From: gprosti
Message: 71587
Date: 2013-11-13
>True, but "can be shown to have divergent origins" contains a potentially vast spectrum of plausibilty. E.g., it's one thing to show that OHG kuning and Finn. kuningas could each be formed from clearly discernible elements or through clearly productive morphological processes within their respective languages. It's another thing to derive one or both of these words using rare, semantically obscure roots or affixes and use this as a basis for doubting their common descent.
> > ---In cybalist@yahoogroups.com, <gprosti@> wrote:
> >
> > I'm not sure what you mean by "regular system of diatopic variation", but if you have a
> > set of words with a sufficient amount of shared phonetic material, plus matching
> > semantics, this overrides the criterion of regular phonetic correspondence when drawing
> > a connection between two or more forms.
>
> Which itself can be overriden if the similarities can be shown to have divergent origins, of course. In this case that'd require a whole bunch of corroborating evidence for the model of forming numerals as "largest numeral not yet named", though. The best precedent I can think for anything along these lines is from the set-theoretical construction of ordinal numbers, a bit advanced for a supposed pre-HG origin :)
>
> I'm going to wonder if analogy may be again at work, perhaps this time starting from the rather similar _Dutzend_. Does this have related /t/-initial forms that could have influenced the appearence of the same in _tausend_?
>
> _j.
>