Re: [tied] Re: Caland [was -m (-n)?]

On Thu, 27 Nov 2003, Miguel Carrasquer wrote:

>
> It just occurred to me that there may be a problem here...  Suppose the
> rule (let's call it the V/R-rule) worked for quite some time, separating
> V-roots with suffix -ró and R-roots with suffix -ú.  In the end, we're
> left
> with 90% separation according to the rule, and 10% doublets, exceptions
> etc.  All of this, I suppose, before, maybe long before, the breakup of
> PIE.

Yeah, that's what I meant.

> But then it's hard to explain why Tocharian replaced _all_ u-stem
> adjectives by ró-adjectives.  I mean, what possible connection could
> there
> be, in the mind of a post-V/R-rule Indoeuropean, between adjectives in -ú
> and adjectives in -ró?  Can a mere 10% relict forms keep such a
> phonologically non-obvious historical connection alive synchronically?

You seem to forget about all the many ro-adjectives that were in keeping
with the rule. That raises the percentage to a lot more than ten.
Lithuanian has done the opposite: *dhub-ró- -> dubùs, so there must have
been an association.

> What this suggests to me is the possibility of an alternative scenario,
> where _both_ forms, -ú (< **-úr) and -ró (< **-ur-ó-), were still fully
> functional and productive in pre-breakup PIE.  They were also essentially
> synonymous, which is why there was a tendency over time to eliminate one
> of
> the two.  In Tocharian this took the form of eliminating all the u-forms.
> Elsewhere, and I mean all groups independently of each other, a natural
> tendency to avoid too cumbersome consonant clusters, and the general
> tendency to prefer thematic forms, may have conspired to filter out
> ro-forms in the R-roots, and u-forms in the V-roots, respectively.

I will assess the probability that all known IE language groups (or all
except some) acted in parallel fashion to produce the picture af
near-complementary distribution between ro- and u-adjectives as very
close to zero.

Jens