Abstract: In 1928 the Polish anthropologist Jan Czekanowski published in the ethnographical quarterly Lud' a study of the Indo-European languages in which he employed the method of differential diagnosis by quantitative correlation determinations which he had long been using with success in physical anthropology and ethnography. This method, whatever its field, rests upon the recognition of isolable and definable features or traits, which we shall hereafter refer to as elements, whose presence or absence can be determined for a number of populational groups or territorial entities, such as races, tribes, cultures, castes, or, in the present study, languages. The distribution of these is tabulated in terms of plus for presence in a particular group, minus for absence, and the question mark for unknown. Then each group is compared with each of the other groups in terms of the four-cell segregation familiar to statisticians. That is to say, four values are determined: a represents the number of elements common to both groups, b the number present in the first but absent in the second, c the number absent in the first but present in the second, and d the number absent in both. In other words, a and d are agreements, positive and negative respectively; b and c are disagreements. These four values are then substituted in a suitable formula, and a coefficient of similarity between the two groups results. When the coefficients for each pair of the groups being considered are assembled, we get a classification of the relative degrees of

Abstract: The theory is that sl is the original sound, retained by some languages, but which in some other languages became stl by a post-IndoEuropean intercalation of t. 1. A survey of the vocabularies of the various Indo-European languages, ancient and modern, discloses a remarkable, and, it seems, hitherto unnoticed fact: The northern belt of languages--Gadhelic, Gaulish, Teutonic, Baltic, Slavic (except Bohemian), Armenian, and Indo-Iranian-have, and always have had, words in sl in some form, but none in stl. The southern belt-Greek, Latin, Romance,2 and Brythonic-have not, and historically never have had, any word in sl, but many in stl and its derivatives. Moreover, sl is found in the reconstructed Indo-European and stl is not.3 As a tentative explanation of this geographical cleavage we may assume that the original Indo-European speakers in their northern home spoke sl; and that those offshoots who migrated southerly to the shores of the Mediterranean Sea lost the facility for pronouncing initial sl as the result of contact with some autochthonous people to whom sl was unknown, and who pronounced it as stl.4 This result of contact with the sea-folk5 is borne out by the fact that

Abstract: On June 26, 1935, the expedition of the Oriental Institute of the University of Chicago found at Persepolis a number of inscribed stone tablets, made in the shape and fashion of clay tablets, and bearing an important document of King Xerxes. Two tablets bore the Old Persian text, one the Akkadian version, and one the Elamite. A translation with a certain amount of historical commentary was given out by the Oriental Institute as a newspaper release, appearing in The New York Times for Feb. 9, 1936, and also in The University of Chicago Magazine 28.4.23-5 (Feb. 1936). This formed the basis for my own remarks on the inscription, in JAOS 56.211-5 (June 1936). A photographic view of the tablets at the place of discovery, with some explanatory remarks by Eric F. Schmidt, of the Expedition, was given in the Illustrated London News for Feb. 22, 1936, page 328; an abstract of this was presented in Archiv fiir Orientforschung 11.91 (first semester 1936). The actual text in the original languages was not accessible until it was published by Ernst Herzfeld in Archiologische Mitteilungen aus Iran 8.56-77 (Nov. 1936); he had previously published in the Revue de 1'Histoire des Religions 113.21-41 (Jan.-Feb. 1936) an address on Die Religion der Achaemeniden, originally delivered as an address at the Sixth Congress of the History of Religions at Brussels, Sept. 18, 1935, in which he drew upon the new material contained in these tablets. The latest bibliographic item which has come to my attention is a study by Hans Hartmann in Orientalistische Literaturzeitung 40.145-60 (March 1937). By the courtesy of the Oriental Institute and of its Director, Dr. John A. Wilson, photographs of the two tablets in Old Persian characters have come to me, with permission to use them in the preparation of this article; and for their kindness I here express my hearty thanks to the Institute and to Dr. Wilson. I wish also to thank Dr. Herzfeld for information given in a personal letter. The Akkadian and Elamite versions are each in 50 lines, covering both sides of the tablets. The Akkadian is complete, but the Elamite

Abstract: Under breaking I include here the change of a front vowel in a stressed syllable to a diphthong by the influence of following velar elements whether these be consonants in the same syllable or vowels in the next. We have to do then with OE consonantal breaking and with ON breaking, and also with that phase of velarization of front vowels in OE which, since the eighties of the last century, has come to be called u-o/a-umlaut. The last is, of course, a velar-vowel breaking.' My thesis is that these three phenomena are not only similar in nature but directly related in origin.2 In ON the phenomena are somewhat more clear-cut than in OE, in that it is only the short front vowel e that is affected. In OE any front vowel may be affected, and not only a short front vowel but, within very definite limits,3 even the long front vowels. It may be well too to observe at this point that OE breaking of the type: beorg, healdan, is held to be a purely consonantal influence, and that it took place very early, whereas such forms as heorot, and liomu, are of much later date, and were due to the vowels of the endings. There is in ON no such division of things, with centuries of intervening inactivity of the velarizing process. In ON also breaking is, to be sure, divided into two periods but on a different basis: in the first period breaking is bound up with syncope, whereas in the second period break-

Abstract: In the following I will discuss a few features of the language of the Dakota Indians which seem to have a wider linguistic interest. The first refers to the phonetic rendering of a close association of ideas. The initial vowel of Dakota words is preceded by a glottal closure. When, therefore, a word with terminal consonant precedes a word with initial vowel there is a decided break following the consonant. The consonant does not become glottalized but the glottal closure follows it. When two such words become intimately associated and form a unit concept the break disappears: napo'g na' 'a handful', for nap-'o'g.na; wali'top'e 'an oar', from wa'l-'i-top'a' (boat-rowinginstrument); hq'pap'a'-'ec't'pi 'moccasin game', for hqp-'ap'a' 'moccasin striking', 'ecz''pi 'they do'. A distinction is made between verbs that take the prefix wawhich expresses an indefinite object and nouns which contain the same prefix. The latter are unit concepts, the former express an indefinite object for which a definite object may be substituted: waa'wqyaka 'he stands guard', wa'wqyaka 'a guard'. When verbs of this type assume a special meaning they may also be contracted: waa'gli 'he brings something back home', wa'gli 'he comes back successful from a hunt';wayu'oa 'he separates something from its covering'; wo'4a 'he husks corn'. The same phenomenon occurs in the possessive pronoun, intimate possession being expressed by contraction: t'ao'wj 'his earrings', i.e., those he made, or those he happens to wear; t'o'wj 'earrings he always wears and that nobody else has a right to wear'; t'awo'wagte 'his occasional good acts', t'o'wa gte 'his goodness' as a permanent quality. These examples show a close parallelism between the concept of psychological and phonetic unity. According to a communication of Dr. Gladys Reichard similar phenomena occur in Navajo: hoyan cajan2

Abstract: Sturtevant in HG' 144, speaking of relatively transparent nouncompounds in Hittite, goes on to say 'One may be almost equally sure that tayukas (ta-a-i-ui-ga-as', da-a-i-4-ga-al, ta-a-u'-ga-a") 'two years old' beside yukas (i-i-ga-a`) 'one year old' is a compound whose prior member is akin to tan (ta-a-an, da-a-an) 'iterum, second'; but our ignorance of the etymology of the second member makes its classification impossible'. This article is an attempt to explain the word yukas 'yearling'.2 In SH 93 Hrozny' suggests an etymological connection of yukas with Goth. juggs 'young' and Lat. iuvencus 'bullock', which Sturtevant LANG. 6.215 accepts as a possibility, although he finds a difficulty in the loss of the nasal, since a nasal before a guttural is apparently

Abstract: fortunately, all three inscriptions in which this formula occurs have been lost or destroyed, so that it is attested only in transcription. But since these transcriptions are the work of different copyists, and since the three words of the formula have been preserved in several independent sources, it seems to me, as to others, that they may be regarded as trustworthy authorities. The authenticity of these inscriptions has often been impugned, but at the date at which the transcriptions were made, knowledge of Messapic was hardly sufficiently advanced for them to have been forged-they are far too convincing as Messapic. The full formula occurs in 371 and 436 b; in 474 a 1 we find only klaohizis denO-n, with a lacuna which some scholars fill in with -avaso as to give denGavan. Klohizis by itself occurs in 476 and 515. In 371 the actual reading is not klaohizis, but klaohi, and I can see no reason, as I shall later explain, for altering the reading. First of all we must settle a very fundamental point-whether klohizis should be regarded as a noun or a verb. In order to do so, it is necessary to examine the word carefully in every context in which it occurs. Ribezzo regards it as a noun with the meaning 'princeps', but such tautology as we get from his version of 371,2 'Noster princeps regnans dux', seems very unusual in official documents, which are generally terse and concise. In 476 klohizis odatis ozar and 515 klohizis avilos Gotorridas ana aprodita apaogrebis it might equally well be regarded as either. When we come to 474, Ribezzo's attempt at translating the word as a noun, 'I1 sovrano imperante alla citth',3 seems very awkward. In 548

Abstract: An examination of the verbal forms derived from Indo-European roots with the gradations ei/i and eu/u (and sometimes ai/i) and ending in stops will show many thematic present stems with the strong grade of the radical vowel1 and not a few others characterized by a nasal infix in the root2. The nasal type is well represented in Latin and Lithuanian, and also in some of the Sanskrit verbs belonging to that class called by the Indian grammarians the sixth (e.g., muicdti, limpdti). The seventh Skt. class3 is not with any certainty represented outside of the Aryan group, and its relation to the thematic type is disputed, but it is made principally from roots of the phonetic pattern described, and must

Abstract: Is phonemic procedure arbitrary? It can be, but I submit that it need not be. If one defines a phoneme as 'one of an exhaustive list of classes of sounds in a language'1, one admits an endless variety of treatments and the choice of one or another treatment is arbitrary. The ideal of exhaustiveness is not in itself sufficient to define a unique scientific procedure. But if we also take the ideals of simplicity and self-consistency, we have the basis for a non-arbitrary method, particularly in view of the essentially systematic nature of phonetic totalities. The aim of a phonemic analysis may then be stated as: the smallest number of elements having the largest number of permutations and the most systematic relationships. To attain this ideal, it is necessary to consider always the totality2 of phenomena in the given language. A formulation as to any detail must be made on the basis of all pertinent and comparable data anywhere in the totality of the given phonetic system. In the present paper I propose to demonstrate this method in application to a selected type of phenomenon, long consonants, in different phonetic totalities.3