AMERICAN JOURNAL OF PHILOLOGY VOL. XXV, 1. WHOLE NO. 97. I.-ON SOME ALLEGED INDO-EUROPEAN CHARACTER. To the north of the great Iranian region, in the Russian steppes about the Black and Caspian seas, there roamed in antiquity Iranian tribes, known as Scythians. The Greeks used this name with convenient vagueness for all sorts of northern barbarians; for the wide region indicated surely contained peoples of more than one ethnic character, especially offshoots from the ThracoPhrygian region to the south. Karl Müllenhoff, in the third volume of his Deutsche Alterthumskunde, pp. 101 ff.', has collected the Scythian remnants of speech from Herodotus and other ancient authors; also from the Greek inscriptions at Olbia which date from the first and second centuries A. D. Though they are mostly proper names there is no difficulty in recognizing their Iranian character, or a character closely allied to Iranian.' Herodotus (IV 67) explains the Scythian tribal name 'Evápees by ȧvopóyvvo; Hippocrates (De aëre, § 106) by avavdpiées 'unmanly', a good etymological reproduction of the original compound made up of Aryan a privative and stem nar, nara 'man'. Names like 'Opóvrns, from a stem equal to Avestan aurvant 'swift' (Vedic arvant 'steed'); 'Apiμaσnoi, composed of stems akin to Avestan airyama and aspa, 'having tractable horses'; Bavádaσños, Corre 1 Edited after his death by Max Roediger, Berlin, 1892. 'Penka, Mittheilungen der Wiener Anthropologischen Gesellschaft, 1893, p. 62, has suggested that the Scythians form the transition between Aryans and Slavs, a theory likely enough on geographical grounds. ་ sponding to an Iranian compound vanat-aspa 'winning horses'the Scythians were a people of horsemen-are quite unmistakable. Similarly Ovapdávns, identical with Vardanes, mentioned by Tacitus, Annales XI 8 ff., as a Parthian, is closely akin to Avestan vareða 'growing' (Skt. root vardh); Xódawos is Av. hu-daena 'religious'; Χοροαθος= Αν. hu-raoa of good figure'; and Χώδαρζος=Αν. hva-derezi'strong of one's self'. A river name Bopvoeévns, though much affected by Greek popular etymology, is Av. vouru-stāna 'having a broad base or bed'. Herodotus I 110 reports as Median the word σnáκa 'dog',' equal to an Iranian stem span =Skt. fvan with suffix ka: this has been borrowed into Russian sobaka, marking at one and the same time the geographical continuity of the Scyths-Iranians and Slavs, as well as their marked linguistic individuality: purely Slavic laws of sound could never have developed a b in this word. Thus the Scyths are Indo-Europeans who connect the Asiatic East Indo-Europeans (Aryans) with the Slavs in the North of Europe, and, through them, with the remaining North-Europeans, the Teutonic and Celtic families. In this great belt there are no problematic Indo-European peoples. With all the intricacies of ethnic and geographic interrelation between Celts and Teutons, and Teutons and Slavs, there is in the north of Europe no IndoEuropean dialect whose broader family traits are obscure, and no claimant for membership in the Indo-European family that is not freely admitted. Beginning where France joins Italy in the ancient region of Liguria; stretching from there and adjoining Etruria across Venetia to Illyria, Thrace, and the rest of the Balkan peninsula; from there again, across Hellespont, Propontis, and Bosporus, into and clear through Asia Minor, including Armenia, until we touch again the Asiatic Iranians-that is the region within which existing records of some sort point to the presence of varieties of Indo-Europeans outside of the well-known families. That is to say, the region adjacent to the northeastern Mediterranean, which came most directly under the influence of Greece and Rome. But for the jingoism of the Ancients, which made them look upon others than themselves as Barbarians, and their all-round ineptitude as linguists, we might have had real records of all these peoples, instead of having, e. g., to rely in the main 1 Cf. Hesychius gloss, παγαίη· κύων σκυθιστί. upon the scant Neo-Phrygian Pigeon-Greek epitaphs for our knowledge of the great Thraco-Phrygian family of languages. The Greeks regarded the púyes somewhat as we do the Dagos, and their knowledge of them was about as exact. In the majority of cases we have not even as much as there is of Phrygian: scant glosses, and proper names, distracted by the unfeeling popular etymologies of foreign reporters, furnish filmy composite photographs where there might have been clear pictures. Nevertheless it is significant that the regions of which the ancients knew most, have up to the present time furnished the minor Indo-European peoples: we should not need to be surprised if new records springing from this massive geographical domain should at any time disclose further candidates for independent membership in our family of languages. In fact, recently, three candidates for Indo-European sisterhood have put in their appearance in a very unexpected quarter, namely the Cuneiform records of Western Asia. The first is the language of the Kassi, the Kossaeans (Koσσaîoi) or Kissians (Kioσio) of the Greeks, long ago the subject of a wellknown little book of Friedrich Delitzsch, Die Sprache der Kossäer (Leipzig, 1884). The home of the Kossaeans was in the valleys of the Zagros mountains in Elam, between Media and Assyria. They seem to have been skilled bowmen who originally lived on war and robbery in their native mountains. But in addition they preyed early on the inhabitants of the Babylonian plain, and finally made more permanent inroads into Assyria and Babylonia, so that they could assume the part of conquerors and impose their rule upon Babylonia. A Kossaean dynasty or dynasties appear to have ruled that country for nearly 600 years, from 1700-1100 B. C.'; at the time of the Tel-el-Amarna letters, the Babylonians were designated in Canaan as Kassi. But, somewhat like the Varangian Norsemen in Russia, they finally were absorbed in the superior Semitic culture around them. Scant record of their language is preserved in the proper names of the Kossaean dynasty, and in a curious Kossaeo-Babylonian glossary of about the tenth century B. C. This is the kernel of Delitzsch's treatise: neither Delitzsch nor his successors have succeeded in making clear the character and the relationships of the 40 words 1 1 According to Sayce, A Primer of Assyriology, pp. 14, 120, the Kossaean dynasty lasted 576 years and 9 months, from 1806-1229 B. C. or thereabouts on record: the limited amount of the material may be said, at least negatively, not to disprove the most natural suggestion that occurs in connection with it, namely that it is a dialect of Elamitic.' 2 Now comes an attempt, fascinating, to say the least, on the part of a younger scholar, J. Scheftelowitz, to show that the Kossaean language is Indo-European. In KZ. XXXVIII 260 ff. he establishes to his own satisfaction I. E. etymologies for practically all Kossaean words on record. The resulting dialect would have to be, as I gather, a satem-language, because šukamuna, an epithet of Nergal, a Babylonian god of the mid-day sun, is identified with Vedic fucamāna 'shining'; šir 'bow' with Vedic faru arrow, etc. Also because nazi, 'shadow', assumed to be identical with Vedic rajas, in which z is supposed to continue labio-velar g (g") Skt. j, shows no labialization, and may be contrasted with Gr. peßos, Goth. riqis 'twilight'. Secondly, it would have to be an Asiatic language, like Sanskrit, Iranian, and Scythian, and not a European language, like Armenian and all the rest, because it eschews the European triad of vowels e, o, a, exhibiting a in the place of European e and o: šuriaš = Skt. sūryas Gr. os; nazi= peẞos. Thirdly, it cannot be an ordinary Iranian language, because initial s changes not to h: šuriaš = Vedic sūryas 'sun' but cf. Avestan hvare; šuvalia or šivalia, name of a goddess = Vedic süvari 'giving birth'. Fourthly, it is a language closely allied to the Vedic dialect, because the majority of the etymologies are founded upon words of that dialect. = We may as well realize at once that this geographical discontinuity of the language is an obstacle of the most serious kind, and to my mind at least, makes the theory in that form antecedently improbable. The home of the Vedic dialect is in longitude 70°; the Zagros mountains are in longitude 45°. It is unfortunate that the Kossaean materials do not contain numerals, personal pronouns, or nouns of relationship, the best lexical criteria; nor can we apply the more delicate tests of morphology, because the glossary is amorphous. With one exception that does not advance the cause: there are, according to Scheftelowitz, numerous nouns of the second declension. Part of them, from Scheftelowitz's point of view, are given in the nomina 1See Hüsing, Mittheilungen der Vorderasiatischen Gesellschaft, X, 1898, p. 313 ff. tive: šuriaš 'god of the Sun', bugaš 'god Bhaga', buriaš 'lord', and 'god Rāmān', viriaš 'earth', and ṭakaš ‘star'. Another lot end in a, i, and u: šukamuna 'Nergal, god of the mid-day Sun', burna 'servant', kamula 'god of the waters', kukla 'servant'; mali and meli 'man, servant', saripu 'foot' and ilulu 'heaven'. Those in aš are noticeably mythic and cosmic: the Babylonian author explains the element aš as being identical with Kossaean iašu 'land'. Since the same authority 'could scarcely have been a Panini', whence did he obtain the apparently abstract stem form in a, which the Greek and Roman grammarians dodged all their lives? Again those ending in i and u are supposed to be given by the Babylonian author with Assyrian endings, an assumption which loosens the backbone of this bit of tradition in a pretty serious fashion. In general we may say, adopting the second half of Voltaire's well-known gibe, that the vowels count for nothing at all in these etymologies; that too, notwithstanding Scheftelowitz's grave attempt to make out a system of vocalism for the Kossaean. Scheftelowitz divides his matter into two main divisions: words which coincide phonetically and semantically with Indo-European words, and such as approach them closely. Surely there is a good deal of external resemblance, but if I remind my readers of Dr. McCurdy's effort to prove the identity of Indo-European and Shemitic (Aryo-Semitic speech, Andover, 1881), and many similar efforts before and after, they will readily assign the correct value to comparisons based upon lexical assonance. Under the first-and best-class figures kamula 'god of the waters' Skt. kamala 'water'. What Sanskritist really knows the very late poetic color-word kamala in the sense of water? It occurs a single time in the Kirātārjunīya V 25, a semantic nonce-act of an ecstatic poet. Or buriaš 'lord' is said to be Skt. bharus 'lord', a very late gloss-word which occurs as one of the numberless epithets of Çiva and Visnu. There is also a word burna 'servant' which is identified with Lith. bérnas 'servant'. Scheftelowitz places much emphasis upon this pair as containing the same root bher 'support' with really differentiating I. E. suffixes ('supporter' and 'supported'). But we note that the u in the radical syllable bur of both words, is altogether accidental and unfit from the I. E. point of view; that buriaš bharus is not entitled to an i, whereas šuriaš=sūryas is; and that Lith. bérnas from the evidence of bernelis 'boy' and Goth. barn (Engl. bairn) means = |