No CrossRef data available.
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 31 August 2011
In an article entitled “The Role of Loki in Germanic Mythology” published in this Review, the late F. Stanton Cawley, acting upon a suggestion made in a Harvard dissertation of 1938 by Dean Gilbert T. Hoag of Kenyon College, attempted to show that þjazi, the name of the giant who in an Old Norse myth carries off Iðunn and her apples, could be etymologically connected with Tvaṣṭṛ, the name of one of the chief Vedic gods. The philological reasoning which led him to this conclusion he published also in slightly more extended form, and for a rather larger public, in Paul & Braune's Beiträge under the heading of “Loki und 
, ein bisher unbekannter indogermanischer Gott.” The conclusion of these studies was stated as follows: “An Indo-European myth of the theft of the divine drink by the god , of the culture-hero type, was transmitted in India and in Germanic territory to the literary period.
1 Volume XXXII, No. 4, October, 1939, 309–327.
2 Entitled Two Norse Myths of the Thieving Hawk. (Typewritten copy deposited in the Archives of Harvard University, Widener Library.) Nothing in this article is intended to dispute the findings of Hoag's dissertation. The denial of any etymological connection between þjazi and Tvaṣṭṛ in no way disturbs the relationship between the two Norse myths.
3 63 Band, 3 Heft, 1939, 457–465.
4 The Harvard Theological Review, XXXII, 325–326.
5 I deeply regret not having expressed my long-held doubts regarding this matter to Professor Cawley, who was my teacher, and a kind friend. It was not, however, until a few weeks ago that I found it imperative to examine scientifically the connection of Norse þjazi with Sanskrit Tvaṣṭṛ. Had Professor Cawley been alive, this article would have gone first to him for a reply. His untimely death has prevented that duty.
6 Cf. Walde-Pokorny, Etymologisches Wörterbuch, 1, 717, for other derivatives of IE 
, among them Sanskrit Tvaṣṭṛ.
7 Cf. Brugmann, Kurze Vergleichende Grammatik, page 332, paragraph 409, 3.
8 Cf. Cawley's note 5 to page 462 of his article in Beiträge: “Fr. Kluge, Nom. Stammbildunslehre # 30: ‘Das Bildungselement in ags. ealdor, bealdor, “Fürst” — an. baldr (mit gramm. Wechsel zu germ. balÞa - “kühn” gebildet) enthält wahrscheinlich das in den verwandten Sprachen weitverbreitete Suffix tṛ, das zur Bildung von nom. agent. diente; dazu auch ahd. smeidar “artifex”?.” But even here [my comment, M.F.] the perfectly regular development and inflection of the noun Baldr, likewise the name of a god, should be observed (i.e. gen. Baldrs, dat. Baldri).
9 The Harvard Theological Review, XXXII, 323.
10 An example of “u-breaking” may be found in the interesting divine name Fiǫrgynn (Lokasenna, 26, in the genitive case); or in ON fiǫr compared with Gothic faírhnis ‘life.’ The phenomenon is a common one: cf. Heusler, Altisländisches Elementarbuch (3 Auflage), page 24, paragraph 76; Noreen, Altisländische Grammatik (4 Auflage), page 87, paragraph 89. The purely hypothetical form *þiǫhst(u)ra(z), written (thus) with all sounds remaining, is impossible at any stage of the language. It is used here only to follow Cawley's argument at this point. (No such form is imputed to him.) The proper complex of sounds which would occur at the period of breaking is given below in the discussion of the sequence of sound-changes.
11 Analogy is, even at the very best, a weak and despairing last resort. The strictest philological method will not admit it as proof unless the conditions which would bring it about can be so completely described as to show its inevitability. In this instance, since the suggested adaptation is of one hypothetical form (*þehsturaz) to another (*þunr-aR), these conditions cannot possibly be depicted.
12 At this point there are many difficulties: (1) The exact quality of the sound represented by the IE symbol þ after 
is not known (cf. Brugmann, Grundriss, I, 2, page 790, paragraph 919), and, consequently, whether in the Germanic languages its derivative is to be classified with other s-sounds remains quite speculative; therefore, (2) it is not certain whether the h in *þehst- is to be treated precisely as if occurring in any other -hst- complex wherein the s is of different origin. (3) Whether h before -st would fall without compensatory lengthening of the preceding vowel is likewise debatable on the grounds of insufficient evidence. (Cf. Jakob Sverdrūp, IF, 35, 1915, page 163; Noreen, op. cit., page 163, paragraph 222, 2; Heusler (who expresses greater certainty), op. cit., page 52, paragraph 168.) The most likely assumption is that the h would fall without any compensatory reaction elsewhere; but it remains only an assumption. (In the concluding argument of this paper allowance is made for the possibility of lengthening of the preceding vowel at some stage in the process.) For the dating of the change see the reference to Noreen above.
13 For the normal sequence of ON syncope cf. Heusler, op. cit., page 34, paragraph 108, and the further references there given. It should perhaps be noted that, according to this order of changes, a would be lost without causing umlaut of the preceding u because the primary accent would not be on the u. The parallel is exact in ON jǫtunn (quoted below). From the resulting form *þesturr an appeal to adaptation to *þunr-aR can, of course, no longer be made.
14 Cf. Noreen, op. cit., page 87, paragraph 89.
15 I record my indebtedness to the American Council of Learned Societies for a study aid which has made possible (among other things) the preparation of this note.