continued by the poets that followed, and extended indefinitely; and geniuses like Lucretius and Catullus made good use of the example. But, as in English, by non-use the facility of such composition was rapidly diminished, and the language lost here much of its flexibility.

Remarks were made by Professor A. G. Harkness and by Dr. Knapp.

Adjourned at 12.45 P.M.

AFTERNOON SESSION.

The Association reassembled at 3.15 P.M.

Professor Albert Harkness reported that the Auditing Committee had examined the account of the Treasurer, compared it with the vouchers and found it correct. The report was adopted.

15. Notes on the véκva of Peisandros, Aristophanes, Aves, 15531564, by Professor B. Perrin, of Yale University.

This is a parody on the véκvia of Odysseus, full of Homeric reminiscences. The Murn is the 'keavós of x 508, X 13, w 11, and the unwashed Socrates as ψυχαγωγός answers to the Hermes of w. Peisandros takes the rôle of Odysseus in A, and goes down to the confines of Hades to consult his own spirit, not that of Teiresias. As a sacrifice, the ram and black sheep of Odysseus, Peisandros has a camel-lamb. He sacrifices it, then like Odysseus turns away (ảπ0€ 1561), when up there comes to him from the lower world - Chairephon the Bat (so called as the spirits of the suitors are compared to bats w 6 ff.).

In 1561, noting the surprising lack of correspondence with A 36 ff., commentators almost universally suspect or correct ả#îe. Kock's earlier κaσто, and his later kaтe are fair examples of corrections made to restore correspondence with Homer. But the ȧæλ of Aristophanes is faultless in its tradition, whereas A 35-49 shew clear signs of alteration from an original context which corresponded minutely with κ 526-536, the forerunner-passage. At just the point where the minute correspondence is broken, we find in λ a group of verses to which the Alexandrian critics took exception. They are memorable and faultless verses in themselves, but they are not adapted closely to their present context.

At some point in the tradition of the Homeric poems, probably the time of their committal to writing, the desire to preserve this avos led to its substitution in A in place of verses corresponding minutely to x 528-530, where there is an expression of which the άe of Aristophanes is a faultless parody.

It is most natural surely that a brilliant parody of the véκvia of Odysseus should base itself on the main account of that vérvia rather than on its forerunner-passage, i.e. on A rather than on x. Either, then, the manuscript of Aristophanes' Odyssey had at A 38 something corresponding to the drovoo pɩ тpañéσbai of x 528, or the oral tradition of the poem, kept vivid by public recitation, had this. We thus get a glimpse of a written or oral status of a famous passage in Homer

K

which was superior, in point of exact correspondence, to the subsequent vulgate which became the basis for Alexandrian criticism. The Homer of Plato has been shewn to differ in more than mere minutiae from this vulgate. In one case, at least, the Homer of Aristophanes so differed. It was probably a long and gradual process by which the vulgate won its ascendancy.

Remarks were made by Professors Gudemann and Elwell, and by the author in reply.

16. NOTE in the Orators, with special reference to Isocrates, by W. A. Eckels, of Baltimore.

Professor Gildersleeve (A. J. P. XIV. 241) records his conviction that "it is safe to speak of stylistic effect within the range of ore," and indicates as a marked source of such effect the use of wore with or without a preceding correlative (OUTWS, TOLOÛTOS, etc.). The same writer (A. J. P. VII. 171) notes Isocrates' effective use of wσte correlative in the construction of long periods. Seume (De Sententiis Consecutivis Graecis) speaks of the large use of the wore sentence in the orators, and its "great oratorical force."

This paper aimed to give a few results of a study undertaken in pursuance of these suggestions, with a view to testing the value of wore as an index of style in the orators. Only the more general and obvious results could find place in so brief a summary.

Isocrates is an author in whom we especially look for the conscious use of rhetorical effects. The plan of this study has been to compare Isocrates' use of σTE (1) with that of several other orators; (2) in the different classes of Isocratean writings; (3) in different orations of the same class; (4) in different parts of the same oration. The points especially studied have been (1) the use of the moods, (2) the employment or omission of the correlative, and (3) the frequency of occurrence of wore in general. The conclusion was reached that (3) is of minor importance as a stylistic test, while (2) is of decided value.

Taking the average occurrence of wore to the Teubner page in six orators, the curve runs thus: Antiphon .28; Lysias .95; Isocrates 1.00; Isaeus .69; Demosthenes .49; Aeschines .30. These results are interesting, but it is not easy to connect them with characteristic differences of style. We should hardly expect, e.g., so close a correspondence in authors differing so widely as Isocrates and Lysias.

But the test of correlation brings out a real difference. In Isocrates, correlative Te greatly preponderates over non-correlative. In Lysias it falls a little below it. Isoc. correlative: non-correlative :: 2 : 1; Lys. 1: 1. Here we

seem to have a true norm of style the free wore in the simpler, less periodic style of Lysias, the correlative wore in the more complicated structure of Isocrates. Isaeus marks a further gain for the non-correlative type- -cor. I; non-cor. 1. The closer "grip" of argument in Isaeus still had need of wore as a logical instrument; it could better dispense with the rhetorically effective οὕτως — ὥστε, which was brought into frequent service by the narrative of Isocrates and Lysias.1

1 The ratio of correlative to non-correlative in the six orators studied stands thus: Ant. 1: 23; Lys. 11 Isoc. 21: 1; Isae. 1: 1; Dem. 1: 1; Aesch. 1) : 1.

Again, compare two works of Isocrates. The Adv. Euthynum is a strictly forensic work close argument, almost devoid of narrative, so concise and plain in style as to be denied to Isocrates by some. The Helen is an epideictic speech of the most ornate type, full of flowing periods and involved sentences. In the average of the occurrence of wore they both occupy a high place, — the Euth. first of the twenty-one works, the Helen sixth. But these wore sentences differ widely in rhetorical effect. Out of 15 cases in the Euth., 12 are of the noncorrelative type; and almost all these represent what may be called the “σTe of logical inference," -a sort of "therefore," introducing an opinion or conclusion. Out of 18 examples in the Helen, 16 are of the correlative type — the outws often at the head of its clause, giving an effective balance. For the two types, compare Euth. 5 and Hel. 37. The predominance of these two types in ratios of 4: I and 8: I respectively seems an excellent index to the widely differing styles of these two orations.

66

Attention was now concentrated on Isocrates, and a table presented showing the average occurrence for each oration, for each class, the number of correla tives and non-correlatives in each oration, and the ratio of correlatives to noncorrelatives for each class. The works are classified according to Jebb, thus avoiding any tendency to twist the classification in the interest of a theory. The ratios of correlative to non-correlative for the several classes are as follows: (1) Epideictic, 3: 1; (2) Philosophical (Essays on Education), 2:1; (3) Political, 2:1; (4) Forensic, 1:1; (5) Hortatory, 1:1.

Assuming that the correlative wore is suited to a dignified, elaborate, and consciously rhetorical style, this order of classes is much what we should expect. Epideictic discourse is the natural home of this kind of writing, and the Philosophical and Political, in Isocrates at least, are much tinged with it. Hortatory and Forensic work involve an opposite tendency (on the yévos dikavikóv, cf. Panath. 1), and are unfriendly to elaborate periods and correlative structure.

The test of correlation was next applied to the individual orations within each department. When we find a work differing widely in its use of wore from the normal usage of its class, we inquire whether it is in other respects abnormal whether it is a fair representative of that class. In nearly every instance wide departures from type in the use of wore were found to coincide with lack of conformity in other respects. The Archidamus, e.g., stands lowest in the Political class in use of the correlative type, correl. 1: non-correl. I, as against 2:1 for its class. But this speech was noted by the Pseudo-Longinus as an instance of poσwronolía-feigned speech of another; the speaker is a young man and a Spartan prince. In closeness of argument and earnestness of tone it approaches the forensic class, and recedes furthest from the epideictic coloring which marks the Areopagiticus and Panegyricus-orations which stand at the opposite end of this class in respect to correlation, with ratios of correl. 41 : noncorrel. I, and 3: 1, respectively.

The widest variations in style in any one class are found in the Forensic, and here are seen the widest extremes in the use of wore correlative and non

1 Euth. 5: Νικίας τοίνυν Εὐθύνου πλείω μὲν ἔχει, ἧττον δὲ δύναται λέγειν· ὥστ ̓ οὐκ ἔστι δι' ὅτι ἂν ἐπήρθη ἀδίκως ἐπ ̓ Εὐθύνουν ἐλθεῖν.

Hel. 37: Οὕτω γὰρ νομίμως καὶ καλῶς διῴκει τὴν πόλιν, ὥστ ̓ ἔτι καὶ νῦν ίχνος τῆς ἐκείνου πραότητος ἐν τοῖς ἤθεσιν ἡμῶν καταλελείφθαι,

correlative. Here, too, the question of genuineness has been oftenest raised. The non-correlative extreme is represented by the Euthynus, whose peculiarities have already been noted, and which may be called the "ultra-forensic" specimen of Isocrates' style. At the opposite poles stand the Trapeziticus and De Bigis. The former was regarded by Benseler as a school exercise; and Dionysius of Halicarnassus lays especial stress on its "epideictic" character. As to the use of wore, it shows a ratio of correl. 33: non-correl. 1.

The De Bigis, ostensibly a court speech, is in effect a glorification of the elder Alcibiades. Jebb notes its "thoroughly epideictic" character. In its use of correlation it stands highest of the Forensic works, and second in the whole Isocratean corpus: correl. 4: non-correl. I.

But the tone is not uniform throughout the whole fragment, and an analysis was here presented showing its composite structure. A speech assigned to one department may show clearly-marked strata of material belonging properly to another. The divisions made were: (1) (secs. 5-21) narrative mixed with argument the tone partly apologetic; (2) (25-38) pure narrative bold and unqualified panegyric; (3) (39-50) argument and personal plea. In use of WOTE, (1) shows correl. 3: non-correl. 1; in (2) (having 10 examples in all) every one is correl.; in (3) all are non-correls. This analysis and the estimate of the stylistic features of the several divisions are closely confirmed by Drerup (Neu. Jahrb. Suppl. Bd. 22, h. 2). He concludes that “this oration proves the forensic diction in Isocrates to be different from the epideictic."

An analysis was also made of the Panegyricus, one of the longer speeches and one of the most consciously artistic. The sections which are clearly of the epideictic order, - pieces of "fine writing," recounting the ancient achievements of Athens, -aggregating one-half of the speech, show a ratio of correl. to noncorrel. of 4: I, as against 3: 1 for the whole work, and 1: I for those portions which could be described as the opposite of epideictic.

Analyses of a number of other orations showed similar results, not always so striking as those presented, but in the main consistent with them and tending to show the stylistic belongings of these two types of wore sentence.

The use of the moods after wore appears not to be, per se, an index of style. It is true, eg, that Lysias has a larger ratio of finite verbs to infinitives than Isocrates, - - 2:1 as against 1: 1; but this loses independent significance when we recall the larger use in Lysias of the non-correlative type of wore sentence; for in this type the finite verb, for obvious reasons, almost crowds out the infinitive. To get at the separate value of the mood test, we ought to eliminate the factor of correlation, i.e. compare the use of moods within each type. The differences are found to be very slight. In non-correlative examples we find that the ratio of finite verbs to infinitives is as 3:1 for Lysias, 33: 1 for Isocrates. In correlative examples, Lysias has 7 finite : 6 infinitives; Isocrates, 11:9. So, in different speeches and parts of speeches of these authors, so long as we study the correlative and non-correlative types separately, statistics of the moods have no story to tell of a difference in style. That the use of the finite verb is not a mark of negligentia, its preponderance in Isocrates would seem to show with sufficient clearness. The fact comes out still more clearly in a comparison of Isocrates and Xenophon. In Isocrates finite verbs stand to infinitives as 1:1; in Xenophon, as 1:1 (according to Wehmann's figures).

17. Plato's Studies in Greek Literature, by Carleton L. Brownson, of Yale University.

External evidence regarding Plato's literary tastes and studies is not wanting, but it is far less adequate and trustworthy than the internal evidence. It is the latter, therefore, which the present paper aims to collect and analyze, considering first, the comments which are to be found in Plato's dialogues upon poetry and the poets, and second, his citations from the works of poets of his own age and of the earlier centuries. The few famous passages, however, which condemn so sternly the moral teachings of Homer, Hesiod, and the dramatists are reserved to be the subject of a later paper.

I. Plato regards the poets as the earliest sages of Greece," our fathers and leaders in wisdom" (Lys. 214 A.), their mission corresponding to that of the philosophers in later times (cf. Prot. 316 D., Theaet. 152 E.). They differ, however, from the philosophers in that their wisdom is the product of inspiration, not of reason. This difference is everywhere (e.g. Apol. 22 C., Ion 533 D. ff., Leg. 719 C.) strongly insisted upon, as marking the superiority of the philosopher. In general, Plato sometimes speaks kindly of the poet (cf. especially Symp. 209 A. ff., Phaedr. 245 A.), but more often slightingly. He is one of the unnecessary additions to a state (Rep. 373 B.), his aim is merely to flatter and give pleasure (Gorg. 501 ff.), and he is rated among the very lowest as regards his comprehension of truth (Phaedr. 248 D.).

Plato has nevertheless made the art of the poet a subject of careful study. This is shown by the well-known passage (Rep. 392 D. ff.), in which he marks the boundaries between the various types of poetry. The first lines of the Iliad are referred to as illustrating a combination of the narrative and mimetic methods. Change the direct to indirect discourse, and the result is simple narrative; or drop the lines which intervene between the speeches, and we have tragedy. It follows, then, that poetry may be either simple narrative or imitation or a combination of both. Clearly Plato is here preparing the foundation upon which Aristotle builds in the Poetics, while at the same time making it evident that he might himself have reared the superstructure.

Plato also proves by very frequent allusions his full knowledge of everything pertaining to the art of the dramatist. Not only does he resort to the theatre for illustrations and comparisons in almost all the dialogues, but he employs in at least two instances (Symp. 194 B., Rep. 373 B.) technical expressions of the playwright which are found nowhere else in classic Greek literature.

II. Plato's references to individual poets and his citations from their works are so numerous that they can only be treated in the most cursory manner. Even the earliest, half mythical bards of Greece - Amphion, Marsyas, Olympus, Orpheus, Musaeus, Eumolpus, Thamyris- are all known to him. Orpheus, indeed, is directly quoted in several instances. Coming to the more real names of later times, we find Plato gleaning over the entire field of Greek literature. In all his citations he is seeking primarily such passages as will serve his purpose in philosophical argument, either by way of proof or of illustration. This fact explains why he cites Theognis more often than Sappho, and Euripides more often than Sophocles. On the other hand, he often quotes what seems to him false or injurious doctrine in order to refute it, while again, laying aside any utili