clearly an adverb formed with the ablatival -ta = Skt. -tas from a pronominal stem ama-, corresponding to the rare Skt. ama-"this” occurring in a formula of the Atharva Veda and the Brāhmaṇas, amo 'ham asmi, sā tvam "he am I, she thou" (AV 14. 2. 71, with Whitney's translation), and further attested by the adverbs amā "at home" ("chez lui') in RV, AV, etc, and amāt “from this place, from here” RV 5. 53. 8.

11. 6,8. Smith transcribes tyamiy but utāmaiy. The text has in both cases moiy, not m'iy. Herzfeld transcribes tyamiy, utamiy, the first incorrectly according to his system, since there is no a-sign after the ya.

1. 7. Smith omits bagānām in his transcription and even remarks that the Persian version has “the greatest” simply, in contrast to the other versions. But his own copy, confirmed by the photograph, shows bagānām. The phrase hya mabišta bagānām is identical with that in Dar. Pers. d. 1, Xerx. Elv., Xerx. Van, and the Elamite and Babylonian versions conform to it. The corresponding Elamite phrase is the same as in Xerx. Elv. and Xerx. Van. That of the Babylonian version differs from that in Xerx. Elv. and Xerx. Van, and also from that in Dar. Pers. g 1, but is equivalent to them in meaning. Smith says that "the obscure word alla here seems to mean the other part, remainder."" On this my colleague Professor Luckenbill furnishes the following note:

ša rabû allā ilānipl of 1. 7 is the equivalent of ša rabû eli ilānipl, "who (is) great above (i.e., greater than, or greatest of) the gods." alla is undoubtedly the cuneiform rendering of the Aramaic 'al. The use of eli in Assyrian for "over, more than" is well established. In Xerxes Elvend and Xerxes Van the same idea of the superlative is expressed by the more common rabû ša ilānip, the "great one of the gods," while in Darius Pers. g we have a phrase, which has the same force as eli. al-la occurs in the Babylonian version of the Behistun Inscription, 1.29 (Š XVI).-"Darius the king thus speaks: al-la ša anaku adūku ana Gumātu Magušu, after I had slain Gaumāta the Magian." Here alla ša is the Aramaic 'al še.

‘ See also Peiser, Babylonische Verträge 230 and Tallqvist, Die Sprache der Contracts Nabû-nā'ids 42, for the use of alla in the

Neo-Babylonian and Persian period contracts. 1. 8 mām auramazdā pātuv, as in many parallel passages, while others have mām Auramazdā pātuv hadā bagaibiš. In this inscription the Babylonian version agrees with the latter, the Elamite with the former.

dam, just like mittumanna in NR a 23. It is the kuiš, the regular equivalent of OP yātā, which here in both cases answers to the whole Persian phrase amata yālā ā.

CATELLUS, CANICULUS. A CASE OF SEMANTIC

INTERCHANGE

A. H. SCHUTZ

UNIVERSITY OF MISSOURI

Of special interest in the Romance of Gaul and Spain are the descendants of catellus and caniculus. Within each group, as we shall see, there is considerable fluidity of sense, to such a degree that they come to be used interchangeably for 'kitten' and 'puppy.' Of this state of affairs, the grammarians and lexicographers seem to be unaware.

At least as early as the third century A.D., catellus! was a serious rival of catulus, if indeed it did not threaten the existence of the latter. Probus records catulus non catellus, though Baehrens at one point rather curiously puts both in brackets in his edition of the Appendix, p. 6. Judging by the subsequent fortunes of both, catellus seems to have won its battle, certainly in the territory we are considering, for the -ulus variant seems to have left no descendants. On the other hand, to list all the forms of catellus would require not a little space, as the sources referred to will readily show; a certain number of specimens are here cited: cael, chael, keel, chiau, chaon, chaillon for the Old French examples; among the patois of the northern regions, chaé, ché, chiau, chiaule, quiaule. We may add a few forms that occur in the South of France:

1 The problem of the single or double t in cat(t)us is discussed by Garcia de Diego, Contribución al diccionario hispánico etimológico, g111.

2 A curious juxtaposition of the two is found in the following account (Osbertus de Miraculis Sancti Dunstani, cit. Du Cange) of an exorcism: 'et quod in modum parvuli Cati discurreret Francigena lingua dicentibus, ille contra, qui linguae ipsius omnimodis inscius erat, subridens, eadem lingua similiter verbo diminutivo consonanter respondebat dicens: Non ut catulus, sed ut catellus'.

3 As to the form in -ellus being the less classical, Baehrens says: “Unberichtigt ist die Zurückweisung von catellus” (Sprachlicher Komment. 2. App. Probi, 121). The Thesaurus Ling. Lat. gives both, with catellus marked 'a catulus deminutive.' We have in Varro, LL 9. 74, both forms: ‘ut est cista cistula cistella et canis catulus catellus' cit. LANGUAGE 2. 187 (Kent). See below for association of canis and catulus. 4 Mémoires de la Soc. de Ling. de Paris 14.211 (Sainéan).

chadel, kàdèl (in Pyrén. Orient.), kàdèu (in Bouches du Rhône), tchàdèl (in Lozère), kòdèl, kòdèo (Cantal). In Spain we find Old Cast. cadiello, Aragonese cadillo, Catalan cadell (and Spanish America has cadejo, related to these forms)."

All these are listed under the general sense 'little dog,' but the meaning of catellus and catulus as far back as classical Latin was much wider in scope, for it could signify the young of the wolf, lion or similar animal; extension of this idea even permitted application of the term to the young of the rabbit. These numerous interpretations lived on. Godefroy (Dict. Anc. Frçs.) speaks not only of the cayeau du lion and the chael of the fox (see also Renart 896-7, ed. Martin), but refers in this manner to children:

Par la bataille ez vos poignant Borrel

O lui estoient si XIII chael,

Tuit chevalier adoubé de novel. (Aleschans 6256)

as also La Curne de Ste. Palaye: 'De cibis delicatis pascebant catulos suos quos de turpibus concubinis, ipsi turpiores procreabant' (J. de Vitri). Catulaster is defined in the Codex Cassinensis as 'Iuvenis duodecim annorum'. In Old Provençal, cadelet means 'young dog' but also 'young lion cub'. Whether so great a diversity of meanings has been carried forward into modern dialects is difficult to say with our present facilities, but many things lead one to believe that it has indeed been the case. García de Diego (§109) defines Catalan cadell 'flores de algunos árboles' and Arag. cadillo as 'flor de olivo'; Mistral (Trésor) renders cadelas as 'jeune et gros chien' and 'grand jeune homme qui fait l'enfant' and gives cadeliero as 'vache portière qui porte chaque année comme les chiennes.'

Canis (aside from numerous metaphoric values" that it may have,

Boucoiran, Dict. des idiomes méridionaux.

Atlas Linguistique 1789.

7 García de Diego, op. cit. §109: 'significa un animal fantástico que la gente supersticiosa se representa como un enorme perro negro'. Cadejo<*catic(u)lu(m), just as caneja<canic(u)la (Roman. Etym. Wbch. §1586).

* Varro, RR 3.4, speaking of rabbits and their prolific reproduction says: 'etenim cum habent catulos recentes, alios in ventre habere reperiuntur.' 'Corpus Gloss. Lat., 5.550 (Götz).

10 Levy, Supplement-Wörterbuch; see also Roman. Etym. Wbch. 1763.

11 E.g. Sainéan's list of terms derived from the idea ‘dog' and applied to machinery (loc. cit.). In Spanish gato similarly takes on a number of meanings, being applied to 'pocketbook', 'iron hook', 'block and tackle' (Dicc. de la Real Acad. Española, 1925).

resembling cattus in this respect) can be applied to other animals. That Phaedrus uses it as 'wolf' is natural, considering the close relation of the species. It is more curious to note the Spanish folk use of perra

12 chica and perra gorda for five and ten céntimo coppers, referring to the lions that adorn them. As in the case of catellus, there exist a number of derivatives of canis in Romance, e.g. French chenet, OF chenel (Godefroy), chienetel (ibid.), Fr. dialect chienneton, Prov. canilho, chenilho, (Mistral), Sp. canijo.13 A glance at Fr. chenille shows that, like the derivatives of cattus, possibly even to a more diverse extent, the formations on canis show a great variety of significance: Roman. Etym. Wbch. mentions kanaya (in Tessin canton, Switzerland), rendered 'Kinder' and Tyrolese kanai 'Knabe.' Thus the two major groups, the one based on canis and the other on cattus, have, since early times, followed, from the semantic point of view, a similar and at times common road. Already the grammarian Virgilius, in discussing gender, speaks as follows: '. . quidam simpliciter dixerunt quod masculinim hoc esse debuit quod secundum habitudinem corporis ostenderetur et ita femininum ut vir et mulier, taurus et vacca, aries et ouis, canis et cata et cetera animalia.'14 Hence an expression like Mistral's La chino dou segnour a cadela 'the seigneur's dog has puppies' need not seem extraordinary. It is, however, more curious to find under the heading Petit Chat, in the Atlas Linguistique, Map 1498, the forms kanālos (fem.), kànile (id.), chenil, chnil, all obviously from canis. That such interchange of meanings should have occurred is the almost inevitable consequence of semantic kinship aided by the lack of fixity within each group

One more remark, in this connection, apropos of Mistral's translation of fa de-catetos 'en baissant la tête, en rampant avec un air d'humilité.' Was he aware that this is more the manner of a dog than of a cat? Certain it is that cat-de-mar is given as 'chien de mer' without comment.' In the last instance, the confusion may be of long standing, because caniculus and catulus are both rendered 'piscis' by the Latin glossators.16 THE INSTRUMENTAL AND THE COMITATIVE IN THE

ALEUT LANGUAGE

WALDEMAR JOCHELSON

There are in Aleut' three endings to indicate the number of nouns: t' for the singular, x for the dual, and n (in the Atka dialect, s) for the plural.

A'dax' father, a'dax two fathers, a'dan (Atka a'das) fathers.

The singular has two cases, the absolute and the relative (or the nominative and the genitive). The absolute may serve in the sentence as subject or object (direct). All other relations between nouns and other words (with the exception of the instrumental and the comitative forms, for which see further) are expressed by the relative case and by postpositional pronouns added to the end of the noun. The dual and plural have only one case, the absolute, as the duality or plurality of the compound is expressed not in the noun but in the postpositional pronouns. The suffix of the relative case is m instead of the r' of the

x absolute case.

Absolute case: a'dax' father. Relative case; a'dam of the father. a'dam-i'lan of the father to him,'i.e. to the father. a'dam-ilā'n 'of the father from him,'i.e. from the father. a'dam-ili'kin to two fathers. a'dam-ili'ñin to the fathers. cam-ku'gan 'of the hand on it,' i.e. on the hand. cam-kugā'n 'of the hand from it,' i.e. from the hand. Instrumental. The element sa, expressing the instrumental, is suffixed

i The Aleut is a dialect of the Eskimo. The author studied the Aleut language (in three dialects) while spending about two years (1909–10) in the Aleutian Islands on behalf of the Russian Geographic Society. Archaeological Investigations in the Aleutian Islands were published by the Carnegie Institution of Washington, 1925. The Aleut folklore (texts and translation), grammar, and vocabulary, are in preparation for the Publications of Columbia University, New York, under the editorship of Prof. Franz Boas. The material on the somatology, ethnology, and sociology of the Aleut is also in preparation. In this article g is used instead of g with a dot beneath to represent an uvular (g).