H. Jacobsohn, a pupil of Leo, in a Göttingen dissertation of 1904, ‘Quaestiones Plautinae’ declared that Plautus allowed Hiatus and Syllaba Anceps at two places in the line (these places are now called loci Jacobsohniani):—

(i) at the fourth arsis of the Iambic Senarius (and the corresponding part of the Trochaic Septenarius),

(ii) at the second arsis of a Trochaic Septenarius.

In this paper I am merely analysing the meaning of Phaedo 99–107 without discussing either its historical significance or its value as a contribution to the logic of science. The paper attempts in fact to be little more than a paraphrase of the Greek, which aims only at accuracy of statement, but I am adding in the third section a note on the relation of this passage to the discussion of method in Republic VI, mainly in order to minimize its relevance. The first section of the paper deals with the linguistic analysis of the text, the second, very sketchily, with the general doctrine.

The problem suggested by this passage cannot be properly appreciated unless it is shown first of all that the treatment of poetry and art in the Laws fundamentally agrees with, though of course in some respects it provides a welcome supplement to, the attitude set forth in the Republic and elsewhere by Plato. The demand that music and poetry should ‘imitate’ the good; and that this ‘imitation’ should have meaning and accuracy, and be free from mere emotionalism (655–9, 667 sqq.) directly recalls the doctrine of the Republic (e.g. 395, 401) and other dialogues on the two kinds of imitation, which I have explained in the Classical Quarterly XXII p. 16, XXVI p. 161; the reason for the demand is the same—the imitation of evil has the same effect on character as the companionship of evil men. As for the effect of this ruling on comedy, the Republic (396e where παιδιᾶς χάριν leaves a loop-hole, 606c) is perhaps not over-clear. But the Laws explains that the comic art is to some extent valuable; its pictures of the ludicrous serve as a warning, and also help to make goodness more intelligible by setting it in relief; hence comedy may be performed for the edification of citizens by slaves or hired aliens (816) under certain restrictions such as that a piece shall not have too long a run (England on 8i6e 9) and that no ridicule of free men is to be permitted (935–6). On tragedy the two dialogues are in full agreement; it must become once more a hymn (Rep. 607a), using for a good purpose those forms of dramatic representation which are allowed to the poet in Rep. 396, and avoiding the mournful and complaining language (and music) of profane tragedy, which merely harrows the feelings and which cannot be used without blasphemy at a festival of the gods (Rep. 383bc, 605–6, Laws 800, 817).

The mistranslation by Mr. J. D. Duff of nox ubi sidera condit as ‘where night hides the stars’ is also the interpretation of many commentators from Sulpitius in the last decade of the fifteenth century to Lejay in the last decade of the nineteenth. Lucan is clearly speaking of East and West in 15, of South in 16, and of North in 17–18. How can night be said to hide the stars in the West? Burman saw the difficulty and expressed himself thus: ‘…dubito, an recte dicatur, nox condere sidera, id est, Stellas, quae sole cadente prodeunt, et se spectanda praebent, obscurare et occulere: neque nunc occurrit alius ex veteribus locus, unde ita locutos fuisse Poetas appareat. Nox enim adveniens prodit sidera, praecipitans uero, aurora adveniente, potest recte dici condere, et quasi auferre ex oculis hominum sidera.’ Burman then was tempted to understand sidera as the sun, but could not parallel this use of the plural, although he admitted the use of sidera solis. He cited Ouid. Met. 14, 172–3 caelumque et sidera solis / respicio, as did Haskins, who took the same view with hesitation. Ezra de Clerq van Jever in his Specimen Selectarum Observationum, which he published at Leiden in 1772, definitely understood sidera as the sun, though he could parallel only sidus in the singular from Ouid. A.A. 1, 723–4 aequoris unda / debet et a radiis sideris esse niger. But, it may be said, these Ovidian passages are such that no ambiguity is possible, and are not quite relevant to Lucan's phrase.

In the tragic senarius the divisions of the sense normally coincide with the main divisions of the metrical structure. Punctuation is most frequently found at the end of the line, or at the penthemimeral or hephthemimeral caesura. There are few traces of any desire to produce a persistent clash (as in the Vergilian hexameter) between verse structure and sentence structure. Thus at Med. 446–50 and 709–13 five consecutive lines, at Med. 364–71 eight consecutive lines, are more or less self-contained in sense. But this principle, while valid in some degree for all three tragedians, is not equally valid for all. Sophocles, as is well known, is far suppler in his iambic technique than Aeschylus and Euripides, and far more prone to write sentences which over-run the main metrical divisions.

The remarks I wish to make upon the Thucydides scholia in Bâle University MS. no. E III 4 will be clearer if prefaced by an outline history of the scholia to Thucydides in general.

The editio princeps (Aldine, May 1502) contained no scholia. They made their first appearance in the following year as an appendix to the Aldine Gemistius and Herodian. The same scholia were then reprinted in the Juntine edition (Florence 1526) as a frame to the text. In the Hervagian edition of Camerarius (Bâle 1540) they were relegated to the end, and Camerarius claims in his prefatory epistle to have expended some trouble in emending them. Henceforward the same corpus of scholia, unchanged save for occasional emendation, passes on from edition to edition, either framing the text, as in Stephanus ([Geneva] 1564), or beneath it, as in Stephanus ([Geneva] 1588), Hudson (Oxford 1696), Duker (Amsterdam 1731), Gottleber-Bauer-Beck (Leipzig 1790–1804).

The metre of Olympian II is still a matter of some difficulty. It has commonly been recognized as differing from Pindar's other metres, but many opinions have been held of its character. An understanding of it is, however, not merely essential to any general theory of Pindar's metric but vital to the textual criticism of the poem. Without some coherent theory we cannot say where ‘Responsionsfreiheiten’ are allowed and some important cruces remain unsolved. In recent years three theories have been put forward. In his Res Metrica pp. 153–5 W. R. Hardie put forward the view that it is an example of a μέτρον μονοεɩδές composed in Cretic verse. He holds that it is written in Cretics or their recognized metrical equivalent, Paeons. This kind of verse is known from the Delphic Paean to Apollo and the Paean of Limenius, and so far as the equivalence of Cretics and Paeons is concerned, there is no difficulty in it. But as Hardie himself admitted, his theory does not account for all the facts. It may be said to suffer from the following defects. n In Str. 1, 3, 5, 7 and Ep. 1, 4, 5 he has to assume the existence of an extra single syllable at the beginning of the line. Since all attempts to defend ‘Anacrusis’ are extremely doubtful, this can hardly be maintained here. 2. The clausula of the Strophe cannot be reduced to Cretics or Paeons. 3. He has to admit that in Ep. 2 and 3 there is ‘some Trochaic movement’ and presumably he would have to admit that in Str. 3 there is either an Iambic, or, as he would prefer, a Trochaic movement with Anacrusis in ἤτοɩ πɩσα μεν. There is certainly some truth in Hardie's theory, but since it does not consider all the facts or co-ordinate them into a whole, it is not really satisfactory.

Sidgwick describes this strophe as ‘locus corruptus, coniecturis nondum sanatus.’ Mazon, who prints απεύχεται in 1. 625, leaving the rest as it stands, says ‘texte douteux.’ Of the three principal attempts to amend or otherwise interpret it, that of Hermann (ed. 1852) is too radical and far-fetched, requiring an excessive parenthesis. That of Headlam (C.R. 1900, p. 196 f.) involves the strange theory that the chorus suddenly divides itself into two, and that one half indulges in a sort of disorderly interruption of the other. That of Wilamowitz (ed. maior, 1914), while leaving the worst textual difficulty to stand, supposes that the chorus accuses itself of weakness and subservience to Clytemnestra, which, in view of Wilamowitz's own convincing interpretation of the Parodos, is highly improbable.

In R. C. Flickinger's ‘Greek Theater and its Drama’ (3rd ed. second impression 1929) we read on pp. 171–2, ‘In the iambic trimeters written by Aeschylus a trisyllabic substitution (tribrach, anapaest or dactyl) for the pure disyllabic iambus occurs only once in about twenty-five verses.’ A similar statement is found in other earlier writers, e.g. Haigh (‘Tragic Drama of the Greeks’ 1896 p. 372). The authors of the error appear to have been Rossbach and Westphal, who are quoted with approval by J. Rumpel (Philologus XXIV (1866) p. 407), and subsequently repudiated by him (Phil. XXV (1867) p. 54). When Rumpel found that he had been misled, he counted for himself the trimeters and the trisyllabic feet in Aeschylus, but his results are of dubious value owing to the fact that he counts as ‘trimeters’ lines like;

δαίμονά τε τὰν ἂμαχον ἀπόλεμον ἀνίερον (Ag. 769) and

αὐτοϕόνα δίμορα τέλεα τάδε πάθη. τί ϕῶ; (Sept. 850).

The name of Democritus can claim a place in any discussion of indivisibles. Yet its introduction in this paper seems to depend on the lucus a non lucendo principle; for Democritus did not believe in the existence of indivisible lines. Nowhere is the belief ascribed to him and in at least one place it is implicitly denied, the scholion on De Caelo 268a 1, which says he made his elements indivisible solids, as contrasted with lines or surfaces. Two passages, one from Plutarch, the other from Simplicius, will show why he could not believe in indivisible lines.