Re[2]: [tied] Re: Brugmann's Law

From: Brian M. Scott
Message: 51638
Date: 2008-01-20

At 10:08:52 AM on Sunday, January 20, 2008, fournet.arnaud
wrote:

> From: "Piotr Gasiorowski" <gpiotr@...>

>> On 2008-01-20 15:05, fournet.arnaud wrote:

>>> I compare :
>>> Latin lassus to Arabic laghab "to be tired" : hence H2
>>> Latin Lentis to Arabic lax "slow" : hence H1

>> Then you're guilty of the practice of "reaching down".
>> Before you compare the Latin words with any outgroup
>> material, their status within IE must be established. If
>> IE cognates (in this case, specifically, Germanic *sle:p-
>> and *le:t-) rule out *h2 and suggest *h1, then the PIE
>> reconstruction has *h1, whatever the Arabic facts.

> I don't feel guilty at all. We are talking about
> methodology. I consider that **any** laryngeal in PIE has
> to be checked against PAA data and especially against
> Arabic.

Since no genetic connection between PIE and PAA has been
established, it is obviously methodologically unsound to
favor PAA evidence over PIE evidence. Relying on one
particular PAA language merely compounds the error.

Brian