cf. the use of Bl after parenthetical or anticipatory yap-clauses, Denniston 70f. An example from tragedy is Rh. 611 (not in Denn.). I should add, however, that there are no real parallels for this use of ws, nor, for that matter, for the use of causal ws to explain a vocative. To be sure, ells occurs quite frequently after a vocative but then it is always exclamatory, which means that it has intensifying force, as in the following examples, where 4>(>..TaTE" precedes: S. Ant. 572: 4>l>..rn8' Alµov, ills a' dnµa(n TTaTI)p, S. El. 23: 4>l>..rnT' dvBpwv lTpoaTT6>..wv, ws µot aa11 I O'Tlµe-1a alvns fo8>..os E"ls T)µas ye-yws, Hee. 990: 4>l>..rn8', e-Zi Kdetws Xlyns al9e-v, Or. 217: cf>l>..Ta9', µ' T1licf>pavas E"ls liTTvov TTrnwv, Tro. 1167: cf>l>..Ta9', aOL 9avaTos ~Me- 8uaTuxfls. Note that in S. Ai. 14ff.:

w

w

a

w

w

w

ws

ws

w

ws

w

w

w

w

63. Other examples of this use of yap to explain the very form of the preceding vocative are: Ba. 1316: w 4>l>.TaT' dv8pciiv -Kal yap oiiKlT' C:,v ISµt&IS' / Twv 4'LhaT11iv lµoLy' dpL8µ~011. TlKvov, IT 828: iL 4'lhaT', oliBtv dUo, 4>l>.· TaTOS' yap El, Med. 465: iL TTa-yKaKLwVT'Iµ' aKOVW KTA, where thews-clause does explain the form of address, ws is, again, exclamatory. 64 As for the intensifying force, this is clearly present in all cases; e.g. at S. Ant. 572 ws expresses the idea that in lsmene's (or Antigone's) view Kreon wrongs his son-to a very high degree. If ws at Ba. 178 is exclamatory as well, which is strongly suggested by the parallel passages discussed above, it should be connected with cro#Jv: 'how wise is the word I heard from a wise man'; ws ... ao#)v is used predicatively, like ws ... 6uaTuxfls at Tro. 1167, quoted above. The position of aocf>11v at the beginning of 179 before crocf>oD, rather than at the end of 178, may be explained on the principle of rhetorical juxtaposition, cf. K-G 2, 602. Observe that ws can hardly be taken as a modifier of ,jcr86µT)v, since a verb like 'hear' cannot easily be intensified:65 'how much did I hear your voice' is rather odd. Note that in the parallel construction Tro. 1167: ti> cf>(XTa8', ~ croL 8dvaTos ~X8e6uaTUxfls, ws cannot easily be taken as a modifier of~>.ae- either.

184 1roi 8E"i xopE6nv ••• ; To the remarks of Dodds and Roux one may add the more comprehensive note of K-G 1, 545, A. 4, who discuss both the use of ,ro'i: instead of troD and that of troD instead of ,ro1.

185 lf11yo0 Not 'expound' (Tyrrell, Dodds) or 'enseigner ou interpreter les rites en qualite d'exegete' (Roux) but 'give me instructions how to act'; cf. LSJ s.v. II, 2, and instances like A. Ch. 118: Tl cf>w; 6(6acrK' cfoE"Lpov lfrr youµlVT'I, Hdt. 4.9.4: on XP'l'i TTOLfE"LV letiye-o av. Kadmos wants to have specific information concerning his two questions, not just general information on the Dionysiac rites. 64.

Sometimes a causal nuance would seem to be present, e.g. S. El. 23, but the exclamative meaning clearly predominates. - Other examples of a vocative followed by exclamatory ws are A. Pers. 249,472,515,518, S. El. 504f. ()yr.), 1126, 1199 (perhaps with a causal nuance), E Ale. 1068, Andr. 71, Ba. 608-9, 1024-7, 1352, Hee. 716 (perhaps with a causal nuance), Hel. 315-1, 862, Med. 328, 1363, 1364,

Or.211. 65. Portus' 1')86µ11v forucr86µ11v eliminates this problem. - Verdenius4 thinks that '{)cr86µ11v 'is probably used absolutely in the sense of "to be right"'; he compares Or. 752. There, however, alcr8aV1J is Orestes' reaction to a question by Pylades, confirming the latter's hypothesis: 'you have got it' (rather than 'you are right'). Since the words preceding Ba. 178 are not a question, Or. 752 is not really parallel.

30

191 Els- lSposAs appears from these words, Kadmos' questions of 11. 183-4 must be understood as 'where ?' That it was known at Thebes that the Theban women were somewhere in the mountains is clear from 11. 217ff. The audience, of course, have been told explicitly where the women are, cf. 33ff. and 63-4. 191 o1)ico0v Roux, following Paley, takes line 191 as a statement, not a question: 'Cadmos songe tout naturellement a utiliser un char, non seulement a cause de son age, mais parce qu'il est un roi.' There are two mistakes here. First, we have just been told that Kadmos does not any longer consider himself old. Second, it is true that princes and kings normally move around in chariots, as Roux' parallels show; however, the circumstances are not normal here, as Kadmos himself has indicated (180ff.): he is no longer behaving as a (former) king. Yet I think that Kadmos may, indeed, be thinking quite naturally of using a chariot, for this will enable them to join the women as quickly as possible. I prefer reading I. 191 as a question (so too Lacroix and Verdenius4; Lacroix points out that I. 191, if taken as a question, conforms to Kadmos' behaviour as described in lines 193 and 195), and more precisely as a question introduced by ol'.,icouv (Lacroix) rather than by ouicoOv. Verdenius thinks that the more quiet ouicoOv suits better here. But this is doubtful: Kadmos is not quiet, as clearly appears from 180: i\icw 8' lTOLµos, 184: 1TOL 8n xopd,nv ... ;, 187: OU icdµOLµ' icpoTwv, and especially 197: µaicpov TO µl>J.nv. He is impatient and but all too eager to leave. To this attitude ouicouv is better suited than ouicoOv. 66 The focus of the question is 6xounv: 'Shall we then not go by chariot to the mountain?', 'We surely shall go by chariot ... ?' One may even be tempted to take ol'.,icouv ... ,re-pdaoµe-v; as iussive: 'Shall we not ... ?'= 'Let us .. .'. 67

ws

av ...

66. Cf. for the intricate problems connected with ouKoliv and oliKovv Denniston 430ff. and especially K-G 2, 166ff., who connect oliKouv inter alia with an 'ungeduldiges Gemilt'. Friis Johansen and Whittle note, on A. Supp. 300: 'Neither tragedy nor comedy offers a clear case of the interrogative inferential or progressive oucoliv ... so it is reasonable to accent the word oliKow in all questions in drama'. 67. Cf. K-G 2, 167. It is true that K-G only give examples with second and third person fut. indic., but this seems arbitrary. Observe that at I, 177 they have some cases of iussive ou + first person singular fut. ind.

31

192 dU' oiJx 6µoCra>S" 'But then not in like manner' viz. as when they did not go there by chariot; the implication that they will therefore go on foot is taken up by 1raL8a:ywy{)C7w a'. 193 iraL8aywY11aw a' lyoo; This ambiguous form should probably be taken as an aorist subjunctive; for details I refer to Appendix 4. 194 dµoxOd LP: aµox8l Elmsley, Murray, alii LP's dµox8d can probably be retained. See Risch (1972: 65-73, = 1981: 167-75). 19S This line is taken by Dodds as 'an expression of doubt', rightly, provided that this not be equated with 'dubitative'; see, again, Appendix

4.

173-98

173: [Tw, dadyyE>-M; 184: ,rot 8d xopf{,fLV, ,rot ica8LO"TdVaL; 185: te11yoO; 197: lxov; 198: evvairTE,

EvvwpCCov

Most imperatives and semantically related verb forms ((&1) xopd,nv, Ka8urravm) in this passage are formed from the present stem, whereby Kadmos and Teiresias emphasize that the respective states of affairs should be carried out straight away. For this so-called inceptive value of the present imperative cf. Rijksbaron (1984: 46), and especially Ruijgh (1985: 24ff.), Moorhouse 218f. and Bakker (1966: 31-66). 68 Bakker (1966: 61) aptly remarks that an aorist imperative expresses an advice or a proposal, whereas the present imperative is used when immediate action is called for. The theoretical foundation of the inceptive value of the present stem is a difficult matter. Bakker (1966: 65) rather vaguely speaks about 'a connexion between the existing situation . . . and the action ordered', 68. There is one aorist inf., E concerne le futur immediat qui s'etend a partir du moment present' (italics Ruijgh). In other words, he views the inceptive value in terms of the temporal value 'present' of the present stem. However, in both approaches the value of the present imperative is disconnected from the uses of the other forms which are derived from the present stem. To name only one objection: in the imperfect the present stem obviously does not have the (temporal) value 'present'. I would therefore prefer a different explanation, which sees the inceptive value as a manifestation of the general (aspectual) value of the present stem, viz.: 'presents a state of affairs as not-closed, i.e. as being carried out.' The present imperative, then, focuses on the state of affairs being actually carried out; it could be paraphrased by 'be carrying out this state of affairs.' The aorist imperative lacks this feature, it could be paraphrased by '(just) carry out this state of affairs' (the focus being perhaps on the closing, the terminating point, of that s.o.a.). A nice illustration of the inceptive value of the present imperative is provided by the use of dKouE in Euripides as opposed to that of dKouaov. "AKOUE occurs 23 times (not counting the fragments), 18 times it is modified by a particle or other expression calling for the addressee's immediate attention, viz. 611 (Denniston 216: 'may usually here be rendered "come" or "now"'), 611vuv (Denn. 218: 'expresses an increased urgency in command or appeal'), Tolvuv (Denn. 571: ' ... the answerer ... plunges at once in medias res'), or d.>.M (Denn. 14: ' ... transition from arguments for action to a statement of the action required'). 69 The aorist imperative, on the other hand, which occurs 29 times, is never modified by 611, 611vuv, Tolvuv and only once by d.>J.a..7° This may admittedly also be due to metrical factors, since the sequences clKoul 6ii, clKoul 6iivUv, clKoul Torvov could be fitted much more easily into 69. 8fi: Ion 336, 1539; 8fivuv (also 8i'i vOv): Cy. 441, Hee. 833, Hel. 1035, HF 1255, IA 1009, 1146, IT 753, Or. 237, 1181, Pho. 911, 1427, Supp. 857; To(vvv: Ion 936, 987; cillci: Ale. 781, IT 1322. The five remaining instances are: Cy. 175, Hipp. 1296, IT239, Supp. 428, 1183. 70. The aor. imper. occurs at Ale. 400, Andr. 1079, Cy. 253, 323, 519, El. 226, 619, Hee. 273, 788 (+ dllci), 1137, 1217, Hel. 1049, Held. 643, HF 279, Ion 589, 633, 1609, IA 1374, IT 1012, 1313, 1436, Or. 565, 854, 1131, Pho. 1070, 1586, Tro. 394, 635 (del. Dindort), 906.

33

(spoken) iambics than the sequences clKouaov 6if etc. On the other hand, the fact that Plato has 12 instances of dKOUE 811 as against 3 instances of dKoooov 611 71 indicates that the metre is not the only factor to be reckoned with. In Euripides dKouaov usually expresses a (polite) request, a plea or a prayer, rather than an order. (Cf. also below, on 309). It is especially at home in situations where the speaker is in no position to simply give orders to his addressee, as when children are addressing their parents, e.g. Ion 589: wv 8£ yt yll(OOKW, TTO.TEp, I d.Kouaov, 633, Ale. 400: VTTO.KOOOOV dKOOOOV, ~ µci.TEp, d.vna(w. Cf. Ruijgh (1985: 35f.), Bakker (1966: 11-17, 98-127). A concomitant difference between the two imperatives is that usually the order given by the present imperative has been prepared. An example from lyrics is Pho. 911, where dKouE follows upon Teiresias' request at 909 j3ov>..i:i ... cf>paaw; and Kreon's positive reaction to this request at 910. It are precisely these preparative elements which create the appropriate framework for the present imperative: after possible obstacles have been removed, the addressee can be asked to give his attention forthwith.

200ff. oii8lv aocf>L{6µ£a8a Toi:aL 8aCµocw ICTA.. The text of lines 200-202 presents a number of problems, which are summarized by Roux as follows: (1) the 'brutal' asyndeton between 199 and 200; (2) the construction of crocf>(Crn8m with the dative; (3) the asyndeton between 200 and 201; (4) the pleonastic pronoun auTa at 202, which 'dans la meme phrase' (ital. Roux) refers back to a feminine plural antecedent. In my opinion, however, there is nothing amiss with the text. First, the asyndeton between 199 and 200. Many scholars have tried to get rid of it by reading ou6' lvaocf>t(6µEcr8a, with Musgrave; this is, of course, hardly a conjecture, and it gives good sense, although lvaocf>l(oµat ('to use cleverness on') is not found elsewhere. However, the asyndeton may be acceptable, in view of passages like A. Ch. 911-2: Op. Kal T6v6E TOtvuv Motp' ltr6pauvEv µ6pov. K>... ouBE-v aE~lCi:i yEvE8Mous dpas, TlKvov;, Hee. 1273-4: Ilo. (your tomb will be called) Kuvos Ta>..a(VTIS cri\µa, vauTl>..ots TlKµap. - EK. ou8iv µlML µoL, O'OtJ 'YE µOL 86VTOS 6l101V, Or. 1114-5: Op. wa8' 'E>J.as auTij aµLKpov olK11T11PLOV. Ilu. ou6Ev TO 800>..ov trpos TO µ11 71. dKOUE Bfi: Cra. 413 d 7, Tht. 201 d 8, Phdr. 230 e 5, Chrm. 173 a 7, Grg. 458 e 3, 506 c 5, 523 a 1, Hp. Ma. 289 a 8, 291 d 6, R. 338 c 1, 595 c 5, Tim. 20 d 7, also [Pl.] Just. 374 e 2; IIKouaov 6{): Smp. 214 b 9, Grg. 453 a 8, R. 477 c 6.

34

8ouAOv ylvos. 72 In all these instances the speaker reacts to something said by another person; and since there is no formal connection between the two lines, the idea is conveyed that the reaction follows immediately upon the preceding line, cf. Campbell's 'urgency' (see footnote 72). Similarly, at Ba. 200 the asyndeton signals that Teiresias does not hesitate one moment to put forward his own view concerning the matter at hand. As for the interpretation of aocf>LC6µEa8a TOLOL 8a(µoaLv, I think Verdenius4 is right when he considers the dative a dative of interest; more specifically it is a dativus incommodi. 13 To the parallel adduced by Verdenius (Med. 1292: Baa l3poTots lpE~as t\811 KaKa) one may add s. Ph. 1019: 6AOLO' Ka( OOL TTO>.A0.KLS T68 11u~aµ11v. Third, the asyndeton between 200 and 201. This can best be accounted for by taking it as explanatory; so Verdenius. In that case a colon after 8a(µooLv seems better than a full stop. Finally, atiTa in 202. This, too, is defended by Verdenius, again rightly, I think. 74 He refers to K-G 1, 61; especially relevant is K-G's remark that a neuter plural pronoun which takes up a masculine or feminine noun denotes 'der ganze Bereich (dieses) Begriffes'. So here aliTa (or TauTa) is: the traditions and everything which they stand for. As to the fact that the ' ... pronoun comes surprisingly soon after the noun which it resumes' (Dodds) I refer to K-G 1, 660f., where a number of similar occurrences of auT6s and other personal and demonstrative pronouns in the same sentence ('des rhetorischen N achdrucks wegen') are given, e.g. Hdt. 3.85.3: TWV °'1Xlwv fotrwv µ(av, -rliv b dapE(ou 'lTTTTOS lOTEPYE µaALOTa, TaUTllV ciyaywv ES TO TTpoaanov KaTe-8flcrE. 1

av

72. Cf. also, outside stichomythia, S. Ph. 553: ob8£v 1rou KciToLa8a T.MTpLa aocj,lCTJ), 721 (TOuTOyl aoct,lCET«L), Av. 1401 (xaplEvrci y' ... foocj,law Kal aocj,ci). The fact that in Classical Greek ao4>lCoµaL is never construed with 'real', external objects pleads strongly against Roux' proposal to take 1rapa8oxcis as the object of (tv)aoct,LC6µEa8a. 74. Although Ta0Ta is perhaps preferable, since this would resume 1raTp(ous1rapa8oxcis- etc. far more emphatically than auTci; and such emphasis would not seem inappropriate here.

35

I.,ine 201 contains, however, one further problem, which to my knowledge is not recognized as such in the commentaries, viz. the coordination of the relative clause as etc. with the adjective TTaTptovs (if this is the correct reading). 75 The standard grammars do not discuss this phenomenon, nor is there a special monograph on this and related subjects comparable to Ottervik's monograph of 1943. On pp. 206-7 Ottervile presents a number of cases from prose authors where an adjective is coordinated with a relative clause, mostly introduced by olos or Boos. He adds that this type is rather uncommon.76

202-3 oiJ8ds- ...

1CC1TC1PdllEL

>.cSyos-, / oiJS' d ...

E-0p1lTClL

KaTal3d>J..EL is the reading of L (P has the impossible Karnl3d>J..11); El'.JP11TaL is the reading of both Land P. Some editors change KaTal3a.>.>.Et to KaTa~1, following Scaliger; most editors follow Elmsley in reading 111'.JP11TaL, probably rightly, see note on 1024. Verdenius4 argues that KaTal3a.>J..EL 'expresses a stronger confidence than the future KaTaf3a>.E-1 '; he clearly thinks, then, that KaTal3a.>J.n has future reference. He refers to K-G 1, 138 for support, incorrectly. As appears from their examples the use of the present indicative for the future 'mit rhetorischem Nachdrucke' occurs typically in the apodosis (in the strict sense) of conditional clauses with the subjunctive or with afuture indicative. Oearly the order of main clause and subordinate clause at Ba. 202-3 pleads strongly against adopting Verdenius' view. Also, his view entails that we take 11iip11TaL as a subjunctive: if we do not, KaTal3J.f"L cannot have future reference. However, 110p11Tm should al75. As often, connective TE may convey the idea that conceptually the two items belong closely together, cf. Ruijgh 170-1. 76. As Ottervik notes, such coordinations are a quite normal phenomenon in French and other Romance languages, as well as in Latin. This may explain the silence of grammars and commentaries alike on examples like Ba. 201. In the Germanic languages such coordinations are hardly acceptable, if at all. - Roux views TE in lls TE as an instance of epic TE, but this is impossible, since the relative clause does not denote a 'fait permanent'; there are, in fact, only three certain instances of epic TE in Euripides, all in lyrics. Cf. Ruijgh 1002f. I should add, finally, that there is another way to analyse lis TE etc., viz. by taking it as referring to a different set of 1rapa8o,caC from those referred to in the first member. In that case we must supply 1rapa8o,cds with lis. This is not impossible in itself, but the creation of two different sets of 1rapa60,caC, first the ancestral ones and, second, those that are as old as time itself, is rather unnatural. Cf. also fn. 75.

36

most certainly be taken as an indicative, with passive meaning. 77 Also, the indicative in conditional clauses often conveys strong scepticism on the part of the speaker as to the likelihood of the fulfilment of the condition.78 This seems appropriate here: Teiresias does not really believe that such clever arguments have been found. 79 IfriOprirnL is, indeed, an indicative form, Karnl3.Xn cannot easily be retained, since on a natural reading it must have present reference: 'no argument is destroying them (right now), not even if ... ', which does not make sense. It might be argued that KaTal3a~L could be taken as a generic present. So apparently Roux: 'Tiresias ne fait pas une prediction; il constate un fait d'experience'. The 'fait d'experience' must be that ancestral customs are never lost. The indicative is, to be sure, occasionally found in 'general suppositions', to use Goodwin's term, see his§ 467; cf. also Moorhouse 282, who points out that in such cases 'the generality is regularly marked ... by the use of d ns'. However, in view of the extreme rareness of this use, and of the observation made by Moorhouse, it is very unlikely that El rivpriTaL expresses a general supposition. Scaliger's KaTal3a>.E1 removes all difficulties in an unobtrusive way. 80

206-7 o-0 ydp 8L'dP11X' 6 8El>s- dTE Tl>v vtov El XP1' xopdnv dTE Tl>v yEpatTEpov. The text of the mss. has often been suspected; for various emendations see the appendix of Kopff (who himself adopts Matthiae's o-!hE ..• oC,TE; so, too, Murray). It is, however, defended by Roux, Lacroix and Verdenius4, who follows Lacroix' explanation. Roux remarks: 'Bien 77. '11fipTJTaL might be middle (with the personal subject implied in >..6-yos-); but this form is usually passive' (Dodds). As for the subjunctive, the synthetic subjunctive of the passive pf. of diplaKw is not found elsewhere. In fact such subjunctives are extremely rare, see Kiihner-Blass 2, 100. Incidentally, if EUPTJT«L (11u-) is taken as a perf. subj. (Hermann) it should be written EUpfjTaL, K-B 2, 101, A. 5. In itself a bare subjunctive in conditional and other subordinate clauses is not impossible, in drama. For an extensive discussion I refer to Bers (1984: 142-165). I should add, finally, that EfipTJT«L could also be a subj. aor. middle, but this form too is quite uncommon. 78. Cf. K-G 2,466, A. 1, Rijksbaron (1984: 68-9). 79. Hermann, who retained ElipT)TaL, preferred to take it as a perf. subjunctive (cf. fn. 77), believing, wrongly, that the indicative was not appropriate,' uL in re incerta'. 80. IL is, moreover, more in accordance with Euripides' tendency to avoid a long third anceps after a resolved fourth longum, cf. Korzeniewski (1968: 58). (Although in the Bacchae this tendency is less prominent than in Euripides' earlier plays, Korzeniewski, ibid.). - The forms of 13d>.>.L.i and its compounds are often confounded in the mss. Cf. Stork (1988) for a detailed study of such phenomena in Herodotus.

37

qu'avec dTE ... dTE led soit pleonastique, ii me semble devoir etre conserve (sic Gregoire)'. In her opinion, then, dTE ... dTE apparently is an indirect question, with 8t1Jf>11X', and d would simply be a repetition of the first ELTE. 81 This seems to me rather far-fetched, and she does not provide support for her analysis. Lacroix, on the other hand, takes d 8Et (read: XPTI) xopdinv as a conditional clause and dTE ... ELTE as a double indirect question; he translates din his commentary as 'lorsque' and in his translation as 'quand'. But this is hardly possible. First, d + indicative does not have temporal meaning. 82 Second, Et TE .. . ELTE is not made into a real indirect question, since the alleged question(s) has/have no verb(s). And in his translation, in fact, Lacroix apparently considers ELTE vfov ... dTE -yEpalTEpov simply the object of8t1JPTJx': 'Le dieu ... ne distingue pas entre le jeune homme et le vieillard'. This is not impossible in itself but then I would like to advance an alternative analysis. I suggest taking d XPTI xopEVELV as governed by 8ti.lPTJX' and dTE . . . dTE as disjunctive coordinators: 'either ... or' .83 The first member precedes the dependent clause, just as it does with Matthiae's ovTE Tov vfov d ... , i.e. as a prolepsis. Compare S. Ph. 444: ToilTov otaa• d (wv KtJpE1;, below II. 859-60 note, K-G 2, 577ff. All in all this yields: 'The god has not determined whether it is (either) the young who must dance or the older man', or, more literally, respecting the prolepsis: 'The god has not determined, either with respect to the young man whether he must dance, or (whether) the older man (must dance)'. A more conventional order would be: ou yap b BEDS 8t1JPTJK d XPTI XOPfVELV ELTE TOV vfov dTE TOV -yEpalTEpov. EtTE approximates, then, to,;; cf. also Denn. p. 506 (d), Ruijgh § 180. 1

211 11'pocf>'liTTIS' Mywv D.: 'an interpreter in words to you (of what is happening)'. R.: 'Le TTpocf>rrrr1s (est) "celui qui parle ouvertement au nom d'un autre". II est done necessairement le "prophete des paroles" de quelqu'un. (-) 81. Surprisingly her translation runs: 'Le dieu, a l'heure de la danse, ne marque point de difference entre le jeune homme et le vieillard', which is rather similar to Lacroix' analysis; see below. 82. Verdenius' general reference to K-G 2, 481-2 is not very helpful. K-G argue there that d sometimes has a causal value. Incidentally, this claim seems unwarranted: in the cases presented by K-G d may come close to hrd, but it always indicates, be it for rhetorical purposes, that the speaker has some reservations concerning the reality of the state of affairs involved, a value which hrd lacks completely. 83. With 8LaLplw El one may compare 8LaKpCvw El, Hdt. 7.54.3.

38

lei, Cadmos veut simplement dire que c'est lui qui jouera le role habituellement devolu aTiresias: ce dernier decrit l'avenir aux hommes qui ne le voient pas; Cadmus va decrire aTiresias la realite qu 'il ne voit pas'. V. 4 , who rejects the interpretation proposed by D.: 'a "proclaimer of words", where "words" is used instead of "happenings" as a rhetorical contrast of "sight"'. I think Roux' and Verdenius' solutions are basically right; cf. also Wecklein's 'Vermittler des Vorgehenden'. Yet these words may also have a mildly ironical tinge: Kadmos will act as a proclaimer of just words, viz. his own words, in sharp contrast to Teiresias himself, who is the ,rp4>ip11s dL6s (Pi. N. 1.60). Teiresias may be the expounder of the will of Zeus, but in the present situation he must rely on Kadmos, although the latter has only his own words to 'proclaim'. 84 So Kadmos will assume Teiresias' role, but only for the occasion, on a rather plain level. For the general idea of this sentence one may compare S. OC 138: 4>wvij . . . bpGi: just as Oidipous sees by the voice of others, or, as Jebb ad loc. puts it: 'in sound is my sight' ,85 so Teiresias will see through Kadmos' voice. 213 8{8wµL Cf. the remarks above on T(KTEL, 1. 2. Cf. Dodds: 'the present is used because the effect of the action persists'.

215- 6 lK8T1µ0S' '3v µtv . . . IT'l1yxavov ... , I d.dw Bi ... Kirk translates: 'I chanced to be away from this land, but now I hear ... ', perhaps rightly. However, I would venture an alternative analysis. The structure of the sentence may indicate that Pentheus got his information about what was going on in Thebes while he was absent. 86 See KG 2, 232, on the phenomenon of 'rhetorical parataxis'. On this analysis the first member is semantically equivalent to a (temporal) subordinate clause, and KAUW must be taken as a historic present, as at Rh. 789. Ob84. Cf. also Dodds: 'Cadmus perhaps jests on the point that he, not Tiresias, is the 'seer' on this occasion'. 85. Which means, in this particular context, 'I know your presence by your voice' (Jebb). 86. The other translations are not very clear on this point. Roux has: 'Je me trouvais absent . . . et j' apprends •, Lacroix: 'J' etais absent . . . et j 'app rends ... '. Their commentaries are silent, as is Dodds'.

39

serve,.moreover, that if icX.uw does not have past reference the transition to 226 &ras µe-v oliv dX.T)cf>a is rather strange: Pentheus has only just now heard about the situation in Thebes, but already a number of women have been caught. If icMw has past reference, we may assume that in the meantime Pentheus has taken such measures as were necessary to catch the women.

224 ,rptSC,,aaLv R.: 'accusatif absolu, «sous le pretexte apparent de ... »'. V. 4 : 'stands in apposition to the unexpressed internal object of the preceding lines'. Roux' terminology is rather unfortunate: absolute accusatives are formed from verbs. Also, 'apparent' is redundant. Verdenius' preoccupation with internal objects (see 11. 9, 71) here leads to a very bizarre analysis indeed. 1Tp6cf>aaw is an adverbial accusative, i.e. a pseudo-accusative. Observe that such alleged accusatives (for other examples see K-G 1, 317, A. 20, e.g. dipos, vi/sos- etc., lSvoµa, ylvos-) differ in three crucial respects from real, nominal, accusatives: (a) they cannot in any meaningful way be made to depend on a verb, either directly or in apposition; (b) they cannot be modified by adjectives, or if they can, the adjective cannot be omitted (e.g. yvwµT)v tµ11v); (c) they cannot be pluralized. For the difficulties involved in distinguishing between 'accusative' and 'adverb' see the discussion of Tl; in Rijksbaron (1976: ch. 2).

224-S ,rptS4>aaw µtv ... I Bt There is a parallel in Hom. II. 19.302: ws lcf>aTo icx.aloua', brl 6e aTe-vcixovTo yvvaiice-s, I ITciTpoicX.ov 1rp64>aatv, acf>wv 6' aiiTwv tJynv ot6s nE) est une apposition ... '. Verdenius4 is inclined to follow Lacroix. There is, however, one difficulty: the position of d.vfip forbids us to take 8vvaTas Kal >Jynv ot6s TE as an apposition to 8pacrus d.vfip. The correct way of construing 8pacrt'Js is, I think, to take it predicatively, as in Roux' translation: 'Reckless(= if he behaves recklessly), an able and eloquent man turns out to be a bad citizen, since he lacks good sense'; vovv ovK lxwv elaborates upon 8pacrUS": his 8pacrt'JTT)S' consists in his having lost his senses. For comparable cases of predicative adjectives cf. Ion 14-5: d.yVWS' 8e- traTpl ... / yacrTpOS' 8LiivEyK' oyKov (temporal meaning), Med. 737: MyoLS' 8e- crvµ~as Kat 8Ewv d.vwµOTOS I lAOS" ylvoL' dv, Or. 163: d8LKOS d8LKa ... lMKEV (sc. b Aoecas), 456-8: ylpovn 8E0p' ciµLAAQTQL tro8l . .. Tvv8dpEWS', µEMµTTETTAOS ... TE ... KEKapµlvos, s. El. 521: (e-eELTTaS' WS') 8paO"E'la Kat trlpa 8lKT)S' / dpxw, Pl. La,. 182 b: oliT' av im6 'YE EVQS Els b TOOT' e-mcrTaµEvos ov&v av mi8oL (conditional meaning: 'if he is on his own'). Cf. K-G 2, lOlf. on adjectives (mostly preceded by modifiers like KaltrEp, an) used predicatively without also 1, 273ff., as well as Letoublon (1988), on 'adverb-like' predicative adjec-

wv

wv

wv;

44

tives of the type KaTEABE1v inr60"1Tov8os, and, finally, K-G 1, 614 on the predicative use of adjectives in general.99

272 f. O~TOS' 8' 6 8a{µwv KT~. A prime example of a noun phrase with thematic function. Cf. K-G 1, 47, 6 on what they call 'rhetorische Anakoluthie'.1 00 Observe also that we have a second proleptic constituent in µlyE80S', which succinctly introduces the topic of the first part of the ensuing speech (275-285). The whole lay-out of this sentence emphatically announces what will follow: 'Now about this new god: I could not describe his power, how (powerful) he will be'. 101 Cf. also on Tnprn(as, l. 173. 102 99. Apart from a brief remark by Gildersleeve(§ 627), the grammars ignore the predicative use of adjectives with indefinite expressions (as at Ba. 270) and concentrate on the type lvfoplJaav TCLS" O'ICTJVCLS" lpfiµous- ('because they were deserted'), cf. Th. 1.49.5. Observe that as soon as the article is omitted an ambiguity arises: (a) they set fire to tents because they were deserted; (b) they set fire to deserted tents (attrib.). Actually, however, there is a strong tendency among commentators to take the adjective in anarthrous phrases as attributive. A good illustration is Hom. Od. 7.51-52: 8apaci>-E"oS' -yap civi'tp h TTBO'LV ciµdvwv / lp-yOLO'LV TE>.l8EL, .... This probably means 'a bold man', but an interpretation 'a man, when bold ... ' is by no means impossible. It might be objected that semantically these two interpretations are rather similar. Perhaps so; there is, nonetheless, in principle a difference: real attributives primarily denote permanent qualities, whereas predicatives denote such qualities and characteristics as are present during and relevant for the state of affairs denoted by the finite verb. (Cf. for this topic Pinkster (1990: ch. 8)). It is this latter value which is clearly present at Ba. 270: an able and eloquent man is a threat inasmuch as he is(= behaves) reckless(ly). Observe, finally, that a generic article at Ba. 270, which in itself is fully possible, would yield an unambiguously predicative 8paaus-: 8paaus- 8' b 8waTOS" Kal >.l-yELV ot6s- T' civrip / KQKOS" ,ro).l TTJS" -yl-yVETaL. (Bothe conjectured 8paaus- 8' b 8waT6s-). 100. Lehmann (1984: 348ff.) has a full discussion of thematic and related constructions. 101. Syntactically µl-yE8oS', which in a first analysis might be taken as the object oH~ELTTE'i:v, is rather a pseudo-object, because it is replaced, so to speak, by lSaos··· foTaL. If µl-yE8oS' were the real object, Baos- ... foTaL, which is not, of course, a relative clause with µl-yE8oS', could not be construed. 102. OUToS' b 8alµwv differs pragmatically from TELpEa(as-, inasmuch as the latter introduces a new topic, whereas ohos- b 8alµwv takes up old information. In fact, theme constituents usually involve 'shared knowledge' of speaker and hearer. At Ba. 272 the word order gives prominence to the theme, to the (new) topic (µl-yE8os-) and to the subordinate clause, the latter being a specification of µl-yE8os-.

45

279 duriviyKaTo LSJ s.v. da4>lpw II Med., 3: 'bring in with one, introduce'. However, since the active has the same meaning, cf. LSJ I 3 and 4, it seems best to assign a proper middle meaning to e-laT1vlyKaTo 'introduce to the mortals in his (Dionysos') own interest', e.g. to have his own cult. Cf. also below, 590n. 286 KaTayE)JlS- VLV D.: 'not elsewhere with the accusative, but Aeschylus uses the passive

' That KaTaye->,a.w is used in the passive is inconclusive, for quite a number of verbs governing the genitive may form a passive as well; cf. K-G 1, 124, 6. The differences between the various constructions of KaTaye->,ciw are perhaps as follows. The genitive 103 expresses the idea that the laughter is directed against a person, 104 cf. the use of the preposition KaTci + the genitive of persons. In this case, then, the genitive depends on the preverb KaTa-. In the construction KaTaye->uiw TLVL (a favourite of Herodotus' 105 ), on the other hand, the dative constituent apparently is a dativus causae: someone or something makes me laugh. 106 This use is analogous to that of ye->.ciw + dative; observe that the dative cannot possibly depend on KaTa-. The preverb expresses perhaps that the state of affairs is realized to a high degree: deride utterly, to consider s.p./s.th. completely ridiculous; cf. LSJ s.v. KaTci E.V. Finally, in the exceptional construction KaTaye->,ciw nva the accusative 103. E.g. Hdt. 5.68.1; Ar. Ach. 1081 (not in Todd's index); And. 4.29; Pl. Grg. 482 d, Ap. 26 e, La. 200 b; Lys. 3.9, 14.46, 15.10. 104. In fact, the genitive constituent always refers to a person or to a personified inanimate entity (Ti'is 1r6>.Ews, Lys. 15.10). 105. (Inanimate entities:) 3.37.2, 3.38.1; (persons:) 3.155.2, 4.79.4, 7.9.1. In 4.79.4 i..dw occurs in a 'Bacchic' context. Herodotus tells us how the Scythian king Skyles starts behaving like a Greek; one of the things he wants lO do is to have himself initiated in the Bacchic rite of the Boryslheneans, the Greek settlers of the Crimea. The Scythians greatly disapprove of this plan: 2:icu8aL 8£ Toii ~aicxEuELv 1rlpL "E>J,.11m 6v£L8lCouaL' ou ydp 4>aaL olicos dvaL 8Eov t~Euplai.aµ~VEL" vOv ot,TOS' 6 8a[µwv ical TOV bµlTEpov ~QO'LXla >.E-M~l]KE, ical ~KXEUEL TE ical {mo TOO 8Eoii µalvnaL. 106. The dativus personae with icaTayEMw may also be considered a 'Dativ der Beteiligung' (S-D 140ff.), like that with µlµ4>oµaL, Eiivolw, lmnµdw etc.

46

cannot be explained on the analogy ofyEMw + acc., 107 nor does it depend on KaTa-. 108 In other words, KaTayEMw has acquired a syntax of its own, in which KaTa- may have the meaning 'down', as in KaTal3oaw Ttva, Ar. Ach. 711, Eq. 286, and perhaps in KaTapovlw, below 503. The transitive use may also have been facilitated by the existence of 8tayEMw + accusative (e.g. above 272), where the accusative does not depend on the preverb either; 8ta- probably = 'asunder, in different parts', as in 8tal3d>J.w. Cf. also K-G 1, 366, A. 13.

286 ICaTayEX4S' .•• ws- ... R. (transl.): 'tu le tournes en ridicule, parce qu'il a ete, dit on, ... '. L. (transl.): 'tu te moques parce qu'on dit .. .'. 'Dit on' and 'on dit' apparently are added to convey the idea that ws turns the reason adduced into a subjective one. Cf. also Kirk: 'you laugh at him and the story of his being sewn ... '. Whereas in principle this is right, it is not other people's words that are involved here but rather Pentheus' own: 'You deride him, putting forward that .. .',' ... for which you produce as a reason that .. .'. Cf. Rijksbaron (1976: 120) for this value of ws with finite verbs; cf. also WS' + participle, ibidem pp. 151ff., and below 729.

2 91 ota 81'! 8£65' D.: 'in a manner worthy of a god'. R. (transl.): 'un artifice tel qu'en peut machiner un dieu'. L. (transl.): 'un strategeme tel qu'un dieu peut en inventer'. V. 4 : 'Not "in a manner worthy of a god" (D.), but "as was to be expected from a god": a mortal would not have dared to oppose Hera'. Roux and Lacroix apparently take ota as the object of ..uaLv from l&r,Ke- by T6v8' Bµ.11pov e-K&8ous does not favour this solution. Also, T(&r,µl TLva 8Ld>..uaLv 'to make someone (a means of) appeasement' is as far as I know unparalleled. Normally, if in this construction the object noun refers to an animate entity, the predicative noun denotes an animate entity as well, as in 8ftval n.va alxµ11TT1v etc., LSJ s.v. B (1). 113 Of the other suggestions Usener's vdKfO"LV is rather attractive. Dodds, to be sure, rejected it, and the interpretations he proposes are, in fact, barely tolerable. Reading vdKfO"LV we can, in his opinion, translate either 'he made this into a Dion., surrendering it as a hostage to Hera's jealousy', or 'he made this a hostage to Hera's jealousy, putting Dion. out to nurse'. He adds: 'but on the former view the word-order is hopelessly contorted, 114 and on the latter e-K8L8ous is an oddly ambiguous verb to describe what Zeus did with the true Dion.'. There is, however, a third way to analyze the text as it stands, while reading vdKfO'LV. This consists in taking AL6vuaov predicatively with the implied object of e-K8L8ous, viz. the µtpos TOU al8epos (the anaphoric pronoun (T6v&) has the gender of the predicative noun, as often, K-G 1, 74). This yields: '(Zeus) made this (i.e. the µepos Toii al8tpos) a hostage, 115 by surrendering it as Dionysos to the hostility ofHera'. 116 This is substantially the view of Paley, who translates 'palming it off as the real D.' For this construction of E"K8(8wµL cf. Ar. Av. 1634-5: Tf)V 6E BaaLMlav Tf)V K6PT1V yuva1K' e-µol / E"K8oTfov e-O"Tlv. Such predicative nouns are 113. This objection could be met by talcing l8rjKE 8LCi~uaLv as = 'brought about appeasement', cf. LSJ C.4. 114. Cf. my remarks in foomote 112 on Roux' and Lacroix' interpretations. Observe also that in this translation, too, the participle is simultaneous, which makes, again, no sense. 115. Viz. of his future good behaviour (Dodds). 116. "Hpas vdKEULv = "Hpq: VfLK06011, abstractum pro concreto. The plural suggests perhaps that Hera expressed her feelings more than once, 'her fits of hostility'. Cf. K-G 1, 16f., Gildersleeve § 42 ('distributive' use of the plural of abstract nouns).

49

quite often found with the object of 8(8wµt, cf.: El. 1249: IIuM81J ... 'HMKTpav 8os d>.c>xov; Supp. 204: dyye->..ov I yMiaaav Mywv 8ovs; Ale. 1024: 8£8wµt TT)V8E OOLOl TTp6crTTOAC>V (LP: TTPOOTTOAELV BOY) 86µots; (passive) Andr. 4: cicj)tK6µllv / 8dµap 8o8e-tcra ... "EKTOpl. Observe also that the semantic value of these predicative modifiers, which can be paraphrased by 'to be .. .' (e.g. at El. 1249: 'Give Electra to Pylades as his wife', i.e. to be his wife) is very appropriate in the Ba. passage: Zeus handed over the piece of the ether to be Dionysos. As for lK8l8wµt, the meaning 'give up, surrender' is a very common one, cf. LSJ, who specify 'esp. something seized and detained unlawfully'. Finally, on this interpretation the present participle makes sense as well: Zeus' making the µlpoS TOtJ al8lPoS into a hostage could only be effected by surrendering it as Dionysos. For the instrumental (or 'coincident') use of a present participle with an aorist main verb see Hettrich (1976: 78ff., 117); see also fn. 119 and below, 351n. All in all I think there is no need to assume, with Murray, Dodds and Kopff, a lacuna after l. 293. xp6vl\l Bi vw

294-7

IJpoTol ~acf>'ftva( cf>aaLv lv Jl'llP~ AL6S', c5voµa JI.ETaan\aavTE"S', c5n 8E"4 8E"OS' ·upq. ,ro8' C:,µ~pE'UOE, auv8iVTES' Myov. R. (transl.): 'Le temps passa, et les humains, dupes d'une confusion de mots, racontent qu'il fut cousu dans la cuisse (meros) de Zeus, du fait qu'un jour il fut remis en otage (homeros) - lui un dieu! - a la deesse Hera. Ainsi prit corps la legende'. L., who prefers LP's Tpacj)fivm to Pierson's pacj)fivat, (transl.): 'Avec le temps, on s'est mis dire, en modifiant un mot, qu'il avait ete nourri dans la cuisse (meros) de Zeus, et parce qu'etant dieu, il avaitjadis servi d'otage (homeros) on a compose cette legende'. Kirk: 'So in time men say he was sewn in the "thigh" of Zeus, making up the story once they had changed the word, because he had been "hostage" to Hera, god to goddess'.

a

Roux' translation is barely tenable. First, 6voµa µnacrTT)craVTES must denote a conscious, not an involuntary action. Second, her taking the lht-clause as a modifier of cj>acrt is not supported by the word order; moreover, it hardly makes sense: 'people say he had been sewn in Zeus' thigh, because of the fact that he had been surrendered to Hera as a hostage' (?). Third, 'Ainsi prit corps la legende' is misleading: the aorist

50

ow8EVTES signals that this state of affairs is anterior, not posterior, to

another state of affairs. Lacroix' translation is in my view basically correct, 117 although the 'et' before 'parce que' is not in the Greek text. In fact, this 'et' blurs the issue somewhat, for it is precisely the establishment of the exact order of events which makes this sentence problematic. I would suggest the following translation: 'In time men say 118 he had been sewn in the thigh of Zeus, after they had changed (or: by changing)119 the word, having constructed the story because he, a god, had been hostage to a goddess'. Observe that the order in the text does not reflect the temporal sequence of events, which is: (1) Dionysos had been a hostage. (2) Therefore, people constructed a story (viz. to explain away this fact). (3) This led to their changing the word which aroused their suspicion. (4) They say now that he had been sewn in the thigh of Zeus. 120 In syntactic terms: µETaO'TT)C1aVTES modifies a.aLv, this group in turn is modified by avv8i VTES, which is modified by 6n ... wµT)pEvaE. To bring out the function of the 8n-clause vis-a-vis avv8€VTES I would prefer omitting the comma after wµT)pEvaE _121

305 µavCa 8t ical TOOT' taTl ... R.: 'on attendrait auTTI; mais il arrive que !'attraction ne soit pas faite quand le substantif attribut precede le demonstratif sujet'. 117. As for Kirk's translation, Verdenius rightly observes that Kirk is mistaken in construing ~voµa µETacrr1'aaVTES with auv8lVTES. 118. Xp6Vf¼i ... ~aw must be interpreted, I think, as 'After some time men say, although I could not find parallels for the use of xp6vl¼I + a habitual present. 119. Mnacrr1'aaVTES might, in fact, also be taken as instrumental (cf. Lacroix' 'en modifiant') with lj,amv, although in this use, the so-called 'coincident' use, an aorist participle is mostly combined with an aorist indicative. Cf. for this non-anterior use of the aorist participle Hettrich (1976: 78ff.; also 118, cb), for some examples where an aorist participle is combined with a non-aorist main verb) and Rijksbaron (1979: 247f.). The classic example of such participles is the Homeric ells d,rwv c:!TpuvE µlvos Kal 8uµov l°KciaTou. Cf. also above, 293-4n., and below, 351n. All in all it is perhaps better to take µnaa~aaVTES as anterior to lj,aaLv: 'Having changed the word they said ... '. 120. For a comparable complex structure see e.g. Hdt. 3.1 irlµlj,as ... Al-ywTl¼I. See also K-G 2, 104. 121. Roux finds the new version hardly more plausible than the first one. There is, however, a parallel in the birth of Athena from the head of Zeus.

51

Rome is mistaken both in saying that one would expect afrrr1 and in explaining the occurrence of TOUTO by referring to the fact that the predicative noun precedes the demonstrative. Attraction may but need not occur not only when the predicative noun precedes, but also when it follows the demonstrative. See the examples given by K-G 1, 74-5 (e.g. of nonattracted demonstratives: fonv ~ T00To Tupavv(s, Pl. R. 344 a; v1re-pl30Xi'i yap d6uc(as T00T6 ye-, Dem. 18.16); see there, too, on possible semantic differences between attracted and non-attracted predicatives.

309-13 dU' lµol ... in8o0· µ:ij •• cdfXE'L ••• µT16(t) ..• 86icEL ... Btxou .. . a,rlv8E •.. pdicxEvE .. . aTl ♦ou Of the present imperatives, the first two probably have a continuative value: 'do not proceed with the state of affairs you are currently involved in', 122 whereas the last four rather stress that Pentheus should start carrying out these (new) states of affairs immediately; at the same time they may also imply that he should repeat them, thus making a habit of them (as opposed to the aorist imperative, which would both lack the inceptive value and might refer to a single occurrence of the state of affairs): 'accept the god straightaway and repeat your accepting him', i.e. accept him on a permanent basis, 'make libations straightaway and repeat making them', etc. Cf. above on 173ff.1 23 As for ci>J..' lµol m8o0, this clause marks, after Teiresias' rhesis, the 'transition from arguments for action to a statement of the action required' (Denniston 14 on ci>J..ci).1 24 The aorist probably indicates that Teiresias implores Pentheus to obey. Note, in this connection, that m8o0 is often accompanied by intensifying expressions like >J.crcroµm, ,rpos Twv 8e-wv etc., e.g. S. OT 650 and 1064 (>..loooµm), 1432-4 (1Tp0S' 8e-wv), Ar. Av. 661 (vi'! at'),125 V. 760 (1TpOS TWV 8E"wv), Pl. 103 (civTLl3oAlii). Cf. also S. OC 1441, where Antigone with ci>J..' lµot m8o0 resumes her initial appeal (1414): lKe-Te-uw Of ,rno&fivaL Tl 122. Cf. for this value Balcker (1966: 35ff.), Rijksbaron (1984: 43f.). 123. For a similar case of Bixou see Ar. V. 736: 8£ 1rapwv Bixou; note the presence of 1rapwv. 124. Similarly Cy. 309, Hel. 323. 125. Incidentally, the next lines nicely illustrate the inceptive use of the present imperative. After Peisthetairos' request to the Epops iKf3lf3aaov ... Toiipvl9Lov· iK· IY~aov ... ,rpos 9Eiiiv, the actual order as given by the Epops to the little bird is: lK~Ll,lf (666).

av

52

µoL. 126 Also, m8o0 is mostly spoken by someone who is in a subordinate position vis-a-vis his/her addressee, e.g. S. OT 650 (Chorus to Oidipous, dva~ (650)), 1064 (lokaste to Oidipous), Cy. 309 (Odysseus to Cyclops), Hee. 842 (Hekabe to Agamemnon,@ 8fo1roT' (841)), Hel. 451 (Menelaos to the Old Woman, who is 'in charge', 443: lµol ... TOUTO ,rp6aicnTm ... / µT18lva ,re-MCnv ... ), Ar. V. 760 (Bdelykleon to his father), Pl. 103 (Chremylos to Ploutos). Ild8ov, on the other hand, 127 is never modified by such intensifying expressions, and is often spoken by someone who by his status is in a position to give orders, in which case it has inceptive meaning, asking the addressee to obey forthwith, Hel. 1393 (Theoklymenos to Helen), Rh. 993 (Chorus-leader to the Chorus). If the speaker has a subordinate position, ,rd8ov rather has conative meaning, Hel. 994 (Menelaos to Theonoe; clearly not 'obey' or 'listen to' but 'be persuaded'), Antir. 233 (Chorus to Hermione, 8fo,rmv' (232); note the qualification provided by 'as far as an easy opportunity offers' in 232, if this is what this line means),128

a

126. At 1442 Polyneikes reacts with µ-fi irEte' µ11 6Et, which apparently combines continuative meaning with conative meaning: 'stop trying to persuade me ... ', thereby cutting short possible further attempts of Antigone to persuade him. 127. 1rd8011 is the sole reading of the mss., and is printed by Murray and others, at Hel. 994, 1393, Andr. 233 (where m8o0 is excluded metrically), and Rh. 993. It is also found at Rh. 594, where, however, most edd. accept Wilamowitz' ird8ns. In four cases both ird8ou and m8o0 are transmitted, both being acceptable metrically, viz. Ale. 792 (where P, however, has 1rC8011, not ird8ou), 1101, 1109, and Or. 1101. See also next note. 128. At Ale. l 101 and 1109 I would prefer (inceptive) m,l8ou to m8o0 (which is read by most editors). Herakles has been trying for some time to persuade Admetos to accept 'the woman' into his house, but Admetos is very reluctant I think that after Herakles' 8l:,cou at 1097 a blunt ird8ou both at 1101 and at 1109 is more forceful than m8o0. Note the impatient µ6vov at 1109.

53

As Ba. 3()1) shows, the 'imploring' aorist imperative can very well be combined with the inceptive present imperative. They each in their own way convey a sense ofurgency.129

315-6 tv i1J

♦-6aEL I ••• lvEOTLV Dodds thinks that this, the mss. reading, can only be retained by assigning to lvE"OTL the meaning 'depends on', but because in his opinion this meaning is 'unexampled', he prefers Bergler's E"l to lv, putting a comma after ad (whereby E"l ..• ck( becomes the object-clause of O'KOlTE"Lv); this is also Kopff's text. Thus lve-aTL apparently (Dodds is not explicit on this point) means 'is present in'. Although Dodds is mistaken as to the possibility of lvrnTL lv meaning 'depends on' ,130 there is no need to give it this meaning. It may just mean 'is present in', i.e. 'is a d,aracteristic of', as in Ar. Eq. 1132: d O'OL 1TUKv6T11S' lve-OT' lv Tlf) Tp61TC\>, and Ba. 269: lv TOLS' >hyOLO'L 6' oliK lVE"LO'L O'OL tf>plve-s.

129. Cf. S. OT 1064: ISµti>S' m9o0 µoL, MaaoµaL, µl'i 8pdv Td&. - Interestingly, while Plato has no examples at all of the formula d.Ua m9o0, there are six instances of the formula d.Ua ,rd9ou (Cri. 44 b 6, 45 a 3, 46 a 8, Euthd. 287 d 3, Grg. 486 c 4, Phd. 111 a 3). Four of these are of the form d.ll' e-µol 1rd8ou (or ,rd9ou µoL), three of them from Crito, 44 b 6: d.ll' ... ln Kal viiv 1rd8ou Kal aw!hJn (where Burges conjectured m9o0, needlessly, I think), 45 a 3: d.ll' Eµol 1rd8ou Kal µl'i 4llws ,ro(n, 46 a 8: d.lld. 1raVTL Tp61Tl\l, ~ J;wKpaTES", 1rd8ou µoL Kal µT)8aµiiis 4>Jw,s ,ro[EL. The latter two examples occur at the end of a speech. The use of the present imperative is in accordance with Plato's general practice in such transitional formulae, cf. the examples given by Denniston, pp. 14-5. Usually such imperatives have an inceptive value, sometimes with the secondary semantic value 'and continue doing so'. Observe that Krito (who was of the same age as Sokrates and is included in Xenophon's list of true Socratics (cf. Burnet, Introd. note to his Comm.)), more or less orders Sokrates to listen to him. The urgent tone is in conformity with Krito's passionate appeal at 46 a 4ff.: ~ou>.Euou - µdllov Bt ovBt ~ouMurn9m ln ~pa dlld. ~~ouMiia9aL - µ(a BE ~u>-1'· TilS" yap E1TLOUalJS" vuKTOS" 1rdVTa Ta0Ta &t 1TE1Tpcix8aL, El 8' ln 1TEpLµEvoiiµEv, d.8uvaTov Kal oucln ot6v TE. Note that the tone of Krito's appeal becomes ever more urgent: d.ll' ... ln Kal vOv ,rd9ou (44 b 6) is followed by dll' Eµol ird9ou Kal µl'i 4llws ,rolEL (45 a 3), and this, in tum, is followed by dU.ov 8£1CT1V µiTE"LµL Elmsley's 8llCT}v for LP's 8llCT} is generally accepted and gives, in fact, a fine idiomatic construction. As Dodds and others suggest, 8llCT}V should be taken as an internal accusative; for µnlpxoµm as a legal term= 'prosecute', cf. LSJ s.v. IV 2. One may also compare 8llCT}v VLKQV etc., K-G I, 305-6. See also on 65-6 and 247. More difficult to explain are cases like Cy. 280-1: ~ (sc. -f\KETE) ~ KaKlO'TT!S' oi. µETI)A8E"T' a.ptrayas / 'E>J-vris ~Kaµav8pou ydTOv' 'DJ.ov tr6>.Lv; and Or. 423: ws Taxv µni\>.86v a' atµa µTJTE"pos 8ml. Here, a.ptrayas and atµa are not just internal objects but rather refer to the cause which makes the subject of µETlpxrn8m pursue the person or entity referred to by the second accusative. Syntactically the double accusative should perhaps be analysed as a blend of µETlpxoµat TL 'seek to avenge', cf. rr 13-4: TOUS 8' iJj3pLa8evms yaµovs / 'E>J-vris µETE>.8€1v, and µETlpxoµal nva 'prosecute'. Compare also the parallel construction ofnµwpoi:µal nva Tl, Ale. 733: El µTj a' a&>.tpiis atµa nµwpTjanm. Greek rather liberally allowed such ditransitive constructions, involving an accusativus personae and an accusativus rei, alongside monotransitive constructions involving either one or the other accusative. See further K-G I, 327, A. 9.

346, 348 aTE'L,CiTra> TLS', TpLa(vou For the transition third ps. imperative : second ps. imper. see on 173.

3 51 s,1eoµaL 8pdaas- Td8E". Verdenius4 refers for this, the so-called coincident or instrumental, use of the aorist participle to K-G I, 199, S-D 301 and Barrett on Hipp. 289. However, the remarks of K-G and Barrett are somewhat misleading, since they are both inclined to view the 'normal' aorist participle as 56

expressing pastness. 131 Actually, however, it expresses anteriority, which, in turn, is a realization of the semantic value 'closedness' of the aorist stem. The peculiarity of the construction exemplified by Ba. 351 lies in the fact that 6paaas is not anterior to 6ii~oµaL but coincides with it, denoting the instrument by which the 'biting' comes about. The differences between this and the 'normal' use of the aorist participle can perhaps best be demonstrated by the pair:

iJy; d,rwv 6TpW€ µlvoS Kal 8uµov E"KciT' l1T0Tpuvw11, e.g. II. 20.364, where it has iterative value. 134. Verdenius remarks: 'called an attribute by Roux, but her translation shows that the word is a predicate'. There is, however, no contradiction between her terminology and her translation, for in the French grammatical tradition 'attribut' is used where other languages have 'predicate' or 'predicative', the French equivalent of 'attributive'

57

So too other commentators and translators, e.g. Kirk: 'after seeing a bitter end to bacchanals in Thebes'. But these translations are deceptive, for it is precisely the definiteness marker 'his' (Dodds) which is absent from the Greek text, nor does the sing. j3a.Kxe-vow permit us to translate 'bacchanals' (Kirk). IfmKpciv is predicative our translation should run: 'having seen a bitter end to a bacchanal in Thebes', which makes no sense. Incidentally, the 'end' in 'bitter end' is not present in the Greek either. Nor are the parallels adduced convincing. In Ba. 634: mKpoTa.Tovs l86vn 8rnµo1JS' TOUS' e-µous, where TTLKpOTO.TO'US' is clearly predicative, the presence of the definite noun phrase 8e-aµous T. l. makes all the difference. Similarly, in Med. 1388 mKpas TE"AE"VTpe-vwv ('you lost your head') and µa(VlJ d>..yLom (326, 'you are behaving madly'), and may express 'a permanent state of derangement', it primarily expresses, I think, a higher degree of folly ('intensifying' perfect, so Verdenius4 ); 136 note also the presence of-f\8TJ. So: 'you have

ws

being 'epithete' or 'qualificatif. Some consequences of this deplorable confusion are briefly discussed by Letoublon (1988: 162f.). 135. I should add that mKpdv might be predicative if j3dKXEVaLv h 9T\j3aLS' is taken together as one 'Gesamtbegriff, in which case the article may fail both with the head noun and with the attributive modifier, if the latter is a proper name. Cf. Th. 2.2.1: µna Ellj3o(as d>.waLv, K-G 1, 607, k. In such cases, however, the noun phrase refers to an event or entity which is either known from the context or has unique reference, as at Hdt. 3.26.1: KaTa 'EUT}vwv y>.waaav. Neither of these features would seem to be present at Ba. 357. 136. Lacroix ('contrairement a Dodds, nous pensons que µlµovas (sic) indique une folie plus complete et plus definitive que l'aoriste l~foTTJS' cj>pEvwv') has misunder-

58

now reached a state of complete derangement'. In fact, from the two statements quoted above we may infer that µalvoµaL was conceived of as a gradable verb. As with other verbs denoting, in the present stem, an emotional or intellectual state, the perfect may denote a state which is more complete, so to speak, than the state denoted by the present, thereby emphatically conveying the idea that the subject cannot go further up the scale. This may, but need not, imply that the state is of a more or less permanent nature. A good example is Tl cre-crLW'ITT)Kas, Ar. V. 944. I:Lw1rav itself, of course, denotes a state as well; cre-crLW'ITT)Kas, then, expresses that the subject is 'dead silent'. Observe that crrnLw'ITT)Kas at V. 944 does not refer to a permanent state, auJmqs, too, would have been possible; cf. the semantically equivalent Tl crLy4s; at Ra. 832.137 As for µalVJJ ws dAyLOTa, when compared with the perfect, this expresses a lower degree of madness, again implying that the madness may be of a temporary nature. To bring out his nuance µalVJJ may indeed, as suggested by Dodds, be rendered by 'you are behaving madly'. 138

360-2 lfaLTWµE8a I tmlp TE ,rd ~E" ws

TOVTOtJ

•••

I tmlp TE

The middle indicates that Teiresias is hoping that he and Kadmos, too, will profit from the action they want to undertake for the sake of Pentheus and the city.

364-6 1TE'Lpw ••• 8otJAEtJTlov D.: 'the old men are deeply discouraged, and in their discouragement the Dionysiac illusion of youth (187ff.) has abandoned them'. Roux, who is followed by Verdenius4, rejects this psychological explanation and prefers taking lines 363-365 as referring to the logeion, i.e. the steps which connect the skene with the orchestra. This may be all stood Dodds' note, for Dodds, too, thinks that µlµTJV«S' expresses 'une folie plus definitive'. 137. Note that of verbs which in the present stem are non-stative, and more specifically tenninative, only the perfect denotes a state, e.g. dalpxnaL 'he is coming in' vis-a-vis ElaE~~~u8£v 'he has come in and is now present'. Accordingly, such perfects do not have intensifying force. 138. Elsewhere the present stem has a comparable meaning, e.g. Ba. 130: µaLV6µEvoL ~dTupOL; HF 952: iraCCn irpos -fiµiis 8rnir6TTJS' 'f\ µaCvET«L; Cy. 168: lSs "(E ,r(vwv µT) yl"(TJ8£ µa(vET«L; Med. 873, IA 389 etc.

59

right, but nevertheless it can hardly be denied that the difficulties voiced by Teiresias are in sharp contrast to the self-assured and vigorous tone of his own (and Kadmos') words at 187ff. On the basis of these lines we might expect the two old men to descend these steps effortlessly; if they do not, this can only mean that the statement lm>.EAfiaµe-8' T)6Ews / ylpoVTE"S' 6VTE"S' is now void. Having failed to bring over Pentheus to their side, they feel very uneasy about the consequences and, as the wording suggests (tTw 6' Bµws, especially 6011>.EuTlov, 366), 139 they not only feel discouraged, or disillusioned, as I would prefer, but are also quite unwilling to leave. Dodds' interpretation of these lines must therefore be accepted.140

367 lhrws- µ~ Cf. for the adhortative/prohibitive construction B1rws (µfi) + future indic. Amigues (1977: 64-78). On p. 69 she observes that the instances from drama 'se rencontraient tous dans des passages de style soutenu et generalement a un moment crucial du drame'. 369 TOLS' 1rpdyµaaw D.: 'by the facts'. R.: 'L'experience, le sens commun enseignent qu'un dieu outrage se venge'. Dodds is certainly right; 'the facts' relate to Pentheus' behaviour, asappears from the explanatory µwpa yap µwps Myn: fact is that Pentheus is a fool, from whom anything may be expected.

369 µ.wpa ydp µwpos- ~iyn R.: 'ses propos fous sont d'un fou'. L.: 'car cet homme est un insense et ses propos sont ceux d'un fou'. K.: 'for he who speaks folly is himself a fool'. Roux' and Kirk's translations are wrong, inasmuch as they suggest (a) that this is a general statement about 'a' fool, and (b) that µwps is some 139. Significantly Dionysos is called, at I. 366, b ~L6s: his power derives from 2.eus'. 140. It is not clear to me why Verdenius thinks that 364 refers to 'the miraculous power of the god'. Roux' treatment of these lines is methodologically questionable; seen as a mere 'indication d'un jeu de scene traditionnel', they are reduced to some kind of intennezzo, and deprived of all structural meaning.

60

sort of predicative noun. Lacroix' translation is much better: the subject of >.lyn is Pentheus, who is both a fool and says foolish things. Syntactically µwpos functions as a predicative adjective to the subject of >.lyn; keeping more closely to the syntax I would translate: 'Being a fool he speaks foolish things'. For predicative adjectives see also on 270 and 792.

378 6s- Td8' lXEL D.: "'this is his kingdom"'. R.: "'qui possede cela", qui a cela en apanage'. L.: 'qui possede ce pouvoir'. V. 1: 'lxn is equivalent to TTaplxn '_1 41 Lacroix' translation is better than those of Dodds and Roux, who apparently take lxn simply as possessive, thereby ignoring the fact that Ta& announces three infinitives. Ve,rdenius' view rests upon taking the worshipper as the subject of8tacrd,nv and ye->.acrat. 142 But this will not do, for the subject of dTT01raOcrat can hardly be anyone but Dionysos himself; cf. lines 280ff., of which the present lines are a partial echo. If Verdenius were right, the result would be a syntactically highly implausible change of subject. Furthermore, the parallels adduced by Verdenius for EXEL = TTaplxn all involve a concrete noun (µaVTLICT)V at Ba. 299, cre-µv6TT)T' at Ba. 486, it8ova.s at HF 732). We are dealing here with the construction lxw + infinitive, 'have means or power to', LSJ s.v. A. III. If an infinitive which normally would depend directly on a finite verb is prepared by a kataphoric pronoun like Ta8e-, the subject of the infinitive must be supplied from the finite verb, as appears from the examples presented by K-G 2, 4, c. Cf. e.g. Pl. Smp. 192 d: clpa yf. TOO& E"TTt8uµe-1TE", E"V Tcj> aim'ji ye-vfo8at;.143 141. So, too, Kirk: 'these are his gifts'. Something similar was proposed by Dalmeyda: 'qui possede, c.-a.-d. qui nous donne' (where the 'c'est a dire' betrays a rather remarkable view of 'possession'), and by Tyrrell: 'of whose gift it cometh'. 142. This was already proposed by Tyrrell, who supplied 'men' as the subject of8LaaeueLv. 143. As to the question of whether 8LaaeuELv is intransitive or transitive, I think this matter can hardly be settled. However, the parallel adduced by the proponents of the transitive view (e.g. Dodds), viz. Ion 552, is false, for there an explicit object is present: l'>s µe ... l8uiaeue. It is perhaps best, then, to take it intransitively at Ba.

379.

61

379-81 8LaaEVELV ••• yE>.d.aaL ..• d,ro,raOaaL The aspectual differences may suggest that 8Lacrdew is the framework within which the other actions occur: while the 8LacrEunv continues Dionysos now laughs and then again puts an end to the worries. 144

389f. 6 8t Tiis- f)auxCas- JUoTos- KTA.. Compare for this idea Hdt. 7.18.3: yvwµT)v Elxov dTpEµlCoVTci aE µaKapLcrTov Elvm ,rpbs iraVTwv dv6pw,rwv. Dodds notes: ' ... orgiastic religion is not, to our thinking, particularly ijcruxov, and one is tempted to discover here a secondary reference to the contemporary wartime controversy about 'quietism' or the 'ivory tower' attitude ... '. But we should not connect b Tiis riouxtas j3toTOs with 'orgiastic religion' in the first place. This phrase rather refers to the life of someone who does not meddle with other people's affairs, and is the positive counterpart of dxaX(vwv crToµciTwv etc. This 'someone' clearly is not Pentheus, and the implicit message is, therefore, that only by not interfering with other people's affairs, i.e. with Dionysos', Pentheus' life will remain unshocked. The use of the attributive genitive TOS' ricruxtas instead of the adjective ijcruxos is discussed by K-G 1, 264. They observe that 'wahrend das Adjektiv nur die Eigenschaft eines Substantivs ausdrtickt, der G en et iv vielmehr das Wes en desselben nachdrticklich bezeichnet'.

ircSpow ydp c5µws- al8lpa va(ov-

392-4

TES- 6praow Td ~pOT(dV oilpav(8aL. R.: '&µws, construit proleptiquement, porte sur bpwaLv (Kuhner-Gerth, II, p. 280, 4)'. Neither the note, nor the reference to K-G is very illuminating; this peculiar use of &µws is discussed by K-G at 2, 85, A. 8. In those cases where Bµws precedes rather than follows the participle, this would seem to convey that the contrast between the content of the participle and that of the finite verb is rather striking. Some examples

are:

Hdt. 7.165: Mynm ... ws Bµws Kat µlhXwv dpxEcr8aL imb AaKE6mµovtwv b rl>.wv ll3ofl8f1crE av ... , d µ11 ... 144. Cf. for the semantic differences between present and aorist infinitives after ability verbs Stork (1982: 325-347, esp. 346).

62

Pl. Phd. 91 c: };Lµµ(as ... 4>o~e-tTaL µri ,; \ltvXTI &µws Kal 8n6Te-pov Kal Ka>.>J.ov ov Toii awµaTOs TTpoaTTo>J.."'1TaL. Here we should probably construe the first Kal with ~v. 145 An example without Kalis: Hdt. 8.74.1: ol 8e- EV lli>.aµtvL Bµws- TQUTQ TTVv8av6µE"VOL appw8rnv.146 See further K-G 2, 85, A. 8. Such strong contrast would, indeed, seem to be present in our line as well: even though the gods live far away (as is common knowledge; the addition ofKal would seem possible), yet .... As for the syntax of TT6paw . . . lSµws- al8lpa valoVTE"S, the position of &µws- between two constituents of the participial phrase is also found in Ion 734: lyw Bl a' ... I 8foTToLV' Bµws olia' aVTLICT}8d,w TTaTp6s. 145. At Hdt. 7.165 and PI. Phd. 91 c the front position of llµws- is perhaps also to some extent due to its contrasting directly with the preceding context. The participle restates, for clarity's sake, that information from the preceding context with which the finite verb contrasts. Observe, in this connection, that the possibility of Gelon being ruled by the Spartans has been mentioned at 7.163.1. The additive-concessive adverb Kal emphasizes the contrast between the two parts of the sentence. It would seem that Kal occurs in particular when the content of the participial clause is, or is considered to be, known from the context or situation. It is for this reason that I view the first Kal at Phd. 91 c as modifying !iv: it is, of course, taken for granted, in the Phaedo, that the soul is more divine and more beautiful than the body. The same applies to d Kal (cf. Denn. p. 300 (i): the conditional clause may express 'an admitted fact'), to Engl. even though and Du. ook al. Compare also e.g. Hom. fl. 8.125: Tl>v µtv lirEu' daa£, Kal axvvµiv6s- TTEP halpo11, which comes after "EKTopa 8' alvl>v dxOS" TTVKaai ,plvas- T)vL6XoLo; similarly 8.317, 15.133 (refers back to 110, ,rijµa, and 114, l»,o,11p6µivos-), and, with ou8l ... TT£p, 13.419 (cf. 417). That Kal need not be present appears from 22.424, where axvvµiv6s- ,rip is preceded, in 422, by d~:yi' l8rJKE. On the other hand, when no mention whatsoever is made in the preceding context of dXOS" and the like, we do not find Kal. Cf. fl. 1.241, 588, 17.459. See for the uses of Kal and Kal (... ) TT£P with participles Bakker (1988); he ignores, however, the relationship between the context and the occurrence ofKal, and thinks, wrongly in my opinion, that 'the presence (or absence ... ) of kai is mandatory on metrical grounds' (1988: 140-1). 146. Here, the information presented by the participle is new; Kal would, indeed, not seem acceptable.

63

406-8 IIdcf>ov 8' 411 haT6aToµoL flapfldpou ,roTaµoO ~oat icap1rCCouaL11 clvoµflpoL This passage is, as Dodds puts it, 'a standing puzzle'. Among others the following analyses have been put forward:l47 (1) The above text, which is that of the mss., is retained; thus e.g. Murray. 8' coordinates K(nrpov (403) and Ilciq,ov. Thus Dodds and Kirk. This entails two problems, which taken together are to my mind fatal to this analysis. First, although in itself the idea might perhaps be defended that in some way the Nile was sometimes conceived of as being the mother-river of the springs of Cyprus (see the extensive discussion by Dodds), I think it is meaningless in connection with Paphos. What may be the relevance of stressing this quality of Paphos in particular in the present context? That is, if it is a quality of Paphos, for this town appears not to have been renowned especially for its fertility. Second, the coordination of Ilciq,ov and Kt'.mpov strongly suggests that besides Paphos the Chorus wish to come to other, unspecified, parts of Cyprus as well, which s~ems slightly irrelevant. What we would rather expect is a specifying apposition, as in Hom. Od. 8.362: iJ 6' dpa Kutrpov YKaVE" q>L>..oµµn6ris 'Aq,po6lTll, I ES Ilaq,ov.1 48 (2) Read cl>cipov (Reiske) 8' av; thus e.g. Roux. cl>cipov has then to be taken metonymically as = Egypt. Although a reference to Egypt would seem entirely appropriate (see also below), I strongly doubt whether in this Dionysiac context Pharos is the most suitable candidate for being the carrier of this metonymia.1 49 Moreover, the qualification of Pharos as 'fertile' is quite unexpected (cf. (1)).150 147. In discussing these proposals I start from the assumption that ,roTaµoO ~oa( refers to the Nile; accordingly the emendations which aim at getting rid of this reference are not taken into account. 148. Denniston remarks: 'Greek sometimes employs connexion by TE where English prefers an appositional construction' (Denn. p. 502 (e)). But in the cases discussed by him, the two nouns or noun phrases denote, generally speaking, two different properties of one and the same referent, e.g. IA 1454: 1raTlpa Tov dµov µ'l'i ov, lfv 8'. Thus e.g. Kopff, after Shilleto and Verrall. ITd.ct,ov is then the object of vlµovTm, which is in conformity with the use of vlµoµm elsewhere in tragedy (cf. Dodds).1 51 "Av 8' has then to be taken either as an autonomous relative clause, referring on its own to Egypt, or as a relative clause to a head noun which has to be supplied. But the question is: to be supplied from where? If, with such a 'headless' relative clause a head noun has to be supplied, this should always be present in the context. Two clear examples are: Hel. 1048: -fiv yap dxoµEv 8aMO'O'' txn, SC. vaOv, mentioned at 1047: a.XX' ou8E µriv vaus foTLv 'ft aw8E'LµEv dv I ct,d,yoVTES', Pl. R. 327 a: OU µlvTOL ~rrov l-ct,atvno trpltrnv fiv ol 9pov 8' liv ov must then be considered a gloss on 11. 402-405, which has ousted the word for 'land'. Perhaps this points to yai.av 8' av, since palaeographically this comes closest to Ilaq>ov 8' av (fAIAN8AN "'IIAON8AN).157

408-9 ... 4voµ~pOL. I o~ 8' d. Thus, i.e. with a full stop after dvoµf3poL, e.g. Murray, Dodds and Kopff, and apparently also Kirk; oli is Schone's conjecture for LP's 8trou, which is metrically impossible. With this punctuation, oli 8' 154. Ilciipov, itv 8' is also objectionable metrically, since it makes I. 406 a telesilleum with a 'resolved' first element. This, it is true, is also what we have at 421, if LP's taa 8' is retained, but since the second strophe and antistrophe do not contain any other telesillea, it is perhaps best not to introduce one per coniecturam at 406 (and to read, with/, taav 8' fortaa 8' at 421). 155. Cf. for the connexion between Dionysos and Egypt also Hdt. 2.42.2; H. hymn in Dion. 1, 28-9. 156. Willink (1966: 222) would read -f\ x86v' llv, arguing that'-/\ is positively required unless the Chorus are proposing a tour. The gist of the strophe is "Oh to escape to Cyprus or Egypt; but my real longing is for Pieria"'. This may be true (see also below on 408-9, Bl), but do we really need ,1\ to arrive at this interpretation? 157. For the responsion yarav 8' "'raav (421) cf. /on 184"' 194, 205 == 219. Note that x86va would yield a telesilleum, for which cf. fn. 154. - Theoretically there is one more way of analyzing Ilcicj>ov, itv 8', viz. to take TE as an instance of epic TE. However, in view of the rarity of epic TE in Euripides, such an analysis is, I think, indeed a theoretical one. Cf. footnote 76. Also, of course, the objections raised above against viewing the relative clause as a modifier of Ilcicj>ov are not removed.

66

functions as a pre-modifying locative clause to d:ye- at 412; it is taken up by h:e-iae- at 412. Cf. Kirk: 'where is lovely Pieria, thither bring me.' But this can hardly be right, for it is not clear why the Chorus should ask Bromios to bring them only there, and not also to the other places mentioned, viz. Cyprus and Egypt. The oli-clause is, rather, governed by lKolµav: 'may I come to Cyprus ... to Egypt ... and to where Pieria is'. Thus e.g. Roux and Lacroix, although they read oli 8', with Schone, needlessly, I think, cf. Denniston 513. The mild contrast conveyed by Bl possibly suggests that the Chorus consider Pieria more important than the other two locations; observe that Pieria (and Olympus) play a prominent role in the second stasimon (560ff.). For locative clauses as complements of verbs of coming and bringing see K-G 2, 443; some other examples are S. El. 380, S. OC 188-9 (dye- vvv µe-, trai, 'iv' ... ), 273 (vvv 8' ... lK6µ11v '(v' lK6µ11v), Th. 2.11.9; also below Ba. 1381-2: d-yn' ... µe- ... 'iva AllliJ6µe-8', 1383-4: l>,8oLµL 8' 8trou µftTe- KL8mpwv uylj n68a

Dodds and others supply ~v to npcios, although this, as Dodds himself notes, is a 'relatively rare ellipse'. In fact, K-G 1, 41, A. 2 give only one Homeric instance (Od. 5.477), and two from classical Greek (Pl. R. 503 a (read: b) and Aeschin. 3, 71); Friis Johansen and Whittle note (on A. Supp. 296) that 'omission of ~v ... appears not to occur in stichomythia'. Roux proposes an entirely different analysis; in her view we should supply EaT(v with npcios (and also with wxp6s in 438), because the guard, 'tout excite par les evenements extraordinaires', intersperses his narrative with statements about the present situation. This analysis is erroneous on at least two accounts: (1) there is no indication whatsoever in the text that the guard is 'tout excite'; (2) the proposed syntax is impossible. I venture to suggest that the adjective once again is used predicatively (cf. on 270 and 369); accordingly, ov8l is adverbial = 'not even, not so much as'. The result is: 'This wild beast here, being gentle to us, did not even withdraw his foot'. vyfj should probably be taken as a modal dative, lit. 'in flight', as in Hee. 1065: not Kat µe cf>vyq. 1TTWOOOVOL µvxwv;, and Supp. 930f.: . . . nplv M1TWV Ka6µov 1T6A1.v

I cf>v'Yfl 1Tf)OS' "Apyos 6Lal3a>.Etv au8mph41. On irrrla,raaev Roux remarks: 'aoriste inchoatif: "il n'a pas cherche a

tirer secretement (im6) son pied par la fuite"'. This is, again, impossible. First, 'ii n'a pas cherche a' should rather correspond to a conative imperfect in Greek; second, only aorists of verbs with a stative meaning (like l3aaL>.Euw, yeMw etc.) can have ingressive value.

68

438 ra\xpos Again, used predicatively. One should follow Murray, therefore, and not put a comma after xtpa.s . 158 Observe that -yE">-wv in the next line is without any doubt a predicative modifier; it would seem natural that the other activity verbs (imfo,raaE"v, l6wKE"V) are modified by 'behavioural' adjectives as well. 440 dirpE"irts iroLo-6µ.E"vos D. (who reads drrprnts 1TOLouµE"vos): 'the use of the middle ... seems merely to stress the Stranger's personal participation in the action, as in 1r6>.Eµov 1TOLE"L08aL and similar periphrasef. R.: 'rendant en ce qui le concerne (d'ou l'emploi du moyen 1TOLouµE"vos)'. Both comments are wrong. First, we are not dealing here with a periphrastic construction, but with an adjective used predicatively with an object. Periphrastic constructions involving 1TOLE"La8aL consist typically of a form of 1TOLE"'tcr8aL and a bare noun, i.e. a noun without a modifier, like, indeed, 1r6>.Eµov, or dpTJVT]V, 6lm Tav, vavµaxlav etc. Second, 'en ce qui le concerne' is not a meaning normally expressed by the middle. The correct analysis is that of Tyrrell (who reads dmprnts): 159 'turning for himself my task to seemliness', i.e. the Stranger acted in his own interest by making my task a dignified (or: easy) one'. 445-6 cf>poii8a( y' licdvaL ~d.vµlvaL 1rpos 6pyd8as

OICLpTciJOL R. (transl.): 'elles sont dehors celles-la, libres de leurs liens, et bondissent vers les halliers sacres ... '. K.: 'they are gone, those women, released, and to the mountain glens they are dancing their way ... '. L. (transl.): 'elles sont parties; leurs liens ont disparu et elles bondissent dans la campagne ... '. 158. Dodds mysteriously remarks that 'the adverbial use of wxp6s is not so natural as that of 4Kwv ', and for that reason he prefers supplying ~v once more. Regrettably, Dodds does not amplify his statement Cf. for predicative wxp6s also Theoc. 14.5-6: TOLOVTOS'

,rpwav

TLS' dcf,(KETO Ilu&ayopLKTQS', /

wxpbs

Kciv111T68TJTOS'.

159. Roux incorrectly says that all editors have accepted Canter's EVTpEirls.

69

In their commentaries Roux and Lacroix are silent on the syntax of these words (as is Dodds). There are, I think, two ways of construing them. First, cf>poOBat -y' EKELVaL may be taken as a verbless sentence. In that case a semi-colon should be printed after EKELVaL (as far as I know only Bruhn punctuates in this way), and trpos 6p-yd8as depends on O'KLpTWO'L, so literally: 'they are gone, those women; having been released they are now dancing to the glens' (Lacroix' 'dans' is impossible). Alternatively, cf>pou6a.L may be taken as a predicative modifier with M>..uµivm; this as a whole modifies O'KLpTwaL: 'having been released and gone away to the meadows they are now dancing ... '. With a finite form of>..uoµm we would have cf>po08m >.l>..uvTm: 'they have been released and are gone'. For this use of cf>po08os cf. Ar. Pax 197: cf>po08oL yap tx8£s dow fe(¼>KLC7µEvOL; 160 and for the directional complement with cf>po08aL Andr. 73: cf>po08os 8' fo' auTov MEVEMWS 86µwv dtro, Med.1110:cf>po08os is "AL8ou 8avaTos.161

451 µ.a(vE"a8E"· XE"Lpraiv To08' Iv clpicuaw yelp ~" This, the reading of LP, is defended at length by Tyrrell, who takes Toii8' as referring to Pentheus himself; cf. the addition in L oflµoii supra T008'. He objects to Burges' µi8rn8E xnpwv Toii8', which is adopted by most editors and commentators, ' ... is it not ludicrous to suppose that the attendants were holding on to the hands of their gentle and unresisting captive?'. He correctly points out, moreover, that µi8rn8E xnpwv Toii8' could not mean 'let loose his hands', which would be µi8ETE XELpas roii&. In fact, µE8lEµm + genitive of concrete entities is (cf. LSJ) 'free oneself from, let go one's hold of', negatived 'cling to', e.g. to protect a child, Hee. 400. Note, in this connection, that elsewhere in Euripides µE8lEµm + gen. of concrete entities is always negatived. 162 Dodds' plea for µi8Ea8E: 'an order to unhand the captive is required here in view of the order to seize him at 503', is unconvincing. If the attendants did never hold on to the Stranger in the first place, the 160. Where v. Leeuwen and Coulon, needlessly 1 think, put a semi-colon after yd.p. 161. 4>pou8os is not discussed by K-G. Roux wrongly gives 'qui a fui, qui a disparu' as a general translation of cj,pou8os; cj,pou8os is never used attributively. LSJ s.v. observe: 'Rarely found in any case but nom. sg. and pl.: gen. sg. once in S. Aj. 264'. This is correct, but LSJ fail to observe that this situation is due to cj,pou8os being used only predicatively. In S. Ai. 264 cj,pou8ou is, again, predicative: 'when it is gone'. 162. At Or. 172 µelkµlva is followed by the action noun KTUTTou.

70

order at 503 is perfectly appropriate.1 63 Finally, as Verrall observed, 'to drop µalvE"CJ8£ makes P. pass in silence the escape of the bacchanals and the man's miraculous explanation', upon which Dodds comments: 'this is plausible'. LP's text presents, to be sure, one major difficulty, viz. the awkward position of yap. We could get rid of yap, however, by reading 1rapwv, while keeping µalvE"CJ8E. The asyndeton is then explanatory and the interpretation is the same as with the reading yap, as discussed above. For 1rapwv cf. 511, 1rdpn (subj.: the Stranger), and for 1TdpELµL lv Hel. 944: olKTpov µEv ol 1rap6VTES µfol¼> Mym, 1184-5: daopw yap o\Js 8LwicoµEv I 1rap6vTas lv 86µoLaL Thus: 'You are out of your mind; being present here in the nets of my hands, he is not so swift as to escape me'. With these words Pentheus reacts to the implication of the attendant's words at 447-50, viz. that the Stranger, too, might mysteriously escape. With 'the nets of my hands' Tyrrell compares lyyi,s ~8ll y' foµEv cipicuwv ~lcj>ovs at Med. 1278, and Ale. 984: ical a' Ev ci4>u1CTOLOL XEpwv El.>.£ 8£a. &aµo1s. The plural µalVEa8£ probably has 'sociative' meaning: 'you and your companions'; cf. the 'sociative' use of the 1st person plural for the 1st person singular (Bers 1984: 49ff.). Note also the presence of the plural 1raprnµEv at 434. Moreover, Tyrrell may be right in observing that ' ... the whole posse ... no doubt by looks and signs indicated their approval of the tone of their spokesman'.

ev

ws

454 llf,' c'J,rEp R.: 'elliptique, reprend l'idee de seduction implicite dans le membre de phrase precedent'. Possibly so (although there is nothing elliptical as far as l' B1rEp itself is concerned). More specifically, the antecedent of 81TEp is: TO µ11 dµopos Etva.L ES yvva'ticas: your end in having come to Thebes is to be beautiful to the women. Compare for the use of l' &1rEp referring to a preceding state of affairs as a whole rather than to a particular head noun K-G 1, 62, top of the page.

ws

4 5 5 ou ,rdA.T1s fl,ro D.: '"proving you no wrestler'', lit. "by not wrestling"'. 163. Roux loo believes lhal 'L'imperatif µl8Ea8E ... justifie l'ordre que donnera ulterieuremenl Penlhee de le saisir anouveau'. However, 'a nouveau' is not in lhe Greek.

71

R.: 'litteralement "du fait de la non-Jutte"'. I do not think this is correct. The problem is: how can Pentheus take for granted that the Stranger is no wrestler, as suggested by 'du fait de'? Also, Dodds' 'proving you no wrestler' is not just another way of saying what is literally 'by not wrestling'. 'Proving' itself may be all right, however, for it suggests that Pentheus, from what he sees, is making an inference as to the way of life of the Stranger, and this is, in fact, what we need here. But from this it does not follow that we must take ou as a modifier of ,ra~:r1s- rather than of 1ra>,:r1s- Oiro (= irrro 1rci>..T1s-). The inference is a sarcastic one. So our translation should not be: 'Your locks are long by not wrestling' but: 'Your locks are long, not by wrestling'. So, too, Schone, if I understand him correctly (Schone remarks: ou ira>..T'IS- £mo, sc. TOLOUTOS- yEv6µEvos-). K-G, in their discussion of adverbs modifying nouns (1, 609-10), do not mention Ba. 455 (nor 458, see next note), rightly, I think_l64

458 oi.lx -tiMou po>.atau,, dU' fi"ll'o aicLCtsD.: 'This seems to be like 455: oux in effect negatives 13>.a'LaLv, despite the following d.>.M. '. R.: 'meme tour negatif que v. 455, "par le fait des non-rayons du soleil, mais par le fait de I' ombre'. K.: 'You keep your complexion fair by careful contrivance 165 - not in the sun's rays but under the shade hunting the pleasures of Aphrodite , Dodds is here more hesitant, and understandably so, for the presence of

d.>.M. makes all the difference. Observe that Roux' bizarre translation is

caused by her ignoring the contrast expressed by ou .. . a.Ma; her 'mais' is completely in the air. So we should follow Kirk, taking oux ••• .ov, IT 1339: ~µEv f\µEvot, 1368: nv-yµal 6' ~aav E-yKpoTouµEvat,168 Pho. 66: Cwv 6' foT' fV OLKOLS', Supp. 527: Tt ... EOTlv ou KaAWS" lxov;, 767: 6nvov ... ~v l3dcrrn-yµa Kdtcrxt'iVT)v lxov.1 69 In actual usage there would seem to be but little difference be-

u

166. Surprisingly, these constructions are not discussed at all by Moorhouse, although they are very frequent in Sophocles. 167. In a few cases the periphrasis is based upon an activity verb: S. Ai. 1324: Bpliiv yap ~v T0t.aOTa µe, Hee. 1178: el ... vOv >.lywv fonv ns- (sc. icaicws-). 168. Here, ~aav might also be taken as presentative, and lYKpoTouµevaL as atbibutive: 'There were fists hailing blows'. 169. Aerts (1965: 12f0, following BjOrck and others, argues that many of these and other cases traditionally called 'periphrastic' are rather combinations of elvaL and an adjectivised participle. This may be true in itself (cf. e.g. ,rplirov, which, like many regular adjectives, has an adverb in -ws-), but does not alter the fact that these combinations often compete with monolectic verbs, which means, in tum, that we still have to explain their (dis)similarities. Compare, by way of example, the difference between Engl. The results were amazing as against The results amazed everybody (not:

73

tween the synthetic and periphrastic forms. 170 However, the stative value may come to the fore more clearly in the construction dµt + predicative participle, the emphasis being put on the participle. A case in point is S. ar 747: 8nvws a:8uµw µ11 ~>J-trwv b µcivns fl, where the synthetic ~AETTlJ would be rather a run-of-the-mill expression. 171 So at Ba. 471, too, by choosing lcrrl ... lxoVTa rather than lxn, Pentheus may stress, sarcastically no doubt, the permanency of the form of the rites, that is, the permanency in the eyes of the Stranger, om being a dativus iudicantis, cf. K-G 1, 421: 'What form are these rites provided with in your view?'

475 h:LPB~~EUaas-, tv' ... et>w> For the use of the subjunctive after a historic tense see K-G 2, 380, 3.

479 KOiJ8tv ~'YWV With Verdenius (1962) I think icat does not have concessive value; the icat-phrase adds an explanatory circumstance. More often we find ical mum+ participle in this use, e.g. Ar. Pax 35, V. 1184. 172 483 4>povo0aL yelp KdKLOV 'Ell~vwv ,ro~'l1 The same thought in Hdt. 1.60.3: ltrd 'YE" a:TTE"Kpl&rl lie tra>.aL Tlpov TOV ~ap~pov l8VE"OS' TO 'E>.AT)VLICOV lov ical BE"eUllTE"pov ical dr T)8lT)s 11>.i.8tou d:1TT1>J.ayµlvov µa>.Aov. 492 T( µE" TO 8nvov lpyddTvaLaLv D.: 'in the neighbouring stables'. R. (transl.): 'pres des rateliers, au fond des ecuries'. L.: 'trl>.cxs est pris adverbialement'. Dodds apparently takes TTEMS' attributively. While this in itself is fully possible (cf. K-G 1, 594-5 on the attributive use of temporal and loca174. As does Lacroix; in view of his commentary C-1:is in his text is a mistake. 175. Dodds' and Kopfrs information on P. Ant. 24 Ibis inaccurate. According to Roberts (1950: 55) 8pq.s is not wholly certain, for he prints ~pas. 176. I fail to see what insight is gained from calling this 'a late-fifth-century mannerism' (Dodds, repeated by Roux), apart from the fact that it is not a late-fifth-century mannerism (cf. the ex. from Hdt.; other exx. from Hdt. are: 6.102, 9.2.1).

76

tive adverbs, and cf. ol T6Te- 4v8pw,roL etc.), ,re->.as can hardly be attributive in our line, for 4>a.TVJ], as Verdenius (1962) points out, is not 'stable' but rather 'manger', or, as I would prefer, 'hay-rack', since the a.T"'l is especially connected with horses, cf. l1rmKa1s in 509. 177 With ,rl>.as taken attributively we would get: 'shut him up in the nearby hayracks', which seems impossible. So we should rather follow Bruhn and Roux in taking ,rl>.as as a preposition governing the dative, for which cf. e.g. Ale. 366, Pho. 845, IT 460. Observe also that at 618, where the same kind of imprisonment is referred to, we find ,rpos 4>ci.Tvms.

510

~

4v ... daop4

ws

Like B1rws 4v + subj., 4v + subj. may express 'la confiance du sujet (i.e. the 'sujet parlant' -A.R.) en la realisation du resultat attendu', Amigues (1977: 163). Simple B1rws and (+ subj.) lack this nuance.

ws

513- 4

xe-tpa

,ravaas

8ov,rou To08e-

Kal

f3vPO'TIS

KTVlTOU

D.: '8o{J1Tou Tou8e- is explained by f3vpCTT)s KTV1rou: for Kal connecting words in apposition cf. 919 (Denniston, Particles, 291)'. R.: 'Kal .. . KT1J1Tou developpe et explique 8ov,rou; Ka( explicatif (Denniston p. 291, § 5)'. L. (in his translation): 'apres avoir mis leurs mains dans l'impossibilite de poursuivre le vacarme et de faire resonner !'instrument de cuir'. Apart from the fact that Dodds and Roux refer to Denniston in a rather careless way (Denniston has: 'appositionally related ideas are occasionally linked by Ka('), I do not believe their analysis is correct (this also applies to 919, see below); Ka( is not simply 'explicatif'.1 7 8 As to Lacroix' translation, note that he wrongly has 'le' instead of 'ce 177. Cf. Roux' 'rateliers'. Cf. also Hom. /l. 10.567: hnrous µEv KaTl8ricrav lVTµTJTOLcrLv lµlicrL / 4,dTVU l4>' lTnrdlJ, IS8L TTEP dLOµTJ8eos tTTTTOL / foTacrav wKlmo8ES' µEMT)Bla 1rupov l8oVTES'. Both LSJ and Chantraine Diet.

Etym. only give 'manger', 'mangeoire'. 178. More in general I think that most instances of Denniston 's 'appositional Ka(' can be explained differently, cf. also footnote 148 on allegedly appositional TE. As in similar cases of TE, Ka( sometimes couples two nouns or noun phrases which denote two different properties of the same referent, e.g. at HF 15; alternatively, they may express a whole-and-part relationship, as in IA 751: ij~EL ... 11µ6eVTa Kat 8(vas dpyupoEL8ets. Cf. the use of 'and' in Engl. 'visit Greece and its monuments', which is obviously not the same as the appositional 'visit Greece, that is, its monuments'.

77

vacarme-ci '; also, it is not clear how he views the relationship between 6o{nrou and KTUTTou; in his commentary he is silent on this point. Following a suggestion of Dodds', I would propose an entirely different analysis.179 Dodds suggests, rather hesitantly, that the Chorus may have called attention to themselves by thumping their kettledrums (as Tou8esuggests) to emphasize the Stranger's warning at 508. I do think, indeed, that the Chorus have called attention to themselves at 508, not, however, by thumping their drums, but by beating their breasts. For the use of 6o0TTos in connection with breast-beating, as a sign of sorrow, cf. S. Ai. 632-4: xe-p6TTAT1KTOL 6' / lv aTlpvotat TTrnoOvTm I 6o0TToL, and especially Ale. 104: ou ve-oACI(a 6ouTTe-t xe-tp yuvmKwv,180 where TTlv8n is found in the preceding sentence. 181 What happens, then, at 508, is that the Chorus are accompanying the Stranger's reference to Pentheus' ill-boding name with the sound appropriate for expressing TTlv8os. BupO'TIS' KTUTTou, in turn, refers to the sound of the tympana, which the Chorus were invited by Dionysos to produce at II. 58ff., note especially ciµ4>l 8wµaT' l>..8o0aat Ta8e- / KTUTTELTE" Ile-v8lws, 60-1.

515 aTE£XoLµ' clv 'I am ready to go' (Dodds), a common formula, signalling that an order or request is about to be carried out; note the inceptive value of the present stem. Cf. Ba. 845, El. 669, Hee. 1132, Ion 668, Or. 640, etc. 182

515-6 lS n

yap

µ1'

XPE~v. oihoL XPEWV I ,ra8Eiv

MfJ in a relative clause expresses the idea that the state of affairs denoted by the relative clause is of a general rather than a particular nature. Compare K-G 2, 185: (µii occurs) 'wenn der Nebensatz eine Bestimmung enthalt, die our in der Vors tell u n g vorhanden ist, daher, wenn der Nebensatzeineallgemeine, unbestimmte Angabe von Perso179. Which was apparently also considered by LSJ, since they range Ba. 513 with S. Ai. 634, see below. 180. The text presents some difficulties, but these do not affect 8ot11TEL xdp yvvaLKwv.

181. Note also the frequent use in Homer of 8ov1rlw to denote the heavy thud of a falling corpse. 182. Interestingly, we only find >.lyoLµ' dv in such formulae (as at Hee. l 132, Or. 640), never ELTTOLµ' dv. At/A 2, after the imperative w 1rpfof3u ... OTELXE, the Old man reacts with onlxw; this means: 'I am coming' and is spoken while he is on his way. Cf. the sequel: Ag. OTTEUOELS; (or: oTTEiioE), Old man: 01rEu8w.

78

nen oder Sac hen, von einem all gem e in en (nicht konkreten) Begr i ff e oder von ofter w iederkehrenden F iillen enthiilt (-)'. With & TL µri XP£WV we should supply tra8e'iv. 183 All in all, in interpreting this sentence we should follow Schone's excellent note: 'Der Sinn ist: "Denn ich werde doch nicht erleiden, was mir nicht zu erleiden bestimmt ist"' . 184

519 On the question of whether before line 519 something is missing or line 537 is spurious, see on 537. To the testimonies for the important role played by the Acheloos in Greek thought one may add Hdt. 2.10.3: dal 8e- Kal d>J.oL TTOTaµot,

OU KQTQ TOV NE"LAOV MVTE"S µeya8rn, OLTLVE"S lpya citro&!aµE"VOL µe-yaXa dot· TWV E'YW 4>paaaL lxw ouv6µam Kal dAAWV Kal OUK finaTa 'Axe-¼ou, .... 523-5

lSTE

µrip~ ••• ,ipiraal

vLV185

D.: 'snatched ... (and hid) in his thigh'. R.: 'datif de destination, "pour sa cuisse", c. sa cuisse"'.

a. d. "pour le cacher dans

Although the locative dative with verbs of movement is rare (cf. K-G 1, 441-3, b), µT)p4i of the place 'wohin ein Gegenstand gelangt' (K-G) 183. For some reason I cannot fathom Dodds and Kirk apparently supply dvaL with µ11 x~wv, cf. Dodds: 'for what is not to be I have not to suffer'. 184. Verdenius' remark: 'the force ofµ~ is ... simply emphatic' is meaningless. -A good illustration of the differences between ou andµ~ in relative clauses is provided by Pl. Smp. 200 d 8 - e 5. There we find, first, a relative clause with ou: ovKoOv ToilT6 y' laTl11 lKe(vou lpav, 8 olhrw hoLµov almil faTLII ou~ lxeL, Tb els Tbv lTrELTa xp611011 TailTa elvaL auT4'i al\)(6µe11a Kal 1rap611Ta. Here, Socrates is summarizing the preceding discussion, in which he carried on a dialogue with an imaginary interlocutor on a number of concrete, particular, objects of lpws. Then he passes over from this specific person to every person who is longing after something, thereby giving the conclusion he has just arrived at a much wider, indeed a universal, application. In the sentence concerned we findµ~. not ou: Kal o~TOS dpa Kal dllos 1ras b lm8uµcii11 Toil µ11 ho(µou lm8uµet K«l Toil µ11 1rap611Tos, Kal 8 µ11 lXEL Kal 8 µ11 lanv auTbs Kal o~ lv~s faTL, TOLailT' dTTa foTlv c!iv 'fi lm8uµ(a Te Kal b lpws laTLv; l further note that whereas some relative clauses with µ~ can be paraphrased by a conditional clause containing an indefinite pronoun, e.g. a µ11 lxw = et nva µ11 lxw, this is excluded for relative clauses with ou: a ouK lxw can only mean: the particular, concrete, objects I do not have. (In writing this note I have made use of an unpublished Amsterdam paper by P. Kemper). 185. I notice in passing that Kopfrs !Sn, at 524, must be a misprint.

79

might be taken as a variant of the construction c:ipmi,w ds. which we find at/on 1598: c:ipmiaavT' ls d:yK.as / ... f3pe-os. I should add, however, that ifµ11p4> is taken in this way (cf. Dodds), lines 521-525 are quite odd, since on a natural reading they say that Dirke received the infant after Zeus had put him in his thigh (where he was to stay for some time, cf. 11. 526- 7 and 11. 94-100). The alternative proposed by Roux, however, viz. to take µ11p4> as a directive dative, is not unproblematic either. To be sure, an interpretation of these lines in which the infant, after Zeus snatched him out of the fire in order to hide him in his thigh, was first received by the streams of Dirke, 186 is plausible enough, but µ11p4> is difficult, as is the (final-) directive dative as such. Most instances of this use adduced by K-G (1, 406f.) and others, e.g. S-D (139f.) and Moorhouse (80f.), can be explained differently, 187 e.g. as a dative of interest (Hom. fl. 5.174: aLl xe-tpas avaaxwv). as a locative dative (II. 11.593: OOKE" wµoLOL K>J.vaVTfS") or as an instrumental-causal dative (Ion 1467: J.ou 6' .i"trEL >.aµn 6wµ' d:ve-[3>..e-4>'). The final-directive use is probably restricted to a few isolated cases involving an action noun, as Hom. fl. 7.218, npoKa>JaaaTo xapµ1J_188 Observe that µ11p4> cannot easily be taken as a dative of interest either: the use of this dative is confined to persons (including inanimate entities referring to persons, e.g. aQ otK4> in Hom. Od. 2.186: 04> o\'.K4> Bwpov nonSlyµEVOS" ). All things considered I incline to think that there is something wrong with µ11p4>. Possibly it is a gloss which has ousted a word like npwTov (lSTf npwTov 'when once' or possibly 'as soon as', cf. Hee. 631, Dem. 18.141). 1

528-9 dvacf>aLvw ... 6voµd(£LV D.: 'I reveal thee to Thebes ... that she call thee by this name'. R.: 'je te revele Thebes pour qu 'elle t'appelle de ce nom'.

a

186. In his commentary on 11. 94-5 Dodds interprets our lines in this way ('the child was first washed in the waters of Dirce (521f.)'). This does not harmonize with his remarks at 523-5, where the addition 'and hid' makes it impossible, I think, to arrive at the interpretation advocated by Dodds at 94-5. 187. As is admitted by Moorhouse himself. Cf. also his cautious conclusion on p. 81: 'There seems, then, to be some support for final-directive use in Sophocles'. 188. Cf. the final infinitive after TrpOKa>.loµm in Hom. fl. 3.432-3: TrpoKd>.rncraL dpT1tl/>L>.011 Mivi>.ao11 I .. . µaxfoacr6aL.

80

This can hardly be correct.189 The transition from 'come to this womb' to 'I reveal thee to Thebes' is rather harsh, since it more or less suggests that Dionysos is revealed to Thebes by coming to Zeus (which was not the case). I therefore prefer, with e.g. Schone, Tyrrell and LSJ, to take ava4>a(vw as = 'proclaim', and efi[3a.LS as a locative dative: 'I proclaim that they call thee at Thebes by this name'. For ovoµaCnv having future reference after a verb of saying see K-G 1, 195, A. 7 ('bei den Ausdrticken des Weiss a gens').

537-40 otav otav 6pyd11 d11a4'aC11EL ,c8cS11Lo11 ytvos ilccfn1s- TE 8pd1COVT6S' ffOTE Il£118£US'' The words o\'.av o\'.av op-yav correspond to nothing in the strophe and were deleted by Bothe, who was followed by many editors, e.g. Dodds (in his additional note on 537-41), Roux, Lacroix, also by Verdenius (1962) and Kirk. Now Dodds and Roux allege that the omission of these words leaves a good idiomatic sentence: 'P. reveals his earthborn origin, that he is descended from the dragon'. On this analysis we should construe ava4>a(vn first with an object, viz. x86vLOv -ylvos, and then with a complementary participle, lKcf>VS. (Incidentally, I would prefer to take x86vLOv predicatively, supplying 6v). However, this is improbable. For the participial construction Dodds refers to Pl. Criti. 108 c, but there the interpretation of ava4>a(vnv is uncertain.1 90 The text runs: XP'll ... TOUS TTaA.mous noMTas, somewhat zeugmatically, as the object ofvµvE"1v. However, the Thomas Taylor translation (Princeton Un. Press 1944, 189. 'A11acf>a(11w is Hermann's correction of LP's d11acf>a11w, which is metrically impossible. Elmsley proposed d11acf>d11w (subj. aor.), which, as ScMne observes, should be d11acjl1'111w, cf. Ion 860, and BjOrck (1950: 164ff., 254-5)). 190. Dodds also refers to S. El. 1359, but there we have lcf>aLllfS', not dvlcj>aLVfS'.

81

19633) has an entirely different view of this text. This translation runs: ' ... to exhibit and celebrate ancient citizens who were excellent men'. (I would prefer 'the ancient citizens' and 'since they were excellent men'). This apparently rests upon talcing ci:ya8ous ~vTas as modifying Tous ,ra>.a.Lous ,ro>.l TaS, which is fully defensible. Since this passage from Plato is the only parallel adduced for ci.vaq:,a(vw + ptcple, the case for ci.vaq:,alvn lKcf,vs in Ba. 538 is rather weak. On the other hand, the construction of ci.vaq:,a(vw + nominal object is very common, cf. LSJ s.v. 2b, e.g. Hom. fl. 20.411: ci.pE"TTJV ci.vaq:,alvwv. I prefer therefore retaining otav otav 6p-yav and assuming a lacuna before 519, which might be filled by Barnes'@ KallicrTa Kpavav.191

544 r3aT£ D.: '=wcrm:p'. R.: 'ooTE, au sens de 001rEp, est usuel dans le style tragique'.

This is not quite correct. wcrTE is especially used when the comparison involves a permanent characteristic. Cf. Ruijgh § 810: ' ... les trois tragiques emploient normalement WOT£ pour marquer une comparison non verbale (i.e. without a finite verb) exprimant une notion permanente'.

552 lv d.µ(UaLaLv dvdyicasD.: 'at grips with oppression'. V.1: 'The genitive seems to be a genitivus qualitatis'. R.: 'ci.vayKas est un genitif objectif: " ... luttant contre la menace ... "'.

Roux' analysis must be the right one. This objective genitive represents either a dative, in which case avalvELv. If this is the case, the appropriateness of these words qua meaning as well as qua metre is quite striking. 192. With the dative of inanimate entities aµt>.Moµat means 'contend in or with a thing', cf. LSJ.

82

preposition phrase K-G 1, 335 (Th. 1.108.5: ev d.,ro(3op€ts '. R.: 'accusatif de l'objet interne'. Roux' wording is not very clear; she probably means that 8upao- is the 'accusatif de l'objet inteme'. However, since there is no accusative but 8uiaous-, her terminology is rather confusing. (8uiaous is of course an external, not a cognate object). Dodds' analysis should, then, be preferred.194

588

ffE~µaaLV

D.: 'modal dat., though in English we say "shall be shaken to its fall"'. R.: 'datif abstrait marquant l'aboutissement de l'action verbale, comme s'il y avait WOT£ lTE"O"E"'iv'. Dodds' modal dative is correct. There is no need to give it a consecutive value (apart from the question what an 'abstract dative' is). IlfoT)µa, like other nouns in -µa, may function as an action noun, cf. 139 above, note on atµa, 169: O'KLp~µaaL; also HF 1007: K(ova ... &s ,lTE"~µaaL OTfYJlS &xoppayiis lKE"L To, and S. Ai. 1033: 6>.La>Xf 8avaalµ'¼) lTE"~µaTL. The plural possibly refers to the various parts of the house which come down; cf. K-G 1, 16. These lines may echo 11. 389-92: b BE- Tds -fiMya Cf. for the relative clause with the so-called enclosed head noun (antecedent) K-G 2, 417, band Rijksbaron (1981: passim, esp. 247). This construction can best be analysed as the active counterpart to the construction with an attributive passive participle, e.g. TT)V MLcf>8E'Laav cf>M-ya. Note especially the position of the finite verb in between the 'relative pronoun', which actually functions rather as an article, 199 and the noun. The active construction has a finite verb because a parallel construction with an active participle was, of course, not possible. 604-5 lic:irE1T~Tl'YµivaL ••• 1TE1rTCiS1CC1T 1 ••• "Ga&r]aO' The perfect forms indicate that, while Dionysos is speaking, the Chorus are still under the influence of his way of acting. "HLaBlla8E- provides a particularly nice illustration of the fact that the Greek perfect combines past and present meaning: the actual perception occurred in the past, cf. the aorist participle 8LanvaeaVTos, but the effects are still felt after the perception is over. It is difficult to render this nuance in English; 'you are still feeling the Bacchic god having shaken apart the house of Pentheus', would seem better than e.g. Kirk's 'you felt the Bacchic god shaking apart the house of P. ', which is rather the translation of an aorist.

611 t:>s- ••• irEao11µEVOS' This must represent the thoughts of the subject of d:cf>lKe-a8', i.e. the Chorus, not those of the persons «qui pretendaient enfermer Dionysos dans un cachot», Dalmeyda, followed by Roux. 198. But the general semantic value given by Kannicht to 6pcia8aL ('zu sehen bekommen, ansehen miissen') is wrong. 199. llv is not a real relative pronoun, since it has no antecedent.

86

613 TUXWV D.: 'Editors construe d.v8p6s with Tuxwv: "But how were you liberated after chancing upon a man of sin?"( ... ) But the effect ... is oddish and weakish ( ... )'. Dodds' objections are shared by many other scholars but not by Roux, who takes d.v8pos d.vootou as the complement of both Tuxwv and TJMV8E"pw8rts. This can perhaps be defended, but the translation in her commentary cannot stand: 'Comment, alors que tu avais affaire a un mecreant, t'es-tu libere de lui?'. 'E>.Eu8E"poOµaL -::1:- 'liberate oneself', but = 'be liberated' .200 Also, the contrast with 614 would be lost.

619 T~8E" ••• y6vaaL ical XllMlLS' 1ro8ci>v Jacquinod (1988: 135ff.) has a detailed discussion of such 'whole and part' constructions, which always involve a relationship of inalienable possession, notably that of the body and its parts. Rejecting earlier explanations, which roughly say that the case of the part is the same as that of the whole because the part stands in apposition to the whole (cf. e.g. K-G 1, 289), he argues that the original case of the whole was a (possessive) genitive (e.g. To08e") and that this case form was adjusted to that of the part. By this adjustment the whole has the same syntactic status as the part, whereas the genitive rather indicates that the whole is dependent on the part. The overall semantic effect of the syntactic 'promotion' of the whole is that the relationship of inalienable possession is emphasized. 633 8CISµaT' lPPTIEEv Strangely enough, Roux considers the asyndeton 'remarquable'. On the contrary, it is exactly what we expect, 8wµaT' lPP'TleE"v being the first member of the series announced by Ta.8' d>J.a AuµatvETaL. A connective particle would have linked Avµatvnm and lpplleev, which is, of course, impossible. 200. Her translation in volume I, on the other hand, suggests that she accepts F.W. Schmidt's XEpwv: 'Mais comment as-tu &:happe aux mains d'un homme pour qui rien n'est sacre?'. Incidentally, this translation is wrong too, for HEullEpoOµaL does not mean 'escape from' either.

87

661-2 ... KL8aLpli>v' ... , tv' 0~1TOTE" >.Evicfis- XL6vos- dvEt aav Eilayds- ~o>.aC D.: "'Where the white snow's glistening falls never loose their grip'". R.: 'aoriste d'habitude'.201 Dodds adds: 'If this means, as some suppose, that it never stops snowing on Cithaeron, the exaggeration is monstrous; it is still considerable if we take it to mean that the snow lies in places all the year round (I found none when I climbed the mountain in April). We may have here nothing more than a conventional poetic commonplace'. 202 Gregoire's experiences on Cithaeron, however, apparently were different, for he remarks 'pendant de longs mois, de la Beotie et de l 'Attique, on voit la neige briller sur ces montagnes; personellementje me rappelle avoir ete surpris par une brusque et abondante chute de neige sur le Citheron, en plein ete'. Such meteorological observations are needless. 'AvEioav is not a habitual aorist, or at least does not have present meaning - and the clause is not, therefore, a generic, omnitemporal, statement - but rather a normal past tense, which denotes that during some unspecified period of time before, and possibly up to the moment of, the messenger's departure from Cithaeron it snowed incessantly. 'AviLaav may perhaps be considered a habitual state of affairs, but that state of affairs is in any case located in the past. Cf. for the habitual, or iterative, use of the aorist S-D 278, 5. 203 The same use of ouTTOTE + aorist indicative occurs in Hel. 4O5ff.:

... xwrnv

f'Y'YUS-

wmhpas,

TTot[36s TE" ica.yw >.atvous m,pyous irlpLf bp8otaw l8€µe-v icav6cn.v, olirroT' EK pe-vwv e-fNot.' 0.1TE"pvywv ir6XEL · Ef ofi

Recall, furthermore, that the 6pnf3aafo, which is described in the parodos and in the narrative part of the present messenger-speech, took place in winter; cf. Dodds' Introduction p. xiii.

677-679 This messenger-speech is of the type that opens with a main clause, which gives time and place of the events described in the story proper. The information presented in the opening sentence is new, and does not refer back to earlier information. Also, the messenger himself is 'new' as well, i.e. he has not been on the stage before. By these features such messenger-speeches differ from those of the second type, that open with a temporal lird-clause. Such tird-clauses typically do pick up earlier information, and the messenger has been on the stage before, e.g. as an attendant to one of the main characters. An example is the second messenger-speech of the Bacchae, see 1043ff. For details I refer to Rijksbaron (1976a: 299ff.). 680 6pca A nice example of the use of the historic present to signal that the event in question (here, the 'seeing') is of prime importance for the remainder of the story. Observe that the content of the seeing runs up to I. 713, where the description of what the messenger saw is closed; note the presence, at 713, of daLooiv Ta.Se-. Similar historic presents in this passage are: l>J.ox(CoµE"v (722), icvpd (728), xwpo0aL (748); note that in three cases (680, 728, 748) the historic present has front position, which underscores its highlighting function. We are to understand that the herdsman-messenger considers these four event,; decisive moments in the scene he is reporting about. In the passage 680-713 we find one other historic present, viz. at I. 705, lKTTTJ8q. This present probably just indicates that the event concerned is remarkable as such. Cf. for an extensive discussion of the historic present in Euripides De Jong (forthc.).

89

683 ac.Sµ.aaw Moorhouse p. 87 presents a number of such 'datives of respect' from Sophocles. They make clear that Roux' remark 'usite surtout, semble-til, avec le mot awµa ou des noms designant des organes' is not justified, for the nouns concerned involve icci>..ut >..6-yo,;, yvwµ11, yovfi, and the neuter plural aµLicpci. 701 VEOTdlCOLS" R. (transl.): 'c'etaient celles qui, meres depuis peu, le sein encore gonfle, avaient abandonne leurs nourissons'. L. (transl.): 'c'etaient celles qui, jeunes meres, avaient encore le sein gonfle et avaient quitte leur enfant'. K.: ' ... those who had newly given birth, whose breasts were still swollen, and who had left behind their babies'.

These translations are rather inexact. In spite of its position directly after Bams I think VEOT6icoLs must be taken as a, prominently placed, predicative modifier with >.J. ,roooms, which itself has causal meaning: 'those whose breasts were still swollen, because they had left their babies having newly given birth(= immediately after the delivery)'. 701 ~v a,rapyc3v Cf. the note on I. 471. 705 h1T1184 See the note on I. 680.

722 lUoxlCoµEv, 728 icupEt Used like bpw in I. 680, q. v.

723

QUTOUS-

The use and the frequency of the 3rd person reflexive pronoun for the 1st and 2nd person are the subject of the recent monograph by Woodard (1990). 727 o-08tv 8' ~v dicCVl'lTOV 8p6µ'-\) D.: 'and with their motion all things moved'. R. (transl.): 'rien qui ne filt entraine dans leur clan'. L. (transl.): 'tout se mettait en mouvement et participait ala course'. K.: 'and nothing remained unmoved by their running'. 90

These translations, which rest upon talcing 8p6µqi as an instrumental or causal dative, 205 are wrong. They malce this clause a mere repetition of the preceding clause, miv Se- avve-f3a.KXEV' 6ps / Kal 8fipe-s. It is worth noticing, in this connection, that in Roux' translation 'rien qui ... ' etc. is preceded by a colon, indicating that oii8e-v ... 8p6µqi should, in fact, be talcen as an explanatory summary of the first clause. But in doing so she ignores, of course, the connective 8l. Actually 8p6µqi is a modal dative with ovBEv 8' 1iv aKlVflTOV (= lT.wv does not differ from the use of ws + participle which we find in X. An. 1.1.11: Ilp6ee-vov 6e- ... £Kl'>.Evae- ... 1rapaye-vfo8m, ds I1Lal6as (3ovMµe-vos- aTpaTe-urn8aL ('alleging that', 'because, so he said') or Th. 2.59.1: TOV ... Ile-pLKAEa lv al Tlq. e-lxov 1rflaavTa acf>cis ,ro'>.Eµe-tv Kal 6L' EKfLVOV Tats evµcf>0pats 1Tfpl1Tf1TTWK6TfS' ('on the ground that'), upon which Goodwin remarks(§ 864): 'Here Thucydides himself is not responsible for the statements in the participles, as he would be if ws were omitted'. However, whereas in these two cases we are dealing with third person narrative, the peculiarity of our line is that the agent of

ws

ws ws ws

205. It is not clear exactly how Lacroix construes the dative.

91

the main verb and the narrator are identical. Since the herdsman is speaking, then, about his own behaviour, he is obviously responsible for the 'statement in the participle' and this may explain why ws 8l>.wv seems to differ but little from simple 8l>.wv.

7 30 lKp-6,rToµEv 8iµasThus LP. 'EKpt11TToµev 6lµas a:: lKpt11TToµev 11µas auTovs. Many editors, e.g. Murray, Gregoire, Kopff, prefer E"KpU1TT6µ11v (which apparently occurred in the margin of Barnes' edition), but LP's text is certainly sound. (a) The active Kpt11TTEW 6lµas is also found below, 954, as well as at HF 1010, Hipp. 1290-1, IA 1340, IT 106 and Tro. 777; (b) It is difficult to see which nuance the middle might convey; Kpv,rToµm elsewhere has only passive meaning. See also below 1. 955, note. As for the plural, this refers either to the whole group of herdsmen, or, more probably, to the messenger, cf. 669, with Dodds' note. 738 lxovaav iv xipotv 8Cica Thus LP. Reiske's (or Scaliger's) 6(xa, for the meaningless 8(Ka, has found favour with many editors. (Reiske also proposed to read l>..Kovaav for lxouaav, see below). A number of them leave it at that, and retain lxovaav; so Murray, Gregoire, Roux, Kirk (although he translates: 'tearing ... to pieces'). Roux explains 6(xa as 'a l'ecart du troupeau'. While 'a l'ecart de' may be right (however, not from the herd, see below), lxouaav l-v XEpoi.v can only mean 'en la retenant entre ses deux bras', which in the present context is rather lame (cf. also Dodds). To be sure, Roux further notes: 'L'une des bacchantes ... cloue sur place une jeune vache', but her 'cloue' cannot, I think, be got from lxoooav. Dodds would rather read l>..Koooav (Reiske) ... 6txa, remarking that 'the sense required is surely "wrenching asunder"'. But the question is: can l>..Knv 6txa mean 'wrench asunder'? I think not, at least not if 'asunder' is taken as a:: 'in pieces'. For while 6txa can, in fact, have the meaning 'asunder', this always a:: 'in two pieces'. Cf. the examples in LSJ s.v., especially Pl. Sph. 265 e: TlµvE 81') 6uoi.v ofuaw &xa EKaTlpav, where the sequel makes clear that the result of the cutting of the two objects is that there are four parts. So if l>..Kew could mean 'tear', the result would be that the bacchant was single-handedly tearing the young heifer in two pieces, which is rather bizarre (unless one considers a frenzied bacchant capable of anything). However, with concrete and non-liquid objects EAKELV normally a:: 'drag (away, along)', not 'tear' or 'wrench'. LSJ, it is true, also give (I, 3) the specialized 92

meaning 'tear in pieces', but this is open to objection. In Tro. 280: l>.Jc' OVUXEO'O'L 6l'TTTlJxov traprniv the meaning is rather 'rip up', and in Hdt. 1.140.1: ... nplv liv im' 6pvL8oS fi icuvbs t>.Jcua&ij (sc. b vlicus), l>.Knv means 'maul, manhandle', cf. Powell's Lexicon. 206 So on a natural reading l>.Kovaav lv XEPOLV 6lxa means either 'dragging it apart (= aside) with both hands' (but this is perhaps not violent enough), or, as I would prefer, 'mauling it with both hands, apart'. For adverbial 6lxa = 'apart, aloof' cf. S. OC 602, Th. 7.81.2. 'Apart', should not be taken, I think, as 'apart from the herd', which is rather trivial, but as 'apart from the other bacchants', i.e. on her own. This makes good sense, for in the next line the messenger passes over to actions that were carried out by the bacchants collectively. Observe also that if the victims are named in order of increasing importance (thus Dodds), this order is matched, on the interpretation proposed, by actions of increasing violence. 207

7 4 6 8LEfopo0vTo lv8vTd D.: 'Best taken as passive (subject TaupoL, lv6uni "retained accus.")'. R.: 'passif plutot que moyen ... lv6vTa., acc. de relation'. K.: 'the garments of flesh were drawn apart ... '. While I agree that 6LEcf>opouVTo probably has passive meaning, and that lv6UTa might be a second accusative of the kind which is found with verbs of robbing and is 'retained', in fact, when such a verb is passivized, Kirk's translation must be right. 'Ev6uTa is, then, the subject of 6LEcf>opo0vTo. Cf. K-G 1, 65, where the exceptions to the rule that neuter plurals have a singular verb are discussed. They argue inter alia, section b, that the verb appears in the plural when 'der Begriff der Vereinzelung oder Mehrheit des aus mehreren Teilen Z u s am men g es et z ten . . . hervorgehoben werden soll'. This per206. The same meaning is found at Lys. 1.12. - LSJ say that in Hom. only O,iclw is used in this way. S.v. O,iclw they write '= l>.icw drag about, tear asunder'. Actually, however, only the first meaning applies. 207. I am tempted to consider a third possibility, viz. to read: l>.icoucrav lv ,cEpotv dKµij (or dKµcits (Nauck)). Although this text departs more drastically from the transmitted text than the other conjectures, it gives excellent sense: 'mauling it with the force of both hands'. Compare: Pi. 0. 2.63: lv XEPOS' dKµ4 (with Tapcicr· crovns), A. Pers. 1060: dKµ4 ,CEpwv (with lpELKE), below 1206-7: M1JK01ri'),cEcrL ,CELpwv dKµai:O'Lv (with i')ypEucraµEv). I must admit, however, that palaeographically the change from the quite common dKµij to the bizarre 8CKa is not easily accounted for.

93

fectly fits our case: the pieces of flesh were tom off one by one. Roux' 'accusatif de relation' is highly improbable.

7 48 r.lSaT' 6pvL8ES' dp8EtaaL 8p6µl\) D.: '"like birds upborne by their own speed"'. R. (transl).: ' - dans l'envol de leur course on cut dit des oiseaux -' K.: 'like birds lifted by their flights'. The translations of Dodds and Kirk suggest that they construe d:p8€1om with 8p6µ'¼), and the whole with 6pvL8Es. While this is not impossible (although the result strikes me as rather recherche), I prefer taking W-.ns Ta.s 'fTlV 'Aalav KaTOLKovaas.

7 s 7 oil xaXK6s, oil aL811pos D.: 'If the text is sound, the "bronze and iron" are not weapons - the maenads' only weapon is the thyrsus, 762 - but domestic utensils'. The opinion (which originates with Wilamowitz) that ou xa>--K6s, ou al8T1pos refers to utensils is shared by e.g. Verdeniusl, Roux, and Kirk; Verdenius thinks that especially heavy utensils are meant. All three accept the text as it stands. Dodds himself, however, is inclined to doubt its soundness and to assume, with Middendorf and others, a lacuna after 756, whereby the sentence beginning with otr6aa ends at trlBov. 'The Herdsman may perhaps have continued "Nothing resisted their assault, not bolted doors, not bronze, not iron"'; he does not specify what bronze and iron refer to in this interpretation. All in all this is not very plausible. Lacroix, too, assumes a lacuna after 756, and interprets tentatively: 'aucune arme de bronze ou de fer n'avait prise sur elles'. This is improbable, since it is only at l. 759 that we hear about the inhabitants using their weapons. Kopff, on the other hand, assumes, with Hartung, 94

a lacuna after I. 754, but he does not say what this lacuna may have contained. Oranje (pp. 181-186) has the most detailed discussion of this much disputed passage. Following a proposal of Jackson's he puts ou xaAK6s, ou crl6T1pos after I. 761, making these words an apposition to (3l->.os-. 208 The result is a little odd: ' ... no weapon, not bronze nor iron'. If the material of the weapon must be mentioned, we would rather expect an adjective, as in Or. 1309, crL6apfoLcrL ~lAEO'LV. Oranje objects (footnote 400) to the proponents of the 'utensils view', notably Roux, that the parallels adduced by them to illustrate this meaning are inconclusive. He is certainly right about S. Fr. 500 Radt (= Fr. 500 Pearson), where xaAK6s and cr(6T1pos must refer to weapons, and about Th. 3.68.3, where the presence of llrm>.a before xwos Kal crl6T1pos, as already pointed out by Dodds, makes all the difference. To prove his point he also mentions Pl. Leg. 956 a (ignored by Roux): crl6T1pos & Kal xaAKos ,ro>.lµwv 6p-yava. On the other hand, his blunt statement: "'bronze and iron" means "weapons'" (fn. 392),209 may just be a little too strong. I:l6T1pos may refer to all kinds of iron artefacts, cf. LSJ s.v. II, and xaAK6s may even have the meaning 'anything made of metal', LSJ s.v. II, 1 and 2. What things are involved is usually made clear by the context, which brings us back to the original question, for it is precisely the lack of clear contextual data that makes the interpretation of our line uncertain. All in all P's text can perhaps be defended. As for the syntax, it seems best to take o,r6cra, with most commentators, as an autonomous relative clause: 'all things that, whatever'; the entities referred to by lm6cra may, but need not, include the children mentioned in I. 754. t'm6cra is the subject of ,rpocrdXET' as well as of lm 1TTEV, and ou XaNC6s, ou crl6T1POS an apposition to &rrooa ... l&aav.

758 trOp lff,Epov, ou8' tKa(d' P's E"KalE8' is generally replaced by Elmsley's lKaLEV (the object being 'the locks' or 'them'), needlessly, I think. Normally ,rup is the object of KalELV (and thus the subject of Kalrn8aL), rather than its subject (one of 208. Jackson, again followed by Oranje, suggested some other rather drastic alterations, which are of no relevance here. 209. Oranje's argument (ibid.) that 'it is unlikely that the women, who had fled from their own pots and pans in the city, would now come to carry off those of the villagers' hardly needs refutation. We must assume that the frenzied women were not too selective as to what they took away.

95

the rare examples is II. 21.343). With lKale-8' we have 'and upon their locks they carried fire, but it did not (really) bum'. For oW = d.>J.' ou cf. Denn. 191 ('holding apart incompatibles'), and for the change of subject cf. below, II. 1096-8: avToii xe-pµd8as ... lppmTov ... , 6{01.crl T' ... 'ftKoVTl(e-To. The fire may be either supernatural in origin, or it may be the fire from the hearths of the villagers. In the latter case too, however, its not really being a fire must be due to divine intervention, just as in the case of the not falling down of the (heavy) objects on the Maenads' shoulders.

761 Tc15' P's TJ>v waTE µ11 lyye-Mv ("tout est bien pourvu que ... ") '. This is only partly correct. While one of the constructions involved in this contamination may well be, indeed, the second one of Roux' note (ica>.6v, being semantically closer to icpe-1aaov, is better than ayaTTT)T6v), the other construction is not KpE1aaov + gen. comparationis (of which there is no trace in l. 842) but rather 1rav icpdaaov ~ WCTTE lyye-Mv lµoL Cf. K-G 2, 503, 8.

843 4v 8oicij Following the Aldine edition, most editors read for P's c'iv; thus e.g. Dodds, without comment. Roux, however, keeps c'iv, taking it as= i\v, but as Oranje observes (p. 86, n. 211), dv never= -f\v in 5th-century literary Greek. Cf. also Ellendt-Genthe p. 51, section 5: ••Av pro Mv ... apud tragicos non dicitur'. On the other hand, = a c'iv) is not unproblematic either. Oranje renders (p. 86): 'I will decide what seems best', but for this translation to be valid the dependent clause should rather have a future indicative223 and an (indirect) interrogative pronoun or adverb, as in Hdt. 3.84.1: lPov>.e-uovTo paaLAla 6LicaL6TaTa a~aoVTaL (cf. also 7.175.1), Pl. Chrm. 176 c: [3ov>.e-vov lSn 1roL1'-

av

av (

ws

221. Oranje (p. 85) thinks that line 837 is an invitation to Pentheus to cause bloodshed. I fail to see how this sense can be given to Dionysos' words. 'AW surely expresses opposition, cf. Denn. 7 (i): 'in answers, objecting to the previous speaker's words or behaviour'. Moreover, it glaringly runs counter to the tenor of the whole passage 8 IOff., which is that Dionysos wants Pentheus to go unarmed to the Theban women. Nor do I understand why Kirk writes: 'At 796f. PENTHEUS seemed to be enjoying the idea of spilling some female blood; now he is deterred by it. Nothing between accounts for his change of mind( ... )'. This is simply not true, for at 812 Pentheus had responded eagerly to Dionysos' proposal to go and see the women. 222. 'yy1:>.iiv Pierson, Reiske: y1:>.iiv P. 223. Or possibly a subjunctive, but these I have not found.

103

ans~ Smp. 174 d: J3ou>..Eua6µE8a lSn lpoOµEv, Thg. 131 a: J3ou>.Eoo6µ£8a Bn 6paaoµEv. Kirk's translation 'I shall decide on what seems best' is better, for av 8otji at Ba. 843 is not an indirect question (since l1. is not an interrogative pronoun, see also note on 859f., and subjunctive + dv is not found in indirect questions), but rather the direct object denoting the topic the deliberation or the decision will be about. Cf. Hdt. 8.40.1: tva ... J3ou>..d,awVTm TO TTOLT1Tfov aiiTo1aL laTaL, Pl. Grg. 527 d: oTTo1ov dv n 'i)µ1v 8oicij, . .. J3ou>..Eua6µE8a: 'whatever we will think fit we will consult on' (= 'whatever topic seems suitable'; oTTo1ov is not an interrogative but an autonomous indefinite relative pronoun). 224 However, if av Boicij is read and taken as in Kirk's translation, teEaTL at 1. 844 hardly makes sense. Kirk renders 'As you wish', and Dodds 'As you please', but these are not attested meanings oHernn. "EernTL = 'it is possible, permitted' i.e. 'granted!', as at Hee. 238: teEO'T', lpwTa·, which follows after the question (234f.): El 8' lO'TL ... leLO'TOpfjaaL. This suggests that Pentheus' words are riv 8otji 'if that is all right (with you)' .225 In the manuscripts dv and 'f\v are not seldom confused, e.g. Held. 1020, IT 980; for other examples see Allen-Italie s.v. Mv. Finally, I agree with Roux that there is no need to assume a lacuna after l>..86v8' ls otKous, if these words are given to Dionysos, and ftv 8otji to Pentheus. 226 By interrupting Dionysos' words Penthei.Js wants Dionysos to believe that he feels in command and that as yet nothing has been decided upon. But Dionysos, of course, knows better, cf. his statement fien BE J3wcxas in 847.

852 oil

µ1'

8d:I\C71J

P's 8E>..fian can hardly be retained, since ou µfi + fut. ind. 3rd ps. is only used in oaths or solemn statements. See App. 3, p. 173 and fn. 17. 224. For direct objects with j3ov>.Eurn9aL cf. also Td crnov8aLfoTaTa in Hdt. 1.133.3: µE900K6µEVOL Bt lw9aaL l3ov>.Eurn8aL Td a,rov8aLfoTaTa TWV 1TP'll'Yµchwv, and 9civaTOV in IA 1102: 9dvaTOV ••. av 1TaTT1p j3ov>.EunaL. Note also that the indirect questions quoted above follow after j3ov>.EUEa8cn, while the direct object clauses of Ba. 843 and Pl. Grg. 527 d precede iL 225. So too Roux, but for her retaining the impossible c'lv 8oicfj. 226. I think fiv 8olCf,j has a condescending ('if it's the same to you') rather than an ironic (Roux) tinge.

104

854 811(:la(OLS' A dativus iudicantis, K-G 1,421, b. Cf. Med. 580-1: lµot yap, 8crns 46LKOS' wv ao4>bs >J-ynv I nlq>VKE, n>.El>.i.aKavfl. 859-60 yv~anaL 8t T6v AL6S' AL6vuaov, 8s- iri♦UKE'V iv TiM"L 8€65',

Such is the text as it is commonly punctuated, and with P's lv Tf>.EL. There are two interrelated problems here, the construction of 3s nlq>VKEV and the construction and meaning of lv Tf>.EL 8€6s. Dodds rejects Elmsley's and Hermann's view that 3s nlq>VKfV is equivalent to TlS' nlq>VKEV or otos nlq>vKfV, 'because the following words then come in awkwardly, as an explanation of 8s '. He inclines, nlq>VKEV. Willink (1966: 228) would therefore, to accept Dobree's delete the comma after 8E6s, thereby making the adjectives of the next line attributive modifiers of 8E6s. Roux, who has no note on these lines in her commentary, in her translation takes 3s nlq>VKfV lv Tf>.EL 8E6s as a simple relative clause: 'Alors, il connaitra le fils de Zeus, Dionysus, qui naquit dieu a part entiere', which is rather flat. Lacroix has something similar. Kirk reads ws (and EVTfA'llS', see below), translating: 'And he shall recognize the son of Zeus, Dionysus, as a god in perfect essence'. In my opinion the equivalence suggested by Elmsley and Hermann, while being objectionable syntactically, is semantically correct. rvwcrETaL is construed with a proleptic ('topicalized') accusative (~L6vvaov), followed by a relative clause which is the real object of yvwcrfTaL. By the prolepsis the subject of the relative clause, Dionysos, turns up as a pseudo-object in the main clause. Being 'topicalized', it has a certain prominence; cf. also above, nn. on 173 and 272, and K-G 2, 577-8. For more examples see Van Leeuwen on Ar. Ach. 118. "Os only seemingly functions as an interrogative pronoun, for in reality it is a normal relative pronoun, K-G 2, 438.227

ws

227. Cf. Ameis-Hentze (ed. of 1903) on Hom. fl. 2.365: yvw.awv, 'lSs TE Relativ, nicht Interrogativ'. Monteil (1963: 64f.), while translating this line, correctly, as 'Tu connaitras ceux des chefs qui sont lac hes, ceux qui le sont des soldats', nevertheless comments, incorrectly: 'La valeur interrogative de la relative doit correspondre au sentiment linguistique de l'usager; mais ii n'est pas exclu qu'elle n'ait etc au depart qu'une subordonnee objet du verbe principal'. There is no syntactic reason to assume that this construction developed from an original object clause into an interrogative one (the use of the definite article in 'la valeur interrogative' of course begs the question). At S. OC 571-2, to be sure, lSs is found coordinated with interrogative pronouns: au ydp µ' lSs dµL,

105

There are numerous other instances of the construction 'Yl yvwaKw (or another verb of knowing) + proleptic acc. + relative clause, e.g. Hel. 818: ou yvwanat µ' lSs dµ' lyw (µ' Blaydes; laudant Kannicht, Zuntz (1965: 46): y' Tricl.), Ar. Ach. 118: Tov hEpov TOUTovl / ly~6' 8s fon, KXncr8ElfTlS b ~l~UpT(ou, ibid. 442: TOVS µtv 8rnTas d6Evm (sc. &t) µ' lSs dµ' lyw. 228 Such sentences are pragmatically motivated variants of the construction 'YlyVWC1Kw (ol.6a etc.) + a single object relative clause, e.g.: s. Ai. 1259: µa8wv El ucrlV, s. oc 1171: l~ol6' QKOUWV Twv6' l>s EO'T' b TTpooTd.TflS,OT 1068: µi)TTOTE yvoll'IS Et. Also, with verbs of saying: Th. 1.136.4: 6T1>..ot ... lSs EO'Tl, X. Cyr. 6.1.46: TTEµTTE"l TTPOS T0V Kvpov dTTWV ~v. That the relative clause has the function object with these verbs can be defended by comparing our clauses with the autonomous relative clauses in e.g. s. OT 530: a ... 6pwa' ol KpaTOVVTES oux bpw, Held. 214-5: a ... O'E &t TELO'Ql >Jyw O'0l TTau1(, Ar. Pl. 932: bpqs a TTOlfL;, Pl. Ap. 22 d: a.XX' ,;TTtO'TaVTO a lyw OUK ,;mO"Ta.µT'lv. On the above analysis sentences like OT 1068: µi)TTOTE yvotlls d should, strictly speaking, not be taken as 'may you never come to know who you are' but as 'may you never come to know the person you are' .229 As for Ba. 859-60, since it is os- TTEUKEV in yvwaETal Tov AL6vvaov os- TTE6v; ,i Tl TO icdU&.ov ,rapd. 8ECiJv ytpas iv ppoTots ii xEtp' b,rlp icopu♦cts Twv lx8pwv icpdaaw icaTtxew; Thus the text of P, but for the correction icpdaaw (P has icplaw, p icpfoaw). Dodds translates and interprets as follows. "'What is wisdom?" The Chorus do not stay for an answer, but seek one indirectly by asking a second and easier question. "Or what god-given right is more honourable in the sight of man than to keep the hand of mastery over the head of a foe?"' This can well be defended (although I think, pace Dodds, that TO aocf>6v = 'cleverness' rather than 'wisdom')236 , but for the impossible T6 before icd.>.>.tov. There are, in fact, two ways of construing this sentence: (a) Tl ... ylpas, Tl being attributive, icd.MLov predicative: 'What right is more honourable ... ?'; thus Dodds. (b) Both Tl and icd.>.>.tov are attributive: 'What more honourable right is there ... ?'; in this case lv ~poTots rather = 'among men' than 'in the eyes of men'. If we keep T6 neither of these interpretations is possible. 237 Cf. 236. Unlike aocj,6s and aocj>la, TO aocj,6v, whether used by the Chorus or by others, never has a positive meaning in the Bacchae; cf. 203 (: Teiresias), 395, 1005. In fact, the substantivized neuter forms of aocj>6s (To aocj,6v, Ta aocj>Ci) very often would seem to have unfavourable connotations. Cf. E. Fragm. 583: !Sans >JyEL µtv E~. Ta 8' lpy' lei>' ots >JyEL / ataxp' foTl, TOUTOV TO aocj>ov OUIC alvw TToTl,JA 1214: vOv Bl, TdTT' l-µoO aocj>d., I 8aicpva 1rapl{w, Tro. 411: Ta aEµva ical 8~µaaLv aocj>ci. Cf. also S. Ant. 365: aocj,6v n ('cunning', Jebb) TO µT]xav6Ev Tlxvas ... lxwv, Pl. Grg. 483 a: TOOTO TO aocj>ov ICQTaVEVOT]ICWS' icaicovpyELS' lv TOLS' >.6-yms, Euthd. 293 d: cl.pa ... To0T6 fon TO aocj,6v; (ironical). I do not, therefore, agree with Cropp (1981: 39ff.), who writes that 'To aocj>6v in itself need not be taken in an explicitly pejorative sense'. This judgment is based, moreover, on the assumption that 'the Chorus are saying that Dionysus' anticipated victory, whose glory they will share, will finally settle the question whether it is Pentheus or Dionysus (and his followers) who is really aocj>6s'. This assumption is to my mind entirely mistaken: there is not the slightest indication that the Bacchants ever considered Pentheus potentially aocj,6s. 237. Of the older commentators only Bruhn has a, very brief, note on T6: 'der Artikel wie in 492', i.e. as in Tl µE TO &wov tpyciO"(l; However, these constructions are not similar at all. The question of Ba. 492 (cf. note ad loc.) is a predicative construction asking for further information on a point raised in the preceding context; it can be paraphrased by: Tl l°aTlv TO &wov IS µ' l-pyciO"(l; For an example of such a predicative construction involving a comparative see S. OC 598: Tl yap TO µE1(ov i\ icaT' dv6pw1rov vooELS';, which refers to 595 and 597: "'id quod maius est quam pro homine"; Tl predicate' (Kamerbeek); 'What, then, is thy grief that passeth the griefs of man?' (Jebb). The question of Ba. 877, however, does not ask for additional information on a point raised earlier; accordingly it cannot be paraphrased in this way.

109

also Dodds, end of note on 877-81: 'As Paley saw, this' (viz. the deletion of T6) 'improves the metre as well as the Greek' .238 As for 1'i (i.e. the 1'i of l. 877), this is the 'corrective' -1' (for which see the note on 556); its full meaning is: 'Or is this the wrong question in view of this other possibility?' 239 All in all I think the reasoning of the Chorus is as follows. 'What is cleverness? Or rather, what more honourable god-given right is there among men than to keep the hand of mastery over the head of a foe?' This must be a rhetorical question, the implied answer being: ov8£v ylpas icdXALov etc. The Chorus, who in the previous stasimon had demanded Pentheus' punishment (555: 0~pLv icaTdCTXE"S'), 240 by now know for certain that Pentheus will, indeed, be punished (cf. 847: &'.>an 8(1CT1v). They can, therefore, quite confidently (cf. the rhetorical question) voice their approval of his imminent punishment, which is referred to implicitly in the generalized formulation of their question. The words XE"Lp' ViTEp icoptx/>ds icaTlxnv perhaps prepare us for what will happen to Pentheus' head; if so, they are, when seen in retrospect at ll.l 139ff., a rather grim understatement. I conclude this note with a discussion of two alternative approaches of our lines. Willink (1966: 229-31) proposes to read T( TO aocf>6v; ~ TL icdMLov ... ylpas ... 'f\ ... icaTlxnv; 241 In his view Tl TO ao238. Pace Dale (1968: 148): 'On no account to be emended'. Perhaps not on account of the metre (but cf. Dodds), but syntactically T6 is simply indefensible. Oranje, however, retains T6 (p. 162), claiming, implausibly, that the second question aims at defining the alternative to TO uotj,6v, viz. uofj,la, i.e. Dionysiac wisdom. He paraphrases as follows: 'What is clever thinking? Or (in other words)' [this is wrong, see next note - AR] 'which is (of clever thinking and of Dionysiac wisdom) the most beautiful gift of the gods in men's judgement: (clever thinking) or [that is to say what is the more beautiful gift ... THAN] with a more powerful arm to keep down his enemies?'. This I do not understand. 239. Oranje, p. 162, takes 1'I as 'or in other words', wrongly, since 1'I has disjunctive meaning. The same mistake is made by Cropp (1981: 39): ' ... or (to ask an equivalent or related question) ... '. Cf. for 1'I also K-G 2, 297: ' "H zu Anfang eines Satzes driickt oft eine B er i ch ti gun g des vorangehenden Gedankens aus = oder vielmehr, oder genau gen om men'. 240. Although syntax and meaning are slightly different, KaTlXELV in 880 may well echo KaTciUXfS' in 555. 241. He grossly misrepresents, though, both the traditional interpretation and Dodds' views. Dodds does not find the Chorus 'inconsistent'. The only time he uses this

110

4>6v; expresses the idea that the Chorus are puzzled by the shifting meanings of aoJ.ov etc. (b) concerning To KciMl.ov Roux remarks: 'le comparatif precede de !'article est employe en poesie avec le sens du superlatif: TO dµnvov, TO KpE'Laaov, TO xe'ipov, TO >.wt:ov'. This is, first, somewhat misleading, for her examples should also have a noun; as her translation shows, she takes KciMl.ov attributively. Second, although the comparative preceded by the article may, in fact, have superlative meaning, the noun phrases mentioned by Roux, in so far as they occur at all, do not have this meaning. A computer check of these words in the dramatists, Homer, Herodotus and Plato produced just four instances (none of the combination article - comparative - noun). 244 Nor do K-G, in their discussion of this use of the comparative (1, 22, 3), have any examples from tragedy. Again, the examples presented by K-G involve either a predicative or a substantivized comparative, but not the combination article - comparative - noun. Moreover, such comparatives are accompanied by a genitive (perhaps a gen. partitivus rather than a gen. comparationis) or a locative-partitive expression. Examples are: Hom. Od. 7.156: ('Exlvrisl, 8rt c2>aLTJKWV av8pwv lTPO'YEVf'TTE"POS' ~EV, 3.362: ots -yap µna TO'iaL 'YE"palTE"POS' Eiixoµm Elvm; X. Cyr. 5.1.6: TJµwv o -yEpalTE"POS' · There are, then, no parallels for TO Kcik >.i.ov -ylpas meaning 'the most beautiful boon'. ' (c) Roux follows Blake in reading~ at the beginning of 879. In principle, ~ is a neutral interrogative particle, implying nothing as to the nature of the answer (positive or negative) expected. Obviously, this is not what is needed, if this question is meant to express disapproval of subduing one's enemy. It is probably for this reason that Roux' commentary after I. 879 adds an implicit negative answer: '(Certainement pas car)', thereby straining the meaning of~. (d) Roux' views not only glaringly contradict the attitude of the Chorus at 11. 991ff.: hw 8lKa cf>avEp6s etc. (cf. also Dodds on similar views of Blake), and at 11. 553ff.: cf>ovlou 8' dv8pos li~pLv KaTciaXES', but also the general Greek ideas on this matter, cf. Dodds' note, and e.g.

'fl

as

244. A. Supp. 596: To µEtov icpELaa6vwv Kp«TUVEL, text uncertain, see Page ('locus insanabilis'); Friis Johansen & Whittle translate 'exercises the power lesser than superiors' (power)'; S. OC 598, see fn. 237; Med. 911: all' ls To >.ijiov aov µe8E1TTT11CEv ictap, where To >.ijiov clearly has comparative meaning; Pl. Ale. I 108 e: T( TO clµewov >.l-yELS, 13 'A>.icL~Ld8f1;, where To clµELvov is an instance of 'mention': 'What do you mean by "the better"?'

112

a maxim like: v6µou Tov lx8pov Bpav, &1Tou M~"QS', icaicws (E. fragm. 1091, v6µos Grotius).

886 Td. 8EWII D.: 'TCl 8Ewv virtually= 8Eoi,s'. This is not quite correct, for the neuter plural in such cases may denote 'everything belonging to the sphere of the entity involved'. Thus, in the context of the Bacchae, and thanks to the verb a{;fnv, Ta 8Ewv may refer to the cult of the gods, to religious performances, etc. In other contexts Ta 8Ewv refer to other phenomena connected with the gods, as in Pho. 382: 8e't cf>lpnv Ta Twv 8Ewv 'what is sent by the gods'. Cf. KG 1, 269. See also note on 1002.

893-96 icoi14>a yd.p 8a,rdva 11011(· Cew taxi.Iv T68' lXELV,

a

TL 1r0T 1 clpa

Tl>

8aL1J,6VLOll 1

T6 T 1 lv ,cp6vl\) µaicpci v6µLµov dd 4>iiaEL TE irE(f,uic6s-. 893a T68' Heath : T' P The construction of the words T68' lxnv ... ll'Ecf>uic6s is much disputed. The central issue can, I think, be summarized as follows: how many different entities are said here to have power? Our answer to this question depends on how we construe the relative clause and especially the two TE's. The following interpretations have, among others, been put forward. SANDYS translates: 'It costs but little to hold that that has (sovereign) power, whate'er it be that is more than mortal, and in the long ages is upheld by law and grounded in nature'. He apparently (cf. Dodds' note on this translation) takes the first TE as coupling TO lv xp6vq1 µaicpl\) v6µLµov d.El cf>van TE ll'Ecf>uic6s with TO 8mµ6vLOv, and the second as coupling cf>oon ll'Euic6s with v6µLµov aEL DODDS rejects the latter part of this interpretation, on the ground that 'the Chorus can scarcely take thus for granted the identity of two things which Eur.'s contemporaries habitually contrasted, TO v6µLµov and TO cf>uon '. Dodds thinks it 'more likely that they assert the identity of TO v6µLµov and TO 4>van '. He has no note on the syntax but translates as follows: 'And to consider what has been accepted through long ages (to be) an eternal truth and grounded in nature'. 113

GREGOIRE translates: 'II en coiite fort peu, en effet, de croire a la puissance de ce principe - quelle que soit la nature de la divinite - qui au cours des ages a toujours eu force de loi et tire son origine de la nature'. VERDENIUS 1, rejecting Dodds' interpretation, connects ad with 'Tl'C q,vic6s. He rejects the interpretation of Gregoire as well: 'I do not believe that v6µLµov and 'Tl'E4,vic6s are attributively connected with T66E (Gregoire). We should rather mentally repeat loxvv lxnv'. WILLINK claims that 'There is a much simpler interpretation' (viz. than that of Dodds) 'if Ta6 ', not T68 ', is read in 893: ". . . the following things have strength, namely TO 6mµ6vLov, TO v6µLµov ad, and TO

q,oon

'Tl'E4'1J1C6r;;"'.

Roux (who reads Ta6', following Willink (see above) for P's T') favours '!'interpretation proposee par Deichgriiber (Hermes 10, 1935, p. 322-348) et par le R.P. Festugiere: loxvv lxnv a deux sujets, annonces par Td6' ... , le premier, lS TL ••• TO 6mµ6vLOv, reprenant les v. 882-890, le second, TO v6µLµov ... q,uon TE 'Tl'E4,vic6s, les v. 890-892'. She translates: 'II en coiite bien peu de croire ala puissance et du divin, quel qu 'ii puisse etre, et de la tradition consacree par les siecles, qui, toujours, est issue de la nature meme'. KIRK follows Dodds. LACROIX, 245 finally, translates: 'II en cofite peu, quelle que soit la realite de la puissance divine, de reconnaitre la valeur de ce qui a re~u d'un long temps une eternelle force de loi et qui se trouve aussi conforme ala nature'. Of these interpretations I prefer, for reasons to be set out below, that of Roux, which is perhaps also that of Verdenius 1 (he is not explicit on the syntax), although I believe aEt should be connected with TO v6µLµov lv µaicp{j> xp6vci> rather than 1TE4,vic6s. 246 On this interpretation the first TE coordinatesT66' ... , lS TL 'Tl'OT' dpa TO 6mµ6vLov with TO v6µLµov lv µaicP4> xp6vlj) q,oon TE 'Tl'E4,vic6s, while the second TE coordinates v6µLµov lv xp6vci> µaicP4> ad with .Ea8aL ical hn8uµe-1v 6LaLpELS' TatJTov 6v, where the subject of 6LmpE1S' is Prodicus. From the absence of T6 before lm8uµe-iv we may infer that unlike Prodicus, the speaker, Socrates, does consider the two concepts identical. On the basis of these examples our conclusion must be, I think, that oon tre-uic6s does not refer to a separate entity. GREGOIRE'S and LACROIX' interpretations will be discussed together below. WILLINK'S interpretation cannot stand, since he assumes the existence of TO vaEL tre-uic6s. Cf. above, sub Dodds. As for his proposal to read Ta6 ', the plural, while being possible, is perhaps slightly out of place, considering that only two items follow. Also, syntactically speaking the relative clause as it is, i.e. with an enclosed 'antecedent', makes a rather inconspicuous first item of the pair introduced by Ta&: 'the following things, namely whatever it be that is more than mortal and ... '. I prefer keeping Heath's T66(e-}.

ws ov

With the exception of Willink's, the above interpretations are based on taking T66' ... B TL TTOT' cipa TO 6mµ6vLOv as a single noun phrase, B TL ... TO 6mµ6vLOv defining T6&. Syntactically, the relative clause can perhaps best be considered a clause with an enclosed 'antecedent', T66', B TL troT' cipa TO 6mµ6vLOv being a variant of: T6& TO 6mµ6vLOv, B TL troT' dpa. Cf. He/. 306: 'EAE'VT), Tov l>.86v8', lSaTLs foTlv 6 Elvos, I µTl traVT' ci>-T18ii 6oEa01Js dplliclvm, Ba. 247: ... Of3pns uf3plCELv, !SaTLS' foTLv 6 Elves (as against Ba. 769: TOV 8alµov(a) ... T6v6' lxrrLS' lCTT'). Alternatively, TO 8aLµ6vLov might be considered an apposition to B TL troT' cipa, as a kind of specifying afterthought: 'this, whatever is its nature, I mean TO 6mµ6vLOv', as in Hipp. 101: TT)V6', fi 1TVAaLC1L aa1s ElC1TTIKE"V KvTTpLS'. Cf. further K-G 2, 416ff., especially 423; also Moorhouse 27lf. on 'Articular antecedent or demonstrative in the relative clause'. Gregoire and Lacroix, however, apparently take the relative clause as an autonomous parenthetical clause. Thereby T66e- is made the 'antecedent' ofT6 TE" v6µLµov ... ,re-uic6s, which is syntactically not very likely. Also, if B TL TToT' dpa TO 6mµ6vLOv is taken as a paren117

thesis, the Chorus exclude the divine from the entities which have power, which is in itself unlikely, and runs counter to the first part of the antistrophe (882-7).

902-11 d8a.(µ(l)V, (tv) a>.p"1, µaicap(C(I) For a discussion of the semantic distinctions between these words I refer to De Heer (1969). On p. 14 he notes about µciicap: 'The name µciicap evokes a sense complex which is associated with the divine world.( ... ) This association with the divine imparts to the word an emotive value when applied to humans'. He renders (p. 95) µaicap((w at Ba. 911 by: 'I think that such a human being is elevated above all mankind'. Cf. also p. 99: 'None of them' (sc. the different types of 'happiness' - AR) 'is thought to be enduring, but Eu8a.(µwv is connected most intimately with the sense-component of permanence; the condition of being l>}43u>s is an aspect, a manifestation, of being Eu8a(µwv'.

909 d1rlP11aav D. (1st ed.): "'withdraw", vanish. But the text is open to serious doubt.' (2nd ed.): ' ... the real difficulty, as Jackson pointed out (Marg. Scaen. 237), is that d.tro~a(vnv is regularly used in the contrary sense, of things "coming true"'. V.1: ' ... it cannot be maintained that the use of dtro~(vnv in the sense of "to come true" was "firmly established" (Jackson). Consequently the traditional rendering ... by "vanish", "come to naught" may after all be correct'. R.: 'le verbe dtro~a(vnv, employe absolument, a toujours le sens de «tourner bien, reussir, se reveler vrai»; ( ... )En consequence, le v. 908 doit etre corrige: on lira soit dvo;\~oL (Denniston), soit dv6;\~ws (Jackson)'. Verdeniusl rightly argues that in three out of the five cases adduced by Jackson to illustrate the meaning 'come true', this meaning is due to contextual factors. Likewise, it could well be argued that at our line the context, viz. at µev TfAEUTWOLV iv 6)4341 in the preceding line, determines the meaning of d.trl~Tloav as = 'vanish'. Another argument to defend this meaning here is based on the semantic value of the gnomic aorist. The gnomic aorist often conveys the idea that the state of affairs concerned is at once completely realized, as in Hdt. 3.82.3: EK 8£ TOV ct>(>vou dtrl~T'I ls µouvapxt'T'lv ('Mit dem Mord ist gleich die Alleinherrschaft da' - S-D 283). This nuance of 'complete realization at one 118

blow' is better suited to the meaning 'vanish', I think, than to 'come true': 'some hopes reach to the end in prosperity for mortals, others are gone at once', i.e. they are hardly there and they are gone already. Conversely, the present TE"MUTwat.v may convey that coming to prosperity is rather a time-consuming process.

al TOIi ,rp68uµov 6118' ICTA.. For a discussion of this use of the accusative in abrupt addresses see Moorhouse (1988: 213-4). He considers such accusatives 'deictic'. There is no need to assume, with Roux, a 'verbe sous-entendu '. 912

917 ,rpi11'ELS R. (transl): 'Tu sembles'. K.: 'you look like'. L. (transl.): 'tu as l'apparence de'.

IIplm.l) is rather: 'to be conspicuously like', cf. LSJ s.v. II. and Homeric ,rpl,rw: 'apparaitre distinctement' (Chantraine s.v. ). 919 8Laad.s 8t 91\~as ical 1r6A.Laµ' t1rTdaT0µ011 For this use ofKal in 'whole and part' expressions see footnote 178 with 513 above.

922 dU.' ~ 11'0T' ~aaa D,1p; D.: 'Were you perhaps (iroTe-) an animal all the time?'. V.1: 'Dodds: "perhaps". But I have not been able to find an example of'7i 1TOTE'. R. (transl.): 'Tu etais done un fauve?'. K.: 'Were you a beast before?'. L.: '(La) construction de l'imparfait avec 1i 1TOTE" (plus souvent cipa) est tres usitee'; (transl.): 'Etait-il done vrai que tu es une bete?'. This is a rather confusing array of opinions. As for Dodds' view, I agree with Verdenius that 1T0TE" = 'perhaps' is without parallel. I note, in this connection, that LSJ's treatment of 1T0TE" in questions is misleading. One reads s.v. III 3: 'with intensive force, in questions', which is illustrated by TlS' iroTe-; who in the world? 251 Among their examples are A. 251. LSJ's division may have been influenced by that of Ellendt-Genthe, who devote a rather large section (III) to a 'significatio dialectica'. 'Nos etwa, denn, wohl, vie/-

119

Eu. 408: Tlve-s 1TOT foTl; Pl. Tht. 187 d (an indirect question): Tl 1TOT la-rt Tou-ro TO mi8os; ('what it can possibly be') and S. OT 335: ouK ... l~EP£LS' lTOTl ... ; The latter example, however, does not belong here, since (ouK ... ) 1T0Tl = '(n)ever', cf. Jebb ad loc.: tandem aliquando. Similarly S. El. 940: ~ TOUS' 8av6VTaS' UavaaTT)aw lTOTl; = 'Shall I ever ... ?'. In fact, intensifying 1TOTE" would seem to be confined to Tls, Tl and other x-questions, just as, in English, the intensifier ever is confined to wh-questions; cf. Quirk et al. (1985: 817, Note b). Incidentally, the pragmatic effect of lTOTE" may very well be the same as that of ever in such questions ('emphasizes the bafflement or emotional in1

1

volvement of the speaker', Quirk et al., ibid.). Lacroix' note is wrong on two counts (Roux, who has no note in the commentary volume, probably shares his opinion, which goes back to Dalmeyda). First of all, the combinations a.>J..' ~ lTOTE", or~ lTOTE", far from being 'tres usitee(s)', actually do not occur at all in Homeric and Classical Greek, with the exception of rare cases like S. El. 940, mentioned above, or Ar. Nu. 865, where lTOTE" = 'ever', and where, more.over, lTOTE" is at a considerable distance from~- Nor are they mentioned by Denniston. Second, the combination ~ dpa (if this is what Lacroix means) is very rare indeed; it is 'mainly Homeric', Denniston, 284. Only dpa + imperfect is, indeed, very common. IToTE should, then, be taken temporally: 'once, on a former occasion, before', or possibly: 'just now'. Cf. for the temporal vagueness inherent in the meaning of1TOTE" Ameis-Hentze on 11. 8.108: ',roTl einmal, ohne Riicksicht auf nahe oder feme Zeit: bier 3 Tage vorher ( ... ), x 290 sogar von demselben Tage' .252 Cf. also S. OT 1122: Aatov 1TOT ~a8a av;, where the context(: Aatov) makes clear that lTOTE" refers to a moment that lies a considerable time before the 'now'. 1

929 oi,x ~S' D.: 'oux, sc. foT11KWS, which is easily supplied from l~lO'T'TlKE '. As the examples of K-G2, 575 show, we should rather supply to oux the verb from the clause of comparison, so oux, sc. Ka&flpµooµlvos. leicht, doch, am Ende'. In reality, however, the examples given in that section, but for those oftroTE in T(s- and other x-questions, rather have temporal meaning. 252. That TrOTE does not necessarily refer to a time long before (or after) the present moment also appears from Aristotle's definition oftroTE, Phys. 222 a 25f., where he notes that TrOTE just involves 'a measurable quantity of time vis-a-vis the now'.

120

949-5 0 ~ XEpotv dvaa1rdaC1> ic:opvcf>ats- f,,ropa~v ~µov ~ ppaxCova; R. (transl.): 'Ou devrai-je deraciner les cimes ala main, en poussant pardessous de l'epaule ou du bras?'. K.: 'Or shall I tear them up by hand, putting a shoulder or arm beneath the peaks?'. L. (transl.): 'Ou bien (dois-je) creuser de mes mains le sommet du mont en le soulevant de l' epaule ou du bras?'. On d.vacrnaaw (fut. indic. or subj. aor.?) see Appendix 4. As appears from their respective translations Kirk takes T..ol/s referring to the pointed tip of a fir-tree. (d) As Willink argues (1966: 232), the smooth cliff makes a strange hiding-place for Pentheus. All in all, then, it seems preferable to connect c'mo ... aK6>.o1ros with ~E"TaL rather than

with 8oKE"OOVTa.254

There remains, however, the question as to the meaning of aK6Aot/J. Just as in the case of 'tip of fir-tree', there are no convincing parallels either for Kirk's 'pinnacle' or for Lacroix' 'pente escarpee'; the latter can hardly be accepted as an extension of the general meaning 'pointed object'. The uncertainties connected with aK6>.ot/J have led several scholars to emend, e.g. ft K>.wv6s (Dodds: ' ... we ... expect here some vision, not necessarily exact, of the tree in which P. sat'; for this fallacy cf: above); ij (this should rather be aK6TTE"AOS (Wecklein, followed by Kopff), which seems rather weak, if not pedantic; C1KOTTE >.wv (Verdenius 1, for ft aK6>.o1ros-, arguing that ft originally indicated a variant; he refers to Ion 274: C1K61TfAOV 'ijµaeav TTETpaS', and 714-5: &Lpa&s IIapvaaoii ,rfTp«S' I lxovaaL aK6TTE"AOv), but the plural hardly makes sense; also, the word order does not favour connecting C1Ko1rl>.wv with d,r6. If emendation is called for I suggest replacing ft aK6Ao1ros by TTpo/3AiiTOS. This adjective occurs as an epithet of aK6TTE"AOS' at fl. 2.396, and of TTETP11 at fl. 16.407, as well as three times as an epithet of dKTal in Od. It is again used of TTETPTI by Archilochus, fr. 49 D. In tragedy, it is found at S. Ph. 936 (plural), as a noun. In these cases, it is true, 1rpol3A~S' is connected with the sea, but at II. 12.259 it is used of aTiiAaL 'posts, buttresses' so it clearly was not confined to cliffs jutting out into the sea. What may have happened at Ba. 982 is that ft aK6TTE"AOS' was added as a gloss on AE"VpclS' d,ro ,rlTpas ,rpo~AiiTOS by a copyist who

q.)

254. Although at Theoc. 26.10 it is Pentheus who is looking d>.Ll3chw mhpas 4tro. But see Gow ad loc.

124

remembered that trpol3~:f1s in the Iliad was used of cric6trE"AOS as well as of trlTpTJ. This gloss was then misread as ,; cric6>..otros, perhaps by someone who took -OS for a genitive ending, and this may subsequently have found its way into the text. The interpolation of'f\ cric6>..otros may have been facilitated, as Tyrrell suggests, by the joint use of trETpa and cric6>..olf, as means of punishment, as in IT 1429-30: ,; icaTa crTu>..ou 'TTETpaS / p(.lf}(l)jlE"V ,; O'IC6>..olf,L miewµE"v 8€µas. For the appearance of two attributive adjectives with one noun cf. K-G l, 277. As for the metre, with the reading trpo(3XfjTOS we get a kaibelianus responding with a kaibelianus at 1003: (983) trpo(3XffT..otros in the way sketched above. 996-1010 996 ycSvov 997 998 999

1000

1001 1002 1003 1004 1005 1006 1007 1008 1009

1010

'Y'l'YEvfi,

6s d8(1Cl\) yvliSµq. trapav6µf\)

6pyq irEpl cad>, BdicxL ', llpyLa µaTp6s- TE acisµavdaq. ,rpdm8L ,rapaic6,rl\) TE >..nµan aTlUE°TaL, Tdv(icaTov ~S' icpaT1'awv p(q., yvwµav awct,pova· 8dvaTos c8'> d,rpocf>daL· aTos ls Td 8e-6>v lcf,u, (3poTE(ws- T' lxnv d>.uiros p(os-. T6 a4>611 oil 4>8011ta· xa(pCa> <pdouaa Td8' he-pa, µe-yd>..a ~ve-pd T' llvT' - dn .611Ta nµciv 8e-ous-. T1

998 mi suppl. Scaliger 999 µavefo~ Brodeau: µav€1o'a P 1002 8' post Dodds addidi 1003 ds P 1004 (3poTdws Murray: (3poTdw P 1005 8ovw edd.: 86vw P an 4>povw? 1006 Ta8' Heath: TQ 8' P 125

1007 T' lSVT' Musgrave: Twv P dyn Dodds: ad P 8' suppl. Dodds 1009 d,ayo0VT' Hermann: d, d:yovVT' P Contrary to the practice followed so far I have first presented the text of this 'locus difficillimus, et qui omnes Interpretes exercuit' (Musgrave), as I think it should be established. An important characteristic of this text is that P's yvwµav awcf>pova in 1002 is retained. Since the textual difficulties are mainly found in lines 1002ff., I will concentrate on these lines, bearing in mind Brunck's warning 'Hos versus praetereat lector, nisi si quis in eorum emendatione ingenii vires experiri velit'. While defending the text as it is presented above I shall discuss various other proposals. 996. With Roux, Kopff and others I put a comma after 'YT'l'YE"V'Tl, thereby making y6vov 'YTl'YE"vii the antecedent of the relative clause. Other editors, e.g. Murray, put a full stop after 'YTIYE"vii, whereby the relative clause is generic. As Dodds observes, the choice depends on what we read in 1002. See there. 1001. TdvCicciTov is Wilamowitz's correction of P's Tdv av(KaTov. Tyrrell defends Ta.v d.vticaTov as an internal accusative, sc. vlicav, 'the fight that cannot be won', but this is implausible; d.vCKT1TOS -:I: 'that cannot be won' but 'that cannot be defeated'. Also, with Ta.v d.vticaTov the line is one syllable too long. 1001. ~S' icpciTJiawv. Not 'as though to subdue' (Kirk), nor 'esperant que sa violence triomphera' (Lacroix), but 'qui pretend maitriser' (Roux) or, as I would prefer, 'who believes he will master'. Cf. note on 729. 1002. yveiSµav aeiS4>pova. Thus the reading of P. Many of the problems connected with this line disappear if we put a comma after ~(q. and a semi-colon after awcf>pova. Thereby yvwµav awcf,pova is made an apposition to Tav(KaTov, specifying what the d.v(KaTov consists of, namely 'sensible judgement'. From the point of view of the Chorus this means: the sensible judgement shown by those accepting and honouring Dionysos as a god like the other gods (a theme which will be elaborated on in the sentence xa(pw 0Jipe-uovaa ... nµav 8e-ous), which sharply contrasts with the d8LKOS yvwµTI of Pentheus. There are, in fact, two sides to awcf>poavVTI as this concept is present in the Bacchae. First, it is used in connection with the (self-controlled) behaviour of the bacchants, cf. 317f. (Teiresias to Pentheus): ical yap lv ~icxe-vµaaw I ofo' ~ ye- awcf>pwv ou 8Lacf>8apiiae-TaL, 683f. (Messenger to Pentheus): TIMov ... aw4>p6vws, 940 (Dionysos to Pentheus): lhav trapd Myov 126

aw4ipovas f3povds, 504 (Stranger to Pentheus): avBw µE µ11 &tv crwcf>povwv ov aw4ipocnv, 1150 (Messenger to Chorus): TO crWv 8Ewv I Ka.>.>.i.crTov, 1347 (Dionysos to Messenger, Chorus, Kadmos): El 6€ aW.6-yoLaL 6' ovK lvELcrl a0t. cf>pa,Es. For the construction of KpaTELV with the accusative of an abstract noun cf. Hipp. 696: TO yap 6a.Kvov crov TTIV &a.-yvwaLv KpaTE'i.

1002-3. 8dvaTos-

daL/aTOS- ls- Tei 8dlv lcf>u.

These words resume, in the form of a sententia, the theme of the refrain (991ff.): hw 6(Ka cf>avEp6s, LTw ~Lcf>TJcf>6pos / cf>ovEvovcra KT~ 'Atrpocf>a.crtcrTOS probably has active meaning: 'bringing forward no pretexts, unhesitating, straightforward', so too Gregoire, Roux, in her translation ('sans detours')25 6 and Verdeniusl. I have added 6',257 which seems semantically appropriate and ensures metrical responsion, the result being: -yvwµ.av awcf>p(jvi1· 8dlvi1Tos 6' iftrplS4xfcrr- 2 dochmii in responsion with 982: µaTfTp trp.E f3p6xov = atpn (rather: l>.E). We may conclude that the case variation &JipaypEVT~ . . . trrn6vm is by no means exceptional and can be satisfactorily explained along the lines drawn by StahI.267

1024 dTVXE"fS' Most editors correct the, grammatically indefensible, reading of P to dTVX£LS', i.e. an impf. without augment, which was apparently proposed by Kirchhoff. Tyrrell and Bruhn, however, follow Heath in printing 11ir TVXELS'. Which of these should be preferred? The facts are as follows. 268 (a) Murray (lntrod. to tomus III, p. iv) defends £VTVXELS' with a reference to Lautensach: 'Die von Zusammensetzungen mit di abgeleiteten Verben zeigen, im Unterschied von den Verben mit dem Stammanlaut £u-, niemals eine besondere Augmentation' (= Lautensach (1899: 146-9)). To illustrate this statement Lautensach adduces two forms of '£~-verbs' found in inscriptions.269 Unfortunately, however, these forms are inconclusive, since they are perfect indicatives (£V£py£T11Ka[cn), £UE[py)dTh,Kdv)), where strictly speaking the augment is not 267. Ba. 1020-23 differs from the passages from tragedy quoted above inasmuch as in the latter a case form of a personal pronoun is involved, whereas our passage has a noun (8JipayP£tmj). This may be accidental; in other genres nouns, as well as proper names and pronominal forms other than personal pronouns, are not uncommon. Cf. fl. 10.187: c:}s Tliiv ~Buµos lhrvos ... 6>.i.'.i>.EL / VUKTa ♦ v~aaaoµlvoun Kaicfiv, Heraclit. 107: KaKol µdpTUpfS dv8pwiroLaLv 6(j,8a~ol Kal @Ta j3ap· ~pous .J,vxas lx6vniv. Th. 1.62.3: ~v & t, yvwµ11 TOO 'ApLaTlws TO µlv µEO' l-avTOii aTpaT6irf8oV lxovn ... Cf. Stahl 710. 268. After I wrote the draft of this note Mastronarde •s article on the same subject appeared (Mastronarde Glotta 61 (1989) 101-106); his line of argument and his conclusions are much the same as my own (p. 101: 'The purpose of this note is to point out that Lautensach's Law is based on a faulty inference from the evidence of Attic inscriptions and to recommend that future editors of Attic drama be consistent in using 11u- as augment and reduplication for all verbs in Eu-'). I have nevertheless decided to keep the present note, since the emphasis in my note is on the internal consistency of Lautensach' s argwnent rather than on the evidence of the inscriptions. 269. Murray writes that they are 'eiusdem aetatis'. However, since they are from 320-330 B.C. this is slightly exaggerated. Lautensach, in fact, more cautiously speaks of 'ungefiihr derselbe(n) Zeit'.

133

involved at all. Threatte, in his Grammar of the Attic Inscriptions (1980) simply writes (p. 385): 'In the case of verbs in initial Ev- (like EiiplaKw, etc.) spellings in EV- in the aorist and perfect also appear by the end of the fourth century, and it is likely that they ... are due to the change of 11v to EV rather than merely neglect of the augment or reduplication ' 270 ( ... ) 'in the Classical Period only T'IVJ>f&!'J, 11\jpT1TaL etc.' (b) Murray mentions the following 'di-verbs' which 'Euripidis editori in quaestionem veniunt': EVTVXELV (12),271 EUM~Eta8m (2),272

EUTprnlCnv (2), Eucf>rlµEtv (1), Eva8EVELV (1 ?),273 Evcf>palvnv (2).274 Add EVM>'YELV (1);275 eliminate drrprnlCnv (in both cases (Cy. 594, IA 1111) we are dealing with a form of the perfect stem) and EVcf>llµEtv (does not occur in impf. or aor. indic.). (c) To the list given under (b) Murray adds: 'In his codices aut omnes aut meliores semper Ev- praebent, nisi Hee. 30111uTuxovv omnes, Iph. A. 1111 11uTpEmaµlvoL LP, Or. 217 11ucf>pavas omnes, 276 ib. 287 11ticf>pavE MVLP: Evcf>paVE M2AB'. Actually, however, this is less impressive than it seems. The fact is that for the other group, viz. the verbs in initial EV- that does not represent d,, the mss. often present the unaugmented form Ev- as well, although on the basis of Lautensach's statement quoted above, we might expect them to have 11v-. The verbs concerned are E08w, Ev8vvw, EvvaCw, diplaKw and EuxoµaL. Thus all manuscripts have d,Bov at Ba. 683, d,pE at Pho. 1570, dipov at Hipp. 699. 277 In other words, in these matters the ms. authority should perhaps not be trusted too much. Cf. also Lautensach 47 on this class of verbs: 'Derartige orthographische Fragen, wie die Augmen tierung der mit EV anlau ten den Verba, lassen sich endgiiltig nur durch Zeugnisse der Inschriften und Grammatiker, nicht der Handschriften entscheiden'. Remarkably enough, however, the ms. 270. Prof. Ruijgh observes that since Modem Greek has T)Upa [iwra] it is doubtful whetherT)u- developed phonetically into EU-.

271. Impf.: Hee. 18, 301, 1208, 1228, IT 329, Fr. 285.20. Aor.: El. 8, Hel. 1030, HF 613, 1221, Or. 542, Tro. 935. 272. Or. 748, 1059. 273. Cy. 2. 274. Or. 217,287. 275. Fr. 347.2. 276. Efic/>pavas- M2, according to Murray's own app. 277. According to Barrett's apparatus (d1pov is not in Diggle's app.).

134

authority all of a sudden is considered an important factor by Lautensach when he tries to defend (pp. 147-8) the unaugmented forms of E"li-verbs. This strikes me as entirely arbitrary. The only thing we can say of all verbs beginning with E"U- is that the mss. very much favour unaugmented imperfects and aorists. (d) As for the grammarians (and lexica), the majority of these favour TJU-. (Herodianus II 789, 7 Lentz, Etym. M. 400, 26, Schol. in Dion. Thr. III 41, 15 Hilg. (= 804, 31 Bekker). The testimonies to the contrary are scanty and inconclusive. Etym. M. 399, 41 is a remark about the (probably artificial) perfects dlct,payKalTJvcf>payKa. The lemmata of the Suda ss.vv. d.1X6'Yf'IO'a and TJVM'YTJO'a belong to the 'glossae marginales ceteraque additamenta recentiora', according to Adler. There remains a number of glosses from other lexica, which mostly consist of just a verb form and an explanatory note (e.g. Hesych. s.v. E"VBoKTJaav· e-l3ov>.oVTo, T)8E"ATJO'av· T)'Y6.'ITTlaav); most often, we are completely in the dark as to the provenance of these forms. Again, from phrases like lTap' T)µLv (Herodian. II 789, 7; Etym. M. 400, 26) or KaTd KoLiriiv 8L6.MKTOV (Herodian. ibid.), which occur in connection with past tense forms in initial E"U-, we get the impression that unaugmented forms of all verbs beginning with E"U- were typically a feature of Hellenistic and later Greek. Herodian's remark runs: e-lTE"KpaTTJO'E"V oliv ... Tl E"U &ct,eoyyas lTap' YJµ'Lv ou TJ>E"lTE"a8aL TO ~ ds To ;y KaTa Tous lTaP'¼)XTlµEvollS' xpovous, otov E"vxoµaL E"Ux6µTJV'. lTapd 8€ TOLS' 'A8TJvalOLS' TpE"lTE"TaL QUTO ds TO ;y, 'riux6µT'lv' ydp Kal 'ribct,paLv6µTJV' MyoUO'LV ol 'ATnKol 8La Tou ;y Kal u. The lTap' T)µLv-part is confirmed by Mayser's findings (1906: 335): 'E"u im Stammanlaut wird hochst selten mehr augmentiert'; (344): 'Wie die mit E"U- anlautenden Stammverba ... , so haben auch die von Zusammensetzungen mit E"l, abgeleiteten Verba sehr selten eine besondere Augmentation'. 1

From the above discussion we may conclude that there are no compelling arguments to treat the augmenting of verbs in initial E"U- < E"l, and that of the other verbs in initial E"U- (or d.1-) differently. Also, since as yet no epigraphical evidence has turned up from, say, around 400 BC of imperfects or aorists in initial rn-, it seems safest to read TJU- everywhere.

1041 8VljO'KEL For the present of 'life-event'-verbs cf. on 2 TlKTfl. Here, the present may indicate that the Chorus consider Pentheus' death of great impor135

tance for their own situation. A different effect is achieved by the imperfect, as in Pl. Phd. 51 a: Kal he->.e-um; ('Describe the closing scenes', Gildersleeve§ 211).

irws

1043f. l,rd ... M,r611T£S lfiP11µ£11 ICTl.. This is an instance of the second type of messenger speeches (cf. on 677), viz. those that open with a temporal clause referring back to earlier information. Here, the reference is to 965ff. From the words >.i. ,r6vTE"S ietl3T'lµE"v we may infer that the messenger was present on the stage before. Exceptionally, this inference is made explicit here, cf. line 1046; normally we have just a 1st ps. plural verb form. By making the messenger one of the attendants of Pentheus Euripides has taken care that the messenger can begin his story at exactly this point, i.e. with a reference to his leaving the stage. Having been on the stage earlier, he knows that his addressee, the Chorus, was present at his and Pentheus' departure. For details see Rijksbaron (1976a: 294-8 and 302-7).

1045

da£Pdlloµn,

D.: "'We began to strike into the hill-country"'. R.: 'sens attenue, comme dans Hipp. 1198, lire-l 8' lpT'lµov

dae-j3a).).oµe-v'.

xwpov

K.: 'we were striking into the uplands of Cithaeron'.

It is not clear to me exactly how Dodds takes the imperfect. dae-j3ak Xoµe-v may either have inceptive meaning ('we marched into the hillcountry right away')278 , or durative meaning, in which case it expresses the idea that Pentheus and his companions were gradually penetrating into the uplands.279 In the durative reading the impf. signals that it took some time before the three men found a suitable spot (l. 1048). I rather prefer the latter interpretation, for the 'right away' value seems slightly out of place. Cf. also K-G 1, 143f. As for Hipp. 1198 (which is somewhat different, since dae-f3a.AAOµe-v occurs in a temporal clause), here a 278. Cf. s. Ant. 1203f.: ... Tuµl3v ... xwaavn:s aMLs ... daEf3a(voµEV, Tr. 46f.: TOL«lmiv lµol 80.Tov Al.,rwv foTELXE; see Moorhouse 191, who calls this use of the impf 'the imperfect of consecutive action'. Also e.g. Lys. 1.41: o\)s 8' ot6s TE ~ Ml3wv ll3aBLCov, 12.8 l~dBLCov. 279. Cf. the use of the impf. of y( yvoµaL in locative expressions like HdL 1. 70.2: l,rdTE dy6µEvos ls TO:S :Edp8Ls b KPTJTTlP ly(vno icaT.E~oLircSTa

D.: 'a thyrsus which had ceased (to be a thyrsus)'. This seems plausible enough but Dodds does not indicate how he arrives at this interpretation. Grammatically speaking I think we must supply KLaa6v from KLaa4) in the next line. Cf. K-G 2,575, 3: 'Wenn in Einem Satze neben dem Verbumfinitum auch das Verbum infinitum steht, so konstruiren die Griechen haufig das Substantiv, das der Rektion des Verbi infiniti unterworfen sein sollte, mit dem Verbumfinitum'. A parallel is X. Cyr. 7 .3.11: av Be: Ml3Daa Tota& lmK6aµn avTov Tots ,rap' lµoD. The full transitive construction occurs at Ion 1433ff.: aTl-

cf>avov l>.atas ... I .. ./ Bs- ... oihroT' lK>.ffon xM11v, / 8aMEL 8'. 1063

6pcii

As in the first messenger-speech, the historic present is used very sparingly and thus with great effect. The other instances are: 1112 trlTTTfl, 1115 ,rpocrnlTvn, 1117 >Jyn. In all cases it clearly 'hebt entscheidende Momente heraus', to use S-D's words once more (see on 1. 2). Observe 280. Barrett ad loc. creates an artificial distinction between lird c. impf. and !Sn c. impf. Note that the main clause to hrd . . . daEf3d>.>.oµE11 is not ci~ TLS" lO"Tl KT>.. in I. 1199, since the presence of the headland is not subject to temporal modification (note the omnitemporal present), but rather 1201-2: lv8Ev TLS" fixw ... f3p6µ011 µE8fjKE. If the impf. has the value discussed above these lines express that Hippolytos' setting foot onto the abandoned tract was simultaneous with the occurrence of the sound. ElaE~>.>.oµE11 has a distributive, and therefore durative, value in IT 260f.: hrd ... / 13ous- ... 1r611Tov daE~d>.>.oµE11, which is due to the verb having a plural object. Incidentally, here too the real main clause is not ~v ns- 8Lapp~ in I. 262 but rather l11Tali8a 8Laaous- El8l ns- 111:av(as- / ~ucj,opf3os- fiµlii11 in I. 264f. ('When we were plunging the oxen into the sea ... one of us saw .. .'). Cf. Platnauer ad loc. for more examples of this construction.

137

that from 1. 1137 (KE"1Tm)281 onwards the present indicatives are actual presents: TIJyxa.vn (1140), 4>£pn (1142), xwpe-1 (1144).

1066-7 ICUICA.00TO 8' ~CJTE" T6eov ii ICUpTOS' TpoX6ST6pV6µ.Evos- 11'EpLcf)opdv IA.Kn 8p6µov. 1066 ICUKAc>UTO Musurus: KUKAc>VTaL P 1067 TTE"pLOpciV P 'Ce passage, locus desperatus, est l'un des plus difficiles de la piece' (Roux).282 In a first analysis Dodds retains the transmitted text (but for Musurus' correction of the metrically impossible KUKAOUTaL and the, anonymous, correction oftre-pLopciv); he translates as follows: 'And it was bent as a bow (is bent) or (as) a rounded wheel drags its course when it is having its rim (?) marked out with peg and line (?)'. Although Dodds himself raises a number of objections to this interpretation, e.g. the peculiar meaning oftrE"pLopa, he finds little fault with the grammar, wrongly. In his view 'waTE" "as" with finite verb is ... an epicism, and as such appropriate to a narrative passage'. This remark suggests that he takes f AKEL as the predicate of both T6fov and Tpox6s. This is, of course, impossible, since the T~ov can hardly draw anything, and it is presumably for this reason that in his translation Dodds supplies '(is bent)' as a predicate to T~ov, and lets WOTE" have a fresh start after ,1\, with its own finite verb, which is to my knowledge unparalleled. The main flaw of this analysis, however, is that WOTE" + finite verb is definitely not an epicism. In Homer this construction is very rare indeed, since there is only one instance, Hom. II. 17.747; cf. Ruijgh (1971: §§ 810; 469-70). ''EAKE"L can, thus, hardly be retained; Reiske's fALK68poµov or Scaliger's fAKE"8p6µov are plausible alternatives. KuKAc>uTo should be supplied to both T~ov and Tpox6s. The real stumbling-block of these lines, however, is the interpretation of KUpTOS Tpoxos T6pVI¼) ypa6µe-vos. In Dodds' translation T6pvos is taken as a pair of compasses of sorts, but thereby, as Dodds himself observes, the comparison 'is a somewhat lame one ... : the sole point of similarity is the describing of an arc of a circle; we expect a more dynamic image to illustrate (as the bow does) the tension of the bent tree'. It is for this reason that many commentators prefer a different interpreta281. Gregoire wrongly translates 'gisait'. 282. In writing this note I have profited from a number of suggestions by Prof. J.H. Crouwel.

138

tion of T6pvos, taking it as a a lathe, or to be more precise, as a polelathe which is operated by the power of a flexible young tree or a bough. This interpretation, which originates in observations made by 19th-century British civil servants in India, 'peut etre tenu pour acquis', according to Roux. The operation of such pole-lathes is most clearly visualised in Tyrrell's commentary; his drawings are reproduced below (Figures 1 and 2), together with the explanatory notes of E.S. Robertson, one of the civil servants, originally published in Hermathena 3 (1879) 387.

'A stout polo of some elastic woo,! is lixetl into the wall, so as to

I'

FIO. 1,-P, Pole.

S, String.

D, Drum.

W, Wheel.

project at right angles, with its thinner entl free. To tl1ia eutl is attaehcll a stl'iug, which is brought tlowu anti fasteuetl to a pin in tho ,!rum of the lathe. 'l'he workman then attaches the block of timber which is to be turue,I into a wheel ; and ho d1·•RB this rountl in the ,lireetion of the first arrow, until the string is coiletl round the drum as many times "" it will go. This of course bends down tho pole, which is tl1e 111-occss described by Ku«>.ouro. • Fig. 2 represents t 1e lnthe with the polo bent down, and the string coiletl rouutl the tlrum of the lathe. When the workman releases hi• hold of the wheel, the recoil of the polo causes the wheel to revolve in the direction of tho second arrow, antl the

139

workmnn then applies the cl1isel. The process has to be repentecl •• often as tho string becomes completely uncoiled from the drum. 'The slow anti laboured dragging down of the tree, express,•d by HTV)"•, tye,, tr,,, wout.l accurately correspond with the slO\v prcparatiou for the turning process, in a hthc such as is hrre depicted. I'

Fro. 2.-11', role.

S, String.

D, Drum.

W, \Vht.-el.

C, Chisel.

'My suggestion is, that the simile in tho text is taken from the slow bendi11g of' tho JJolc in tho J>rocess of coili1111 the strinf!'.

In spite of Roux' statement, quoted above, at least five difficulties are involved in this approach to T6pvos: (a) Neither our texts nor the archaeological material provide any evidence for the existence of such a pole-lathe in Greece. (b) As appears from Figures 1 and 2, the. pole, being fixed into a wall, has a horizontal position, which does not suit the comparison of Ba. 1066-7. 283 (c) When bent, the pole does not form a semi-circle, see Fig. 2, which, again, does not suit the comparison. (d) In our lines there is no mention of a pole. (e) The meaning of ypacj>ELv is unclear. Because of the last two facts Roux is forced to alter the text drastically. Her text runs: icvpTOS' [3LOS' / T6pvov Tavvµe-vos trE"pL4>pai.s t>.i.ico8p6µ0LS, alleging that icvpTOS' [3L6s = 'mat flexible', but this is impossible: [3L6s only = 'bow', and icvpT6s '# 'flexible'. Willink (1966: 237ff.), too, starts from the assumption that a pole-lathe is meant. Inevitably, this leads, again, to a number of changes. Willink reads: IClJICAOVTO 8', WOTE" T6eov ly1C1JpTOV TPOXOS' I T6pV'¼) ypact,6µe-vos trE"pLopav lAKE"L 8p6µq1, which is rendered as: 'It (sc. the firtree) was bent into a hoop, just as a disk (or wheel) being lathe-turned pulls the bow into an arc by its rotatory momentum'. This must be rejected. (a) WO"TE" should not be coMected with lAKE"L, see above. (b) The meaning given to ypd.E"Lv is entirely ad hoc. (c) Although T6eov is 283. The drawing in Roux, p. 570, misleadingly represents the pole as having a vertical position.

140

translated as 'bow', it actually must be taken as refening to the pole: if it is not there would be no pole-lathe. This seems to me a strange way of interpreting. (d) "EyicvpTOS can hardly mean 'into an arc', for its meaning is 'convex, bulging' rather than 'round(ed), bent'; see below. All in all, then, the view that the T6pvos refers to a pole-lathe has very little for it. It is perfectly possible, nonetheless, that T6pvos does refer to a lathe, since this instrument was certainly known in antiquity. It was operated by two persons; one of these supplied the power by winding and unwinding a cord, see Fig. 3 (taken from Hodges (1973: 170) and based upon a relief from Egypt of about 300 BC).

Fig. 3

If our lines contain a reference to this instrument I think we can make sense of them by changing ypacf>6µEvos to y>.a4>6µEvos, as proposed by Palmer (1946). For yM4>w 'chisel', a by-form of the more common yM4>w, see Dodds and especially Chantraine s.v. y>..a4>up6s. fpa4>6µEvos may be due to someone who mistook T6pvos for a peg-and-line. As regards icvpT6s, this has its regular meaning: 'convex, bulging'. The

141

adjective refers to the slightly rising central part of the wheel, see Fig. 4 (from Piggott (1983: 24)).284

Fig. 4 Often, to be sure, KVpT6s is taken as 'round(ed), bent', but this is less probable, for. it is primarily used of things having mass and rising more or less pronouncedly, e.g. waves, fl. 4.426, and humps, fl. 2.218. 285 Cf. also Hel. 1558: KVpTwv ... vwTa, used of a bull putting up his back as a preliminary to an attack. 286 I interpret the lines as follows: 'The tree was bent as ... a convex wheel, when it is being chiselled in a lathe in a rotatory motion'. I take this to refer to the manufacture of a solid wheel from a square log of lumber. Just as the vertical side of the log by the chisel is slowly rounded until a semi-circle is formed (and eventually, when this process is repeated from the other side of the log, the full circle of the wheel), so the tree was slowly bent from its vertical position, until it reached the ground, forming a semi-circle. For this type of wheel manufacture we have a late, but to my mind, convincing parallel in Verg. G. 2.444: hinc radios trivere rotis, hinc tympana plaustris agricolae 284. This wheel consists of three parts. But I submit that a solid wheel will not have looked differenL 285. Among the examples in LSJ there is only one where the object involved does not have mass, viz. the Roman scutum at Plb.6.23.2. 286. Dodds, who calls Kup-r6s 'a purely ornamental epithet', translates by 'convex', but does not explain which part of the wheel the convexity relates to. The argument of Debidour (1985) is difficult to follow. Reading KupTOS' Tp6xos he translates 'la course decrite en rond par le compas'. But: (a) 'en rond' for KUpT6s contradicts Debidour's own analysis, according to which KUpT6s is used of semi-circular things; (b) KUpT6s never qualifies action or other abstract nouns. Debidour's analysis is further weakened by his reading lXKn.

142

(hinc refers to different kinds of wood, mentioned in the preceding lines). Thomas observes ad loc.: 'have often shaped with the lathe'; tympana are the solid wheels of farm carts. That the terere was performed with the help of a lathe appears from e.g. Luer. 4.361 (if Munro's terantur is adopted) and Plin. Nat. 36.193 alilui torno teritur. I conclude with a few remarks on KUKAOUTO WO"TE" T6eov (sc. KuKAOUTaL ). With the sole exception of Schone all commentators are silent on these words. They must refer to the stringing of a bow (so too Schone). One of the ways to string a bow was to put it upright, and by pressing the right knee against it to bend it in such a way that the string could be fastened around the upper end; see Pauly-Wissowa VI A col. 1852. Depending on the type, the bow, once strung, had either the form ofa C orofaD. To sum up: in both similes reference is made a vertical object is referred to which gradually gets a (semi-)circular form. This fits in with the bending of the fir-tree, since the latter is also bent from a vertical position. Both the stringing of the bow and the making of the wheel would seem to be appropriate items of comparison for the messenger, since they must have belonged to the ever recurring activities in a rural community.

1083 la~pLCE This form, the reading of P, is adopted by e.g. Roux and Kopff. Murray, however, reads l.otJE ••.

8.98.1:

TaV'TO. TE dµa

:E:e"PolS

ls TOV BaTTov

fTTOl£E" Ka\. fTTE"µTTE" .•. 143

Tafrrci Tf iiµa i'ry6pe-UE" ical To lp-yov trpooii'yf·

9.92.1:

In indirect speech: 1.105.4:

dµa >J-yoool. Tf ... vooe-fw, ical bpav ...

9.16.3:

Tatrrci Tf dµa TOV Ile-po11v >J"yflv ical µn1.e-va.t ...

Impf. followed by historic present: 8.5.2:

Tafrrci Tf ciµa -fiy6pe-UE" ical tre-µtrfl ...

Aorist followed by aorist in indirect speech: 1.94.2:

iiµa Sf- TaUTas Tf E~ftJPfeilvm ... ical ... a.trOLiclom.

Cf. also K-G 2, 231. The examples discussed by them confirm the picture presented by Herodotus: there are no combinations of impf. and aorist.

1104

dvEawdpaaaov

D.: 'conative'. R.: 'imparfait d'effort, «elles tentaient d 'extirper les racines»'. It is not clear to me why, of the series of imperfects starting with lppmTov, 1097, this imperfect should all of a sudden be taken conatively. It surely has the same descriptive value as the other imperfects, i.e. the women were (for some time) busy tearing up the roots. That they did so to no avail, i.e. without bringing Pentheus down, is a different matter, which should not induce us to believe that they did not tear up any roots at all. The other actions, of course, were not successful either. Note the presence of (ill') ouic ,'\VtJTov, after'(rn-av at 1100 (it probably relates to lpptlTTOV (1097) and-fiicoVTlCfTO (1098) as well), and of (hfl ••• ) OUK le,'JVVTov after a.vE.E1:v is ... "to embrace"'), probably rightly. On the other hand, his suggestion: 'If ne-pLf3a>.E1:v is sound, TE"K'Vql seems to be needed' (TEK'V(j) was suggested by Kirchhoff; this is the text adopted by Kirk) is not convincing either, for in the meaning 'to embrace' ne-pLj3a>.E1:v would not seem to be found with the singular of xdp, but only with the dual (e.g. Hom. Od. 11.211) or the plural (e.g. Pho. 1458-59; cf. also Pho. 165, with w>Jvas-), not unnaturally, for one does not embrace with one arm. Wecklein' s correc145

tion of P's text: lv atµan crTciCoooav xl pa [1TEpL ]f3a>.€1v TE"Kvou is the most simple way out. It is preferred by Dodds, Verdenius 1 and Roux; see Dodds for arguments. To Dodds' parallel, Med. 1283, one may add Or. 1466: lµ[3a>.oucra 'ITTIXVV OTE"pvoLs. 287 ~Ta.Coucrav should probably be taken proleptically (thus Dodds), i.e. 'so as to make (the hand) dripping?, sc. lv atµan (instr.); for instrumental lv atµan cf. El. 1172-3: vEo4>6VOLS lv atµacrLv I 1rE4>vpµlvoL (alongside of Ale. 496: cpciTVas ... a'lµacrLv 1r£4>upµlvas ).

1170

l~.LKa VEchoµov

D.: "'tendril" of ivy (or vine)'. R.: 'vE6Toµov suggere qu' Agave pense

aune branche de Herre'.

Similarly all other commentators and translators. Dodds continues: 'the head with its curling hair and beard (1185ff.) takes the place of the ritual bunch of ivy leaves on the tip of the thyrsus, and Agaue in her delirium identifies the two'. I find this difficult to accept, in view of the fact that Agaue qualifies V..LKa as µaKdpLov 811pav, (or 811paµ') 'a blessed prey' ,288 which is absurd if she believes she is carrying a bunch of leaves. Moreover, the transition from 'a freshly cut tendril' to 'this young whelp', 289 referring to the same object, is quite unexpected. Roux has noted the difficulty, but believes, following WinningtonIngram (1948: 135, n. 1), that 'Euripide rend ainsi sensible le desarroi mental de la reine'. However, an analysis in which Agaue first takes Pentheus' head for a bunch of leaves, then for the head of a young lion, next for that of a bull (see below on 1185), and, finally, once again for a lion's head (1214-5), introduces perhaps a bit too much disorder into this scene. After all, Agaue was convinced all along that she was holding a lion's head, cf. 11. 1277-78: (Kadmos) Tlvos 1rp6crw1rov 8-f\T' lv ayKciM.LS lxns; (Agaue) AE"OVTOS, ws- y' lwµEVQL. I would therefore like to propose a different interpretation, in which

i~e

287. Kopff, who adopts j3a>.Ei:v as well, reads Tlicvci,. Since ~d.Uw is not construed with a dative of the object hit this is impossible. (Unless one talces Tlicvljl as dependent on lv ... 13«>.etv (and the single dat. al'.µan as instrumental with .i.ica) * 'hunting'. Agaue's 'blessed prey' contrasts with the messenger's 'ill-fated prey' (&fi1>4 8ua1r6Tµljl) at 1144. 289. Reading >.loVTOS' ••• tvw, which is due to Wecklein, and is adopted by Roux.

146

= 'curl', or rather 'curls', the singular having collective meaning (cf. KG 1, 13f.), like e.g. 8pte (LSJ II (a)), and l36aTpuxos (Hipp. 202). "E~ua, being the most conspicuous feature of the head, per synecdochen stands for 'head with curly hair'. Cf. for a comparable case of synecdoche the use of c5µµa for 'face' and even 'human form', LSJ s.v. IV. The meaning 'curl', to be sure, is not attested in Classical Greek, the first occurrences mentioned by LSJ being from the Anthol. Graeca, but can be supported by the use of the related verb t >..taaoµaL in E. Fr. 382.7: /3ooTpux6s ns l],s El>..L"yµlvos. 290 If the above analysis is correct, lALKa VE6Toµov may be a grimly ironical reminiscence of the words spoken to Dionysos by Pentheus at 1. 493: ,rpwTov µtv a.l3pov f36aTpuxov TEµw al8Ev. 291 1185

vios- 6 µ6axos-

D.: 'With the idea of a sacramental feast, Agaue's dream has shifted a little: the head is a bull's head now, though bearded like a young man's'. Similarly Roux and Lacroix. 292 Kirk, however, while translating 'bull', in his note adds ' ... it is not certain that AG AUE strays at this point from her otherwise consistent delusion, during this whole conversation, that what she is holding is the head of a young lion'. Dodds, following Wilamowitz, categorically denies that a young lion could be called a µ6axos, 'particularly without MovTOS' '. However, if the lacuna at 1. 1174 contains a reference to a young lion, as it does when Wecklein's suggestions are adopted, the context would seem to provide sufficient clues as to the nature of the µ6axoS. Observe also that the proper word for a lion's whelp, viz. aKuµvos, may be used for a variety of other young animals as well, see LSJ. 290. "EXL~ may have been used by Euripides because the nonnal words for 'curls', viz. ~6aTpuxos and irMKaµos, are only used of human hair, and were thus unsuitable to express Agaue's delusion. 291. LSI s.v. l>.1.~ (A), i.e. adjectival lA1.t take lAL~ in our line as= ~oOs, comparing the Homeric expressions lALK«S' ~oOs and the like, perhaps on account of µ6axos at l. 1185. S.v. µ6axos, however, they class I. 1185 under 'any young animal'. (This reference has been deleted in the Supplement). 292. On this interpretation we must assume that afterwards her dream shifts back again, cf. 1215.

147

1203-4 TI1v8' clypav, Kd8µou 8uyaTipES" &r]pbs- 'flv 'ii'YPE-6aaµEv D. (on Eh]p6s): 'gen. of definition dependent on dypav, but attracted into the relative clause'. This is of course not correct, since the relative clause starts with -fiv. Nor is the parallel adduced by Dodds, Th. 2.45.1, relevant, for in that sentence (trmal 8' al, ooOL Twv& miP£OTE ... opw µlyav Tov dywva) the genitive Twv&, which in a more regular syntax would modify traurl., has been drawn, indeed, into the relative clause, cf. Classen-Steup ad loc. and on 2.67.4. It is actually rather difficult to determine the syntactic status ofEh]p6s. On the one hand it may depend on dypav, and be separated from its head noun by the prolepsis of Ka.8µov 8vyaTlpES', the subject of the relative clause. As in other cases (cf. 173, 272), the prolepsis gives prominence to that subject (here, we are probably dealing with emphatic focus).293 On the other hand, Eh]p6s too may have front position vis-a-vis the relative clause. 294 There are many parallels, notably in lyrics, for the left-displacement of an oblique constituent from a relative clause salvo casu, i.e. without the case form being dependent on a constituent of the main clause. 295 Cf. icaico'iaw in S. Ai. 473-4: alaxpov yap dv8pa TOU µaicpou xp,jCnv l3lov, I ICQICOLO'LV OOTLS' µ118E"V iea>.>.aaa€TQL, trapa~ICTPOLS in Pho. 1548-9: ,rapa~ICTPOLS / a (sc. d.Xoxos) tr68a aov ... 8€patrd,µaaLV aU-v lµ6x8n, also TOlJTWV in Th. 1.97.2: TOlJTWV 8€ lScrtrEp ical -fitl,aTo lv Tij 'ATntcij evyypa4>fj 'E>.>.tiVLICOS, 13paxlws ... €1TEµvria°'1 (TOlJTWV depends on -fi!liaTo ). A comparable prolepsis, involving a dependent question, occurs in Pl. Grg. 460 a: dtroica>..6!1ias Tfis P11TOpL1ci\s fltrE Tls tro8' Tl 86vaµls l°OTLV, where Tfis PllTOpllci\s has been moved to the left in the case form it would have in the dependent question. If 81lp6s has, 293. Incidentally, the prolepsis of the subject of a relative clause in the nominative is rather rare; it is not discussed in the standard grammars. In Breitenbach's survey (1934: 263) there are two examples from lyrics, Or. 1376: ... 1r6VTov, 'OKEavcs 8v I ... ICVIC~ot, also IA 245. Prolepsis is, to be sure, a quite common phenomenon, but the examples discussed in the grammars (cf. e.g. K-G 2, 577ff.) all involve object-clauses with !rrL (as at Ba. 173, quo vide), ws-, and µfi, or dependent questions. 294. Cf. Wecklein: '&r,pos- hllngt von fiv (dypav) ab'. 295. See further Breitenbach (1934: 263). These prolepseis should not be confounded with constructions where the case form of the displaced constituent is determined by a main clause constituent, as in Th. 1.61.1: ~>& Bl Kat Tots 'A&r,valoLs Eu8us ti dyyE>Ja Tliiv 1r6AEwv lSTL dlj,ECJTdaL. Here Tliiv 1r6A£wv, which is the subject of the lSTL-clause, is governed by ti d'Y'YE>Ja.

148

indeed, been moved to the left, we must conclude that both the subject and an oblique constituent appear in front of the relative clause proper. There are a few parallels in lyrics for the prolepsis of two constituents from a relative clause in the case form they would have in that clause, e.g. Pho. 1519: al>u.vov alayµacnv a / To1a& 1TpoK>.alw, cf. also/A 178-80: .. .'E>.lvav, du' / EupwTa ... / IlapLS' ... av l>.a.13£, so the exceptional position of both Ka6µou 81.ryaTlpe-s and &r,p6s- can probably be defended. Cf. also the following double prolepseis, with dependent questions: Pl. La. 185 d: lhav tuuw xaALvov O'K01TT1Tal TLS' d upoaoLaTfov (observe, however, that xaALv6v might be taken as the object ofaK6TIT)TaL), and Ba. 272 (quo vide): olrrcs 6' b 6al1,1wv ... I OUK 6uvalµT1V 1,1E''.YE"8os e-enue-1v BOOS' / ... foTaL. 296 More common are double prolepseis like the following, taken from Chanet (1988: 70, note 5), where both constituents appear in the accusative: Hel. 85: ou Tdpa a' 'E>.lvnv E"l aTuye-1s 8auµaaTfov (ae- is the subject, and 'E>.lVT1v the object of the El-clause), Ar. Ra. 907-8: Kat µ,;v l1,1auTov µlv ye-, :rnv u6naLv ot6s dµL, / lv TOLO'LV VO'TQTOLS' a indicates that with mip' the sentence starts anew. Thereby both TotaL and K(pKTJ have a certain prominence. 297. Chanel does not treat such cases as examples of double prolepsis but rather as a juxtaposition of a proleptic constituent sensu stricto and an 'anteposed' constituent, the latter simply appearing in the case form it would have in the subordinate clause. Thus, at e.g. Ar. Ra. 907-8 'TTIV 1r6r]0Lv would be anteposed, while lµaUT6v is really proleptic, since the accusative does not depend on ♦pd.ow nor on any other verb. Similarly, she would consider Kd8µou 9uyaTlp£S' an anteposed, not a proleptic, constituent. See also on 173, and Moorhouse (1988: 213-7). Since these distinctions are immaterial to my argument I have just used the conventional term 'prolepsis'. 298. Observe also that the syntactic prominence of Kd8µou 9uyaTlp£S' and &r]p6s is mirrored by their metrical features. Since Kd8µou 9uyaTlp£S' fills the space before the main caesura, and &r]p6s that between the penthemimeral and the hephthemimeral caesura, both constituents are set off. Cf. Korzeniewski (1968: 47).

149

order, e.g.: n'Jv8' dypa.v, fiv Ka8µou 8uyaTEpE"S' ei,pos 11ype-ooaµEv orTI')v8' dypa.v ei,pos, fiv Ka8µou 8uyaTEPES' wpe-ooaµEv.

1218

µoxecav

D.: 'Since the CT1T"1 µaTa are now past, we expect an aor. ptcp. If µox8wv is right, it must imply µox0fiaas CTITllµaaL Kal µox8wv lTL. But we should perhaps read µ6x8wv, gen. pl. noun (Wecklein)'. Roux retains µox8wv, and takes it as suggested by Dodds in the second sentence of his note. But it seems doubtful whether the dative with µox8wv, which probably is a causal-instrumental dative, can refer to a cause (or instrument) which is not present simultaneously with the state of affairs denoted by µox8wv. Cf. Pho. 1548-9: 1rapal3a.KTpOLS' ... 1r68a ... 8Epa1rd,µaaLv ... lµ6x8n, Hom. II. 10.105-6: d.>.M. µLv otw / K"18EaL µox0fianv, oc 350-1: 1TOMOLOL 8' 6µ!3poLS' tJMou TE KauµaaL / µox8ovaa. Verdenius 1, who reads µox8wv as well, rightly points out that the present participle can refer to a past (or rather: anterior) action. Such participles are mostly found in contexts which provide some clue that the participle is, indeed, anterior to the main clause, e.g. an adverb like 1rp6TEpov (with the participle) or d Ta (in the main clause). In our case it can be argued that such a clue is provided by CT1n'JµaaL, since this unequivocally refers to a past action, cf. also Dodds. But this still does not explain why the present participle has been preferred here to the aorist participle. The present participle is probably used to stress the idea, which is also conveyed by the plural CT)TI')µaaL and by the attribute µuplOLS', that the toil of Kadmos lasted for a considerable time. This interpretation ofµox8wv is based upon Stahl's discussion of such participles, pp. 209ff. Stahl calls them 'priorisch dauernd' (as opposed to the aor. participle, which is rather 'priorisch an sich').

s.

1232

>.e-vaaw yap airrtjs- c'SilsLV OUK Eu8aCµova ainfis- P : aim'iv Scaliger

D.: 'ai,TI'Jv foraUTlls (P) is certain'. I do not think it is. Dodds argues that 'with airrfis the sense could only be "I see her face" ... or "I see a vision of her"'. Actually, however, the sense can also be 'I see that she presents a quite unhappy appearance' (lit. 'I see of her an appearance which is quite unhappy'; oiJK Ev8alµova is probably a litotes), aVTTlS being a possessive-pertinentive genitive: 150

'she presents a sad appearance'. 299 This seems unobjectionable. For 6.,,LS' = 'appearance, aspect' cf. LSJ s.v. (I), and cases like Hom. II. 6.468: traTpoS' .8€v Elmsley

The second person can hardly be retained, cf. Dodds. Elmsley's conjecture is adopted by e.g. Murray and Kopff, while Roux, Lacroix and others prefer ~).8' ES'. The latter is probably right, since ~>.aov with a bare accusative is virtually confined to geographical entities, e.g x86va (Ale. 485, 560), tr6>.i.v (He/. 105), "Apyos (HF 1285), traTpl8a (He/. 1092), et saepius. An exception is Hipp. 1102-3: lhav JnJ, but this 299. The genitive as in TO Et86s nvos, "ti µopc/,(J TLVµ' dvi~>.e-4>' avl~>.E,re-v P : corr. Elmsley304

301. vooTov at IT 1112 is a cognate object. 302. Who, in tum, refers to Monro, who made the same observation. Bers, however, (1984: 75ff.) inclines to accept the construction of a bare accusative of words designating persons or body parts with verbs of motion. Among his examples, not counting Ba. 1286, are three instances where a form of 0.8- is involved: S. Ph. 141-2 (lyr.): al (aol T (i.e. Triclinius)) 8' ... T68' tA:nAu8Ev I irav Kpchos, E. El. 917: lA8wv 4>pvyas-, 1281: ~Mev tpvyas-. Since the first of these examples is from lyrics, and the latter two involve nouns referring to geographical entities rather than to persons (cf. also Denniston on El. 917), the case for tµas ~A&Ev xlpas- at Ba. 1286 is rather weak. With other verbs of motion, though, especially fjKw and the forms from the stem lK-, things may be different. 303. I should add that it is not altogether certain that dp8pa here = 'sockets', since KOTUAT18wv normally means 'socket' (esp. that of the hip-joint) as well. Cf. MacDowell ad locum: •... Ar. is evidently not using the precise technical terms of anatomy, and we should translate more vaguely "in my limbs each joint is moving supply"'. 304. Of recent editors only Lacroix retains dvl~>.E,rev, but this form can hardly be correct. See Dodds, Roux and Tyrrell.

152

D.: '"by whom (instrum.) the house had recovered sight" ... we must abandon the traditional rendering "to whom the house looked up"; dval3>J1rw never elsewhere means "look up to" in the sense "respect"'. R.: 'ii> ... datif de direction'; (transl.): 'lui vers qui notre maison elevait tous ses regards'. K.: 'Through him the house saw the light again'. Dodds is probably right in condemning the traditional view of this clause. The parallels adduced by Roux to support it rather support the meaning 'to recover sight' than 'lever les yeux', cf. Ion 1466-7: 8 TE" ... 86µ.os oucln vvKTa Slpice-Tat, I ae->J.ou 6' dvaf3>.fon >.aµmiaLv, where dval3>JTTn >.aµmiaw, with Dodds, is best taken as 'recover sight in the rays of dawn', and A. Ch. 807: d, dvL&tv 66µov dv6p6s, which is, again, best taken as 'recover sight' (a gloss in M has dva13>.llf,aL); cf. Groeneboom ad loc., who refers to 11. 51f.: d~ALOL l3poTooTUye-ts / 6v64>L KQA'UTTTOUOL 66µous. Roux may be right, however, in taking ii,, i.e. Pentheus, metaphorically as 'the light', see below. On the other hand, Dodds' (and Kirk's) view of this clause is not unproblematic either. '¼l, to be sure, can very well be an instrumental dative, cf. K-G 1, 436. Two clear examples of the use of this dative of persons are Held. 390-2: dv6pa ... xpe-wv, I ... I ovK dyyl>.oLaL Tous lvaVT(ous bpo.v, and S. Ph. 494-5: Tots lyµlvms foTe-llov / auT6v ('I used to send for him by means of those who came', Moorhouse 89). See further Luraghi (1989: 300f.). In this use the persons involved are the intermediary agent through whom another agent carries out a certain action. 305 I\>, therefore,= 'through whom' (i.e. via whom) rather than 'by whom', cf. Kirk. Pentheus should then be conceived of as the eye of the house. This is, in fact, Dodds' view, as appears from his remark 'The heir of a family is often thought of as its eye'. But then the result can only be that the house recovered sight via Pentheus. This strikes me as rather bizarre, for the implication is that the house itself did not see. 306 Also, both Dodds' 'had recovered sight', and Kirk's 'saw the light again' ignore the fact that avll3>.E' is an imperfect.

6os

305. Luraghi wrongly writes (pp. 301-2) that the plain dative is not found as a marker of Intennediary constituents. 306. If possible at all, we would rather expect that they saw by means of Pentheus, i.e. that they depended on him, cf Held. 392, quoted above. But this meaning cannot be got from dvt~>.rn£.

153

I suggest a different analysis, in which, first, the referent of T68'

lpvos ~. i.e. Pentheus, stands metaphorically for 'the light', and, second,~ is not an instrumental but rather a causal dative: '(Pentheus) by whom (i.e. by the light spread by him) the house was recovering sight'. This statement of Kadmos refers to the fact that the royal house had been 'in the dark', because he had no male children, cf. his words at 1. 1305. His grandson taking over power promised well for the future, but with Pentheus' death this promise can no longer come to fruition. The imperfect civll3>.£1re- indicates that the house had not yet fully recovered sight, naturally enough, since Pentheus was only a ve-avl as. For the metaphorical use of persons as= 'light' see LSJ s.v. cfxios, 4>ws II, and cf. e.g El. 449: Tpl(/>e-v 'E>.M8t cf>ws / 8lTL8os ... y6vov, IA 1063: 8rnaa>..(q. µlya (/>ws (: Achilles), HF 531: ~ (/>cios µo>.wv traTpL This view is further supported by Ion 1466-7, quoted above, 307 where it is said of the house of Erechtheus, when the lost heir is found, that it is no longer in the dark but is recovering sight cie->..lou ... >..aµtrdatv. Owing to Ion's arrival 'the sun is shining again' upon the house. 308 Observe also that the contexts of these two lines are very similar, cf. Kreousa's words at 1463: cfoat8e-s OtJKE"T' taµh ou8' dTE"KVOL. Like Kadmos', Erechtheus' house is 'Y'l'YE"VETas, 1466.

1312

8CKT1V yap df(av lMµl3ave-v lMµl3ave-v P : l-Mµl3ave-s Hermann, Murray, alii

D.: 'It is hardly safe to reject P's l>,.aµl3ave-v. Although 8lKTtv >..al3e-1v commonly means 'to exact justice', it is used of suffering justice not only by Hdt. (1.115.3) but in a fourth-century Attic inscription ... '. R.: '8lKTtV >..aµl3civnv signifie habituellement «tirer vengeance», «punir»; mais !'expression a aussi parfois le sens de 8lKTtv 8o0vat, «recevoir sa punition» '. There are two problems here, the meaning of 8lKTtv >..aµj3avnv, and the precise semantic value of the imperfect. As appears from the above quotations, Dodds and Roux think that 8(KTtv >..aµl3civnv can have the meaning of 6(KTtv 6o0vat. But this is impossible. To start with, it is highly improbable that the common terminus technicus 8lKTtv MIµl3civnv 307. Dodds wrongly adduces this line in support of his view of~ as = 'eye'. 308. Apparently, Ion here metaphorically = 'the sun'. According to LSJ s. v. 'fl>.i.os (I) 5 the use of 'sun' of persons occurs only in Hellenistic and later papyri.

154

'to exact justice' might simply receive the exactly opposite meaning of its counterpart 8l1CT1v 8oOvaL. 309 And, in fact, the parallels adduced by Dodds and Roux to prove their point are not really parallel. Two of them do not have 8l1CT1V Mµj3civew, but rather n)v 6l1CT1v Mµj3civnv, viz. Hdt. 1.115.3: olm>s 8£ ... Myov e-txe- ou8e-va, es o lXa~e- TTIV 8llCTlv,310 and Dittenberger, Syll.3 1.167.37:311 MavlTa ... TTIV 8l1CT1V M/36VTOS. The presence of the definite article makes all the difference: TTIV 8l1CT1v means 'the punishment someone deserves, the punishment appropriate for the crime'. Similarly in TTIV ci~lTIV Mµlj,e-a8aL (Hdt. 7.39.2), TTIV 8(1CT1v lxnv (Supp. 434, X. An. 2.5.41). Roux also refers to Dem. Phil. 3.4 (= 9.4), but at that place (TT1V) 8lK11v does not occur. We do find &Kllv XT}$E"TaL in § 2 of that speech, where the sense is clearly 'to exact justice'. It might be objected that, in drama, the presence or absence of the article is an unreliable criterion to go by. Perhaps so, although the phenomena discussed at lines 893-96 suggest otherwise (cf. also S-D 23 fn. 1, and the literature quoted there). On the other hand, if there are no examples from prose without the article, I believe the case for (anarthrous) 6lK11V Mµl3avnv 'be punished' is very weak indeed. There are thus three different technical expressions: (a) &Kllv Mµj3civnvlM~E"LV 'to exact justice', subject: the prosecuting party, which forms a pair with (b) 8(K11V 8L86vaL/6oOvm '(to give justice =) to be punished', subject: the wrong-doer, and (c) -rliv 6(1CT1V Mµ~avnv/M~E"LV 'to get one's deserts', subject: the wrong-doer. The semantic differences between expressions (a) and (c) reflect the two main senses of Mµj3cive-w: take and receive. In view of the above facts I think Hermann's emendation should be adopted. The second point to be dealt with is the interpretation of i-Mµ~ave-s. There are, I think, two ways of interpreting this imperfect. First, it may be taken iteratively, referring to Pentheus' behaviour in other cases of outrage: 'Nobody was willing to insult me, for it was your habit to exact the appropriate punishment'. Thus apparently Dodds, who notes: 309. Cf. 8Cicas ... l-8(8oodv TE teal l>.ciµf3avov at Hdt. 5.83.1. 'It is ... extremely improbable that the poet interchanged the two senses in the present passage' (i.e. Ba. 1312) (Sandys). 310. LSJ's information, s.v. 8[1CT), on 8[1CT)v ).aµf3civELv is singularly inaccurate. At IV 3 they note '8[icas ).aµf3civELv sts. = 8. 8o0vaL •, referring to Hdt. 1. 115. There, however, cf. the main text, we find niv 8[1CT)v, not B[icas. Again, at Heracl. 852 the verb is d1roTE[aaa8aL, not ).aµf3civELv. 311. Not 165.37 (Dodds), 167.35 (Roux), 1,65,37 (Lacroix).

155

'lMµj3ave-s' - which is read by Murray - 'sc. d ns u(3p((OL '. Obseive that this must refer, indeed, to cases of outrage which did not concern Kadmos, since we have been told explicitly in the preceding clause that nobody dared insult the old man. This point seems to have been missed by e.g. Roux, who, reading lMµj3ave-v, translates: 'Ma vieillesse, nul ne songeait a la persecuter en te voyant: sinon, il etait chatie selon son du!' This is logically impossible: if nobody insulted Kadmos there was nobody to punish either.312 Alternatively, and more plausibly, lMµj3ave-s may be taken as a variant of the conative imperfect: 'you were going/you were likely/you were already preparing to exact the proper punishment'. The conative use of the imperfect is usually illustrated by verbs like Trd8w or 6(6wµL, i.e. verbs which can only be successfully realized if the second participant involved in the state of affairs co-operates. The impf. of such verbs (as indeed the other forms of the present stem) may denote that the end-point was not reached (the persuasion or the giving did not come about), and thus that the state of affairs concerned did not get beyond the stage of an attempt, whence the term conative. Typical examples are X. An. 7.3.7: fon8ov a.TroTplTrrn8a.L · ol 6' oux UTl'TlKovov and Hdt. 9.109.3: Tr6>.i.s TE" l&6ov Kat xpooov Kat 6-pos. The term conative, however, is too restricted; in many cases the imperfect rather denotes that the state of affairs concerned did not get beyond the stage of being intended or planned; in others, again, the impf. denotes that the state of affairs threatened to occur. Cf. Gildersleeve§§ 213: 'The imperfect is used of attempted and interrupted, of intended and expected actions'; Goodwin§ 38: 'The imperfect sometimes denotes likelihood, intention, or danger ... ', Stahl 406: ' ... das Vorhaben oder das Bevorstehen (ist) wirklich, die Ausfiihrung nicht wirklich', also K-G 1, 140, 7. Some other examples from Euripides are: Med. 592: l~ll3aLVE" 'was likely to end' (Page), HF 537: Kat Tdµ' l8VQ8E"ta' lKaLvqLTJV ~l4>e-L · / ID' l~lK>.EIJ,e-v ... I "ApTe-µLs, 314 Tro. 312. The difficulty was noted by Heath, as we may infer from his conjecture

l-Mµl3av' dv. This conjecture should be rejected for the impf. without clv can very

well be defended, see below. 313. HF 465 dµ.4,ll3a>.>.t-, mentioned by Gildersleeve, is better taken as an iterative impf. 'he used to throw the lion's skin', cf. Bond on 462-79 and 462-5. 314. Diggle puts a full stop after {((j,£L, thereby creating a rather heavy, perhaps too heavy, break between lines 27 and 28.

156

397-9: d 8' fiaav &KOL (sc. the Greeks), XP1lOTOS «:>v lMv8aV£v (sc. Hector). I IT6.p1.s 8' l'Y'lµe- TTJV aL6s· 'YTlµas 8£ µ11, / crLywµe-vov TO Kfi8os e-txe-v lv 86µ0Ls. 315 Nor are such imperfects confined to poetry. Cf. e.g. Pl. Smp. 190 c: al TLµal ... i'ict>avt,e-To 'were going to disappear' (sc. if they destroyed mankind) (Dover); Th. 8.86.4: wpµrr µtvwv ya.p TWV lv :lliµJ.d.- and with d µ11-clauses after impf. + dv, as in Hom. fl. 5.22-3: oW ya.p ovBE KEV airrbs irrrlKct,uye- Kfipa µO,mvav, I a.>J.' ''HaLOTOS' lpUTO, and 7.273-5: Kat vu KE" 8~ eLct,frcrcr' avTOaxe-Bov oirra,ovTo, I El µ~ ~PUKES' ... / fi>.aov. Unlike the counterfactual, lKaw6µ11v leaves open the possibility that the state of affairs just might be realised. 317 Obviously this is rhetorically far more powerful than a 315. Diggle needlessly adopts Burges' and Schafer's conjectures tMv9av' 4v and etx• 4v. The same impf. occurs at S. El. 913-4: d.ll' oW µlv Sri µTJTpos- olie' b vo0s- cj,L>.d / TOLa0Ta ,rpcfoouv olin 8pc'iia' lMv9avev. Campbell, correctly: 'was likely to escape notice'. Heath conjectured tMv8av' 4v, again needlessly, although his conjecture has been adopted by many editors, including LloydJones and Wilson in the new Oxford text. Cf. also footnote 312. 316. Cf. also on 11. 173-89. 317. The 'likelihood' use of the imperfect shows some similarities with certain uses of lµellov + infin., where a counterfactual-like meaning may be involved as well. Cf. Basset (1979: 173ff. 'eventuel passe', 177ff. 'valeur d'irreel avec lµellov'). Sandys already explained tMµj3aves- as= lµelles- Aaµf3dvELv. The Latin impf. may be used in the same way, cf. Kuhner-Stegmann 2, 404. Cf. also Quirk et al.

157

plain counterfactual construction. A similar effect is present, I think, at Ba. 1312, where the impf. conveys the idea that Pentheus was already

preparing countermeasures in case somebody would dare to insult his grandfather after all.

1343

d8aLJJ.ovotT' clv P: T1i.&uµoV£tT' Musgrave, d&uµove"tT' Hermann

For the variation T1V- I fV- see on 1024. Against the optative Dodds argues that in the combination of a past indicative in the protasis with an optative c. dv in the apodosis (which in itself is possible) '(a) ... the protasis almost always ... expresses an open condition ... ; whereas the condition here is clearly unfulfilled. (b) The sense required by the context is not, I think, "you could, or might, be happy", but "you would be happy" ... The optative would fatally weaken the rhetorical contrast'. It is especially the second part of Dodds' remark which is relevant here, for there are a few cases of the combination d + past tense with opt. + dv in the main clause, where the condition is clearly not fulfilled, cf. Supp. 764: cf>alTIS civ d trapf\a8' lrr' T}yatra VE"Kpous, Lys. 9.11: fl 6' oliv l(T1µlwaav µtv ot& trpOO'TIK6vTws, lKvpwaav 6' lv uµtv 'riiv E"'TTLj3oAT)V, TWV TaµullV acplVTWV flK6TWS av TOU f'YKAT)µaTos ci'TTTl>J.a.Yµlvos dllv (Schottus: -µlvot fkv L), where it appears from the context that the condition is unfulfilled, cf., in the preceding sentence, oln-f ... Ta trpaxelvTa ~ KVpt.a KaTE"O'TT'laav.318 (1985: 214ff.) on a similar use of the present and past progressive in English, e.g. I'm taking the children to the zoo (on Saturday), I was meeting him in Bordeaux the next day ('arrangement predetermined in the past'). The basic meaning of the progres-

sive is: future arising from present (past) arrangement, plan, or programme. They observe (p. 219) that the past progressive, just like was/were going to+ infinitive, 'favours an interpretation of non fulfilment'. 318. Note, however, (a) that the El-clause virtually has concessive meaning: 'even if they had been correct in imposing a fine ... ; for 8' ow cf. Denniston 465: '(8' o~v) is particularly used when a speaker hypothetically grants a supposition which he denies, doubts, or reprobates'; (b) that ElK6Tws llv ToO lyK>.fiµaTOs 0.111JAMI'Yµtvas d11v may also be taken as the main clause to the causal genitive absolute, the result being: 'Even if they had ... , now that the Treasury clerks have remitted the fine, I should stand fairly cleared of the impeachment'. Supp. 763 and Lys. 9.11 are from Stahl 408. Stahl has many other examples, the majority of which, however, should be analysed differently, viz. as open conditions, e.g. Pl. Prot. 343 c-d: E00vs ydp Tb 1rpiiiTov TOO 4aµaToS µavLKbv llv 4'avd11, El f3ouMµEVoS MyELV lSn dv8pa dyallov yEvla8aL xa~1r6v, hELTa lvt[3a~ Tb µEv ( Simonides has added µtv; the focus is on the relation between the El-clause and the intention: 'The opening of the ode would at once be quite crazy, if, while intending to say that it is hard

158

As for the optative + dv, there is no need to take it as such a strict expression of possibility as Dodds seems to believe. Indeed, in English a translation by 'would', referring to the present, is often perfectly appropriate, as in the following combinations of d + indicative and optative + dv: 319 S. El. 797-8: tro~v civ iiKoLs, 320 @ flv', dfws TVXE"LV, I d TT)V8' ltravaas Tiis tro;\vy>.waaov 13oiis,321 And. 2.12: d ... µE"ya>.wv dya8wv alna iiµiis dpyciaaVTo lKE"tvoL ('if they have done'), µlpos l-yw OUK civ l-MxLa-rov 8LKalws Ta.Am nvis Tlilv avTcilv ,rpayµciTwv ,rpa6TEpov l1rEµE>,:{i&!Jaav ('if they have taken care'), ElK6TCa>S' llv l'iµtv lmnµiiiev· El BE µ"{iTE Tofn-o -yl-yoV£ µ'fi8' ol6v T' lUTl KT>.. Also Is. 1.30. See further K-G 2,472, who correctly analyse the example

from Isocrates as an open condition ('rein logisch'). However, K-G go strangely astray in applying the same analysis to Ba. 1343. According to them d st awcf,povE'i:v l-yvwTE = 'Wenn ihr zur Vemunft gekommen seid, oder: sein so 11 tet (nicht: w llret)'. This is, of course, impossible, if alone because of the future reference assigned to fylll&ITE. 319. Cf. also K-G l, 23lff. and 2,467 for instances of opL+ clv in independent sentences and in regular conditional periods, i.e. with d + opL in the protasis. 320. 'It would seem that you have come' (Campbell). 321. 'If you have really stopped Electra' (referring back to 796, 1rrna6µE8' l'iµds). As often, the indicative conveys scepticism on the part of the speaker. 322. As observed above, Dodds seems to believe that optative + clv must have future meaning (cf. also his footnote on p. 237). Similarly, and in more outspoken tenns, heureux dans I' avenir ,. ' Dalmeyda: 'ce potentiel signifierait « vous pourriez (italics Dalm.). But this belief is quite mistaken. One out of many counterexamples is. S. El. 797-8, quoted above. See further Wakker (1986). 323. According to Gildersleeve § 442 opt. + clv 'so far from necessarily denoting uncertainty, ... is the combination most frequently used to indicate moral certainty'. If this were right, the opt. at Ba.1343 could very well be the correct fonn. Actually, however, it seems better to follow K-G, who write (1, 233) that the potential opt. was especially suited 'dem Vortrage fest begriindeter und bestimmter Urteile oder sicherer Thatsachen die Farbe des Zweifels und der Unentschiedenheit zu geben '. 1bere may, indeed, be 'moral certainty' involved, but this does not show. 324. Taken from Wakker (1986), who has an extensive discussion of the (dis)similarities of potential and contrary-to-fact conditions. On p. 237ff. she discusses possible pragmatic and rhetorical differences between them.

etre

159

vw

8' .... Here El+ optative is combined with optative + dv; in principle, then, the protasis denotes a state of affairs that may be realized. But the context makes it clear that it is actually very unlikely that it will be realized: vw Bl 'but in reality' has exactly the same function it has after unreal conditions, as in Hdt. 1.120.6: d cf>o[xp6v TL lvwpwµ£v, 1Tav dv aoL 1TpOEpaCoµ£v. vvv Bl .... On the basis of such cases we might postulate that the conditional period at Ba. 1343 is a mixture of these two constructions, combining an unreal protasis with a potential apodosis. On the other hand, the result is, again, hardly satisfactory, for we do not expect Dionysos to make a cautious assertion here, for politeness' sake. 325 All things considered the imperfect should probably be preferred to the optative. 326 I should add, however, that Supp. 764327 and Lys. 9.11, quoted above, can very well be analysed along the lines sketched above.

1350

8/8oic:TaL ... Tl,1µov£s- uya(

D.: 'The so-called axfiµa IIw8apLK6v or BoLwnov ... uyal is added in explanatory apposition to the impersonal subject of 8l8oKTaL '. R.: 'Kuhner-Gerth (I, p. 69, An. 2) et Dodds ont probablement raison de considerer uyat comme une exclamation, en apposition au sujet sous-entendu de 8l8oKTaL impersonnel'. Roux apparently has misread Dodds, for Dodds rejects the exclamation view, perhaps rightly, since there are no formal indications that uyat is exclamatory. Cf. also Gildersleeve §§ 117-8. So it is perhaps best to take uyat as just an apposition. Dodds also points out that 8l8oKTaL has no available plural form, and that uyat is singular in meaning. For uyal = cf>~ cf. also e.g. Hipp. 1043: lKTnva Tot a' KOU 4>u-

av

325. Cf. footnote 323. 326. Wakker suggests (pers. comm.) that d,&uµovo'i:T', being the lectio difficilior, might be retained if taken as a past potential: 'you could have been blessed'. Again, however, this is rather weak. Also, it is very doubtful whether Attic knew of a past potential, see next footnote. 327. In his note ad loc. Collard would rather analyse (/)alT)s clv as a past potential,= dixisses, referring to Goodwin §§ 442-3 on similar expressions in Homer. Note, however, that Goodwin writes (§443): 'Such examples are extremely rare in Attic Greek'. Moreover, one of his two examples can probably be discarded, Med. 568, where we should read, I think, 'icvlCEL (KvlCEL LP : KvlCoL 0). The resulting syntax (ouS' 4v afJ (/)alT)s, d crE µ1') 'KvlCn >.lxas) is exactly parallel to that of Supp. 764. I therefore prefer taking cj,a(T)s dv as 'You would say now, if you were/had been present then'.

160

yats l(T1µ(ovv, El. 233: troO 'Y'TlS' 6 TXiJµwv TXiJµovas cf>uyas

lxwv;, LSJ s.v. II.

1351

T(

8-fiTa µlllEO' iltrEp dvayica{(a)S lxn;

Gregoire: 'Que tardez-vous, si telle est la fatalite! '. R. (transl.): 'Pourquoi done differer devant l'ineluctible?'. L. (transl.): 'Pourquoi done tardez-vous executer l'ineluctible arret?'. K.: 'why then delay over what is inevitable?'.

a

Dodds and all other modem commentators are silent on this line, thereby suggesting, quite wrongly, that the syntax of this line is transparent. Of the older commentators only Bruhn has a short note: 'sc. 8pdv'. So we have only the translations to go by. While all four translations agree in taking µi>.>.w as 'delay' or 'hesitate', there is considerable disagreement as to the syntax of the dtrt:p-clause (as far as this can be judged by the translations). Lacroix apparently assumes that this clause depends on an unexpressed 8pciv or trpacrcrnv (cf. Bruhn), while Kirk rather seems to take the dtrt:p-clause as some kind of acc. respectus. Gregoire's 'si telle est la fatalite' cannot be got from the Greek. Finally, Roux' 'devant' in 'devant l'ineluctible' I simply do not understand. But how should we construe? Can dtrt:p d.vayKalws lxn be made to depend on µl>.MTt:? The data presented by LSJ s.v. III suggest that it cannot. 328 In the meaning 'delay, put off, hesitate' µl>.>.w is construed in just two ways: it may be followed by an infinitive, or it may be used absolutely, but it is not construed with an object or object-like constituent. If µlX>.w is used absolutely, it is usually clear from the context which particular action the hesitation relates to, cf. Ale. 255: T( µlXXns; ltrdyou, spoken by Charon to Alkestis. So Tl µl>-XELS'; ;;; Tl µl>.AELS' llvaL; similarly Hee. 1094: Tl µl>-XETE; sc. llvaL, to be supplied from 1093: ti> LTE µ6XETE trpos 81:wv. 329 Obviously the construction found at Ba. 1351 is in no way parallel to expressions like Tl µl>.AELS'; (where T( is adverbial, 'why'), and it would therefore be entirely ad hoc to supply 8pciv or the like to µl>.AETE. The only possibility left is to take litrEp etc. as the direct object of µl>.AETE. Although there are no parallels for the construction µlX>.w TL, the transitive use can perhaps be defended on the analogy of d.va~a>.>.w 328. This use is not discussed by Basset in his monograph on µlUw (Basset 1979). 329. A strictly absolute use occurs at Held. 722-3: "AprJS' crruyet / µllloVTaS' ('people that waver').

161

TL. Also, the existence of passive µlXMTaL, cf. X. An. 3.1.47: ws µri µl:UoLTO ... Ta 6foVTa, and Dem. 4.37: lv ~crc.i> Taiha (sc. lµ~lVE"LV, O.VTEµ~Lf3aCew) µlAAf"TaL, suggests that µl>Jw> may have been used transitively after all.

1373

l8dicpuaa

Roux quotes Dalmeyda's excellent remark: 'Une action presente est representee comme ayant commence anterieurement; ainsi etendue dans la duree, elle semble prendre plus de relief et d'intensite'. The aorist specifically conveys that the speaker had begun to feel the emotion concerned even before his interlocutor finished speaking. So here Kadmos indicates that he had started weeping for Agaue and her sisters even while Agaue was saying O"Tlvoµat cre. That the emotion is still present at the time of utterance is usually left implicit, but may for emphasis' sake be expressed explicitly, as in Hel. 673: icaTe&iicpvcra ical ~Mc/>apov vypalvw I 6dicpvcrw 'I burst into tears' (probably when Menelaos said (669): Tls cre 6alµwv fi 1r6Tµos crv>..q.. 1rdTpas;) 'and am (still) weeping'. This use of the aorist is best taken as a variant of the ingressive aorist, cf. K-G 1, 164.330

1382

tva ... ~11"1dµ£8a

To be connected with dye-r{E) as a complement of direction: 'to where'. Roux, in her translation, is uncharacteristically wrong in taking'lva as a final conjunction. 330. Moorhouse holds, p. 195, that the aorist rather brings to the fore 'the instantaneous nature of the occurrence' (sc. of the feeling) and that 'so far as the time of the occurrence is concerned, it should be taken as contemporary'. But this is less probable, in view of examples like Hel. 673, where the actual feeling is referred to by a present tense fonn. Cf. also Ar. Pax 1066: (Trygaeus) all3ml3oi:. (Hierocles) Tl yt:)4s-: (Trygaeus) ,'fa&i,v xapo,rotaL m8~KOLS-, where ,'fa8TJv xapo,rotaL m~KOLS- explains why Trygaeus had burst out laughing in the middle of Hierocles' recitation.

162

Appendix 1. Differences with Murray's text The following survey presents the departures from Murray's text: 21 68 102 135 140 144 150 154 191 193 206-7 262a263 270 293 294 406 408 440 451 506

523

ICaJCE"L

with LP

µe-M8po1.s; 0rJp6Tpocf>ov YJ8us l'lpe-crow b B' l~apxos. Bp6µLoS' icairv6s 136o"rpuxov XAL&i oiiicouv

punctuation Roux accentuation Rijksbaron with LP, Verdenius with Schone punctuation Verdenius with LP, Kamerbeek, Roux with Earle with Victorius Rijksbaron punctuation Dobree with LP

TIS' b&i>; TIS' lJ&i>; TIS'

1TaL8aywy{J(Jw a';

ehe- ... ehecfri,· Til> 8uooe-(3E"las 8pacrus' VEllCEaLV

ya1av dvoµf:3pot, e-inrpe-irEs µru.VEa6E"• ,rapwv

Cw

oOO' ... oOO' tµTlpci>t

Rijksbaron with LP no lacuna after this line with Usener with Thompson punctuation Roux with LP, Tyrrell with LP, Tyrrell Rijksbaron with LP, Kirk (first instance) with P. Ant., Kirk; (second instance) with LP, P. Ant., Kirk Rijksbaron (irpci)Tov, exempli gratia)

553-4 µ6>.E, xpuaiirrra nva.aawv dva 8vpaov icaT' "OAuµ,rov, punctuation Roux 558 '1.L6vua'; punctuation Rijksbaron 738 lAKovaav with Reiske 163

758 791

EKalee'

808

ta&

843

i\v

withP with Schone withP with Musgrave punctuation Musgrave Rijksbaron

ICLVOUV TL

f3aicxas

, TOUT6 y' la8L,

860 868 869 877 959 983 1002 1006 1007 1022 1024 1067 1083 1147 1157 1164 1232

164

AL. t>.86vT' ES' ollCOUS' . . . ITE.

i\v 8otcfl

l3ovMoooµaL

line division Roux punctuation Rijksbaron AL6vvaov 3s' trlcf>vicEv, cf,of3Epov with Musurus with P, Maas 8fpaµ' [TO] with Paley, Dodds of;icovv ... , Rijksbaron with Campbell; trpol3}.frros ffl aic6>.otrost Rijksbaron, exempli gratia yvwµ.av awcf>pova· with P; semi-colon Rijksbaron 8avaTos withP trEa6vTa with Heath TIVTVXELS' y).acp6µEVOS with Palmer with Reiske t>.i.ic66poµov €empt{€ with P with P, Roux, Kirk fl "AL8a with Pap. Ant., Reid with Wecklein [trEpL ]l3aMLV with P, Tyrrell alrrfis 1



Appendix 2. The speeches of Pentheus and Teiresias, 215-262 and 266-327 Several commentators note that these two speeches differ formally: whereas Pentheus' speech is an 'angry tirade' (Dodds), Teiresias' rejoinder is rather modelled on oratorical practice; see Dodds ad 266-71 and Roux p. 337f. It is not sufficiently acknowledged, however, that they are also very different qua content. Pentheus' speech focuses on three points: first, the behaviour of the Theban women (217-225) and his measures against them (226-232), second, the appearance and behaviour of the ~lvos (233-241), as well as the things said by him about Dionysos (242-245), which is followed by another reference to the stranger (246-247), and, third, the behaviour of Teiresias and Kadmos (248-260), which gives him an occasion to return once more to the behaviour of the women (260-262). Thus, on a total of 48 lines, 16 + 2 1/2 lines deal with the women, 11 lines deal with the stranger, 4 lines deal with Dionysos and 12 1/2 lines deal with Teiresias and Kadmos. Not only, then, are the Theban women uppermost in Pentheus' mind in terms of the number of lines devoted to them, he also begins and closes his speech with a reference to them. As for Teiresias' speech, it is certainly true, as Roux remarks, that he 'refute methodiquement les accusations de Pen thee'. But this is only half of the truth, for the attention Teiresias pays to Pentheus' main concern, the Theban women, is negligible when compared with that paid to Dionysos. Teiresias' speech contains the following elements. After an introduction of 6 lines, in which he deals with Pentheus' state of mind, partly in general terms, the bulk of his speech is devoted to Dionysos: the properties of the god (272-285), an attack on Pentheus' disbelief concerning the birth-myth (286-297), further properties of Dionysos (298-309). Then follow his appeal to Pentheus (309-313), a general remark on the behaviour of women (314-318), a second general remark, in which the behaviour of the women is implicitly explained as a tribute to Dionysos (319-321),1 a short passage about himself and Kadmos (322-325) and two final lines in which Pentheus' state of mind once more comes to the fore (326-327). On a total of 62 lines, not a single line is devoted to the 1. This explanatory function of 319-321 is lost if these lines are put after 313, as e.g. Metzger and Kopff do.

165

stranger, while no less than 37 1/2 deal with Dionysos. The second most important item is Pentheus (12 1/2 lines), while only 5 (or 8, if we include lines 319-321) lines are devoted to the women, and that in a very general way, and even less (4 lines) to Teiresias himself and Kadmos. These two points are, then, treated very disparagingly. This also appears from their position in Teiresias' speech: at the very end. The message is clear: rather than occupying himself with the Theban women and with Teiresias and Kadmos, Pentheus should be concerned about his own behaviour (note that Teiresias opens and closes his speech with a reference to Pentheus) considering the importance of the new god.

166

Appendix 3. The constructions of o-0 µ.~ The most extensive discussion of the use and origin of ou µ11 + future indicative, and its relationship with ou µfi + subjunctive, still is Goodwin, Appendix II. Whereas his treatment of the subjunctive construction (pp. 390-94, also§§ 261-6) is very convincing, his views of the future indicative are untenable. The gist of Goodwin's argument concerning ou µfi + subjunctive is as follows. (I slightly rephrase the terminology and add some explanatory observations). Ou µfi + subjunctive expresses a strong denial of a future state of affairs, and occurs in declarative utterances. It is the negatived variant of µfi + subj., which is not seldom found used declaratively as well. The latter is especially frequent with 3rd person subjunctives, but there are some cases with 1st and 2nd person. These utterances, which have the form of an order/request or a prohibition, very often cannot meaningfully be interpreted as expressing a real order/request or prohibition, e.g. because the addressee of the order etc. is not present in the speech situation; this explains, in fact, the preponderance of the third person.1 In such cases µfi + subj. expresses apprehension, coupled with a request to an unspecified addressee to avert the object of fear. 2 Some examples are: Hom. /l. 16.128: 3 µT) 8T) vfias- l>.wm Kat ouKETL 4>uKTC1 1rl>.wvTaL (Goodwin translates(§ 261): 'may they not (as I fear they may) seize the ships and make it no longer possible to escape'; ob1. The conditions on the use of regular 3rd person imperatives and prohibitions are rather intriguing and deserve a full investigation. Generally speaking, such imperatives have the following features. (i) Unlike 2nd person imperatives they need not have an animate subject. (ii) They express an order to some person who is not the subject of the imperative to see to it that the state of affairs concerned is carried out. (iii) This person may or may not be present in the speech situation. E.g. Ale. 999: (Tvµ~OS') 8EotaL 8' l>µollllS' I nµda8w 'must be honoured' sc. by people. If both the subject of the imperative and the addressee are present, a 'two-stage' order is given: the addressee is ordered to see to it that the subject of the imperat. carries out the state of affairs requested. See e.g. Ale. 1104: fi 'YV"'l 8' chrE>.eln,>, spoken to Admetos by Herakles. 2. Goodwin's words run(§ 261): ' ... apprehension, coupled with a desire to avert the object of fear'; 'desire to avert the object of fear' is somewhat misleading, since the speaker himself typically has no say in this. 3. This construction is rather common in Homer, Goodwin§ 261.

167

serve that the second verb itself is negatived by ou); Or. 776: µ'i) M[3wat a' cfoµe-voL; Hdt. 7.10.111: µ'i) ofhw ylv,,TaL. 4 As for 1st and 2nd person, this same 'apprehension-value' is present when the predicate is non-agentive: 5 such predicates typically cannot occur in orders/requests and prohibitions. Compare: Hom. Od. 16.381-2: µfi TL KaKov {>£twaL Kat -fiµlas lte-MawaL, / ... d>.>.wv 8' a.4>LKwµe-8a 8f)µov; 6 S. OC 174: ~ te-1voL, µ11 8ilT' a.8LKT)8w; 7 with µ11 ou: Tro. 982: µ11 1rE"l011s

ao4>ous. 8

Although the above cases are not prohibitions in the strict sense, since they do not directly concern the addressee, there is clearly a prohibitive value: in all cases the speaker simultaneously expresses fear and a general request that the state of affairs he fears shall not obtain. The latter value is virtually absent, however, when the predicate has an inanimate subject. In such cases µfi + subj. expresses rather a cautious assertion that something will be, and µT) ou that something will not be the case. This use of µT) (011), which is first found at Hdt. 5.79.2, is a favourite of Plato's. Two examples: Pl. Grg. 462 e: µ'i) a.ypoLK6Te-pov ti TO a.>..11BE-s d tre-1v ('I am afraid the truth may be too rude a thing to tell' -Goodwin,§ 265); Ap. 39 a: a.AM µ'i) ou Toirr' ti xa>,nr6v, 8civaTov lK4>uye-1v ('but I suspect this may not be the hard thing, to escape death' - Goodwin). Observe that in both cases the subject of ti is a state of affairs. 9 To come back to ou µfi + subj., this is, then, the negatived variant of the above (semi-)declarative construction, and expresses a strong con4. Although formally speaking such requests are not directed to anybody in particular, the addressee of these words may feel that they amount to an indirect request to aven, indeed, the object of fear. Thus the indirect illocutionary force - to use the current term - of fl. 16.128 is: 'and you, Patroklos, you better take care that they don't seire them'. 5. Cf. for this notion e.g. Rijksbaron (1984: 127). 6. Mi't dlj,ucwµE8a Bijµov differs, then, from e.g. agentive µ1'! d.1rlwµEv in that in the latter case it is within the power of the speaker (and the other persons included in 'we') not to carry out this state of affairs, whereas in the former this depends on other persons. 7. This is presented by Goodwin (§ 257) under the heading (see § 255) 'negative exhortations', not under the independent clauses withµ~ (expressing fear; § 264). Both analyses can, I think, be defended. 8. oil is due to Seidler and is generally accepted. 9. Ml'l (oil)+ subj. can also be used interrogatively, e.g. Pl. Phd. 64 c: clpa µ1'! 11>.Ao TL ~ b OcivaTOS' -fl ToOTo; ('is it possible that death can prove to be anything but this?'). See Goodwin § 268.

168

viction on the part of the speaker that the state of affairs concerned will not be realized. It occurs with all three persons; I give an example of each, referring for further examples to Goodwin § 295: (1st person): A. Th. 38: ical Twv8' aicovcras OU n µri AT14>8w 80).(j) ('I shall not be caught by any trick'); (2nd ps.): Ar. Lys. 704: icovxl µfl 1TaULcrµaTWV TOUTWV, 1Tplv c'iv ... ; (3rd ps.): HF 718: b 8' 1Tape-crnv, ov6€ µri µ6>..-o 1TOTf. On p. 394 Goodwin emphatically and convincingly argues against assuming an ellipsis of 8fos or the like to explain this construction. Observe also, in this connection, that the proponents of the ellipsis analysis inconsistently fail to assume an ellipsis to explain the construction of simple µfi + subj.

ov

But what about the future indicative? Here, as I said above, Goodwin's analysis is untenable. He remarks (p. 395): ' ... it was only natural that the Attic Greek should soon begin to use the future indicative ... in the same sense' (viz. of a 'simple future denial'). Why this was 'natural' is not clear. As examples of this 'future denial' Goodwin gives S. El. 1052: a>..>..' dcrL8'· OU O'OL µri µe-ellJ,oµal TT0TE" and Ar. Ra. 508: µa Tov 'A1T6'>Jwl µfi a' l-yw I ne-pL6$oµa1Te->..86VT'; at§ 295 he gives three further examples. Such cases are, then, very rare. 10 Nevertheless Goodwin continues, p. 395, middle: 'When µfi with the future indicative had been established as a regular form of future denial, the second person singular probably began to be used as a form of prohibition'. According to Goodwin, then, these prohibitive cases of µfi + fut. indic. should be seen as declarative utterances; in §§ 297 and 298 he presents a considerable number of such cases, among them Ba. 343, which he translates: 'do not bring your hand near me; but go and rage, and do not wipe off your folly on me'; J3a.Kxe-ucrns 8' lwv is apparently taken by him as an affirmative command, while 'µT18l continues the original prohibition as if there had been no interruption'. All this is highly implausible on at least three counts. (i) The prohibitions and the command would seem to come from an instruction-book rather than convey an urgent appeal. (ii) The future indicative 2nd person may, indeed, have a iussive value, but this is rare; also, the future is not just an alternative to the imperative, as Goodwin holds (p. 395 and § 69), but has some values of its own, as in Ar. Nu. 1352: navTws 6€ T00To 8pacrns 'you absolutely shall do this' (wrongly translated by G. § 69: 'but by all means do this') and in an imprecation like otµ~rn8E" 'you'll

ov

ov

ov

IO.

See further below, p. 172f.

169

have a very bad time'. Generally speaking the future expresses a strong conviction that the addressee will carry out the state of affairs involved. (iii) It is a priori improbable that J3aKXftJC7ELS' 6' lwv is simply an interruption. In analyzing such cases as Ba. 343 as declarative, Goodwin deviates from the common practice, then and now, which was established by Elmsley in his note on Med. 1120 (= 1151 of modern edd.) and which sees such utterances as questions (I will come back to this below). Goodwin refuses to accept this on the ground that by making such cases as Ba. 343, and indeed the great majority of ou µfi + fut. indic. cases, interrogative we separate them from the construction ou µfi + subjunctive, which, as we have seen above, in Goodwin's view belong together. As he puts it in § 300: 'But this explanation (viz. that of Elmsley - AR) requires an entirely different theory to account for ou µfi in clauses of denial (295), where no question is possible'. Indeed it is not, but G. himself has adduced the evidence to explain why it is not (see above pp. 167f.). Moreover, we have seen that his declarative analysis of Ba. 343 runs into serious difficulties and this applies to most examples of§§ 297 and 298, especially to the latter, which are similar to Ba. 343. Goodwin would recognize only a few cases as interrogative, viz. those where the sentence opens with a single ou + fut. ind. and is followed by µfi + fut. indic. (§ 299): 'Sometimes ou with the future indicative in a question implying an affirmative answer (thus equivalent to an exhortation) is followed by µfi or µT16l with the future in a question implying a negative answer (and thus equivalent to a prohibition). Here there is no case of ou µfi'. An example is: S. Tr. 1183: ou 8aaaov ofons (sc. Xf1pa) µT16' dmcrTTJcrflS' iµot; ('will you not extend your hand, and not distrust me?' - Goodwin').1 1 While there is nothing amiss with Goodwin's interpretation as such, his syntactic analysis is unconvincing, for µfi with the future indicative in questions is not attested independently. (The grammars give no examples). In other words, ou extends in all probability to the second member as well. Apart from this, other objections can be raised: - it seems an entirely arbitrary procedure to separate ou µfi + fut. indic., followed by a 'bare' future indicative (as at Ba. 343) from those cases where ou comes first and is followed by µfi or µT'IBl; 11.

Interestingly, this is an emphatic repetition of the inceptive pres. imper.

lµl3aUE X£Lpa 8£~uiv in 1181.

170

- it is, again, arbitrary, to connect ou µri + fut. ind. 2nd ps. with ou µTl + subjunctive, while disconnecting it from the very common use of ou +fut.ind. 2nd ps. in urgent questions, with iussive value, see K-G 1, 176, 7. It is worth noting, in this connection, that this use is not discussed by Goodwin at all. - at Ar. Ra. 298-9 we find lKETEVw added to µT) Ka>.ds µ' ... µT18£ KaTEpELS Toi'.lvoµa, which points to this utterance being a question: 'I beg you' can hardly combine with an assertion. I conclude that the construction µTl + fut. indic. 2nd person should be taken as a question. As Elmsley argued in his note on Med. 1151 (= his 1120), it is the negative variant of +fut.ind. 2nd ps. in questions. Just as ou µEVELS;, to take Elmsley's example, is a variant of µlvElµE'ivov, meaning 'Will you not stay?', so ou µ"i) µEVELS; is a variant of µ"i) µlvElµT) µElll'(ls, meaning 'Will you not not stay?', i.e. the question elicits a positive answer to the rejection of the state of affairs concerned: 'Is it not the case that you will not stay?' = 'You surely won't stay?' 12 This is basically also the view of K-G 1, 177; they suggest, furthermore, that it may be the interrogative variant of µTl + fut. ind. 2nd ps. (on Hel. 437-9: OUK d,ra>.Meu 86µwv / Kat µT) ... / 6x>.ov ,raplens 8Eo1T6Tms; 'wirst du dich nicht entfemen und die Herrschaft unbehelligt lassen?'): 'Es ist also in die Form der Frage gekleidet, was in der Form der befehlenden Aussage lauten wiirde: d,ra>..>.aen Kal µ"i) 6x>.ov ,raplens'. However, in view of the rarity both of the latter construction (for the two or three examples see K-G 1, 176, 6 and Goodwin §70) and of the iussive fut. ind., one might also consider them as derivative vis-a-vis the far more common construction with ou µTl + fut. ind. 2nd ps. in questions. All in all I think the relationship between the above construction and the 'regular' imperative can be summarized in the following equation: µlvE/µELVOV : OU µEVELS; = µ"i) µEVE/µ"i) µElVQS' : OU µ"i) µEVELS;

ov

ov

ov

Finally, two points deserve notice: (i) In his §§ on prohibitions (297, 298) Goodwin discusses a few exµTl + subjunctive 2nd person, noting that these are the amples of forms found in the mss. These are: Ar. Nu. 367 and Nu. 296; the latter

ov

12. Moorhouse has misunderstood Elmsley's note, for he says, p. 337f.: '{Elmsley's interpretation) ... by which the negatives cancel each other: ou µ'it 'feyel)E'ts- [S. Tr. 978] ... "will you not wake?" = "don't wake"'. If this were correct this should be, of course: 'do wake'. Elmsley, however, correctly says: ' ... Graeci ... particulas ou et µ1' saepe ita conjungunt ut altera alterius vim non tollat'.

171

case is particularly interesting because it is followed by d>.M + imperative: µT} aKWlplJS' µ116E 1TOLT)CJ1lS' ('so all the Mss.' - Goodwin) liirEp ol Tpuyo6a.tµovES' ovroL, d>.X' Ev4>rlµn. Actually, however, the mss. are not as unanimous as G. thinks. At Nu. 367 most mss. have, in fact, A'flPTJ.EeriaE"LV KaKOV ~µap.16 Observe that 3rd person ol'., TOL µ-fiiroTE" . . . TLS den (S. 176), just like II. 10.330, expresses a commitment on the part of the speaker to

oc

Mn

14. is prohibitive; this 3rd person oath in reality expresses a strong commitment on the part of the speaker that he will see to it that this state of affairs will not be realized. The same idea may be conveyed by the asseverative ou µdv, as in Il. 17.449-9: dU' ou µciv uµi:v ... / "EKTwp JlpLaµl6rjs hroxflanm, where the commitment of the speaker is made explicit, for these words are followed by: ou -yap tdaw. 15. may also be followed by a present or aorist indicative, /l. 15.41: (1'.aTW vOv T68£ rata) . .. µ11 & ' lµtiv l6TIJTa JloaeL&iwv . . . 1TT]µalveL TpJ,.w.17 All in all it must be concluded that these cases are entirely different from those with ob µT} + future indicative 2nd person. 17. Ou µfi also occurs in indirect discourse + future infinitive or future optative, in solemn statements: S. Ph. 610-2: &s (: Odysseus) 8-r'I Tei T' dll' alrrotaL 1rciVT' tetamaEv I K«l Tcitrl Tpo(q 1rlpyaµ' ws ou µfi1roTE / 1rlpaoLEv; Pho. 1590: aa♦ws -yap EhE TELpEa(as OU µfitrOTE / aou Tfiv&- ')'iiv olKOUVTOS E~ trpafELV tr6Al.v.

174

Appendix 4. Forms in -aC&> in questions: future indicative or aorist subjunctive? If Ba. 193 should be read as a question, as it probably should (see Dodds' remarks ad loc.), we must decide whether lTaL8ayw'Yl'law is a future indicative or rather an aorist subjunctive. It is often said that these are practically equivalent, cf. e.g. Goodwin § 68: 'The future is sometimes used in questions of doubt, where the subjunctive is more common', K-G 1,223, A.5: 'Wie das aussagende £~6µe-8a ... dem Ausdrucke des Begehrens ct,vywµe-v ... sehr nahe kommt, so das fragende lTOL E"v~6µe-8a; ... dem deliberativen lTOL ct,vywµe-v;' and Verdenius4, in a rather puzzling note on our line: 'Probably a future ... equivalent to a deliberative subjunctive'. Since this is a complicated and intriguing matter I will discuss it in some detail, taking non-ambiguous verb forms as examples, and making a distinction between two question-types: (a) yes/no-questions or sentence questions, which may or may not be introduced by special interrogative particles, and (b) x- or word-questions, introduced by interrogative pronouns and adverbs like Tls; lTWS'; lToO; etc. There is, of course, yet another type, viz. that of disjunctive questions. These will not be discussed separately - in fact, the material is rather scanty - but will be referred to in the notes at the appropriate place.I My examples are mainly from tragedy, but occasionally I have taken into account cases from other writers, notably Aristophanes and Plato. 1. Disjunctive questions can be subdivided into two types. In the first type, both members of the question have theirown verb: {tr6Tepa/-ov) Vii\ V2:: in the second type there is only one verb and the question relates to one or the other of two contrasted constituents, which both function as modifiers of the verb: {tr6TEpa/-ov) XI 11 x2 V;. Compare the English questions: shall we stay or shall we go?; and: Shall we go by train or by car? The second type may be considered a variant of x-questions (cf.: How shall we go? By train or by car?), the first is like yes/no-questions, although they differ from these in that a simple yes or no is excluded as an answer. Cf. for English Quirk et al. (1985: 823ff.); the differences mentioned are ignored by Greek grammars. - II6Tepa/-ov may also introduce single questions; in such cases the second member possibly has to be supplied from the context, cf. K-G 2, 532, Anm. 10. Or perhaps there is a contrast with an implied: or not?

175

1. Yes/no-questions The emphasis is on bare questions, i.e. questions that are not modified by interrogative particles. However, to broaden the perspective somewhat I have added some observations on questions modified by clpa and

~-

1.1 Subjunctive Since the dubitative subjunctive may be considered the interrogative counterpart of the adhortative subjunctive,2 its semantics may be formulated as follows: the speaker envisages a certain action but is not certain of the necessity (inevitability, desirability) of that action. Two types should be distinguished, according to whether or not the question is directed at an addressee, i.e. another person who is a participant in the dramatic action. If the question is directed at an addressee, usually in stichomythia,3 the speaker asks the addressee to give his opinion on the necessity etc.of the action envisaged. The decision as to whether that action should be carried out is, then, committed exclusively to the addressee. The latter's reaction, in fact, often has the form of an imperative. This type of questions is, as far as this can be determined, pragmatically neutral, i.e. there are no particular emotional overtones involved: in the course of an exchange of words the speaker just asks his addressee for information. An example is: (1)

Ay. ipas BE- trpooµevwµe-v; a11. µlµv', ws ...

fi

Tl

XP1l

11'0€1v;

(S. Tr. 390-1)4

Other examples are: S. OT 364, Ba. 819, El. 280, Hel. 812, HF 1109,5 IA 1449, Or. 777, Pho. 728, Supp. 293. 6 The appeal to the addressee 2.

Cf. K-0 1, 221.

3. I have found only one example in lyrics: Med. 1275: irapl>.9w 86µo'US';, spoken by the Chorus to the children. 4. A nice example from Aristophanes is: Ra. 1: dirw n Tciiv dw86-rwv, c!i 8lairoTa ... ;, where the question perhaps involves possibility or permissibility rather than necessity; cf. also note 5, and example (4). 5. Again, cf. note 4, the question lXOw ... irlXas- rather involves permissibility. 6. In disjunctive questions, the subjunctive may alternate with other expressions of necessity, e.g. Hel. 1083: ir6TEpa 8' ls otKo'US' aol avVELaE>.Mv µE XPili I -i, ,rpbs- Tci+ci, T4i8' f\avxoL Ka8wµE8a; answer: alrrou µlv'.

176

may be reinforced by the addition ofl36A"Q, 8eMLS' and similar forms, as in:

op' bn-\.v ii 61XTTT1vos 'IDEKTpa; 8e>.Etsµ£tvwµev airrou Kci1TaKooowµe-v -y6wv; Ila. f,o.OTa.

(2) Op.

(S. El. 80-2)

(3) HA. l36A"Q 81.-yw aov KcivaKoucf>taw 6lµas-;

(Or. 218-9) Op. >.a.130 >.a.130 6fyr', .•• Observe that in (2) ~KtCJTa is the answer to 8e-MLS', whereas in (3) >.a.130 >.a.!300 directly reacts to the subjunctive 8l-yw. Cf. also e.g. S. Ph. 761; Hee. 1042, Ba. 719 and K-G 1, 221f.

Such questions may also be introduced by fl, as in the following examples:7

(4) Av. ID' lxrrtS' b TOITOS'

Ot. vat, TEKVOV, ...

(5) Op.

Ilu.

Tl

M-ywµe-v

fl

µa8w µo>.oooa 1T0L; (S. OC 26-7)

ow a&Nf,6 TaUT' lµfl;

µi1) 1TfXlS' Ofwv.

(Or. 787)

These questions are very similar to those with unmodified subjunctive: they are dubitative questions and ask for the addressee's opinion on a given matter. Cf. also the answers vat and µi1) 1Tf)OS' Ofwv. Similar cases are: S. OC 195,478; Ion 560. In some cases, however, the tone of the question is not neutral. Various emotional shades may be involved, depending on the context, and on the structure of the question itself. For the latter cf.: (6)

A6. 81.-yw, ,rpooe-t,rw CGx,-av

ws- 6' 'Ca8'· .tuva, icd.µTTMicw Tou aou µ6pou;

(S. Ant. 554)

where Ismene, in despair, formulates the consequence of Antigone's decision to die and of her advice awaov ae-aun'Jv (553). A similar case is S. Ph. 908. 8 It would seem that in such cases ~oUAlJ etc. could not easily be added. In monologues such questions often alternate with other questions, of various types, and may be called dubitative sensu stricto, since the speaker is at a loss and has to decide himself whether he shall carry out the action or not. 9 An illustrative example is:

(10)

fiKW

6' d.v6."}'ICT1S ls T66'· om-· lµa1s 4>t>.a.Ls 9fil3aLS lvoLice-1v lxYLov· fiv ~ ical µlvw, ls TTo1ov le-pov fi TTavfi-yupLv 4>t>..wv dµ '; ou -yap chas e-irrrpoa11y6pous lxw. d.AA' "Apyos l>.8w; TTWS, lTTe-l cf>e-u-yw mhpav; fp' d.AA' ls dAATJV 611 nv' bpµ,,aw TT&i.v; (HF 1281-6)10

8. An example of a disjunctive question which is not addressee-oriented is Ar. Nu. 845ff. 9. Here the addition of l3ou>..1J is not possible. 10. Cf. for the 'dialogue with oneself' (hypophora) Denn. lOf. - For the sequence future (ElµL), subjunctive (l>Jlw), subjunctive (probably, see below 1.2) bpµTJCJW cf. below, p. 189.

178

Similar questions of despair are: S. Ai. 466-8 (with d.>.M 8fiTa),ll Ph. 1352 (with d.>.M ... 8fiTa), Andr. 193, 198, HF 1384, 12 Supp. 1095 (with KQ.Ta). For completeness' sake I mention here two further uses of the subjunctive in yes/no-questions. First, it may be used in self-corrective questions, introduced by..,, e.g. A. Ch. 1073-4: vvv 8' ali TplTOS ~X8l ,ro8ev au>7T1p, / µ6pov dTTu>;. Cf. also S. Ai. 1312, and probably also Ba. 949: µox>.ovs 4>lpwµev; "' XEp0LV d.vaamiau> I KTA., where the absence of TT6TEpa points to a corrective rather than to a strictly disjunctive question. Second, in comedy the subjunctive occurs in indignant echo-questions, following upon rather than eliciting an imperative, e.g. Ar. Lys. 530: aua'.ma. - aol y', @ KaT.a,,r~v lBpas I µ6vovs T' 'ATpdBas, 1rt>.ayos Alycuov TTEp.E- &O ... 1T6TEpov xa(pw Kd1ro>..o>..u~11> . . . 11 KMOOII> ... ;. Cf. also, with 1r6TEpa alone, A. Th. 95. 13. At A. Ch. 112 Page and others read hE~oµaL Ta&;, but the question mark is apparently not in the mss., and Murray and Groeneboom read this line as a statemenL Taking it as a question is the more improbable because it contains the particles TOL dpa, which in none of the examples given by Denniston (: 555) occur in a question; cf. also Denniston's footnote on p. 555. Incidentally, the extreme rarity of future indicatives in yes/no-questions pleads strongly against putting a question-mark after ,rws ow in El. 646: 1rws ovv lKdvriv T6v& T tv Taln-(ii KTEv..Ol/luxi'laoµ£v, / TOV KaTilav6v8' bpL. n Kal 1Tp0S {µwv @8' Ep-ilµas, ill etvoL, MLcf>0{poµm 811 KOU< rnOLKTLpELTE" µE; Xo. ISS' EOTlv T)µwV VO.UKpa.Twp b ,ra1s. Ix,-' ofrras >.lYQ aOL, Tawa. aOL xflµE1S cfxiµEv. (S. Ph 1070-3)

av

(14) Op. TI. 8fyra 8pwµEV; µwfp' ~ VEOO'OµEV; HX. µ.wv a' ol.KTas ElM, µT)Tpos d&s Blµas;

ws

(El. 967-8)

These questions express various degrees of scepticism or disbelief; 15 note especially Elektra's reaction in (14): she interprets Orestes' question as a sign of weakness. 16 Cf. also A. Ag. 1362, 1366; Pho. 1673. Observe that the questions of examples (12) - (14) are all non-dubitative. which is not a dubitative but rather a rhetorical question, the suggested answer being: µaxovµe8a. 14. So, too, Ar. Nu. 465; PL Tht. 208 a. Cf. Denniston 46: •apa, by itself, often has a sceptical tone'. 15. Cf. for a clear example with a 2nd person S. Ph. 810: 8cipaeL, µevoOµev. - ~ µevets-; - In S. El. 940 we find an ambiguous ~ TOUS' 8av6VTaS' l~avaatjaw 110Tt; This clearly expresses disbelief and should, therefore, be taken as a future. 16. In (14) many editors read T( BijTa 8pwµev µT)Ttp'; ~ 4'veuooµev; on the ground that in principle ~ should not be in second position. (Cf. also Denniston p. 283). Diggle punctuates after 8pwµev, but his reference to A. Ch. 899 is irrelevant because there is no~ in that line. Be that as it may, we should certainly punctuate after 8pliiµev, for (i) there is a parallel for the postponement of~ in S. OC 26, (4) above; (ii) the postponement of~. very appropriately, gives prominence to µT)Tlp(a). ol, nµrbµevov;

180

Future indicative questions may also be given a sceptical tone by other means. Consider: (15) Ay. dCT() cro· X£Pvlf3cilv yap e:cm'to rre:MJ.S. l. cm')croµEv dp' aµl f3Jy£L, µa&'i.v, at T al, Tou8"· £li yap dpllTaL &rr}.u. Kp. ol TTy,LKol.& Kal &~6µrn0a 811 .i.v; Tlva ee-vwv; Kp. lS,rov X8oVOS Ti'p8' E"KTTo8wV µ6.>.i.OT' €C711. (Pho. 972; 977-8)22 The appeal to the addressee can be made explicit, as in: (24) HX. 8 ~ yvva1KE"S, 8wµa.Twv e-i.iefJµove-s,

rne-l trape-OTe- Ti'Jcr&- TTpor,TpoTTiis tµot troµ.trol, ye-ve-cr8E- Twv&- cruµl3ov>.oL TTE"pL ·

21. Cf. the dubitative expression at Ion 971: T( ycip µE xPi'i Bpdv; d.,rop(a TO BuoTUxav. Cf. also McWhorter (1910: 157f.): •... the speaker, in the midst of a dif-

ficulty, may find himself perplexed, distressed, entirely at a loss, and then the question, whether in the form of appeal to another or self-addressed, amounts to an expression - an exclamation - of grief, of despair. No answer here is expected, - it is already anticipated. This is typical dirop(a and is a commonplace with the Tragedians in working out tragic situations•. Mc Whorter distinguishes such questions from questions that simply appeal to another, the 'real appeal of the l3ou>.EL type', as he calls them. However, contrary to what McWhorter suggests, l3ou>.EL does not occur as a modifier of subjunctive x-questions, but only of yes/no-questions. All subjunctive xquestions, whether or not they are addressed at another person, would seem to express a feeling of dirop(a. 22. Two examples from lyrical passages are Med. 1271-2: otµoL, T( 8pciow; irot ♦uyw µTITPOS- XEpcis-; : : oiitc ots·, d.&>-4>i ♦(XTaT'' 6>.>.uµEoSa ycip, and Or. 1376ff.: irfl. ♦uyw, ftvaL ... ; In the latter case, however, the words of the Phrygian are an outburst of despair rather than a real question, cf. the Chorus' reaction: T( 8' foTLV, 'E>.iVllS- ,rp6o,r0X' ... ;

184

,i ~ xfoooa TOO& io,&lous xoas; TTWS- e-ficf,pov' efow; TTWS- KaTE~wµ.a1. TTaTpt; CT. also Cy. 539, IT 96, 1188.

(A. Ch. 84-8)

In some cases, however, the nature of the question is such that they are not really directed at an addressee, although they occur in stichomythia; these questions are, then, dubitative in the strict sense (cf. p. 178), since it is the speaker himself who has to decide on his action. A clear example is:

(25) Ta. cl>puywv aptOTOU TTplv TTo8' ''EKTopos Mµap, µft µE OTU'YfiCTQS" EKWV yap 0.'Y'YENJJ

oux

davawv TE KOi.VO. Ile-Mm8wv T' a'Y'Yl>.µam. Av. ,i 8' fonv; Cs µot. cf>polµlwv apxu KaKwv. Ta. l&ee- T6v& TTa18a ... TTWS' ELTTW >Jryov; (Tro. 709-13)23

Comparable cases, albeit of a more emotional nature, are:

~o. Tl >.iew; TTOL >.6-ywv aµT)xavwv l>Jlw; KpaTELV yap OtJKfTL y>.wcrCTllS CJ8€11(J.). HX. ,i 8' loxe-s dXyos; (S. El. 1174-6)

(26) Op. E'u

(27) Av. @ ZED, Tl >.iew; mi € 8a.TTT€L; mi VLV €upT)C7W µo>.wv; Ek bpEttiv trap' ciµov ft 'trl Aa.ptcrav cf>lpn T{Jµ.[3ov KaT6tlru efC7T0V EK trpoaaTlou. (Ale. 834-6) (35) Hp. olµot, Tl Spa.aw; mi KaKiiiv EpTJµlav €t'.lpw, TTTfpwTOS' 1' KaTa. X8oVOS' µo}.wv; (HF 1157-8) On (34) Dale aptly remarks: 'supplementary information is required: since she is dead, where is she being buried?'; for Ka( she refers to Denn. p. 312. In (35), which admittedly is not fully comparable, since it occurs in a monologue, we are not dealing with a real question, let alone 26. The appeal to the addressee may be made explicit by elements like 'Al."'fE, IA 153ff.; cf. for the same phenomenon with the subjunctive p. 184, ex. (24). 27. A clear example of a disjunctive question of this nature is Ba. 941: rrlrrEpa U 8upaov &-tL4 M1f3cl>v xepl. / ,i Tij&-, 13 ylpov· lTOL XJrrl cf>vy€LV;

(Cy. 193-4)

which may be compared with: (37) 08. Tl 6f)C41.€V, c!> l:LAT1V£; OOL µ.EVEW BoicE'i;

D.. 6oic€'t. (Cy. 539-40) Observe that Tl 6paaoµ.£v, unlike the futures discussed above but like most subjunctives, shows asyndeton. Compare also: (38)

ws

er,.

OVK d lTQTP'¼)(IS fKToS'

er,.

86µ.0'IJS' fonµL, TflB' rn' al-rt~ cf>uywv; OOTLS yuvaLKwv Auµ.£wvas ~&TaL elvo'IJS' Koµ.tCwv Kal euvoLKOUpo'IJS' KQKWV.

TO,XLOTQ yiis;

hr. tro't 6'fi8' b TATUJ.WV Tplc#loµ.aL; Tlvos flvwv

(39) Ka. Tl µ.' dµ.cf>Lj3cill.f"Ls XEpalv, c!> Tahl.Lva tra't, 6pvLs &rrws KTJ'fiva troAL6xpwv KUKVOS; Ay. lTCi. yap Tpa.lTWµ.aL traTpl.Bos EIC{3€13},.T)µ.rn,; Ka. oiiK ol.6a, TlKVov·

(Hipp.1065-9)

(Ba. 1364-7)

Possibly in (38) the future has been chosen because of the preceding oiJK El ... ;,just as trot 6'fim cf,€uyw; continues cf,£0y' in Pho. 972ff. (ex. (23) above). As to questions where no addressee is involved, the situation is as follows. In some cases the future probably expresses scepticism about the possibility that the action concerned will be realized, e.g. in: (40) Iwv

w4>lA11

µ.fyr£p, tr6T'

apa

Kal abv &lioµ.m Slµ.as; (Ion 563)

(41) Ilu. tµ.riv yap aun'tv, fis >Jxos f1T1JVE"Oa, Kplvw Baµap-ra. Tl yap tpw Kayw lTOT€ y'fiv ~€Xq>l6' t >.awv s' trplv µtv iiµas 600TUX€LV cf,lAOS trap'fi,

vw 6' OVKET' dµl 6UOTUXOWTl OOL cf>l>.os; OUK lOTLV.

188

(Or. 1092-7)

In (40) the question is virtually a sigh, and expresses strong disbelief as to the chances of Ion's ever seeing his mother; 28 in (41) the disbelief voiced by Pylades amounts to his rejecting the course of action involved: 'what shall I say?' = 'I will have nothing to say'; note the presence of oiiK l.i1rw dvalvoµaL + participle dva4'alvw dvtxoµaL d1r6 d,r~a(V(I) d 1rpo4>ciaLaTos dTcip at break-off alrr6s and laVToO referring to the same person

fn. 246 20,102 96

24 33 122 178-9 143 18 153 42 81 98 fn. 52 118 127 122 96

22 l30V~1J with dubitative subj. 177 129 j3poTEllllS -yap after a vocative 28 'YL -yvwaicw + proleptic 106 acc.+ rel. clause 141 'Y>.cicf,w st + imperative 122 st in questions 102 st opposing sb'ongly adversative sentences 127 8lKTJV Aaµj3cill(I) 154, fn. 310 92 Slxa fn. 194 8opucf>oplii 78 8o01ros 24,91,94 8p61J111 d seemingly like t1rd fn. 82 d + past tense, with opt. BalCXEVS

212

+ clv in main clause d ical d86µTJV dµl+ participle d~aETaL dTE ... E[TE tic with concrete entity tic8l81alµL lic4'VEL n,Lf V.ictw lAKlll lµE>J..ov + infin. tv instrumental tv Tf>.EL lvEaTL tv lfEaTL t~-yfoµaL t1rd ts 'in the matter of, with regard to' foTci8T,V laTL veridical use of lTEICE .- lTLICTf Ev8alµwv t4'' ISrrEp lifiuaav lxw + infinitive -fl, corrective, in questions -fl introducing a yes/no-question ~ ~>Jlov with a bare accusative ~v, omission of TJlipTJTQL l!t>.ELS with dubitative subj. &fipaµa

158 fn. 145 85 73 98 38 23 49 3 146 fn. 206 92 fn. 317 27 107 54 104 30 10 128 75 101 4 118 71 3 61 110,179 83 177, App. 4 fn. 16 151 68 36 177 108

8LaaE6111 16 135 8~UKEL 13 8odC111 86oµaL fn. 196 100 fo8L Ka( appositional 119, fn. 178 74, fn. 172 Kal explanatory Kal with participles fn. 145 101 Kal µ~v Ka(111 95 KaTayE>.cilll 46 KflV'll'ToµaL 92 KpV'll'Tlll 121 92 KflV'll'Tlll 8eµas141 KVpT6s-µa, nouns in 84 µa(voµaL 59 µciKap 118 µE8(EµaL 70 µETlpxoµaL 56 µllllll 161 µlv, emphasizing 55 µ~ + future infinitive after 6µvvµL 173 167-8 µ~+subj. µ~ in relative clause 78 µ~ followed by a present or aorist indicative App. 3 fn. 15 147 µ6axos23 vdpO,.,f 118 6)./3LOS' ~

62

OOIIIS'(µ~) 60 ISs- not an interrogative pronoun 105 6aL611l 18 ob + fut. ind., followed byµ~+ fut. indic. 170 ob in relative clause fn. 184 55, 104, 167 obµ~ ob µ~ in ind. discourse + fut. inf. or fut. opt App. 3 fn. 17 obBe = dll' ob 96 OfiKOVV

oliTrOTE + aor. indic. olin ... oliTE after a

31, 122 88

negatived clause f>lj,LS' ,rapd with dative of persons

,rapaKO'll'OS' mipELµL lv ,rd8ov 'll'EpLf3dll111 'll'EVKTI

,rlMIS' m8o0 ,roLloµaL, periphrastic constructions with 'll'OTE 'll'OTE in questions ,rpl,r111 ,rp6cf,aaLv aEf3-

al81'1POS' UK6M>lj, aKi,µvosaocf,(a aocf,6v, T6

aTdxmµ' dv awcf,p111v

Ta 8£Cllv Ta VOµLa9lVTa Tdµci TE TE, appositional TE, epic

Tl(s-) TLS' + dubit. subj. n8rJµL 'cause, Create' Tl ylVlllµaL Tl 1rci811l T(KTEL T6 + comparative ToLci& with verbs of saying T6pVOS' cfial'llS' dv cf,8ovw + acc. -cf,oposcf,po08osxa>.ic6s-

76 151 55 11 71 53 145 23 76 52 69 2, fn. 252 119 119 40 85 95 123 147 110 129, fn. 236 78 126 113, 128 15 fn. 250 115 fn. 148 fn. 157 13 fn. 200 102 App. 4 fn. 19 App. 4 fn. 19

I 109 26 139 fn. 327 129 5 70 95

213

ws exclamatory ws + participle ws of subjective reason with finite verbs clv + subj.

ws

i:saTE COOTE

214

+ finite verb

29,58 86,91, 126

47 77 82 138

Index rerum action nouns in -µa fn. 46 accusative, appositional 13 (see also apposition) accusativein abrupt~ 119 accusative, internal 6, 13, 15, 40, 56

accusalivus cum participio

42 additive-concessive Kal with participles fn. 145 adjective.predicative use of 44, 61, 68, 69, 90, 99, fn. 134, fn. 135, fn. 158 adjectives.predicative use of -- with indefinite expressions 58, fn. 99 adjectives that consist of a substantival first member and a verbal second member 5, 17, 20 adverbial accusative 40 adverbs modifying nouns 72 'anteposed' constituent fn. 297 aorist, constative 68 aorist, gnomic 118 aorist, habitual, or iterative, use of the 88, fn. 203 aorist imperative 33 aorist of verbs of 162 emotions aorist participle, coincident or instrumental use of 10, 56, fn. 119 apposition 6,41 article followed by preposition phrase or adverb fn. 262 article, (non-)repetition

of -- with coordinated 116 nouns 34, 35, 87, asyndeton 177, 183 attracted and non-attracted predicative nouns 52 fn. 134 'atlribut' augment of verbs in initial EV133 augment, syllabic 97 augment, temporal 97 case variation 132 see accusative, cognate object internal comparative preceded by the article 112 dative, causal 154 dative,directive 80 153 dative, instrumental dative, locative, with verbs of movement 79 dative, modal 24, 68, 84, 91, dative of respect dative, sociative (or comitative)

dativus iudicantis

94

90 24 74,105

displacement of interrogative pronoun fn. 220 durative 136 focus, emphatic 102 front position 28 fronting 102 future ind. 1st ps. sing. or plur. in yes/noquestions 179 future indicative 2nd person with iussive value 169 future indic. questions introd. by apa and ~ 180

215

genitive, ablatival genitive,attributive genitive.objective genitive, possessivepertinentive imperative 3rd person imperfect, conative imperfect, inceptive use of imperfect of likelihood imperfect, iterative imperfect, unreal + 4v inceptive value of the present stem indicative in conditional clauses infinitive of purpose internal object

11 62 82

150 App. 3 fn. 1 144,156 136, fn. 278 156 155 157 78 37, fn. 321 12 see accusative, internal 125, 128

kaibelianus, versus left-displacement of an oblique constituent from relative clause 148 (see also prolepsis) locative clauses as complements of verbs of coming and bringing 67 messengel' speeches 89,136 fn. 149 metonymia 16, 18, 27, 46, middle 55, 56, 59, 69, 85 neuter plural with plural verb 93 oath, 3rd person App. 3 fn. 14 159 optative + clv past potential fn. 327 perfect 86 perfect, 'intensifying' 58 personal and demonstrative pronouns in the same sentence 35 plural of abstract nouns fn. 116 possessive pronominal adjectives 98

216

praesens annalisticum praesens tabulare

present, historic present imperative, inceptive use of

2 2 1, 39, 89, 137 18, 32, 52, 98, 122, fn. 125, fn. 129, App. 3 fn. 11

present imperative, continuative use of 52 present indicative with future reference 36 present of 'life-event' verbs 135 present participle referring to an anterior 150 action present, 'perfective' 2, 73 prolepsis 28, 38, 148, fn. 297 (see also leftdisplacement) prolepsis of the subject of a relative clause fn. 293 questions.disjunctive 175, App. 4 fn. 1 179 questions, echo questions, selfcorrective 179 175,183 questions, x- or word questions, x- , directed at 184 an addressee questions, x-, not directed 188 at an addressee questions, x- with the future indicative 186 questions, yes/no175 questions, yes/no- , directed at an 176 aoiressee questions, yes/no-, not directed at an aoiressee 178 relative clause, autonomous 65 relative clause with enclosed head noun 86

relative pronoun instead of demonstr. pron. 144 'rhetorical parataxis' 39 simile 108 singular having collective meaning 147 subjects, two (or more), with a singular verb 123 subjunctive + dv, semantic value of -in conditional clauses 12 subjunctive, dubitative, in yes/no-questions 176 subjunctive, dubitative, in x-questions 183 telesilleum fn. 154, fn. 157 terminative verbs fn. 137 thematic combinations fn. 296 thematic function 45 theme constiblents fn. 102 ropic 102 ropic, new fn. 62, fn. 102 verbal adj. in -Ttos 8, 121 verbal adj. in -TOS' 8 verbs denoting, in the present stem, an emotional or intellectual state 59 'whole and part' constructions 87

217