The filological study of literature seems next in honor to the creation of literature. There ar times of ebb in the tide of creativ power.

"Nature endures not the expense

Of multiplying the immense."

Then the interpreters hav their day. They bring home to their generation the inspiring thought of erlier ages. Their great books, their great institutions of lerning, preserv the world from lapsing into barbarism.

MORNING SESSION.

PROVIDENCE, July 8, 1896.

The Association reassembled at 10.10 A.M.

The President appointed the following committees :

On Auditing the Treasurer's Report: Professor Albert Harkness and Dr. Sewall. On Time and Place of Meeting in 1897: Professors C. L. Smith, Lodge, Clapp. On Officers for 1896-1897: Professors Allen, Seymour, Owen.

Professor Allen then explained the Report on Latin Orthography. After discussion it was moved to print a summary of the Report in the Proceedings.

The chairman gave the Committee's reasons for confining its report to School Text-books proper, the grammars, lesson-books, and elementary editions which serve for the first introduction to the study of Latin. In such books, if anywhere, a conventional standard of orthography may be desirable. The authors first put before young pupils are Caesar, Cicero, Vergil — perhaps also Sallust and Cornelius Nepos. All are writers of the end of the Republic or of the first years of the Augustan age. The Committee thought it wholly impracticable — for the present at least — to print these authors, for learners, in the spelling of their own times, as exemplified in such inscriptions as the Lex Julia Municipalis, and the Lex Rubria (45 and 49 B.C.), and the newly discovered inscription relating to the Secular Games of 17 B.C. It was shown that this would necessarily involve not only such forms as quoius, quoi, equos, relinquont, aestumo, but also servei in the nominative plural, and serveis in the dative and ablative plural. Although the latest of the above-named inscriptions, cut two years after Vergil's death, no longer adheres strictly to this use of ei, still we have reason, it was urged, to suppose that Vergil's own spelling was more conservative than that of this inscription. The report then proceeded as follows: --

Your Committee is clear that, as a standard for elementary books, it is best to adhere to the tolerably uniform system of the first century of our era, and that in particular the spelling of the Monumentum Ancyranum - thought by Mommsen to be that of Augustus himself-may well be to us a sort of pattern, so far as it goes. This system is, in truth, only a little later than that which we have been describing. We shall then write servi in the nominative plural, servis in the dative and ablative plural; optimus, aestimo, lacrima; cuius, cui, cum; vultus

1

and servus. As regards the most crucial point-the use of uu, vu it should be said that this spelling did not prevail all at once. From Quintilian 1 it appears that uo, vo were affected by conservative schoolmasters as late as the middle of the first century, and this statement is borne out by inscriptions for instance by the laws of Malaca and Salpensa in Spain, from Domitian's time, in which divom, vacuom, etc., occur. We should then have warrant for voltus and servos. Nevertheless it is certain from the Monumentum Ancyranum2 that uu, vu were in good use early in the century, and the practical advantage of having uniform endings, -us and -um, is so great that it should turn the scale in favor of this spelling. A middle course-voltus, volnus, but parvus, perpetuus — which we find pursued in several recent schoolbooks, seems to have no historical justification.

It is more difficult to decide what to do with -quu-, as in equus, reliquus, sequuntur. It chances that no words involving this combination occur in the Monumentum Ancyranum. Brambach believed equus to be of equal age and respectability with divus, and thought that its adoption, as part and parcel of the standard orthography of the Empire, was necessary. He was obliged, however, to except quum and quur, as non-existent forms, and he was unable to deny the correctness of the spelling with -cu- (ecus, secuntur) in the other words concerned. Notwithstanding Brambach's defence, -quu- has fallen into much disfavor among Latinists, and, as your Committee incline to think, with justice. The combination is of the rarest occurrence in inscriptions of any period. Newer and more careful researches have made it probable that it was never much else than a theory of grammarians, who sought to remove an apparent irregularity in the paradigms of inflexion. We may refer to the discussions of Stolz (Histor. Gramm. I. p. 254), and of Lindsay (Latin Language, p. 86 f.), both based on Bersu's collections. There seems to be little doubt that -quo- passed, in the course of the Augustan period, into -cu- (sometimes written -qu-), but never into -quu-. The Committee accordingly think that ecus, relicus, secuntur, relincunt, parallel to cum, cui, cuius, are the best forms for our elementary books. In like manner exstingunt will be the third person plural of exstinguo. It will, of course, be necessary to provide in our grammars for this replacement of -quu- by -cu-, but this does not seem difficult to do.

Respecting the assimilation of prepositions, it is clear that no hard and fast rules can be laid down, and that a wooden uniformity would not represent ancient usage. The Monumentum Ancyranum has conlegium and collegium, impensa and inpensarum, immortalis and inmissa. The inscription about the Secular Games has similar doublets. There is room to doubt whether ef- for ex(effugere, efficere) was in use at the time of Augustus' death. At any rate there are significant traces of ecferre and the like in the manuscripts of both Cicero and Vergil. But the evidence is not clear enough to warrant a reversal of the customary spelling. It is hard to know what to recommend about ob, ab, sub before s and t. Yet there can be little doubt that the usage of the early first

1 I. 7

26.

2 annuum, rivum, vivus.

3 P. Bersu, Die Gutturalen, etc., Berlin, 1885 The pivotal point of disagreement between Brambach and Bersu is the testimony of the grammarian Probus. See Bersu, p. 63.

4 So often in the older MSS. of Vergil; equs M, Geo. III. 499; loguntur MR, Aen. I. 731. See Ribbeck, Proleg. p 442.

century is exemplified in apsens of the Monumentum Ancyranum, and that the grammarians' fad which introduced absens, obtineo, etc., against the actual pronunciation,1 was not known at that time, and on the whole we are inclined to recommend the adoption of -ps-, -ft- in these compounds.

The Committee furthermore suggests the use of the contracted genitives conlegi, fluvi, etc., the avoidance of final t for d (set, haut, aput, etc.), and the sparing use of ch, th, ph in Latin words (it will perhaps be safest to restrict this to the four words named by Cicero as those in which he gave way to the inroads of secondary aspiration).2 Incohare should be so spelled. This word occurs in the Monumentum Ancyranum. We may add that the above recommendations require little break with current usage as shown in the most carefully prepared text-books.

A word is perhaps desirable respecting the accusative plural of the third declension. Otto Keller, in his second volume Zur lateinischen Sprachgeschichte has given us a fresh discussion of this subject, with full statistics of the occurrence, in inscriptions and manuscripts, of the endings -es and -is. From these statistics he deduces, for adjectives and participles,3 the rule that is is the only proper form for all words, whether original i-stems or not, which have -ium in the genitive plural, and -es the proper form for those which have -um in the genitive plural. This rule he considers applicable to the Augustan poets and the prosewriters of the Republic. The encroachment of -es upon -is he believes to have begun in the Augustan period, but he thinks that the poets, at least the older poets of this period, were not affected by it. The Monumentum Ancyranum, however, shows evident traces of this encroachment; it has labentes, omnes side by side with agentis, omnis. In view of these facts, the Committee feel some uncertainty; but bearing in mind the advantages of a fixed usage, we recommend, with some diffidence, conformity to Keller's rule, at least so far as adjectives and participles are concerned.

Respecting another question of some practical importance, the use of j and v, the Committee are unable to make an unanimous recommendation. One member is in favor of discarding both these modern devices, and accustoming the learner from the outset to distinguish i and u consonant from i and u vowel as Roman boys were obliged to by sense and surroundings. Another thinks that a system which fails to differentiate volvit from voluit is too hard for the beginner, but he dislikes j and v, as tending to fix and perpetuate the notion that the Romans had separate letters for vowel and consonant, and would like to see in use i and u with some diacritical mark, which might be dropped in all but the most elementary books. A third member of the Committee would adopt i for j, but would continue to distinguish v and u by separate letters. This use of v without has been resorted to in a number of recent schoolbooks. The inconsistency has this practical justification, that the difficulty of distinguishing vowel and consonant is greater with u than with i.

8 In substantives -es is more prevalent, and the rule does not always hold. The most that Keller asserts is that all which have (or can have) - in the ablative singular, either have or can have is in the accusative plural. Substantives in -r have -es. For many individual nouns, however (as aedis, finis, hostis, turris, mons, gens, pars, etc ), Keller allows only the accus plur in is. The Monumentum Ancyranum has, it may be observed, aedes, fines, and gentes.

10. The Origin of Sigma lunatum, by Professor John H. Wright, of Harvard University.

This paper is printed in full in the Transactions. It was discussed by Professors Smyth, Allen, and by the author.

11. A Discussion of Catullus LXII. 39-58,' by Dr. Charles Knapp, of Barnard College.

The author's purpose was to show that in vv. 45 and 56 dum . . . dum are correlatives, to be interpreted as literally "the while . . . the while," i.e. as equivalent to quam diu ... tam diu. This view is not new, having been held by Quintilian (ix. 3. 16), Haupt, Riese, Baehrens, Schmalz, and Hale. Its rejection, however, by such recent editors as Ellis and Merrill justifies a new examination of the whole passage. Further, the author claims to have supported the old view by a line of argument never before brought to bear on our passage. Strong exception was taken to the method adopted by both Ellis and Merrill in their attempts to interpret this passage. Both seek to determine the text and the reading of v. 45 by an appeal to v. 56. This the author held to be a complete reversal of the proper method. He came thus to the statement of his main point, which was that more attention must be paid to the form of the poem than has been accorded to it by recent editors. Several scholars―e.g. Ellis, Riese, and Baehrens- call attention to the amoebean character of the poem, but none of them makes adequate use of this point in its criticism and interpretation. Attention was then called to the fundamental law of amoebean poetry, namely, that the utterances of the second speaker should correspond in form and contents to those of the first. See Conington's introductions to Vergil's third, seventh, and eighth Eclogues, and Page's prefatory note to Horace, C. iii. 9. In Ecl. iii. the amoebean dialogue covers 48 vv., each competitor delivering twelve strains of two vv. each; in Ecl. vii. we again have 48 vv., divided into twelve strains of four vv. each. Every one knows how admirably Horace obeyed the law in the poem referred to.

The author then asked, How far did Catullus obey the law in this poem? The carmen amoebaeum proper consists of vv. 20-59. Originally there were three pairs of stanzas. The second of these is now mutilated beyond recovery; only six vv. (32-37) remain. We may therefore leave this portion entirely out of the discussion. The first strophe and antistrophe contain five vv. each, besides the refrain; no trace of incompleteness can be discovered. The third strophe and antistrophe originally contained, it is probable, ten vv. each, besides the refrain. See Riese and Baehrens on v. 41. We may conjecture, therefore, with much probability, that in the matter of form this carmen amoebaeum obeyed the first law of such compositions.

Turning to the language, we note at once striking correspondences between the several strophes and antistrophes. In vv. 20-24 the girls say, "How cruel thou art, Hesperus, to tear the maiden from her mother." The lads reply (26– 30), "How kind thou art, Hesperus, to give the maiden to her lover." Each utterance consists of three sentences: a question in one v., a relative clause in

1 See Classical Review, X. 365.

three vv., and a second question in the concluding v. These final questions are clearly examples of amoebean "tit for tat." In our passage (39-58) the strophe (39-47) forms a single sentence, composed of two clauses correlated by ut and sic. Each clause falls into two parts, with adversative asyndeton at the joints, i.e. at vv. 43 and 46. In the antistrophe (49-58) the structure is the same, save that in v. 54 the conjunction is expressed. (See further, Carl Ziwsa, Die eurhythmische Technik des Catullus, II. Theil, pp. 11, 12, Wien, 1883.) These resemblances in the language strengthen the hypothesis accepted above, that in external form there was originally complete correspondence between the parts of the song. The author dwelt thus on the amoebean character of the poem because on that he rested his special line of argument. His points were: (1) We have here a good specimen of the carmen amoebaeum; (2) the law of such carmina is that the leader sets the pace to which the other must conform; (3) here the girls lead; and hence (4) their utterances must in each case be perfectly intelligible, when taken by themselves. At v. 49 the lads were bound to reply to the girls; they were bound, furthermore, to do this in ten vv., and the form of their deliverance must be as like as possible to that of the girls. It is self-evident that to accomplish this task at all it was necessary for them to understand in every detail what the girls had said. So in our reading of the poem we must put ourselves in the position of the lads by interpreting vv. 39-47 by themselves, and then we must apply the same line of interpretation to vv. 49-58.

The author then proceeded to analyze vv. 39-47. Vv. 39-44 he paraphrased thus: Dum flos intactus est, carus est pueris et puellis; sed cum tactus est, non carus est, etc. When one reads sic in v. 45, his natural expectation is that the correlating clause will itself be broken into two parts, corresponding exactly to those of the ut-clause. These can readily be found, since dum intacta (virgo) manet = dum flos intactus est of our paraphrase, and dum cara suis est, if taken as Quint. interprets it, is a complete correlative to carus est flos, etc. Again, v. 46, which = sed cum virgo tacta est, corresponds exactly to v. 43, which = : sed cum flos tactus est, and v. 47, which = virgo non cara est pueris et puellis, is correlative to v. 44, which = flos non carus est, etc. If this stanza be interpreted by itself, its parts can be arranged in no other way. The beauty and flawlessness of the poet's workmanship are then self-evident.

The author then presented his objections to the views of Ellis and Merrill. The former says: "SIC may well contain the predicate optata est implied in the protasis of the simile," etc. A sufficient answer is the fact that the protasis of the simile contains not merely optata est, but non optata est as well. If, then, est be supplied at all after sic, we must take as its predicate the whole contents of the protasis, not a part, as Ellis has done. The same argument disposes of Merrill's view, which is thus expressed: "The two dum-clauses are not correlative, but co-ordinate, both modifying SIC VIRGO (Sc. est), while SIC is emphatic, referring to v. 42. Thus v. 45 corresponds alone to vv. 39-42, while vv. 46-47 correspond to vv. 43-44." Sic must refer not to v. 42 alone, but to all that is contained in vv. 39-44, and the predicate to est must, as already urged, be the whole contents of those six vv. Thus, v. 45 would correspond, not to vv. 39-42 alone, as Merrill would have us believe, but to all the vv. 39-44. In that event vv. 46-47 would be wholly unnecessary and therefore weak, and the perfect artistic balance which we obtained before would be wholly destroyed.