¹ Cf. e.g. Dale on E. Ale. 232, Denniston-Page, Agamemnon, pp. 239 ff. I mean by sense-pause (or ‘pause’) a positive and significant break in the sense, not merely ‘metrical’ pause, which means the absence of enjambement, elision or the close syntactical bind given by postpositives and prepositives.Google Scholar

² My use of the terms colon, verse, and period, which departs from the normal contentions, is explained on p. 37 below.

³ e.g. by Kranz, W., Stosimon (1933), 22Google Scholar, 118 ff., 151 ff., 177 ff., 230; Kraus, W., Strophengestaltung in der griechischen Tragödie (SB Wien, 231.4, 1959), §22.Google Scholar

⁴ Read to the Oxford Philological Society in June 1965. 1 am indebted to the late Professor A.M. Dale, to Miss L.P.E. Parker and Dr. D.C. Innes, and to Professors K.J. Dover and H. Lloyd-Jones, for valuable criticism and advice; to Dr. R.P. Martineau for help on statistical method; and above all to the generous and constant help of the late Eduard Fraenkel, my debt to whom throughout will be evident.

⁵ Jespersen, O., Phonetische Grundfragen (1904), §122.Google Scholar

⁶ In the closing bars of The Rake's Progress, Act II, Sc. i, Stravinsky actually uses commas in the score to mark the phrasing. Professor Dover suggests that ‘phrasing’ is perhaps a better musical analogy than ‘rest’.

⁷ Ripman, Walter, English Phonetics (1933), pp. 162–3.Google Scholar

⁸ Jones, Daniel, An Outline of English Phonetics (1936), §1005.Google Scholar

⁹ Modern punctuation is so standardized as to be almost useless as a guide to the finer points of delivery. But before the nineteenth century, though sometimes arbitrary, it may be as important as an actual stage direction (cf. n. 82 below).

¹⁰ Greek Word Order (1960), p. 19.Google Scholar

¹¹ Fraenkel, E., ‘Kolon und Satz’, NGC (1932), 197–213Google Scholar, ibid. (1933), 319–54 = Kl. Beitr. i (1964), 73–130Google Scholar; ‘Nachträge zu “Kolon und Satz”, II’, Kl. Beitr. i. 131–9Google Scholar; ‘Noch einmal Kolon und Satz’, SB München, 2, 1965Google Scholar. Further material of the same kind may be found in Leseproben aus Reden Ciceros und Catos (1968), esp. pp. 201–3, 208–12.Google Scholar

¹² See änder, , Nicanoris reliquiae emendatiores (1877), ch. i.Google Scholar

¹³ See n.ll above.

¹⁴ Zur Wortstellung bei Pindar (1959).Google Scholar

¹⁵ Kl. Beitr. i. 78 ff.Google Scholar

¹⁶ Kl. Beitr. i. 78Google Scholar; cf. Agamemnon, p. 512.Google Scholar

¹⁷ Op. cit., p. 20. Professor Dover tells me that he made this decision partly on the analogy of other languages, partly for convenience. He agrees that context could affect the analysis. (Cf. also Ag. 1028 f.)

¹⁸ I owe this example to Mr. W.S. Barrett e.g.

¹⁹ e.g. ‘I hit him, in the eye’ answers the question ‘what did you do?’; ‘I hit him in the eye’ answers the question ‘where did you hit him?’; see p. 28 above.

²⁰ ‘The enigmatic style in early Greek poetry’, delivered in 1961 (unpublished).

²¹ The second ornamental epithet qualifying the noun is however typical of apposition. Cf. Il. 20.70 f.

²² S. Lauer, op. cit. in n.14 above; Jespersen, , Analytical Syntax (1937), ch. 12. Lauer's use of the term is however slightly different from Jespersen's.Google Scholar

²³ I owe this example to Mr. W.S. Barrett.

²⁴ The phrase is borrowed from Fraenkel's term ‘Erweiterung’ (Kl. Beitr. i. 77 f., 82, 91, 120, 122), but has a slightly different sense. Fraenkel means by the term a further elaboration in epexegesis (which I have included under ‘series’, e.g. Ag. 377–8 ) before which there is pause: I mean by it an (adverbial) phrase having enough rhetorical weight to stand on its own.Google Scholar

²⁵ Dornseiff, F., Pindars Stil (1921), p. 86: ‘Die Sprache der Chorlyrik ist mehr dem einzelnen Wort als dem Satz zugewandt’, usw.Google Scholar

²⁶ Cf. Fraenkel, , Kl. Beitr. i. 98 ff., 137.Google Scholar

²⁷ Ibid., 103 ff.

²⁸ Ibid. 104.

²⁹ For metrical pause and sense-pause see n.l above.

³⁰ By the former I mean the type of problem acutely discussed by Parker, L.P.E. in BICS 5 (1958), 13–24Google Scholar; by the latter, the kind of vague and indecisive structural considerations often adduced by Kraus, , op. cit., (n.3) e.g. p. 83 on Ag. 223; cf. pp. 41 f. below.Google Scholar

³¹ In my use of the terms ‘minor period’ and ‘major period’ I follow Dale (LMGD 11Google Scholar ff., 195 ff.). My use of ‘period’ also corresponds more or less with that of Irigoin, ‘Colon, vers et période’, (1967), 65 ff. But since I have no use for his distinction between colon and verse as separate functional categories (on which see Parker, L.P.E. in Lustrum 15 (1970), 52–3Google Scholar) –cola being in ‘verbal synaphea’ (i.e. having elision or word-overlap), verses in ‘prosodic synaphea’ (i.e. with metrical pause but no proven period-end)–I need only one term for the smaller divisions, and as ‘colon’ is disqualified, ‘verse’ seems the obvious candidate. (Dale in LMGD uses ‘colon’ but not ‘verse’ as a functional category.) For places in the metrical scheme with neither synaphea nor proven period-end, I use Maas's term ‘contact’ (GM, §66); a contact is a potential period-end, not a separate category like Irigoin's ‘vers’.

³² This is the criterion by which A. Boeckh first determined the verses in Pindar (‘minor periods’ in my terminology), and it was presently extended to lyric verse in general by Hermann, G. in his Elementa Doctrinae Metricae (1816), pp. 715–16Google Scholar (III.xviii.7). Despite Koster, W.J., Mnemos. 3 (1950), 30–2, this guide is not infallible, but it is sound enough for my purpose. For its real and supposed limitations, see below, n.33 and pp. 41, 52.CrossRefGoogle Scholar

³² a The traditional term syllaba anceps is however a misnomer: see Rossi, L.E., ‘Anceps, vocale, sillaba, elemento’, RFIC 91 (1963), 52–71.Google Scholar

³³ Kraus, , op. cit., p. 146Google Scholar n.2 (hiatus); Pohlsander, , Metrical Studies in the Lyrics of Sophocles (1964), p. 157Google Scholar (hiatus), p. 162 (brevis), where he lists some supposed cases. (Cf. Snell, , Pindar ³ ii. 173Google Scholar). See below, p. 41.

³⁴ GM, $35; LMGD 94, 179; Theiler, W., Mus. Helv. 12 (1955), 183Google Scholar f.; Pohlsander, , 163.Google Scholar

³⁵ Possibly should be read, but the example will serve.

³⁶ As observed by Hermann, G., op. cit. (n.32 above), pp. 378–9Google Scholar. This is true only of marching anapaests, not of lyric anapp., where the paroemiac is a verse in its own right, e.g. S.El. 193, 233–5.Google Scholar

³⁷ L.P.E. Parker has now given a more precise definition to the concept of ‘catalexis’, which makes it usable even in metres which are not analysable she also shows that catalexis is nearly always attended by metrical pause, and that cases like Ag. 197 are very rare (CQ N.S. 26 (1976), 14–28, esp. 20 n.17). Cf. the postscript of this paper, pp. 64–6.Google Scholar

³⁸ The reading is immaterial to the structure.

³⁹ For anaphora at the beginning of a period, cf. A., Pers. 950–1, Sept. 166–7Google Scholar; S. Ant. 791Google Scholar; E. Ale. 460, Suppl. 73–4.Google Scholar

⁴⁰ But not if it is followed by double short, or by a long which cannot be anceps. Thus in ionics elision and word-overlap is not uncommon (cf. also A. P. V. 183Google Scholar, S. O.C. 135Google Scholar). There are however certain metres, such as repeated bacchii, where is regularly followed by short; see BICS 20 (1975), 84. (N.B. In applying this rule, I treat the first element of an aeolic base as long unless there is a short syllable there in one or both places. If the base is resolved, I treat the first element as biceps wherever possible.)Google Scholar

⁴¹ G V 451 n.2: ‘Elision ist hier undenkbar.'Google Scholar

⁴² See BICS 22 (1975), 88–95, esp. 93–4.Google Scholar

⁴³ Cf. GM, §65. In practice I have admitted periods of eight metra + --, since this is very common. By the count of ctus suggested by Dain, A., Traité de métrique grecque (1965), pp 160 ff., this is equivalent to seventeen or eighteen ictus. (Dain of course allows longer periods than this, since his ‘period’ is my ‘major period'.)Google Scholar

⁴⁴ Annali delta Scuola Normale Superiore di Pisa, Serie II, 23 (1954), 275Google Scholar ff. (= Kl. Beitr. i. 396 ff.). The argument from word-division is persuasive, but not in itself sufficient.Google Scholar

⁴⁵ Kraus, , op. cit.Google Scholar, $10–25; Pohlsander, , op. cit., ch. x.Google Scholar

⁴⁶ Cf. n.33 above. For the effect of the adjectives on the preposition see p. 60 below, with n.76. 4, See JHS 96 (1976).Google Scholar

⁴⁸ LMGD 175 f.Google Scholar

⁴⁹ WS 77 (1964)Google Scholar, 19–20 = Collected Papers (1969), p. 190.Google Scholar

⁵⁰ LMGD 134, 145Google Scholar f.

⁵¹ LMGD 145 f. 51 GM, §59.Google Scholar

⁵² Cf. Bacchylides, 18Google Scholar, on which see Snell⁷ p. 35, GV 263. Wilamowitz remarks that the practice of dovetailing exhibits the conflict of two tendencies, one towards the simpler asynartete structures (e.g. of Archilochus), the other towards more complex structures which desiderate caesura.Google Scholar

⁵⁴ Cf. Held. 914–16 = 923–5, where Murray's division (three e oplians) gives an impossible word-break at 924–5: where we can divide tel. + hipp. Again, to divide Hec. 910–12 = 919–21 as three enoplians (as no modern editor in fact does) would give an impossible word-division in 911–12 = 920–1. (Divide tel. + glyc, hipp.; hipp.)

⁵⁵ I shall treat this matter elsewhere. The low incidence of sense-pause in these refrains may have some connection with the intermediate category of ‘verse’ postulated by Irigoin (cf. n.31 above). It is worth notini that the refrain-pattern has much in common with the stanza-patterns of monodic lyric, in which pause seems to be treated in rather the same way.

⁵⁶ Hermes 92 (1964), 23–50.Google Scholar

⁵⁷ Cf. p. 37 above. I do not however regard a simple vocative as an exclamation, as he does, op. cit. 43 (iii) (a): e.g. Sep. 150

⁵⁸ Change of metre to dochmiae, on the other hand, does seem to be a significant factor; see above, and on Table XII.

⁵⁵ At Ale. 120 an indubitable dochmaic, is followed by (Dale's analysis as choriambic dodrans with resolution can be ruled out in early Euripides.) This is a regular clausula to dochmiacs, and is clearly related in rhythm (cf. LMGD 104Google Scholar; Parker, L.P.E., op. cit. (p. 39 n. 37). It is unreasonable to insist that there Is a change of metre here. At A. Suppl. 843 is followed by then by The intermediate verse links the dochmiae with the hemiepes: it is itself either, or both (cf. LMGD, loc. cit). Conomis himself allows as a form of dochmiae at S. Aj.). 881, 884, and notes that Sept. 891, 892 is shown to be dochmiae by the antistrophe at 904, 905 (op. cit. 35).Google Scholar

⁶⁰ A. Cho. 958Google Scholar (his suggestion is perhaps a slip for read by Page; but the pronoun impairs the universality of the prescription); Eam. 783Google Scholar; Andr, E.. 833Google Scholar (the following is undoubtedly a dochmiae, whatever the correct text at 838: fort. as I shall argue elsewhere). He presumably emends in A. Suppl. 854Google Scholar, and in Ag. 1133; rightly, since the one is required by the sense, and the other is hardly a change. He need not however have emended. Eum. 840Google Scholar since pause is given by the phrase in conjunction (series). (But no reasonable restoration of Ag. 1090–1 will give pause at despite Murray's comma.) E. Ba. 1002 is undoubtedly corrupt; I am less convinced that the brevis is due to corruption in Pho. 177.Google Scholar

⁶¹ A. Suppl. 649Google Scholar (he presumably accepts Weil's conjecture); E. H.F. 1060 (the exclamatory force of which he rightly invokes to explain the internal hiatus, does not extend to the terminal hiatus). He should not treat S. Trach. 1024 as lengthening before mute and liquid (op. cit. 39), since besides the rarity of this feature, which he himself stresses, it implies synaphea, which is here ruled out by the strophe 1004, given any reasonable arrangement and text (cf. Barrett, , Hippolytus, p. 407Google Scholar). If this is a dochmiac, as he asserts, there is brevis without pause in both places. But it could be a type of prosodiac (see JHS 96 (1976), 144 n. 87).Google Scholar

⁶² The one certain exception is O.T. 1350 where at least is sound and correption is ruled out by the strophe. In Table XII count hypodochmiacs as dochmiacs, which Conomis does not.

⁶³ A. Suppl. 649, Cho. 958, Eum. 783, 840; E. Andr. 833, H.F. 1060.

⁶⁴ Suppl, 649, Cho. 958, Eum. 783; Sept. 109, Eum. 149.

⁶⁵ See his apparatus at Sept. 109 (pause at 128 is given by the vocative).

⁶⁶ See his apparatus at Eum. 149 (cf. at ibid. 783). (Piatt's conjecture at Eam. 149, which he recommends, is refuted by the evidence he himself gives.)

⁶⁷ See pp. 70–2 of this issue.

⁶⁸ GA, §49. 68a Zielinski, T., Tragodumenon Libri Tres (1925), pp. 140–1Google Scholar; Ceadel., E.B.CQ 35 (1941), 70Google Scholar. Cf. Dale, A.M., Euripides, Helen (1967), pp. xxiv–xxviii.Google Scholar

⁶⁹ ‘Falling suppliant about his robes with allher heart’, cf. Lloyd-Jones, , CR 66 (1952), 135Google Scholar, and see also CQ N.S. 25 (1975), 11–12.Google Scholar

⁷⁰ See JHS 96 (1976), 121–2.CrossRefGoogle Scholar

⁷¹ This gives see Wilamowitz's note ad loc.

⁷² This would imply in 643; see Euripides and the Judgement of Paris, JHS Suppl. 11 (1965), 74.Google Scholar

⁷³ GM, §135.

⁷⁴ It has been suggested that the breaks n such long-winded ionic stanzas should be reated in a special way, and termed ‘hold’ oather than ‘period-end’; cf. LMGD 124Google Scholar; Collard, Z., Euripides'Supplices, p. 117.Google Scholar

⁷⁵ Examples of pre-and postpositive inks across period-end in Pindar are as follows: O. 6. 17 (‘not a copula’, GV 305 n); O. 6. 53 (? leg. ); O.9. 65 O. 10. 18 (comparative); 0.14. 5 P.9. 99 ) 1. 3. 18 1.8. 23 (cf. GV loc. ch.); Paean 2.25 Cf. B.5. 74 For possible cases in tragedy see JHS 96 (1976), 126 f.Google Scholar

⁷⁶ GM, §135.

⁷⁷ ‘Kolon u. Satz, I’, 206 = Kl. Beitr. i. 83.Google Scholar

⁷⁸ I am speaking of places where there is no clear ground for pause, such as would be given by e.g. apposition or selective contrast.

⁷⁹ Possibly emphatic is at 0.1. 81: it is odd behaviour for a man to kill the suitors and put off the marriage of his own daughter. Cf. A. Ag. 224 where the emphasis is clear (for the colometry, see BICS 22 (1975), 86). O.2. 15 is also emphatic, but introduces a new, limiting idea, which leads by way of contrast to the veiled allusion that follows: ‘ … for future generations; as to the past …’ A similar limiting afterthought in a new stanza leading to contrast is 0.1. 99 But at P. 11. 22 is probably in extra-position, since Clytemnestra has just been mentioned; and in 0.10. 55 is in apposition to Neither emphatic nor syntactically independent, however, is 0.9. 29Google Scholar

⁸⁰ Cf. S. El. 693–4.

⁸¹ Cf. or M Gr.

⁸² Cf. Partridge, A.C., Orthography in Shakespearean and Elizabethan drama (1964)Google Scholar, chs. 14–15: Simpson, Percy, Shakespearean punctuation (1911), p. 16 al.Google Scholar