--- In cybalist@..., CeiSerith@... wrote:
> On the other hand, I would doubt that one person in twenty know
> "nit" is, other than something you pick. (Did that ever have a
> English besides "nits?")
The OE form was 'hnitu', so that should have had a plural 'hnita'. I
don't know whether the plural is attested. The regular phonetic
development would have yielded 'nit' for both singular and plural.
The word seems to be one of a not insignificant set of words with
variation between /k/ and /g/. We have Germanic xnito: = 'not, Greek
konid- = 'dust', but Slavonic (or at least, Russian and Polish)
gnída:, Danish gnid, Norwegian dialectical gnit. Onions also quotes
Old Irish sned 'nit' and Albanian thení 'louse'. Pokorny (
eiden ) gives roots k^nid-, knid-, s(k)nid- and gHni:d-. (Note that
the hyperlink to gHni:d- from knid- is broken.)
Another word with similar variation is 'suck', with roots seug- and
seuk-, both of which are attested in Old English - su:can for seug-,
and su:gan for seuk-. Pokorny views them as alternative extensions
of seu- 'juice', 'liquid'. I an not sure why Pokorny does not give
Russian sosat' as an example of the extended root. Baltic forms are
lett. su\kt `saugen'; apr. suge f. `Regen'.
I don't know whether that excludes seuk^-. The Old Prussian form is
the only evidence I see for seug- as opposed to seug^.
Interestingly, Blust in 'Patterns of sound change in the Austronesian
languages' (in Baldi, 1991) observes that Proto-Austronesian
initial 'k' and 'g' have a 20% chance of showing up as the reflex of
the other in some daughter language, and offers an articulatory
explanation for the confusion.