Re: [tied] Re: IE right & 10

On 9/1/04 2:14 PM, Piotr Gasiorowski wrote:

> On 8/29/04 11:50 PM, Richard Wordingham wrote:
>
>
>>Well, if one starts counting on the fingers of the left hand, *dek^m
>>or *dek^mt '10' might have meant something like 'right hand full'
>>or 'rightmost'. With the former meaning, /mt/ _might_ be
>>*met 'measure'. With the latter meaning, /m/ might be the
>>superlative suffix. However, why then do we have *dek^m or *dek^mt
>>and not *dek^sm or *dek^smt for '10'?
>
>
> Assuming that the *-s- of *dek^s- is some kind of detachable suffix, and
> that *dek^- is an acceptable combinative form, one would expect, in a
> hypothetical compound with *met-, *dék^-mot- in the strong cases, with
> *dek^m.t- as its weak allomoprph. Why then do we have *-(d)k^omt- in the
> decadic numerals? It seems to rule out *-m(e)t-.
>
> Piotr

An afterthought: if one wants *dék^m.t to be an analysable compound, the
only possibility I can see is *dék^-h1m.t- (gen.pl. *dk^-h1m.t-óm,
compositional collective or animate stem *'-(d)k^-h1omt-). The second
element could be *-h1m.-t-, an extended root noun derived from *h1em-
'take, get' (the *-t- extension is normal after root-final sonorants and
laryngeals, cf. *-gWm.t- in compounds), with the approximate meaning
'taking'. What we gain is a natural explanation of the heterorganic
sequence *-mt- and of the early disappearance of the initial *d- in
*dk^-. Before a vowel we would expect a "thorny" treatment of *tk^- <
*dk^-, but if a consonant (here, *h1) follows, the expected outcome
involves the loss of the initial stop! I'm beginning to like this idea.

Piotr