No CrossRef data available.
Article contents
The Origin of the Teutonic Weak Preterit
Published online by Cambridge University Press: 22 February 2021
Extract
The fect that the verbs m the Teutonic languages are divided into two classes according to the different formations of their preterits, gives the inquiry into the origin of the preterit of the so-called weak verb a peculiar interest to the student of Teutonic philology. The regular development of the strong preterit from the perfect active of the Aryan parent-speech has long been recognized, but the origin of the weak preterit is still an open question. Formerly the weak preterit was considered to be formed by the addition to the verbal stem of the Aryan root dhē (or as it was formerly put, dhā) “do.” Since the publication by Wilhelm Begemann of his two monographs ‘ Das schwache Präteritum der germaniscben Sprachen’ (Berlin 1873) and ‘ Zur Bedeutting des schwachen Preteritums der germanischen Sprachen’ (ib. 1874), this composition-theory has gradually been abandoned, for Begemann showed that such formations as mah-ta, kun-pa, wis-sa, etc. on the one hand, and nasi-da, habaida, fullnōda, on the other, did not, as the early theory assumed, originally contain a dh in their suffixes, but rather a t in accordance with the similar formation of their participles. (e. g. Goth. mah-t-s and mah-ta). However, his thesis, though supported with acumen and scholarship, was at first rejected by all who passed judgment upon it But it has come to honor since the partial adhesion of Windisch and the full adhesion of Möller, the former m Kuhn‘s Beiträge zur vergleichenden Sprachforschung, VIII (1876), p. 456f., and the latter in Kölbing‘s Englische Studien, III (1880), p. 160f. But the recognition of the fect that the characteristic of the weak preterit was originally a dental tenuis does not constitute a solution of the problem, but only the beginning of a solution. It still remains to inquire further into the source of this tense characteristic and to explain the striking similarity in the formation of the weak preterit and the weak participle.
- Type
- Research Article
- Information
- Copyright
- Copyright © Modern Language Association of America, 1888
References
1 Since this, Paul, indeed, (Beitr. ge, VII p. 136 f.) once more attempted to save the dk which was believed to have been the origin of the dental of the preterit. But his attempts have been success ully refuted by Möller, (ibid., p. 157, “Kunpa und das t-Prüteritum”).
2 We find the same problem in the case of the Irish t-preterite. But the latter does not concern us here, since there seems to be no historical connection with the Teutonic t-preterit, and the Irish, moreover, will hardly throw any light on the special conditions which pertain to the Teutonic weak preterit. But I should like to emphasize one point that it John Strachan (Bezzens, Beitr. XIII, p. 128, is right, then the t of the Celtic preterif must be interpreted according to principles similar to those which, later on, I employ in the ase of the dental of the Teutonic preterit.
3 Thus Möller, l.c., Kögel in Zs. f. d. Gymnasialwesen, XXXIV, p. 407 (not accessible to me), Kluge in Paul and Braune's Beitr., p. 155, Sikvers, ib., p. 561, Bremer, ib., IX, p. 34.
4 Miklosisch afterwards abandoned this opinion, the correctness of which is becoming more and more evident (cf. e. g. Bopp and Scherer in the passages above quoted, and Osthoff, ‘Perf.’ p. 191). In the second edition of his ‘Vergl. Gramm. d. slav. Spr.’ (III, 125) he declares the form vědě to be puzzling.
5 This is Stokes' opinion (in Kuhn's Beitr. z. Vergl. Sprachf. VII, p. 6.) and I agree with him rather than with Windisch (K. Z. XXVII, 163) who went back to the Old-Aryan medial present. Perhaps the primitive medial endings have been preserved in another place in the Irish verbal-system. The distinction between an ‘absolute’ and ‘conjunct’ inflection in Irish is, if I am right, independent of the existence or non-existence of a verbal particle, but finds its explanation in the fact that the absolute inflection is developed from the old middle and the conjunct from the active. At present it is not possible for me to pursue this discussion further and my remarks are only thrown out as a question, to which somebody else perhaps can give an answer.
6 Soon afterwards the same theory was proposed by Osthoff (‘Zur Gesch. d. Perf.’ 1884, p. 191 f.) and independently of Fick, as he says (p. 609). Osthoff also cites an essay by Speijer (Mém. de la soc. de ling. 5, p. 185 f.), which is not accessible to me. in which the – of the Latin perfect is explained in the same way. The fact that the same idea has been expressed independently by three different persons adds to the probability of its being the correct one. In this connection Fick's hypothesis of the Latin v perfect (l. c., p. 594 f.) may be mentioned, according to which in forms like plèv-i, gnlv-i v is identical with the in Old Ind. pa-prdu, ja-jndu. Different views are expressed by Osthoff (‘Perf.’ p. 250 f.) and Stolz (Iw. Müller's ‘Handbuch d. klass. Altertumswiss.’ 2, p. 231) who rather assume that the v perfect has been produced by late analogical formation according to certain u-roots, without bringing forward a valid objection to Fick's opinion; just as on the other hand G. Curtius (Berichté d. S chs. Ges. d. Wiss., Phil.-hist. Cl. 1886, p. 421 f.) and W. Schulze (K. Z. XXVIII, p. 266 f.) who derived the v-perfect from the old perfect participle in a round-about way without even mentioning Fick's more simple theory. If, as I do not doubt, Fick's supposition is in accordance with the facts, then also the tensecharacteristic of the Latin v-perfect has been developed from a part of the ending. In addition, another remark may be in place. In Old Ind. (that is in the Vedas) the 3d Sing. perfect of stems in -a (as da, pra, etc.) generally ends in -au (thus dadau, papr$au) in agreement with classic Sanscrit, but also rarely the Special Vedic ending -a is met with (paprä R. V. I, 69, 2; jaka R. V. VIII, 45, 37 according to Delbrück ‘Old Ind. Verb.’ p. 59). Hence there is a fluctuation betwen -an and -a without a corresponding difference in meaning, just as in the dual (avau and ava, etc.). At length Meringer, in an excellent monograph (K. Z. 28, p. 217 f.), has given the long desired explanation of the dual-forms. The change between -äu (-äv) and ä belongs as M. has shown, to the Sandhi-phenomena. The first ending is the older one; it has been retained before vowels in the primitive language. Before consonants the second ending has been developed by the elision of the v. That the change between -äu and ä in the perf. should be considered in the same light, seems to be so natural that I should not think it worthy of special mention if I had not seen that Meringer (p. 218, note) asserts that in the Rigveda the au of the dual is different from that of the perfect, and that Brugmann (‘Grundriss,’ p. 490 f.) is inclined to extend Meringer's explanation to locatives like agnau and agnä but not to the forms of the perfect. The difference which Meringer finds between the treatment of the dual and perfect in the Rigveda is easily explained, if we suppose that the falling together of the two forms has taken place in the perfect earlier than in the dual; so that in the Vedas the development which seems to be complacted in classic Sanscrit, can be less distinctly traced in the perfect than in the dual. The reason why the falling together occurred earlier in the one case than in the other is evidently this: that the perfect forms were more seldom used. From the Statistics of Avery, J. A. O. S. X, p. 250 and Lanman, id. p. 340 the following relation is seen:
that is, one perfect to about twenty-seven duals. And Schleicher (‘Die deutsche Sprache,‘ p. 61.) has already remarked that those forms which are most rarely used submit ost readily to analogy and the tendency to simplification.
7 In a similar manner the above mentioned transposition from the active perfect into an inflection composed of active and medial elements, will have to be regarded in Latin. Evidently in the middle l-endings the temporal function prevailed, while, on the other hand, the function of the medio-passive withouttense-meaning was exclusively conferred upon the r-endings and extended the sphere of the latter beyond its original boundaries.
8 Concerning the treatment of final ai in Teutonic, compare further: Scherer, Z.G.D. S. 2, 202, 25, 609: Braune, P.-B. Beitr. II, 161 fol.; Paul ibid. 339 fol. and IV, 452 fol.; Leskien ‘Decl. im Slav.-Lit. u. Germ.’ 126 f.; Mahlow, ‘D. langen Vocle’ 53 f. and 94 f.; Brugmann, in his ‘Grundriss,’ p. 518, has followed the view of Paul, without paying any attention to Schmidt's essay. But the explanation of the weak preterit advanced in this paper, if I am not mistaken, settles the question in favor of the opinion held by Mahlow and Schmidt.
9 The older views about Goth. iddja are found in Scherer Z.G.D.S. 204 — 3242 and note Begemann, ‘Prät.’ p. 67 ff. Since then Möller K. Z. XXIV, 432 note and Kluge, ‘Germ. Conjug;’ p. 125 ff. proposed to identify iddja with Old lnd. yäm, 3. sing. áyät and their opinion in the meanwhile has been adopted by several scholars. (Cf., for example, Bremer P.-B. Beitr. XI, 55 and Brusmann ‘Grundris’ p. 128, 516). The explanation given above avoids the supposition connected with the theory of Möller-Kluge that the old augment was preserved in Teutonic exceptionally in this case alone.
10 It has long been recognized that the first syllable of this form preserves the old reduplication. The explanation of the stem syllable has been so far through the Old. Ind. or Irana, perf. act. (e. g. Bopp, ‘Vgl. Gramm.’ II2506; Windisch, K. Beitr. VII, p. 459; Paul, P.-B. Beitr. IV, 464 f., Kluge ‘Germ. conjug.’ p. 103 f.)or through the reduplicating imperfect of the active (Bezzenberger, Ztschr. f. dt. Philol. V, p. 475; Möller, Engl. Stud. III p. 159 and P.-B. Beitr. VII p. 469).
11 With ved. dadhé, Teut. *dedai, perhaps we are allowed to identify also the form dede “fecit” or “posuit,” which occurs three times in Old Gaulic inscriptions [cf. Stokes in Bezzenb. Beitr XI, p. 124, 125, 128 and 157]. It seems that final ai in Celtic changed at an early date to e [through ae] and afterward to e.
12 In other words: The -ai and -tai in the primary ending of the 3. sing. middle constitute an original difference between the ö- and mi- conjugation. If this view be correct, and it is supported by the Old. Ind. inflection, Brugmann's suggestion (‘Morph. Untersuch.’ I, p. 13, note and p. 147) that in the 1. sing. -mai originally was the ending of “nonthematic,” and -ai that of the “thematic” stems, becomes improbable. In accordance with the endings of the 3. sing. we would rather expect the reverse. In my opinion the 1. sing. is sufficiently explained by assuming that in the middle it had the final -ai (without preceding -au) throughout. The Greek - - would then be considered as a new formation in analogy to - and -, just as -, - (), - of the active and this first appeared after that the of the 3. sing. of the thematic formation had become universal.
13 In accordance with this Begemann's participle-theory is partly substantiated. But it is to be distinguished whether we derive the weak preterit directly from the participle, or attribute to the latter only a limited influence upon the development of the preterit.
14 Cf. examples in Noreen, 'Altn. Gr., §448, note 1.
15 Cf. the passages quoted above p. 198, Note 3.
16 In Aarb g. f. nord, Oldk. og Hist. 1869, as I see from the account of Mrius, K. Z. XIX, 212. The objections made by Paul in his and Braune's Beitr. IV, 464, against this view are not, as it seems to me, of great importance. Paul replies, in the first place, that language tends more to compound old distinctions than to create new ones. My opinion has always been that differation is a factor no less important in the development of speech than analogy, and that both go together in every period of the history of language. But even in case I could agree with Paul's theoretical point of view, that would not hinder me from accepting Gislason's explanation, since the latter implies that the indicative and subjunctive-endings in this instance had been made the same.
17 V. Fierlinger, as far as I know, is the last scholar who discussed this point. K.Z., XXVII, 430 f.