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?
Numerous examples are the best method to prove our points of view...
Irregular development due to frequency was one of the possible explanations
for the irregular development of IE roots *kVp-, *ghVbh-, *ghVb-, *gVb- etc.
in whose relationship you do not believe.
Especially, Gmc. *habe:- and Latin *habe:- are as if unrelated because the
former comes from *kape:- while the latter from *ghab(h)e:-. I reject this
statemant because of their meaning but of course it is not enough. As the
most spread view says, words changed regularily, obeing sound rules
(popularily called "phonetic laws"). In particular, neither PIE *gh- may
have yielded Gmc. h- nor PIE *k may have yielded Latin h-. Well, let's check
this out.
1) Skr. hr.d- 'heart' < *g^hr.d- and Slavic sIrdce, Greek kear ~ ke:r,
kardia, Lat. cord-, Arm. sirt, Goth. haírto < *k^r.d-. Two roots or one?
Note also Skr. çraddha: 'believer' and Avestan zrazda:- 'religious' (g^h :
k^). The usual reconstruction *kr.d- is unsuitable for Greek ke:r, and
especially for kear (< *keH2ar ?). Attested Hittite forms are interesting as
well: nominative gir, genitive kardijas^.
2) Skr. aham 'I' < *H1eg^h- and Slavic (j)azU < *je^zU, Goth. ik, Lat. ego:
< *H1eg^- (g^h : g)
3) Skr. duhitar- 'daughter' < *dhug^hHter- - but Avestan dugEdar- <
dhugHder-, Greek thygate:r < *dhugHter-, Lith. dukte., Slavic *dUtji <
*dUkti < *dhugte:r < *dhugHte:r (g^h : g and t : d)
4) Greek dekhomai and dekomai (dialectal variation) 'I take' and Engl. take
< PIE *dVgh-, *dVk-, *dVg- (the last one "prohibited").
5) Greek klagge: [-Ng-] 'sound, tone', Lat. clangere, Lith. klage.ti 'to
cackle' < *kla(n)g- - but Irish clocc < *klank-. We may consider this root
onomatopoeic but it is present in numerous IE languages. Which is more, OCS
(Ruthenian redaction) shows both klegUtati and klekUtati < *klegh- ~ *klek-
(*gh because of Winter's rule); finally we have k : g : gh here.
6) Polish paprac' and babrac' 'to smear, to stain', cf. Hitt. paprai- (b(h)
: p). Compare also similar examples like badyl 'stalk, weed' - patyk 'stick,
twig', or ba,bel 'bubble, blister' < pa,pyl (attested in Old Polish) and
modern pe,pek 'navel, umbilicus'.
Not all of these examples can be called "expressive", whatever it means. If
we admitted *g^hr.d- < *k^r.d-, why not also *ghab- < *kap-? (Slavic gabati,
Lith. gobti may suggest *-b-, not *-bh-). Really, I cannot see any rational
reason. Please, give me one: why we should allow irregular development of
'heart' but not of 'have'.
Such alternations seem to happen in the vicinity of a laryngeal and a plain
voiced (presumably glottalized). That is why I see PIE *kHVb- for Lat.
capere / habe:re, Gmc. habe:- / ko:p- etc. We should also consider forms
without k-/gh- here: Greek haptós 'tangible, palpable', haphao: ' I touch',
Skr. a:pno:ti 'he reaches', Lat. api:sci: 'to reach, to obtain', Hitt. ap-
'take'.
Compare *g^hr.d- ~ *k^r.d- and the strange Greek kear with a possible trace
of a laryngeal.Compare also Skr. duhitar- < *dhugHter with
"unetymological" -h- (even if we assume *gH > *ghH here, it would not be
enough as *gh should preserve before i < H while we have -h- instead).
Further examples for which one IE reconstruction cannot be made:
7) Slavic koza 'goat' < *kag^ha: (Winter's law) and azIno < *ag^ino-, Gmc.
hak-. It cannot be taken as strong evidence that <koza> is a loanword. Just
regard the irregularity in Greek aig- (where is -i- from?).
Nota bene we can see irregularities in other names of goat:
a) Latin capra but Welsh gafr < IE *kapr- ~ *g(h)ab(h)r-
b) Engl. kid, Norse kiþ and Germ. Kitz(e) < PGerm. *kid-, *kiþ-, *kit- (even
if we took English kid as a loan from Norse, the difference between *kiþ-
and *kit- would not be explained). The word is probably a borrowing, cf.
Albanian qith, Irish cit, but why *k- instead of *g- here? (cf. Latin
haedus, Eng. goat < *ghaid-). Note also alternation in German dialects:
kitz, gitz, hitz, hetz, see Pokorny.
(
http://www.indo-european.nl/cgi-bin/response.cgi?root=leiden&morpho=0&basename=%5Cdata%5Cie%5Cpokorny&first=1&sort=lemma&text_lemma=&method_lemma=substring&text_meaning=&method_meaning=substring&text_ger_mean=&method_ger_mean=substring&text_grammar=&method_grammar=substring&text_comments=&method_comments=substring&text_derivatives=&method_derivatives=substring&text_material=&method_material=substring&text_ref=&method_ref=substring&text_seealso=&method_seealso=substring&text_pages=&method_pages=substring&text_any=goat&method_any=substring)
Note also that when "movable" k- appears, we have problems with vowels as
well, cf. Slavic kamy 'rock' < *ka:m- and Gmc. ham- < *kam- (in 'hammer') -
why is here long a: in Slavic? (cf. Greek akmo:n, Skr. açman etc.)
8) Skr. açru- 'tear' (from eyes), Lith. as^ara - but Greek dakry, Lat.
dacruma and lacrima (zero : d : l). Latin l- may be secondary here but see
next as well.
9) Skr. jihva-, Av. hizva- 'tongue' < *gig^hwo-, Lith. liez^uvis, Lat.
dingua, lingua 'tongue' and gingi:va 'gum', Oskian fangvam, Eng. tongue,
Slavic *e,zy < *ing^hu:, Old Prussian insuwis (zero : d : l : g), possibly
also Greek glo:ssa, glo:tta < *gln.Hkj- (if yes, zero : d : l : g : gl
initial and g^h : Hk internal) and Toch. A käntu (with yet another
irregularity, namely metathesis). If Lithuanian and Latin l- after the verb
'to lick', why the same d- ~ l- in the previous one?
Really, we should not treat the thesis of fully regular development too
seriously.
Grzegorz J.
___________________________________________________________
Yahoo! Messenger - NEW crystal clear PC to PC calling worldwide with voicemail
http://uk.messenger.yahoo.com