Prochain, proche and approcher aren't in contrast with the whole
picture, because they rather proceed from *propea:nus, *propeus and
*adpropea:re. Apart from Gaulish names, all the rest is made of 1)
/kt/ and /ks/ with /k/ > /x/ > /ç/ before dental obstruent (a partial
assimilation), 2) /Vs/ + sonorant > long vowel + sonorant (fricative >
approximant > vowel lengthening), also an assimilation, and 3) <gn>
[ŋn] > [ɳ:], the only proper palatalisation; note that étain, as
opposed to Provençal estanh, Catalan estany, Castilian estaño,
Portuguese estanho, simply shows (after diphthongization of the vowel
preceding palatal nasal, as in montaigne < *montanea) regular
depalatalisation of final palatal [ɳ] as in loin < longe, whose
contrast with long < longus definitely excludes palatalization of
velar nasals.
Areal outputs [jn] of <gn> are therefore to be understood as
innovations from [nj], just like metathesis in /VRjV/ > /VjRV/
sequences; this presupposes <gn> [ŋn] to have gone through
'approaching' *[ŋɳ] > mutual assimilation [ɳɳ] (= attested [ɳ:]) or
otherwise assimilation of the first nasal + dissimilation of the
second one ([ɳj], also attested e.g. in Upper Engadine Romansh stagni
[staɳj]) with subsequent metathesis into [jɳ] (e.g. in rgularly
innovative Lower Engadine Romansh stegn [ʃteɳ] < *[stajɳ], cf. pled
[plet] 'word' < *plait < placitum; note that [aj] isn't
diphthongization of /a/, because this is [au] before non palatal nasal
and doesn't take place before palatal nasal, see the contrast between
suffixes -auna < -a:na and -agna < -a:nea)
2012/10/31, stlatos <
sean@...>:
>
>
> --- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy
> <bhrihstlobhrouzghdhroy@...> wrote:
>>
>> Does the sequence /kt/ > /tʃ/ due to /kt/ > /xt/ > /çt/ > /jt/ > /tj/
>> (suggested by areal dialectology) really imply palatalization of
>> anteconsonantal *velar* /n/?
>>
>
>
> The oddity of KC in Romance is easily seen in Rum. kt > pt , ks > ps
> (octo: > opt , coxa > coapsA); since gn (Nn) > mn (signum > semn) also
> occurred, a correlation w/in Romance is easily seen. In the West, the
> change was in the opp. dir. ( K > +pal., not +round (as lingua > limbA
> implies), both ultimately going forward in the mouth).
>
>
> So, the changes for N:
>
> stagnum = alloy of silver and lead L; stagno = tin It; estaño Sp; étain Fr;
>
>
>> ññ \ yn make intermediate NYn likely; simple gn > yn > ny wouldn't match
>> all data.
>
>
> The similar changes for kt / ks, such as:
>
> sex L; seis Sp;
>
> buxus L; bosco It; bois Fr;
>
> *ni:da:x L; niais = nestling > foolish Fr;
>
> [rake:smos ?] race:mus = bunch of grapes L; [raksi:nos ?] raisin = grape Fr;
>
>
>
> make +pal. for both kinds of K likely.
>
>
> Other ks, from borrowing or (re)learned forms, differ:
>
> * approxia:re > approcciare It; aprochier OFr; >> approach E;
>
> Uxantia Gl?; >> Ousance Fr;
>
> Luxovium = (has thermal spring); Luxeuil-les-Bains Fr;
>
>
>
>