STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. BOOK REVIEWS Untersuchungen iiber die Natur der griechischen Betonung. Von Hugo Ehrlich. Berlin: Weidmannsche Buchhandlung, 1912. Pp. xi+275. M. 8. The unparalleled conservatism and transparency of the Greek vowel system has generally been regarded as a result of the musical accent — or, more accurately, of the level stress — of that language. In late years, however, several scholars have assumed rather extensive influence of the Greek accent upon the phonology. Shall we, then, conclude that Greek had after all a fairly strong stress accent from very early times, or shall we take a skeptical attitude toward the utilization of the facts of Greek accent for the explanation of sound changes ? This is the question that Ehrlich has set himself. Although he clearly prefers the second alternative, he goes about the examination of the evidence quite judicially. He does not even cite the great intrinsic probability that a language with a remarkably conservative vocalism did not have enough stress to affect the phonology at all. As an example of his method we may consider the treatment of apocope in the first chapter. A tabulation of all the occurrences in Homer of the prepositions av(d), Kar(a), and jrap(a) not in composition leads to the formulation of this law: a succession of three short vowels, separated by single consonants, is avoided by the suppression of the middle one if that stands at the end of a word. There are only 53 exceptions out of a total of 1,203. The same prepositions in composition with verbs show 88 exceptions to the rule out of a total of 649 occurrences. The greater irregularity here is plausibly explained as due to leveling within the paradigm : KarQipxy led to KarOyo-ia, etc. Most of the apocopated forms in noun compounds conform to Ehrlich's law. He admits exceptions in Ka/*/xopo? from *KaT-p,p,opos, eiracrovrtpot from *«rav- o-ovrepoi, iirrjTr)S and iirrjTvs from *eir-pjT-. The restriction of the monosyllabic forms to certain rythmical surround- ings is usually supposed to indicate that apocope was a matter of tradition with the epic poets and was used only on account of metrical necessity. Hence it was confined to the thesis where short syllables are impossible. Ehrlich finds this explanation inadequate because it does not account for (1) the lack of scansions such as trdpa A«k - ~ — (like tlvayoprj, vjrelpaAa), and (2) the rarity of phrases like ap. wvpyovs - - | -. He is surely over- 482 Book Reviews 483 stating his case here: there is no difficulty in supposing that the poets avoided itap where they could, but nevertheless preferred it to *irdpd, and the phrase dva irvpyous can be got into the verse by merely putting the preposition in the arsis. Ehrlich's other objection to the appeal to metrical necessity has more weight. Apocopated trap and dv were perfectly good Ionic forms {trap, SGDI 5434, 9, etc.; dv-, Hdt., Herodas, etc.), and the poets would not have avoided them. The Homeric exceptions to Ehrlich's rhythmic law are easily accounted for as the earliest extensions of the shortened forms. The full forms have scarcely spread beyond their original sphere. They can, of course, stand before single consonant and short syllable only in case of metrical lengthening (e.g., KoTaKtiaTot), and this occurs only three or four times. Of the other dialects, only Aeolic-Thessalian is as near the primitive stage as Homer in respect to apocope. All the others show extensive modifications in its use. It was analogy which confined the operation of Ehrlich's law almost entirely to a few prepositions; *irarip for warcpa beside -irarepas could not survive in view of avSpa beside dV&pa?, etc. It is, however, still possible to find a few stereotyped case forms that have lost a final short vowel. Thus dyicas, "in the arms," is plausibly explained as a locative plural of ayKmv. 5pa may well contain *5-v and 'ktq> '/SoASv, and the verb, o-Kopa»a£«r- 6ai, Dem., presupposes the phrase V KopaKas. We must probably admit that in the fourth century certain individuals at Athens used a stress accent. Ehrlich attaches more weight to Kretschmer's well-known argument that the confusion of £ with r) and of o with u> (Ehrlich adds « = I with t) in inscriptions and papyri indicates a breaking-down of the distinction 1 Here and elsewhere, by "stress accent" is to be understood, of course, a fairly- strong stress accent, such as might be expected to affect the phonology. 484 Book Reviews between long and short vowels, and therefore a relatively strong stress accent. The reviewer has shown (TAPA XLII, 45 ff.) that the confusion between e and 1/ and between o and avcys, where the altered vowel is accented! It is difficult to see why we should suppose that the accent had any connection at all with the change. There is no convincing evidence for a stress accent in Greek before the Christian era, except in the speech of some very low social stratum at Athens in the fourth century. Ehrlich treats in detail a number of points which concern his main thesis only indirectly, and these excursuses embrace some of the most valuable contributions in the book. We may examine one which is devoted to an important point in Latin grammar. Our author's elaborate argument in favor of Ebel's hypothesis that the Thessalian genitives in -01 are locatives in origin may or may not win general approval. Much more attractive is a theory which he advances in this connection in order to rescue the old identification of the genitive and locative Book Reviews 485 singular of the Latin o-stems (pp. 7 Iff.): "Wahrend in der altesten (lateinischen) Sprache unbetontes oi und ai gleichwie betontes ei sich in e wandelten, wurde unbetontes ei zu reinem I." In that case early genitives like Latini and urbani (SC. de Bacchanalibus) and contracted genitives like Pomponi in the dramatists are no longer evidence for an original -I. In Celtic also it is possible to trace the genitive ending -i to earlier -ei, and the Messapian -ihi in the genitive of io-stcms may represent earlier -iei. Ehrlich is able to point out several incidental advantages of his new theory: we are no longer compelled to assume that Oscan -eis in the genitive of the o-stems represents a transfer from the t-stems; it is the Italic genitive-locative in -ei plus -s which is the final sound of the genitive singular in all other stems (cf. early Lat. isti-s, etc.). Kent has demonstrated (AJP XXXII, 272 ff. ! ) that Lucilius' rules for the use of ei and i are in general accurate, but he felt compelled to admit an error in the line (369 Marx) : "hoc illi factum est uni," tenue hoc facies i; for illi and uni could not be separated from Oscan altrei. On Ehrlich's hypothesis Lucilius is right here too. Another point deserved rather more space than Ehrlich has allotted to it. Although quoiei, dative singular of qui, does not occur in inscriptions until after the confusion of ei and I had become general, it is usually assumed that the original form of the case was really *quoiiei, with -ei which stood in some relation to the ending of the dative singular of iste, unus, etc. It has, however, been difficult to derive the various historical forms from this (cf . Husband TAP A XLI, 22) . The theory of Sommer, Handbuch, pp. 465 f . (cf. Exon Hermathena XII, 218), that i£ of quoiiei dropped before following -l, which then united with the preceding o to form a diphthong (cf . gen. sing. Pompei from *Pompeivi), could not be accepted while it was supposed that original unaccented -ei was pronounced -e in Plautus' day. Now, however, the difficulty is removed; * quoiiei regularly became *quoiii>*quo-i> (Pl&u- tine) quoi. (So far we have been following Ehrlich.) The spondaic quoi-i which occurs several times in Plautus (Neue-Wagener II, 454; Sommer Handbuch, p. 465) and epigraphic quoiei are re-formations on the basis of the genitive. The reviewer does not recognize the " Plautinische Messung qudji" which Ehrlich cites (p. 77). If there really is any such form, it should rather be spelled qud-i and explained in the same way as el (below). The history of the dative of is runs parallel; *e(iei regularly developed into monosyllabic ei. Spondaic em is due to the influence of the genitive. The iambic form probably arose from the analogy of isli and the other pronominal datives; istum:isto:istl=eum:eo:ei. That the early dramatists used both the spondaic and the iambic forms of the dative of is there is no doubt, but for monosyllabic ei the evidence is not so satisfactory. None of the passages cited by Neue-Wagener (II, 378 f.) really demands a 1 One misses a reference to this article on p. 73. 486 Book Reviews monosyllable; that pronunciation has metrical guaranty only in Cat. 82. 3 (hexameter) : eripere ei noli multo quod carius illi est oculis But if there was a monosyllabic dative of is in Catullus' time there may have been one a hundred and fifty years earlier. The reviewer has shown (Con- traction in the Case Forms of the Latin io- and id-stems, p. 18) that the nature of early prosody prevents our distinguishing a long monosyllable from an iambic word in the iambic and trochaic verses of the early dramatists unless it occurs (1) as the final syllable of an iambic cadence, or (2) as the second syllable of a resolved thesis or arsis, being shortened by the iambic law, or (3) where the meter requires the complete loss of the word by elision. Con- sequently our failure to prove the presence of such a form in the dramatic poets does not create a presumption that it is non-existent. We need have little hesitation about reading ei as a diphthong where, for example, it balances the monosyllabic genitive, as in Ter. Ph. 188: "Efus me miseret, ei nunc timeo, is mine me retinet " The point with which we are at present concerned is that this monosyllabic ei is the regularly developed dative parallel with quoi and hoic. There are, of course, some difficulties in the way of assuming that unac- cented ei became -i in preliterary times, but they are none of them fatal. The locatives Brundisii (Enn.) and Sunii (Ter.) beside contracted genitives such as ingeni were long ago adequately explained by Lachmann: "illud i, quod est ante notam casus, Graecum videbatur, ut more Latino praeteriri non posset." Cato's Lavini is in all probability correctly recorded. Tibet, etc. (sibei, SC. de Bacch.), can no longer be directly connected with Osc. slfei and Paelignian tfei. The Latin forms, like OB. tebe, sebe, contain Indo-European -bhoi, while the Umbrian and Paelignian words go with OPr. tebbei, sebbei, and show the e-grade, -bhei. The reduced grade of the suffix appears in Greek -<£t. Since the dative singular of consonant stems is found with diphthongal spelling several times before the sound i begins to be written ei (e.g., vir- tutei, CIL I, 30), it is necessary to give up Solmsen's (KZ. XLIV, 161 ft.) connection of it with Oscan medikei, etc., on the basis of an Indo-European dative ending -ei. This is the more regrettable in view of recei in the old Forum inscription. If this form contains a dative suffix -ai, we must assume that the inscription is later than the weakening of unaccented ai. Ehrlich is probably wrong in assuming (p. 72) on the basis of Plautine at, dis, i, and Is that in dissyllabic words even unaccented oi and ai had become I in Plautus' day. It is far simpler to suppose that the sound group -ee contracted to -e, which later became -I. The forms dii, etc., were merely orthographical representations of d%, etc. (See my Contraction in the Case Forms of the Latin io- and ia-stems and of deus, is, and idem, pp. 15 ff.) Book Reviews 487 Ehrlich's treatment of his subject is independent and suggestive through- out. Even where his conclusions fail to be accepted, his novel grouping of the evidence will lead to numerous advances in our knowledge. Misprints are very few. The only one of any consequence that has been observed is in 1. 3 of p. 28, where one should read - - | - for - | -. E. H. Stuetevant Columbia University / Fasti Consolari Romani. By Giovanni Costa. Milan: Libreria Editrice Milanese, 1910. Vol. I, Parts 1 and 2. Pp. x+547; 150. L'Originale dei Fasti Consolari. By Giovanni Costa. Rome: E. Loescher & Co., 1910. Pp. 77. These books are a monument of zeal and patience if not of wisdom. In the preface of the first the author states that he proposes to initiate a "radical and fundamental" reform in the study of Roman history by substituting a consistent inductive method for what he regards as the more or less haphazard method of criticism hitherto in vogue. With this laudable purpose in mind he devotes this volume to an exhaustive study of the lists of magistrates of the republic, intending in the second volume to restore the original, which he calls the spina dorsale of Roman history, and to draw such conclusions as may be based upon it. It is perhaps unfair to pass final criticism before the construc- tive part of the work is in our hands, but it is at least proper to express grave doubt as to the surpassing value of this spina dorsale. It is not of great importance, for our real appreciation and understanding of the development of the history of Rome, that we should know whether a Fabius or a Claudius was consul in any particular year. So our eagerness to know what revolu- tionary use Costa will make of such facts when determined is all the more keen. His theory is in brief this: In literature and on the monuments we have two distinct traditions of the list of magistrates of the first three centuries of the republic, both derived from the records of the pontiffs, the Annates pontificum. One, which he calls the chronographic, was that of the revised form of the Annals, published by Q. Mucius Scaevola, and known as the Annates maximi. This was the basis of the chronicles of Nepos and Atticus and of Varro's work, and is preserved for us with varying degrees of accuracy in Diodorus, Cicero, Velleius, the Fasti Consulares and Triumphales on the Regia, Tacitus, Cassius Dio, Pliny, and the Chronograph of 354 a.d. The other tradition was based on the Libri lintei made up from the Annates pontificum in the latter part of the fourth century b.c. This was a simple list, and preserved for the earlier period a greater degree of authenticity than the more elaborate Annates themselves, which suffered at the hands of editors