Abbildungen der Seite


A Study of Races in the Near East. Pp. xiv + 139. By WILLIAM H. WORRELL. New York: D. Appleton and Co., 1927.

Professor Worrell has given us a romantic and popular account of the ethnology and linguistic history of the Jews and their ancient neighbors in Western Asia and North Eastern Africa, with an occasional excursus upon some more distant race or language. The book is entertaining and suggestive throughout; but it suffers from the inevitable shortcomings of popular treatises upon large topics which have not yet been fully subjected to the tedious processes of sound science. The 'general reader' must be warned not to accept any statement made herein until he finds it verified elsewhere. Scholars would appreciate more argument and less dogmatic assertion.

The weaknesses of the book are most striking in the chapters that deal with ethnology and with Indo-European linguistic history. For example, the Cassite dynasty of Babylonia is labelled 'Aryan' (p. 122), in spite of the fact that we have a Cassite vocabulary which includes a few Indic names of gods, but which otherwise yields no trace of IndoEuropean. Witzel is given credit (p. 123) for the decipherment of the Hittite documents, although his book (Hethitische Keilschrift-Urkunden, Fulda, 1924) is seven years later than Hrozný's Die Sprache der Hethiter, and is besides a thoroughly amateurish piece of work which has added but little to our knowledge of the language (see Götze's review in Orientalistische Literaturzeitung 28. 234-9-1925).

Of fundamental importance for Worrell's doctrine is the familiar assumption that the learning of a new language by a community usually, if not always, produces changes in the phonetics of that language. He knows Jespersen's demonstration (Language, Chap. XI) that sure instances of the process are lacking, and he counters (p. 39) with an appeal to the mixed languages of New York City. He seems not to have studied Hempl's brief but penetrating study (Transactions of the American Philological Association 29. 35-8-1898) of foreign influence upon American English; he would find there a convincing argument that such influence is not and cannot be permanent in the conditions prevailing in the United States, except in the matter of vocabulary.

Professor Worrell is a Semitist, and the most valuable parts of his book are those which deal with the history of the Semitic languages and with the relationship of Semitic to Hamitic. He makes out a plausible case (p. 78) for his opinion that 'Hamitic represents a survival of conditions more primitive than those to be found in Semitic.' Nevertheless he is obliged to admit that 'comparative Hamitic-Semitic studies are not yet upon a satisfactory foundation.' The truth is that if Semitic scholars are going to write on linguistic prehistory, they have got to learn the grammatical method which has been perfected in the study of Indo-European, Finnish-Ugrian, the American Indian languages, etc., but which has not yet been extensively applied to Semitic.


Linguistic Evidence and Archaeological and Ethnological Facts. The Sir John Rhys Memorial Lecture. Pp. 18. By J. FRASER. New York: Oxford University Press American Branch (for The British Academy) 1927.

This lecture is concerned with the reasons why Linguistic Palaeontology cannot by itself solve the problems it set out to attack around the middle of the last century. The discussion seems strangely out of date; for there has long been agreement that in the task of locating and describing Indo-European civilisation Linguistics can hold only a subordinate position, even though its contributions need not be disparaged.

The author believes he can point out some 'possibilities of error which have, in the past, not been taken into account', but to the reviewer they seem long familiar. Thus we are told that 'we must allow in the Indogermanic language for a second dimension'; meaning that the Indo-European language had extent not only in space but also in time, though the latter dimension disappears from our reconstructions which seem to bring all facts into a single plane. At least thirty years ago this was stated with all desirable clearness by Brugmann, Grundr.2 1.24 (1897), while for its bearings on the problems of Indo-European antiquities reference may be made to Schrader, Reallexicon1 x-xii (1901) and to Hirt, Indogermanen 1.235 (1905). To show 'a further defect' 'in the current conception of the language of the Indogermans' it is said that 'we are asked to think of it as a finished product which shows no signs of a distant past, of a slow growth; and, at the same time, as something which came into existence in a way for which the history of languages of which we can observe the growth does not prepare us.'

I should say on the contrary that warnings against such views have long been a commonplace. Nor is it a novelty to tell us: 'We must think of the language of the Indogermans as one which developed in the same way as languages the growth of which we can observe more closely, and under analogous conditions'; or that: 'During the common-Indogermanic period this language must, in all probability, have been in contact with languages of a totally different character, and have influenced them and been influenced by them.' Again, when we are told that 'we must allow for the possible existence of languages that were partly Indogermanic', another old idea is being presented in a new terminology. The new term is also bad, for on this principle Sanskrit could be classified as a 'partly Romance' language. What is meant is that there may have been languages which, while not members of the Indo-European family, were nevertheless akin to it. For such languages the world has been ransacked ever since the days of Bopp, though only the recent discovery of Hittite has brought material that is both considerable in extent and amenable to treatment by our present methods of investigation. The form of the lecture must have made it a splendid agonisma es to parachre ma akouein; but it presents no new points of view that will be helpful (positively or negatively) in the study of Indo-European antiquities, and belittles unnecessarily1 the knowledge that has already been gained.


Altitalisches Wörterbuch. (Göttinger Sammlung indogermanischer Grammatiken und Wörterbücher.) Pp. viii+ 583. By FREDERIK MULLER. Göttingen: Vandenhoeck & Ruprecht, 1926.

Until the appearance of this volume we have had but one etymological dictionary of Latin that was worthy of the name, that of the late Professor Walde of Innsbruck. Zimmermann's was unsatisfactory, the French ones were too elementary, others were out of date. We now have two of the first rank, Muller's as well as Walde's, and it is our privilege to examine them and to observe their relative merits and demerits.


1 Compare: 'such statements as that the Indogermans were acquainted with the principle of the wheel, made on the ground that a number of languages, including Sanskrit, Greek, and English, have a common name for it, must not be taken as much more than a conjecture' (12); and 'We may suspect that, for example, the question, "Where was the original home of the Indogermans?" is a meaningless one, and that in the course of the development of the Indogermanic language the speakers of it may have changed their home frequently' (16).

Walde's Lateinisches Etymologisches Wörterbuch appeared first in 1906; the need for such a volume was shown by the fact that four years later it came out in a second edition, entirely rewritten and enlarged by at least one fourth. Those who used it found however that it was too much devoted to the comparative side of etymology, and too little to the relations of words within the Italic itself; especially, it neglected the history of other than radical syllables.

These probably are among the defects which impelled Muller to undertake a new and independent etymological dictionary of Latin. The work was long and fatiguing; in the course of it he found himself obliged to make a special study of what the ancients themselves thought upon the problems of their language. The fruits of this are in his volume De Veterum imprimis Romanorum, Studiis Etymologicis, 1910. An original plan for a series of volumes had to be much restricted; but he issued in 1920 a volume entitled Latijnsche Woordverklaringen op semantisch-taalhistorischen grondslag (Verh. d. Kon. Akad. v. Wetenschappen, XX, 3, Amsterdam).

He was now ready to devote his energies and his knowledge to the present volume. His aim was not to include every Latin word, but only those words which might be called 'old Italic'. Just how old they had to be for inclusion, is a problem to which even he can in his introduction give no absolutely definite answer; but he reduces every word to a pre-Latin or an earlier form, to serve as a caption. Of the consequences of this I shall have more to say later.

Each article consists of caption word, its original meaning, its Latin form (or forms), its earliest meaning in Latin if that has already diverged from the primitive meaning, the place of its first appearance in Latin, and whether it goes on down into any Romanic language. Then comes its phonetic and morphological history, if these offer any peculiarities, including comparisons with other Latin words showing the same features; next the cognates in the Italic dialects, then those in the other branches of Indo-European, with such explanatory remarks as they may evoke. Other Latin derivatives may be added at the end, if this is more convenient, with their special history. The arrangement is subject to variation if the special subject is better treated in some other way. But this outline gives a fair idea of the data to be expected in each article.

Muller has the advantage over Walde that he gives greater attention to the history of the word within the Italic itself, to its first appearance, to what the Romans themselves thought of it, to the history of the

suffixal syllables, to its continued life in Romanic; he has also at hand the writings of more than fifteen years since Walde's work was in final form.

And yet, if I had to choose one of the two volumes and discard the other, I should find myself quite baffled in my choice. For this I have several reasons.

First: Muller has chosen to use as his captions reconstructed forms, which obscure the location of the words. The selected index, in which x has the alphabetic position of cs, is intended to overcome this difficulty, but it is quite inadequate. For example, words which in Latin begin with fa- are listed in the index if they are under captions other than fa-; but I cannot find fābula nor fātum anywhere, nor facinus unless I guess that it is under facio which is in the index as fakio; facilis seems nowhere to occur, and făcundus, which should come under fārī (170) seems to be instanced only 506.2 as a parallel to fecundus, the two showing the same suffix. Reciprocus is not in the index, but is mentioned incidentally under prope 363.20, and vesper 541.30. Successive caption words give curious successions of Latin words, as on page 135, hospes, fovea, grandis, and page 138, hilum, formus, fornus, lēvir, damnum. This is a difficulty inherent in Muller's plan; but it makes also an undue difficulty for the user of the book, and often leaves an uncertainty whether the word may not be tucked away somewhere, although he cannot find it.

Second: The citations of modern articles are rather inadequate; I have often, as on nequeo (289), been inclined to attribute original results to Muller himself, until I looked into Walde, where I found the desired citations. Further, there are a number of abbreviations which I do not recognize, and others which are queer, such as Schulze, ZGLE. (211.20); there is no bibliography with the list of abbreviations used.

Third: The caption-words do not belong to any one linguistic period: For example, lăq"ěios 229, penge 331, trěies 493 are pure primitive Indo-European, while ně+kēd-ti- 284 shows the purely Latin loss of a sibilant before d (-kezd-), beside containing a chronological error. Mădia (lis) is the caption for Latin maiālis (sic) 249, and fŏdiō is that for Latin fodio 183, which shows carelessness in the use of vowel and consonant i; for it is di that gives ii, while di remains. This is typical of Muller's use of i and i.

These three points may perhaps be accepted as relatively objective criticisms: there are other general points which may be subjective with the reviewer. For example, I object to the marking with the

« ZurückWeiter »