Closely connected with the subject of the above paper is Professor Tenney Frank's "Semantics of Modal Construction" (Part 2, in Classical Philology, vol. iii. No. 1, January 1908), the first part of which was noticed in my article last year (p. 98). Here constructions of the type nulla causa est quin det are discussed. Professor Frank holds that expressions of obligation and propriety differ so widely in form and function in the different Indo-European languages, that we are not justified in assuming a proethnic usage as the source of the different idioms. Thus he agrees in some important respects with Professor Morris, to whom he acknowledges obligations; but the investigation is conducted on independent lines. One of the most interesting points is the establishment of a subjunctive expressing "logical necessity' (pp. 15 f.) as in cur minimo declinent? (Cic. de Fato, 46). "Latin versus Germanic Modal Conceptions" is the title of another paper by the same author (in the American Journal of Philology, vol. xxviii. 3), in which the subjunctive of reported speech in the Germanic languages is discussed, and declared to be different in origin from the subjunctive called by the same name in Latin. What the origin of the Latin subjunctive of reported speech is, the author does not say; and indeed it is a difficult question. But until it is determined, it is impossible to say that the origin of the Germanic construction is different. And it is to be noted that, whatever the origin of the constructions may be, in their outcome they coincide in meaning to a large extent. Another work bearing on modal syntax has come into my hands on the eve of writing this article, and I can only briefly indicate its contents: Die Grundbedeutung des Konjunktiv und Optativ und ihre Entwickelung im Griechischen (Leipzig, 1908), by Dr Carl Mutzbauer, already well known as the author of an important work on the Greek tenses. The present work is a valuable collection and analysis of all the Homeric instances of the subjunctive and the optative moods. The author adopts a classification which he calls "psychological," but which Professor Hale would probably object to as in part at least "metaphysical":

1. Objective judgments (Indicative).

2. Subjective judgments:

(a) expressions of Will (Imperative).

(b) expressions of "emotional expectation" (Subjunctive).

(c) expressions of Wish, containing elements of Will and Emotion (Optative).

The idea of "expectation" (Erwartung) is treated by Mutzbauer, as by some others before him, as the fundamental meaning of the subjunctive; that this mood expresses Will, as maintained by Delbrück and others, he denies. Expectation differs from futurity in having in it a "subjective' element. According to this theory, erwuela means properly or originally, not "I will our following," but "I expect our following" (ich erwarte dass wir folgen), TS TIS TOL πρόφρων ἔπεσιν πείθηται ̓Αχαιῶν; he translates wie ist zu erwarten dass deinen Befehlen noch willig einer der Achaier gehorche? It is to be noted that the German gehorche is itself a subjunctive; and can folgen in the other sentence be an indicative? Might not, then, the meaning of eπúμeða be expressed just as well by a simple "we are to follow" as by "it is to be expected that we follow"? I am sceptical as to whether the idea of "expectation" is really adequate to the subjunctive mood. As to the subjunctive form, Mutzbauer has a curious suggestion that the lengthening of ἴμεν and ἐθέλομεν to ἴομεν and ἐθέλωμεν was simply a means of delaying the expression of the idea, and thus, by producing a certain tension (Spannung) in the mind of the hearer, suggested the idea of expectation-in other words, that the short and the long thematic vowels of the subjunctive inflexions were essentially of the nature of a hesitancy of speech. All uses of the optative he traces (pp. 142 ff.) to the idea of Wish.

The Latin Language (Boston, 1907), by Prof. C. E. Bennett,

1 My own view as to the fundamental meaning of the subjunctive will be found, briefly expressed, in a paper read before the Classical Association last October; it will be shortly published in the Proceedings of 1908 ("The Unity of the Latin Subjunctive").

is a revised issue of the author's Appendix to his Latin Grammar (1895). The present volume gives a concise and useful outline of the history of sounds, inflexions, and syntax, for the use of teachers and advanced students. In opposition to Morris's Principles and Methods, the author regards the phenomena of linguistic growth as pointing to the early existence of a fairly definite value for every inflected form. For the Latin subjunctive he accepts as fundamental the "volitive" meaning (derived from the Indo-European subjunctive), and the "optative" and "contingent future" meanings (derived from the Indo-European optative), the old subjunctives of pure futurity having been absorbed by the indicative (e.g. ero). As to the division of words into syllables, Bennett rightly follows Dennison and Hale in rejecting the rule of the Latin grammarians, as out of touch with several sets of facts (§ 35). But I cannot agree with him in regarding the peculiarities of Plautine scansion as evidence enabling us to determine the natural quantity of vowels in syllables which are long by position; e.g. the iambic line quo nemo adaeque iuventute ex omni Attica (Most. 30) is no proof that the e of iuventus was naturally short (§ 36, 2). One might as well (or better) argue that the a of bonas, the ē of vides, the i of pudicitia, the ō of obsecro, the au of audivi, etc., were naturally short in Old Latin, on the basis of lines like Stich. 99, Most. 199 (see my critical note in 2nd ed. 1907), Epid. 405, Cist. 453, Curc. 594-all of which inferences could be easily overthrown by quoting other lines in which the syllables in question are long. It was the employment of this kind of evidence that vitiated many of the conclusions of Anton Marx as to "hidden quantities." Accent is somewhat strangely defined by Bennett in such a way as to include the "quantitative prominence" of long syllables (§ 54); though in the following section the author appears only to mean that long penultimate syllables received a secondary stress. The book must be used with caution; but it is a convenient clue to the labyrinth of modern syntactical theory.

Professor Nutting has made a further contribution to his

valuable studies of conditional sentences: "The unreal conditional sentence in Cicero" (American Journal of Philology, vol. xxviii. Nos. 109 and 110). He distinguishes "indirect inferential" uses of the unreal conditional from "normal" uses: the former are those in which the important thing is not what is stated but what is implied; e.g. si id culpa senectutis accideret, eadem mihi usu venirent (Cic. de Sen. 3, 7) -the intention of the writer being to say, "I have no such experience; therefore (by inference) such discomfort is not due to old age merely." Similarly, we have the “indirect explanatory" (distinct from the above); e.g. ego vero ita fecissem, nisi interdum in hoc Crasso paulum inviderem (De Orat. ii. 56, 227 f.), where the meaning is: "I did not do so, because I am a little jealous of Crassus at times." On this basis Nutting accounts for many sentences of the type si esset... fuisset (e.g. Cic. de Invent. i. 48, 90, pro Mur. 8, 17, Phil. ii. 3, 5, vi. 3, 6, etc.), rejecting as non-proven the current explanation that the past imperfect tense is chosen because the if-clause expresses continuous or repeated past unreality (p. 165). Of course Nutting's theory is only applicable in cases where the unreality of the apodosis (as well as that of the protasis) is implied; cf. Goodwin, Moods and Tenses, § 412. common complaint on the Continent and in America that English scholars take little interest in questions of grammatical theory. How far the reproach is merited in general, I will not venture to say; but we have within the last year a book by Mr H. Richards which contains a number of valuable contributions to syntax-Xenophon and others (Grant Richards, 6s. net). His interpretation of Soph. Aj. 186 ἥκοι γὰρ ἂν θεία νόσος (p. 101) is surely right (even as against Jebb); ef, in addition to the parallels cited by Richards, Aesch. Ag. 1507 (Aor.), and Thuc. i. 9, Herod. i. 2 (Pres.); see also Harry in Class. Review, 1905, p. 150. I had independently noticed the use of ego to denote an "ideal 1st person," as in Juvenal's ubi tu pulsas, ego vapulo tantum (iii. 289); Richards quotes other instances, from both Greek and Latin (pp. 59-61). His views on ar with the future in Attic (pp. 277-90) are worthy of

careful consideration, based as they are on a relatively complete collection of instances; his conclusion is that the construction is illegitimate and due to MS. corruptions. And the reader will gain new light on other questions of syntax, e.g. the meaning of the aorist subjunctive in subordinate clauses (p. 27). The book also contains a good deal of lexicographical matter. The diction of Xenophon is examined, and in regard to many words the question is raised as to how far they are or are not good Attic prose words. This part of the work has, therefore, a direct bearing on the practice of Greek prose composition. Mr F. E. Thomson's Syntax of Attic Greek has appeared in a new edition (rewritten). Professor Grandgent's Introduction to Vulgar Latin (Boston, 1907; price 6s.), though intended primarily for students of Romance philology, is of interest to classical scholars also, as throwing light on questions of vocabulary, as well as of phonology, morphology, and syntax.

The new part of the Archiv (xv. 3) contains, among other articles, a discussion by Havet of the forms and meaning of eluare, "to come to grief." The new part of Glotta (i. 2, 3), edited by Kretschmer and Skutsch, contains many articles of interest by leading scholars. And we are promised a new Russian periodical for classical philology. "Let us hope that this periodical which has arisen out of the ashes of its predecessor may contribute to raise again the banner of classical philology in Russia" (Anatol Semenow of St Petersburg, in the Berl. Phil. Woch., 25th April 1908).

LEXICOGRAPHY.-It may be useful to call attention to the Lateinisches Etymologisches Wörterbuch, by Prof. Alois Walde, though the book appeared in 1906 (pp. 870, published by Winter of Heidelberg). It contains a concise and trustworthy summary of the latest results of research, and will be found very useful by all Latinists. It is unfortunate that the price should have been fixed so high (£1, 4s.). Prof. Gonzalez Lodge has just produced the fifth part of his monumental Lexicon Plautinum (Leipzig, Teubner, 1908; price

1 It has been thought desirable henceforth to include "Lexicography" and "Metric" in this article.

Р