From: dgkilday57
Message: 70046
Date: 2012-09-07
>Krahe used a non-laryngeal formalism, and I suspect that he did not always reconstruct the correct vowel with rhotic suffixes. In his master list in "Struktur", he has two examples of *-ara with explicit /e/-grade in the root: *Eisara (Ysere gen. 859, Isera 1104/1170, Isara 1119/1173, Ysara 1119, now the Ijzer in Flanders) and *Medara (now the Metter in Wuerttemberg). But these names could be reconstructed as well, or better, as *Eisera and *Medera. They could thus be considered substantivized feminine adjectives of the type with *-ero- appended to an /e/-grade root, like Greek <eleu'theros> 'free' from *h1leudH-. On the other hand, river-names with true *-ara and zero-grade roots can correspond to adjectives with *-h2ro'- appended, like <hudaro's> 'humid' (Hes.) from *ud- (full grade *wed-).
> [...]
>
> Memory failed me. It was the Eder, not the Oder, which was anciently called Adrana (Tac. Ann. 1:56). The old references to the Oder are Odagra/Odogra 892, Odera 940/983, Adora 968, and Oddara/Oddora (Adam of Bremen, ca. 1075). My guess is that -gr- in the oldest is graphy for guttural /r/ by a scribe whose native /r/ was dental. I am not familiar with the evidence mentioned by Piotr for anaptyxis in these forms. At any rate Krahe reconstructed the protoform as *Adara in the most comprehensive presentation of his theory, "Die Struktur der alteur. Hydronomie" (Abh. Akad. Wiss. Mainz, Geistes- und sozialwiss. Kl., Jg. 1962, Nr. 5, 285-342, esp. 305-6). Krahe had some other hydronyms in *-ara, so from his viewpoint no anaptyxis was necessary to derive it from the more common *Adra. In principle the *Ad- could represent either full-grade or zero-grade of *h2ad- or *h4ad-. I suspect that the root is indeed represented in Sanskrit, forming the suppletive inst. pl. <adbhis> and dat./abl. pl. <adbhyas> of <ap-> 'water'. (Or perhaps the suppletion was in the opposite direction, due to bizarre regular forms of a C-stem *ad-?)