>Miguel:
>> No. Vocalized */&2/ does not equal */a/. The weak
>> stem is nas- (not *nis-) in Old Indic, therefore
>> there cannot have been a laryngeal between the /n/
>> and the /s/.
Yes, you argument is correct but we don't have here a simple h2.
A PIE *Hneh2s- for Indic naìs- is more probable: because we can well
suppose a laryngeal metathesis next :
*neh2-s- < *nh2e-s- < naìs
See as cognates:
Latin na:ris too.
Lith. noìsis , both derived from *neh2-s
(Also the arguments regarding the existence of the */a/ in PIE are
not very solid)
Best Regards,
Marius
P.S.-1
Please take note that :
1. Indic naìs- is a perfect cognate with Romanian nas 'nose'
too. :)
For other Romanian <-> Skt. cognates see:
2. Romanian apa 'water' <-> Skt. ap- 'id.'
3. Romanian azi 'today' <-> Skt. adyaì 'id.' (attested in RV so
around 1400 BC) both forms constructed with
4. the demonstrative particle Rom. *a <-> Skt. a < PIE *h2e '`this
here' (and not PIE < *h1e as Lubotsky supposed), present also in
different Albanian/Romanian dem. compounds (see Rom. aia 'she' but
also 'this one (fem.)' <-> Alb ajo 'id.' < PAlb a-ja: cognates with
Skt. ay- attested in RV ayaìm 'this here, he' )
Best Regards,
Marius Alexandru
P.S.-2
For a supposed Latin derivation of Rom 'azi' please remember that
Lat. /o/ never give /a/ in Romanian