It is clear, then, that Plato's strictures on Homer ought not to have given any encouragement to allegorical interpretation. The eulogists of Homer ought to have sought other grounds for the defence which he invited them to make; while the allegorizing philosophers, if they persisted in treating interpretation of the poets as an instrument of knowledge, ought to have answered Plato not by multiplying allegories but by producing a defence of the allegorical method. The question with which we are concerned is, of course, not what ought to have been done, but what actually happened. But the foregoing considerations ought to warn us against assuming without evidence that all those who allegorized after Plato did so in order to answer his attack on the poets, and were directly incited by that attack to engage in allegorical interpretation. The fact that the early Stoics were prolific in such interpretations is in itself no proof that they intended in this way to answer Plato's criticism of the myths. Other evidence is required to prove that it was the desire to answer Plato which impelled them to wallow in allegory.

Priscian tells us in his dedicatory introduction that he took his material from many Latin sources—collectis etiam omnibus fere quaecunque necessaria nostrorum quoque inueniuntur artium commentariis grammaticorum. This can hardly mean that he owed everything to his predecessors. At any rate it is unlikely that he copied all his illustrative quotations from earlier grammarians. The problem is one which, for our purpose, does not need to be solved. We can make Priscian responsible for every quotation (unless, of course, we suspect that copyists have had a hand in introducing variations), because he had the opportunity of correcting or commenting on such borrowings as he incorporated in his work. If he could acquiesce in a garbled version of a line, we must take the fact into account in appraising his value as a witness to the text of Terence. And we must assume that the Terence text of his day did not differ from the versions which he quotes. Nowhere does he hint at a discrepancy. This assumption, it will be understood, has reference only to the points of diction for which we are tolerably sure that we have Priscian's attestation, and not to minor details which did not interest him at the moment, and in which we can prove that he was far from conscientious.

No one has as yet done for Italy what Rohde's Psyche did for Greece, and the reason is not far to seek. Rohde had at his disposal a large amount of literary material, of which no one could doubt that it represented Greek feeling and practice of various ages; but the investigator of the corresponding Italian field is met with a twofold difficulty. He must in the first place discard a great deal of the written records, because they clearly reflect, not native ideas, but those which were common to the ancient world in Hellenistic times. He is then left with a very few scanty documents, which refer, not to Italy as a whole, but practically without exception to Rome, or, at any rate, to districts strongly Romanized. To add to his perplexities, he has a great mass of archaeological evidence for all ages, from the palaeolithic onwards, which, as it is unwritten, he must interpret as best he may, bearing in mind the very wide margin of error in all such attempts, and reconcile, if he can, with what the classical authors tell him.

The discussion of the construction with usus est is conventionally bound up with that of opus est, and comparatively little attention has been given to the problem presented by the use of the ablative case with the former phrase.

The tendency to pass lightly over this matter is due doubtless to the manifest etymological connexion between usus and utor, which leads to the assumption that the verbal noun would naturally follow the syntax of the verb.

In English, down to the 19th century, the beast which carried Balaam was generally and almost universally, both in speech and in writing, denominated the ass. It is so no longer: the name ass, except in metaphor as a term of contempt or insult, has disappeared from conversation and from most kinds of print, and survives only in serious poetry and in prose of some solemnity. The name donkey, first printed in 1785 in a dictionary of slang, has usurped its place.

The whereabouts of this ancient town is of interest in the Leukas-Ithaka controversy, but unfortunately there is little information on which to fix it. The data may be said to be one passage in the Odyssey and one in Thucydides. In the former, ω 377–8, Laertes is wishing he were young again,

οἷος Nρικον εἷλον υκτμενον πτολεΘρον, κτν ἠπεροιο, Kεϕαλλνεσσιν νσσων

Anaxagoras’ theory of matter offers a problem which, in bald outline, may be stated as follows. The theory rests on two propositions which seem flatly to contradict one another. One is the principle of Homoeomereity: A natural substance such as a piece of gold, consists solely of parts which are like the whole and like one another—every one of them gold and nothing else. The other is: ‘There is a portion of everything in everything’, understood to mean that a piece of gold (or any other substance), so far from containing nothing but gold, contains portions of every other substance in the world. Unless Anaxagoras was extremely muddleheaded, he cannot have propounded a theory which simply consists of this contradiction. One or the other proposition must be reinterpreted so as to bring them into harmony. Some critics attack one, some the other; some try to modify both. Mr. C. Bailey has recently published a fresh attempt to reconstruct the theory in a form consistent with itself. But the result, as he admits, is arbitrary and (above all) uneconomical. These defects, like all the contradictions and obscurities found in other accounts of the system, can be traced to the second proposition: ‘There is a portion of everything in everything’, or rather to the construction put upon it. The language is crude, vague, and ambiguous. What does ‘everything’ mean? Has the first ‘everything’ the same sense as the second ? If taken to mean that there is a portion of every material substance in every material substance, the proposition leads to a result for which ‘uneconomical’ is an indulgent epithet.

Ennead VI. 4. 1 (Volkmann's edition, p. 363, 1. 7): καὶ ὅμως ἐν μὲν ταῖς ποιότησι τὸ αὐτὸ μεμερισμένον θεωρεῖται, ἐπί δὲ ψνχῆς τὸ αὐτὸ οὐ μεμερισμένον κ.τ.λ. As there is no antithesis to justify ὅμως (praeterea, Ficinus) we should read ὅλως, which neatly sums up the argument: ‘in a word, while in qualities the same thing is seen as divided, in soul the same thing is not divided,’ etc.

Some years ago, when discussing Theocr. 22. 177 sqq. (C.Q. XIII. p. 22), I suggested that Theocritus had been a little careless in envisaging the circumstances which he is describing, and had written as though a duel normally resulted in the deaths of both combatants. That still seems to me the probable explanation of the difficulty with which I was dealing, and, as I then said, the oversight with which I charged Theocritus is venial enough, for in fact two deaths result from the particular duel which he is about to describe. He is concerned with the drama, not with the setting, and if the setting proves to be of cardboard, the play is after all the thing.

There are two ways in which the text as it stands has been construed: (1) Cypris came sweet and smiling (i.e. taking ἁδ. καì ἁ K. as=ἁ K. ἁδ. καì); (2) sweet Cypris also came smiling (i.e. taking ἁδ καì K. as=καì ἁ ἁδ. K.). Of these (1) is at least grammatical. The order of words, too, can be paralleled (Hiller cites 10. 34; 11. 58; 15. 79; 17. 93; 22. 25 and 68). But what I think cannot be paralleled is this curious conjunction of adj. and partic. (2) is frankly ungrammatical. Yet Legrand (Étude sur Théocrite, p. 309) adopts it, adding: ‘Il faut reconnaitre une construction incorrecte au moins dans ces quatre passages: 1. 95…; 1. 109 ὠραîος ξὤδωνις (=καì ό ὡραîος Aδωνις); 4. 49 ῥοικον τὸ λαγωβΌλον (=τò ῥοικον); 29. 33 τàν γένυν ἀνδρειαν (= τὰν ἀ. γένυν). Of the last three we can with confidence say that ὡραîος (if the line be genuine) and ἀνδρειαν are to be taken as predicative; in 4. 49 Legrand himself now (in his Budé edition) accepts the emendation τι for τΌ, as he does in the similar 15. 145 passage. There is not, in fact, sufficient, or indeed any, evidence to convict Theocritus of this solecism. Emendation has been rife, but unhappy. ‘Nescio an ipse rem acu tetigerim, legens ᾖνӨέ γε μὰν, ἀνíα δὲ, καì ἁ K. γ.,’ writes Briggs. We think he has not.

Since the time of Brunck there has been a more or less general acquiescence in his substitution of πνθώμεθα for πνθομεθα, inasmuch as there is no obvious reason to be alleged in support of the optative. Campbell, it is true, found the optative more in accord with the feeling of the blind and weary Oedipus; but who will listen to this nowadays? Therefore it is the more surprising that Radermacher should retain the optative as expressing the eager wish to attain the goal, basing himself on J. M. Stahl's Syntax of the Greek Verb, p. 482. For Stahl, who admits the existence of certain exceptional instances to be justified in this sense, expressly excludes from the category this passage as well as El. 57 and 760, and seeks to limit the scope of his pronouncement to the writings of Aristophanes, Xenophon and the Orators—a curious combination.

The earlier part of this paper yielded the result that the assertion ‘A portion of everything in everything’ has no place or function in the explanation of any sort of apparent ‘becoming’ or change. This conclusion is important because, ever since Aristotle, it has been assumed that the assertion was made in order to explain away becoming and change. But if (as I have shown), according to the best evidence, becoming and such sorts of change as Anaxagoras considered can be explained away without using the assertion at all, it follows that the motive for it must be sought, not in the processes of apparent becoming and change, but in the permanent constitution of the ungenerated and unchanging things which figure as elements in the system. A visible mass of gold or of flesh is formed by the aggregation of smaller particles of gold or of flesh, and it is infinitely divisible into parts, every one of which is gold or flesh. But what is the constitution of a piece of gold of whatever size, and how does it differ from the constitution of a piece of flesh? The proposition we have so far successfully excluded comes in here as part of the answer to this question.

One of the most recent writers on the early history of Rome has shown that the framework of the traditional story is perhaps to be trusted, even though there are many details, inconsistent and self-contradictory, which are obviously to be rejected. In view of this fact, it might be worth while to reconsider the Coriolanus story, the prevailing opinion concerning which is that vouchsafed by Mommsen many years ago: ‘die Erzählung ist ein spät, in die Annalen eingefügtes, darum in alien Stücken denselben ungleichartiges und widersprechendes Einschiebsel.’ The reasons for arriving at such an opinion are sufficiently obvious to warrant their receiving but the barest recital. First, it is incredible that the Volsci would either choose a renegade Roman to be their general, or, even if they did, allow him at the last minute to rob them of the fruits of victory. Secondly, inconsistencies in the version of the story which we possess induce us to suspect its historicity; for example, Dionysius of Syracuse is made to send corn to the starving Romans'—yet Dionysius lived some hundred years later; a youthful Coriolanus is represented as having considerable influence in the senate—yet in those early days the senate was essentially a gathering of venerable men; the Roman populace learns immediately the gist of Coriolanus' remarks in the senate—yet senate meetings were held in secret; Volsci are allowed to attend the ‘ludi’ and to meet at the Spring of Ferentina—yet in the fifth century none but Latini could do this; the Roman Marcius is given an honorific cognomen, Coriolanus, because of his behaviour at the capture of Corioli—yet such cognomina were not granted until the third century or even later and even then only to the general and not to the subordinate; the plebs is represented as wielding great power in the assembly1—yet we know that in the fifth century it did nothing of the kind.

In writing on a subject in which the most significant words have been used in quite different senses by modern authors, I think it most prudent to begin by defining my terms. By rhythmical prose I mean all prose in which the writer consciously follows a definite scheme in order to obtain particular cadences at the close of the period or within it, and this whether the favoured cadences are marked by quantity or by accent. I subdivide rhythmical prose into metrical and accentual. The latter term explains itself. In spite of Aristotle's familiar opposition of ṕÅθμóσ and μτρ¿ν, I use the word ‘metrical’ of quantitative prose, both because the term ‘quantitative prose’ seems to me cumbrous, and because to the speakers of any modern language metre is the natural opposite of stress accent, whether in prose or in verse; and Aristotle's terminology, which takes no account of stress accent, may reasonably be displaced by the modern distinction. (For the beginnings of this later usage in post-classical theoreticians cf.

This paper is (as will be sufficiently evident) a by-product of two systematic inquiries: one into the Strategia and Strategoi of the fifth century, the other into the military expenditure during the Peloponnesian War. The two inquiries cannot be kept distinct, and both have largely to depend on a minute study of the financial inscriptions.

Although much has been written in the attempt to date individual plays of Plautus—too often, unfortunately, an attempt to make bricks without straw—little has hitherto been done to determine the approximate chronological sequence of the plays as a whole. Yet this appears the most obvious necessity if any advance in scientific criticism is to be made. Not till this is done can we see the bearing of the innumerable facts which have accumulated in the extensive Plautine literature of the last fifty years. No one would dream of a scientific criticism of Shakespeare which did not take into account the chronology of the plays. This is equally true (for example) of Tennyson and Browning, though the method has not yet been as systematically applied to literary as to musical criticism. Think of a writer on Bach or Mozart, Beethoven or Schumann, who mixed up indiscriminately early and later work! If we could only arrange Plautus' plays in something like chronological order, a flood of light would be thrown on details of technique which at present escape us; the utility of the Lexicon Plautinum would be doubled.

The revival of interest in Greek metre cannot be without an influence on the text of Pindar. In some ways this influence may have been for the good, but in one respect a theory based on insufficient evidence seems in danger of corrupting the text. The theory is briefly that in his dactylo-epitrite poems Pindar occasionally equated a choriamb — ∪ ∪ — with an epitrite — ∪— — or — — ∪ —. This theory has been stated explicitly by two leading Pindaric scholars, Otto Schröder and U. von Wilamowitz-Moellendorf. In the Prolegomena to his Editio Maior, p. 13, Schröder briefly states his case for what he calls ‘metathesis metrica.’ Moved by the fact that in iambic and Ionic metres a short syllable may be substituted for a long, and vice versa, he is convinced that the same applies to the metres of Pindar.

Testimony, even in the well-worked parts of antiquity, continues to accrue, some from new papyri, some from texts long known but overlooked. In the result the critic is discomfited, honest men come by their own. οιόνδ' ἀπέβη τόδε πρᾶγμα.

This document, which consists of two lives of Homer and the Agon inserted between them, has long been derived from the Μουσεîον of Alcidamas. The conjecture was contested and the survival of Alcidamas denied. The publication of twenty-five lines from a papyrus of s. II.–III. p. C. by Professor J. G. Winter (Transactions of the American Philological Association, 1925, Vol. LVI.), containing a Homeric biography signed αλκι] δαμαντος περι ομηρου put an end to the matter. See on the whole question Mr. Winter's remarks and a paper by Sig. Carlo Gallavotti of Bologna, Rivista di Filologia, 1929, 31.

In Vol. XXIIL, pp. 105–8, Mr. Lockwood criticizes some of the observations which I made in pp. 7–10 of the same volume.

§ 271. I ‘assume’ that τοῦτ' δετι δεινότητα explains ύπόκρισιν καὶ ἀγῶνα and not τὸ διαλελυμένον, (1) because ‘figures of speech in general’ cannot be said to ‘produce’ τὸ διαλελυμένον asyndeton, which is itself one of those figures; (2) because, conversely, τὸ διαλελυμένον is not equivalent to δεινότης but a means of producing δεινότης. τὸ διαλελυμένον must, therefore, be nominative, not accusative, and Professor Rhys Roberts' rendering, ‘lending especially the effect of abruptness, in other words of energy,’ cannot (I speak with all respect) be right. The Professor himself is not happy with it, and adds in a note that ‘a better sense would be obtained by placing μάλιστα τὸ διαλελυμένον before ὑπόκρισιν, i.e. “the figures of speech, and especially asyndeton.”‘ My excision (or, rather, Professor Radermacher's) gives the same sense. Orth, in his translation, omits the words τοῦτ' ἐστι δεινότητα, and interprets as I do.

As I had rejected en bloc references to Bacchanalia as indications of date, it was a mere chance which led me, after the above had left my hands, to look up Cas. 980 ‘ nunc Bacchae nullae ludunt,’ quoted in Schanz, R. Lit.3, p. 78 (q.v.). As this affords an excellent opportunity of testing my conclusions, perhaps a word or two will not be out of place.