From: Piotr Gasiorowski
Message: 6190
Date: 2001-02-22
----- Original Message -----From: tgpedersen@...Sent: Thursday, February 22, 2001 11:11 AMSubject: [tied] Re: Ingvar and Ivar> The existence of two stems <zan>/<zand> in Old High German intrigues me. Are they somehow correlated statistically with nominative/the oblique cases, such that <zan> is preferred in the nominative and <zand> in the oblique ones (and nominative plural)?I'm not a specialist in OHG, but I have friends who are, so I can make inquiries if necessary. My impression is that <zand> and <zan> are just free variants. Zand survives in German dialects and, e.g., in Letzeburgesch. I see no reason to regard this d-loss as more puzzling than the reduction of "kind friend" to "kine fren" in colloquial or regional English. The fact that this particular simplification is not often reflected in the spelling doesn't make it less trivial. English-speakers do it all the time, though there are only a few examples of -nd > -n perpetuated orthographically (but there are some, e.g. ME launde > lawn).What old nominative would <zan> reflect, anyway? The one you quoted (*do:n) is academic fiction supported by respectable but Hellenocentric IEists like Szemerényi (based on the analysis of Greek present participles, though even in Greek we find odous < *odons < *odonts). Even if *do:n were real, however, a form like <zan> could not be derived from it -- the vocalism is wrong. Actually, most comparative evidence points to PIE Nom.sg. *h1donts, with later branch-specific treatment of the affricate [ts]. For Proto-Germanic, one would expect Nom.sg. *tans, Acc.sg. *tanTum, Gen.sg. *tunTs (Germanic innovation), Dat.sg. tundi: (Verner's Law applies), Nom.pl. tanTiz, etc. One could perhaps make a case for OHG zan < *tans, but as there is no other evidence for the faithful preservation of root nominatives in Germanic, this smacks of fantasy, and a more down-to-earth explanation is preferable, IMO.Gothic retains traces of the zero grade (unfortunately, without Vernerian alternations, and the attested forms add up to only a fragment of the paradigm). It seems that a u-stem arose in Gothic due to the reanalysis of Acc.sg. *tanT-un (< *dont-m) and Acc.pl. *tunT-unz (= Gothic tunTuns < *dnt-ns). The generalistion of weak vocalism is untypical here, as is the complete switch to the u-declension, but the _partial_ replacement of root stems by u-forms is otherwise well evidenced in Germanic (cf. Gothic fo:tus 'foot' for "expected" *fo:s, but also OE hnutu 'nut', pl. hnyte < *xnut-u-z, *hnut-iz, or Old Icelandic u-umlaut in the "tooth" word: tönn- < *tanTu-). To sum up, Germanic languages often generalised the vocalism of the Nom.sg., but repaired the irregular Auslaut of that form by inserting a dummy "buffer vowel" (*-u-) after the stem-final consonant. For these reasons one would expect the Proto-NW-Germanic "tooth" pradigm to have the following forms: Nom.sg. *tanT-u-z, Acc. *tanT-u-n, Gen.sg. *tanT-s, Nom.p. tanT-iz, etc.Piotr