points to common Iranian. On the other hand they seem to precede the Iranian dialects, because initial s before vowels is not changed to h as is the case in all Iranian dialects of historical times. With all the desperate slipperiness of the ground it does seem likely that initial s in the names Sutarna, Šatiya, Šutatna, Šuvardata, etc., represents I. E. and Aryan s. The point becomes more significant if we remember that later Cuneiform records. seem to render West Persian names beginning with h=Aryan s with total omission of the h. Esarhaddon (B. c. 680-668) mentions his victories over two chiefs named Šitirparna and Eparna. The first of these is either cithrafarna=Tı(o)σapépvns, or khšathrafarna; the second seems Old Persian hu-farna = Avestan hu-hvarenanh 'having brilliant lustre', 'illustrious'. Similarly Avarparna hvare-farna 'bright as the sun' omits initial Old Persian h=Aryan s.

=

It is possible, however, that the Mitani and other Western Asiatic Iranoid proper names come from a dialect closely allied to Iranian, but yet not exactly Iranian, i. e. a dialect which did not change s to h. From such a dialect may come also Kossaean šuriaš and a few other words Kossaean proper loan-words; see above, p. 7. The Median language of which Herodotus reports σTáκа 'dog' was surely I. E. and certainly, by the very terms of σTáka, closely allied to Iranian. On a broken prism belonging to the Annals of Sargon II, there is a list of Median countries and their princes among which figures the province Musana whose prince is Sutirna. In the Tel-el-Amarna letters 232 and 233 (Winckler's edition) the same prince is called Šutarna of Mušihuna: he sends a request in abject language to some Pharao for an Egyptian garrison to protect his country. This name Šutarna, or Sutirna, is certainly identical with Šutarna of the Mitani dynasty, above, and therefore would seem to show su =Iranian hu at a period much later than the Mitani letter.' Possibly, therefore, the Mitani and Western names are Median, and, finishing the circle, Median is an Iranian dialect which does not change initial Aryan s to h. It would seem therefore well to leave the decision as to whether the Iranoid names of the Tel-el-Amarna letters are ordinary Iranian, or derived from a language closely allied to Iranian (Median?) to the future. But

1

1 Note Rost's doubtful comparison of Sutirna with Utirna, a Median province; see Zeitschrift für Assyriologie, XV 353 and 356.

this doubtful element does not impair the least bit the clear result of my inquiry: there is no record of a new and independent I. E. language in the non-Assyrian Mitani letter, and in the Iranoid names of Western Asia recorded in the Tel-elAmarna letters.

The third of the would-be I. E. languages is also found among the Tel-el-Amarna letters: it was spoken in Arzawa, or Arzapi, the latter being another possible reading. The location of this country is unknown. Northern Syria, Cyprus (Alašia in the Tel-el-Amarna letters) have been suggested, but the author of the most comprehensive treatise on the Arzawa, the Christiania. Assyriologist, Knudtzon' places the land of Arzawa in eastern Cilicia or southern Kappadocia. There are two of these letters; in the longer one of them the Pharao Amenhotep III, who, according to present reckoning, dates from the 14th century B. C., carries on a correspondence with the Arzawa king Tarhundaraba. Knudtzon points out the noteworthy fact that the beginning of the name Tarhundaraba corresponds with numerous Cilician names of men and places which begin with Ταρκυ-, Ταρκο-, ΟΙ Τροκο-, names like Ταρκόνδημος, Ταρκύμβιος, Τροκοζάρμας, etc., and bases upon that his geographic conclusion. Yet he does not fail to point out that the same type of name appears very widely elsewhere in Asia Minor and even outside of Asia Minor.

The Mitani dynasty with Iranoid names shows that conclusions of this sort when based upon dynastic names are not altogether safe: anyhow Knudtzon makes it the basis of his assumption that the Arzawa is akin to the dialect of the so-called Hatti, Hatians or Hettites; that both of these are I. E.; and that the other languages of Asia Minor that have names with Tarku- are Indo-European. I need scarcely remind the reader that the protagonist of Hettite study, Jensen, is the main source of the belief that Hettite is an I. E., quasi Pre-Armenian language, and that critics so careful as Winckler and Messerschmidt do not agree with him. Jensen's most recent summary of Hettite results, 'Hittitisch und Armenisch', IF. XIV, Anzeiger, p. 47, does not seem to me as convincing as to its author;2 at any rate the attempt to

1J. A. Knudtzon, Die zwei Arzawa-Briefe, Leipzig, 1902. Cf. Horn, Anzeiger für Indogermanische Sprach- und Altertumskunde, Vol. XV, p. 1; Kretschmer, Deutsche Litteraturzeitung, March 28, 1903, column 778 ff.

2 See Winckler, Der Alte Orient, Erster Jahrgang, 1900, p. 20.

elucidate Arzawa by assuming connection with other languages of Asia Minor is at present an effort to clear up obscurum per obscurius.

Dr. Knudtzon has called in the aid of his two colleagues, Professors Bugge and Torp, both distinguished scholars, both deeply interested in the remoter and more problematic I. E. languages. After Knudtzon's concrete treatment of the letters come Bugge and Torp with 'Remarks on the Arzawa letters'. The nature of the claim that Arzawa is I. E. may be illumined best by Torp's statement (p. 108): 'The language shows its I. E. character in its forms. But, as it seems, only in its forms. We should expect the lexical materials to be equally I. E., but are disappointed this expectation. Both nouns and verbs are opaque and cannot be compared with I. E. words corresponding in meaning'. Torp explains this on the assumption of an uncommon influx of loanwords. Standard tests for Indo-European, such as the numerals, pronouns, the familiar verbs are excluded by the very terms of Torp's statement. And yet Knudtzon publishes his book under the sensational sub-title: 'The oldest documents in I. E. speech'. The situation would be a curious one: the oldest I. E. would happen to be farther removed from the reconstructed parent speech than the most modern dialect of Germany or India.

The larger letter begins with a kind of an address in Assyrian : 'Thus speaks Nimutria, the great king, king of Egypt, to Tarhundaraba, king of Arzawa'. This is followed by two correlated passages, in the first of which sundry Assyrian words for houses, women, children, warriors, countries etc., mixed into the Arzawa text, are furnished with the ending mi; in the second the same words are furnished with the ending ti. After the analogy of the abundant Assyrian Tel-el-Amarna letters it is evident that the sender states in one of the parallel passages that it is well with him, his houses, women, etc., in the other he expresses the wish that it may go well with the recipient of the letter, his women, etc. The second of the parallel passages, in which the affix ti appears contains the additional word e-eš-tu (ēštu); this Knudtzon identifies with Gr. arw (I. E. estod) may it be '; he concludes furthermore that ti in the same passage means 'thine', and that, consequently, mi in the first passage means 'mine'. A form Labbaian is probably a caseform of a nominative Labbaia (where s is lost): Knudtzon assumes an to be an accusative corresponding to I. E. forms in m. In addition he seems to

have found a genitive in -aš, and some forms which he regards as verbs ending in -t, in -un, and in -ndu.

The repetition of mi and ti after each word, so that the series 'thy houses, thy women, thy children, thy warriors, thy horses, thy wagons, thy lands' is rendered by bitzunti dammešti turmešti zabmešti kurrazunti bi-ib-bi-it-ti kurzunti, to begin with, seems to me not Indo-European, but either Shemitic or Agglutinative. The stems themselves are confessedly not I. E. but Shemitic: if mi and ti really express 'mine' and 'thine'-the reverse, notwithstanding e-eš-tu, is not altogether impossible-we seem to have one of those mischievous accidents which we may call the standard snag of universal linguistic comparisons. The comparison of e-es-tu with I. E. estod is not as compelling as Knudtzon assumes, because of the initial vocalism.' So long as the statement 'To thee I have caused to bring a pitcher of gold as a present for thee' reproduces Arzawian ka-a-aš-ma-ta up-paah-hu-un 1 su-ha-la-li-ia azag-gi-aš dmq-an-ta, or, so long as zi-in-nu-uk hu-u-ma-an-da is supposed to mean 'send thou abundantly' we shall hardly feel justified in counting the Arzawian among the treasured minor I. E. languages.

The Tel-el-Amarna letters have increased the scope and importance of Cuneiform science. Since Grotefend's decipherment of the syllabic Indo-European Achemenidan, or Old Persian, no less than eight languages in wedge character have been discovered, and are in a more or less advanced state of decipherment: 1. Achemenidan, or Old Persian.

2. Elamite (Amardian, Neo-Susian), the second variety of Persian Cuneiform.

3. Assyro-Babylonian.

4. Sumerian or Sumero-Accadian.

5. Kassian or Kossaean.

6. Vannic, or Armenian Cuneiform of Van, of problematic character.2

7. Mitani.

8. Arzawian.

MAURICE BLOOMFIELD.

1Cf. Kretschmer, Deutsche Litteraturzeitung, March 28, 1903, column 780. 2 See Winckler, Der Alte Orient, Erster Jahrgang, p. 28; Sayce, A Primer of Assyriology, p. 36.

II. THE HISTORICAL ATTITUDE OF LIVY.

Very little can be gleaned from later writers in regard to the personality of Livy, and critics are compelled to gather from his work the statements which portray his historical attitude. But few illustrations are needed of the principal elements of his character. Deeply religious (3, 20, 5; 6, 41, 8; 43, 13, 1), but not unmindful that religion might degenerate into superstition (27, 23, 2); a firm believer in the good old days of republican rule and righteousness (26, 22, 14; 39, 6, 6), and a skeptic on some points in early Roman history (1, 16, 4; 2, 10, 11; 2, 14, 3), he undertook to construct, out of official records, earlier annals, and traditional reports, not a mere annalistic account, but a history in accordance with the principles of literary art. Yet it was historical accuracy rather than literary adornment that he had in view. As he says 9, 17, 1 nihil minus quaesitum a principio huius operis videri potest, quam ut plus iusto ab rerum ordine declinarem varietatibusque distinguendo opere et legentibus velut deverticula amoena et requiem animo meo quaererem.

At the beginning of his work he did not comprehend the full extent of the field which he had undertaken to explore. This he frankly admits 31, 1, 5 iam provideo animo, velut qui proximis litori vadis inducti mare pedibus ingrediuntur, quidquid progredior, in vastiorem me altitudinem ac velut profundum invehi, et crescere paene opus, quod prima quaeque perficiendo minui videbatur. Still this failure to predetermine the extent of the field does not affect the value of the work any more than the value of McMaster's History of the People of the United States is affected by the fact that it is more extended than was at first designed. Mistaken at the outset in regard to the labor involved, the value of Livy's work is but little affected because there was no careful preview of the material he utilized.

There are but few statements which indicate an examination of original records. It is stated 6, 1, 2 that most of the early records perished at the capture of the city by the Gauls: etiam si quae in commentariis pontificum aliisque publicis privatisque erant monumentis, incensa urbe pleraeque interiere. Prior to this