Grzegorz Jagodzinski wrote:
> 2) *kan- is also false, the same about semantic, it was not "sing" but "emit
> a rhythmic sound".
To give the best-known example, Lat. cano: certainly meant 'to utter
melodious notes, to sing, sound, play' (Lewis & Short), used also of
animals, musical instruments etc., then, transitively, 'to make
something or some person the subject of one's singing or playing' etc.,
but the concept of singing or playing a tune -- anything perceived as
music -- is always there. The barking of dogs, in particular, was not
included, or there would be no "canis a non canendo"! Same in Celtic
(Old Irish, Middle Welsh and Breton, Cornish): 'sing', 'play an
instrument', and by extension 'prophesy, praise (in song)', but not
'emit a(ny) rhythmic sound'. Celtic *kan-tlo- may mean 'song, singing,
recited lesson', but not 'barking', 'tapping', 'clapping' or the like.
The expected o-grade noun *kono- can be seen in Toch.B kene (= Toch A
kam.) 'melody, tune'.
> Of course I see *kEn < *kHn here as well (or even *kHnH),
> cf.
> a) Latin cico:nia "stork" < *k°koHn- (do storks sing?),
No, which is precisely why I object to this etymology. Storks clap their
bills, of which a reduplicated syllable with [k] is a plausible
onomatopoeic representation.
> b) *hana- "cock" and OHG huon 'hen' (< *xo:niz-) (do hens sing?),
No, but cocks do. This is why Gmc. *xanan- is more basic. It represents
the *kán-o:n 'singer', a common type of agent noun already in PIE, most
likely a substantivised variant of the active present participle. Of
course the IEs had no word for the domestic fowl, but they did have the
agent noun in question and presumably applied it to the singing males of
various birds. At some point, speakers of (pre-)Proto-Germanic derived
their word for 'hen' from that for 'cock'. In our earlier discussion I
suspected a Proto-Germanic vr.ddhi formation, which is a possibility,
but perhaps the simplest solution is the best: the expected feminine
derivative of *kán-o:n would have been *kan-n-ih2. If at the time in
question the ancestor of Germanic did not tolerate geminated nasals,
degemination with compensatory lengthening would have produced *ka~ni: >
*ka:ni:, further transformations withing Germanic yielding secondarily
extended stems like *xo:n-jo:n- (ON hœna) etc.
> c) Latin canis may also be transformed under the influence of canere
> (barking IS a rhythmic sound, isn't it); but *cuonis would be expected
> instead,
> d) Sanskrit kinkini: 'a kind of bell', kinkira 'Indian cockoo' and also
> 'foal', 'bee' and 'deity of lov(ing)' (everything related to rhythm, in such
> or another way), kinkira:ta 'parrot'
Independent onomatopoeic words. They may have been coined at any time.
> e) possibly also Sanskrit khan 'dig', praes. khánati < *kH2án(H)eti, pp.
> kha:tá < *kHnH-tó-, cf. also khanitra 'spade', 'a tool for doing rhythmic
> moves' < *kHanH-tro-
A spade is a tool for digging holes in the ground, not for doing
rhythmic moves.
Piotr