Re: [tied] Greek labiovelars (Celtic)

--- In cybalist@yahoogroups.com, Sean Whalen <stlatos@...> wrote:

>
> > There's
> > no
> > > reason to assume *gWaneti with a>o not *gWeneti
> > with
> > > e>o. In Welsh e>e~>a~>a before nasals explaining
> > go-
> > > vs gwa- here.
>
> I'm not sure of the exact rules but Irish e varies
> with P-Celtic a in many words; I still believe
> *gWeneti with e>o here.
>
> "tooth" de:t OIr; dant W; dans Co; dant Br
> "hundred" ce:t OIr; cant W; cans Co; kant Br
> "step" ceimm OIr; cam W; cam Co; kam Br
> "tongue" tenge OIr; tafod W; tavot Co; teod Br
> "being first" > "perceptive" ce:tbaid "sense" OIr;
> canfod "perceive" W
>
> for "tongue" Ngw > Nw > M > v in W/Co

It is now commonly assumed that PIE syllabic nasals gave PCelt. *aN:
*Hdn.t- 'tooth' > *dant- etc.

In Irish we then have a further development of *a in front of
nasals:

- raising/fronting + lengthening in front of nasal + unvoiced stop
or *s. The outcome is <é>: *dant- > dét /d'æ:d/, *anku- 'death' >
éc /æ:g/, *gans- 'goose' > géis /g'æ:s'/.

- raising/fronting in front of nasal + palatalized voiced stop:
The outcome is <i> (or <e> by analogical lowering?): PIE *Hn.gWn. >
PCelt. *amban 'butter' > *æmbæn > imb /imb'/

- unconditioned raising in front of *n+g: *tangWa:t- 'tongue' >
*tængWa:t- > *tingWa:th- > tengae (subsequent lowering i > e).

OIr. céimm (VN of cingid), W cam 'step' is not really a problem:
*kn.g(H)-s-mn- > PCelt. *kan(x)sman > Ir. céimm (*-ansm- > *-æ:ssm-

> -émm-), > Brit. *kamm (*-ansm- > *-assm- > *-amm-, regularly

without length).

In British Celtic you _can_ tell the difference between PCelt. *aNC
and *eNC, in Irish the two have merged, apart from some positions,
e.g. land < *land-, where nasal + voiced non-palatalized stop does
not cause rasing.

> There's more rounding than that:
>
> pe:NkWe uper- nigW- pro-widhu-
> peNkWe uper- nigW- prowidhu-
> kWeNkWe uper- nigW- prowidhu-
> kWeNkWe ufer- nigW- frowidhu-
> kWeNkWe uxWer- nigW- xWrowidhu-
> kWeNkWe uxWer- nigW- xWrowidu-
> kWeNkWe uxWer- nigW- xWrowidWu-
> kWoNkWe uxWor- nugW- xWrowudWu-
> kWoNkWe uxWor- nugW- xWruwudWu-
> kWoNkWe uor- nugW- xWruwudWu-
> kWoNkWe wor- nugW- xWruwudWu-
> kWoNkWe wor- nugW- xWru:dWu-
> etc.
> co:ig for- dofonuch ruud

cóic is indeed another case of rounding.

for- is probably not a case of rounding, but rather influenced by
*upo > *wo, cf. also British Celtic *wor > gour, gor, war etc.,
which must come from *wor.

You may have to elaborate on the other examples.

>
> and as I previously said:
>
> wekWsperos sept_m gWHenti gWHedHyei
> wekWsperos sept_m gWHeneti gWHedHyeti (analogy)
> wesperos sept_m gWHeneti gWHedHyeti
> wesferos seft_m gWHeneti gWHedHyeti
> wesferos seft_m gWeneti gWedyeti
> wesferos seftem gWeneti gWedyeti
> wesxWeros sexWtem gWeneti gWedyeti
>
> wesxWeros sexWtem gWeneti gWidyiti
> weskWeros sexWtem gWeniti gWidyiti
> weskWoros sexWtem gWoniti gWudyiti
> veskWoros sexWtem gWoniti gWudyiti
> feskWoros sexWtem gWoniti gWudyiti
> etc
> fescor secht(N) gonaid guidid
>

fescor is probably /f'esk&r/, just like lebar, lebor 'book'
is /l'ev&r/ etc. I feel inclined to regard it as a borrowing from
Latin vesper, just like W gosper, Bret. gousper

It is still impossible to get British Celt. *gwan- (W gwan-, MBret.
goan-) from PCelt. *gWen-.

*-ye/o- verbs turn up as *-i/yo- verbs in Irish (for some reason),
hence the 3sg.pres. conj. -guid < *gWidhith < *gWedhith < *gWedyet
(i).There isn't room for the -y- which should have given -iy- and
then and left an extra syllable.

Anders