On 05-02-17 00:05, Miguel Carrasquer wrote:
> What was Pinault's law again?
In a nutshell, the syllabification of postconsonantal laryngeals with a
following *j and their coalescence with it, as in *krewh2-jo- > *krewjo-.
> But you know what I think of Caland-variants: they're nothing but
> **-n- palatalized (*-y-, *-i-), labialized (*-m-, *-w-, *-u-) or
> plain (*-n-, *-r-).
I know. However, I think Jens's explanation is more convincing -- not
without its little problems, to be sure, but I understand that it's
still work in progress. Both models share the merit of invoking just
phonetics and no magic, but Jens stays closer to orthodoxy by making
fewer assumptions about pre-PIE, which is the main reason for my current
> But Caland aside, I think that an /i/ was also present in the
> essive/fientive suffix *-eh1- (*-eh1i-). I see no other way to
> explain the Baltic and Slavic reflexes /e:/ < *-eh1- (aorist,
> infinitive), /i/ < *-h1i- (Baltic present) /i:/ < *-eih1- (Slavic
Do you see any difficulties with *-eh1-/*-h1-je/o- outside Balto-Slavic?
If not, it's a branch-internal problem and whatever the best solution
is, it probably describes a local innovation.
>> To my mind, <patni:> is inseparable from <taks.ni:> [sorry, I meant
>> <taks.n.i:>] and <ra:jn^i:>, which clearly have nothing to do with
>> oblique cases of neuter stems.
> I would agree if we had *pata: like we have taks.a: and ra:ja:.
But since we are dealing with a word of PIE date, not one derived within
Indic, it's enough if we have PIE *poth1on(t)-, which is just what Lat.
potens seems to reflect. I can account for the absence of *pata: in
Sanskrit -- the compositional form was _extremely_ frequent (note the
enormous number of compounds with <-pati->, quite a few of them securely
pre-Indic), so it simply replaced its free-standing precursor.
I've been looking for examples of words behaving like my putative
*poth1o:n. The 'deer' word *h1el-h1-o:n may be another substantivised
participle of a stative verb (from *h1el-u- 'reddish-brown'?), with
*h1el-h1-n-ih2 as its feminine counterpart.