Re: [tied] *kap-

Piotr Gasiorowski wrote:

> Grzegorz Jagodzinski wrote:
>
> [after quoting many examples]
>> And finally, because you do not believe in possibility of irregular
>> development of 'to catch' due to frequency, how do you explain
>> colloquial Polish trzym instead of trzymaj 'catch!, imper. 2 sg.'?
>
> Hey, spoko[*], aren't you fighting a straw man? Where did I say I
> didn't believe in a correlation between irregular development and
> frequency of use? For the record, I do.

My example is an argument that words meaning "catch" are enough frequent to
develop irregularly.

> The difference between us is
> just a matter of heuristic approach: I object to using "irregular
> pleading" if other possibilities haven't been exhausted.

There is no way to exhaust any possibilities if you state a priori that two
words are not related.

> One has to
> remember that the frequency of use can scarcely be estimated for PIE.
> *h1esti and *h1senti were probably amongst the most frequently used
> words in the protolanguage, but they somehow managed to survive with
> little deformation,

Really? English "is" without -t, Czech "je", Polish "jest" instead of the
expected <jes'c'>... The word was too short to undergo more severe changes.
And what concerning *H1s-enti ~ *H1s-onti: how many verbal forms with real
null-grade do you know? "Real" means no vowels, just like in *H1s-. I am
under impression that nearly all CVC roots has a normal or reduced grade
instead of null-grade (<I> or <U> in Slavic, <a> in Latin and Greek), except
instances of reduplication etc. And *H1es- / *H1s- is the only exception,
cf. *H1éd-mi 'I eat' and *H1ed-énti 'they eat' instead of **H1d-énti.

> so I'm not happy when somebody almost takes
> irregular development for granted for any lexeme that _might_ have
> been frequent.

It is not taking irregular development for any lexeme but for a lexeme whose
counterparts do undergo irregular development.

> Germanic *sunuz and *wiraz show an irregularly
> shortened vowel -- why? Because they were frequently used?

Of course yes. See also Greek hyiys ~ hyios (declension tending to be more
"regular"), Avestan hunus^ (short u like in Gmc.) and German Sohn without
the expected -e < -u, French fils [fis] without [l] (--> Engl. -fitz). The
word "son" was used frequently in patronymica of the type "X Y's son" (not
being true compounds). In Slavic another type of patronymica was in use (see
Russian otchestvo, without the word "son") and indeed, we do not observe the
irregular shortening in Slavic.

> But so
> were many other words that don't display any such shortening. And
> *suHnu- yields regular reflexes outside of Germanic despite being
> presumably so common. Perhaps there is some other reason for the
> shortening, e.g. frequent occurrence in compounds, where such
> clipping is regular (cf. Skt. viraps'á- < *wi[h1]ro-pk^[w]-ó-).

Shortening in compounds may also be due to frequency. But compounds are long
enough to be shortened even not being _very_ frequent (their frequency is
not very high but too high to tolerate such long words). And why *wiraz only
in Gmc.? Of course not only, also in Latin (vir), Old Irish (fer). Why not
shortened in Baltic and in Indo-Iranian (except compounds), we can only
guess (perhaps less frequent).

Grzegorz J.



___________________________________________________________
To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com