Re: [tied] Celtic n (was: Greek labiovelars)

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

> Wouldn't *(komangu > komunWgW > kuMung) work to
> raise o also?

It's a matter of relative chronology. As I recall, raising precedes
u-infection. It should be possible to find some examples.

>
> > Isn't this extraordinarily complicated, against
> > assuming that PIE
> > *eNT > PCelt. *eNT, PIE *N.T > *aNT?
>
> I prefer e>a for several reasons that seem to add up
> to an understandable whole.
>

[...]

>
> A similar set of rules could apply to Irish, etc.
>
> 1. e(>e~?)>a between N-N in same syllable
>
> *men-lo- "remain, delay" > OIr mall "slow"; Lat.
> promello: "I postpone"; Grk. mello: "I delay"; Rus.
> medlennyj "slow"; Skt. manda- "slow";

Traditionally: mall < *ml.-no- or *ml.-so-, right?

[...]

> 2. e(>e~?)>a before N-stop+voice with same place of
> articulation

= PCelt. *enR > Ir. an+r,l,m,n,w,y,s?

>

[...]

>
> e(>e~?)>a between N-N in same syllable

You may have lost me again, but are you saying that PCelt. *NeN
gives *NaN in Irish?

Anders