Re: Language reconstruction question

En réponse à John Cowan <jcowan@...>:

>
> As Alexis Manaster-Ramer pointed out ten years ago or so, there is
> simply
> no evidence at all for this claim -- it is something that historical
> linguists, especially IE-ists, have told one another for so long that
> they have come to think it is gospel.  We have a tolerable
> understanding
> of Proto-Austronesian, whose time-depth is probably greater than
> PIE's,
> and with far less in the way of ancient written sources.  Carefully
> carried out (as it has not always been), reconstruction past the PIE
> level is certainly possible.
>
Not if the only source for it we have is reconstructed languages. Of course,
the time depth we can achieve probably differs depending on the language
sources we have, the way they changed through time, etc... But reconstruction
methods don't give a precise idea of the time depth we reach (without
archeological evidence, it's quite difficult to give a good datation of PIE for
instance. Actually, we are quite sure that the features we reconstruct for PIE
were not all concurrent at the same time, but represent different stades of the
language. And we are not always sure how to order them). When we get a
precision of + or - 1000 years, the difference is not that important. So of
course depending on the data we have we can reach different time depths. But
going further is no more than speculation, especially when you begin to compare
data from different reconstructed languages (thus incomplete data, since we
will never be able to reconstruct exactly those ancestral tongues). It may be
interesting speculation, and can lead to interesting hypotheses. But without
some other more certain data, this will never reach the level of certainty we
have about PIE for instance.

So in short, for each language family we can produce a rather good idea of the
ancestral language, and those tongues are reconstructed with different time
depths, some further than others. You can even sometimes get a bit further by
internal reconstruction, although the result can only be less certain than the
data you began with. But as soon as in order to go further you need to compare
the data from different reconstructed ancestral tongues, then you enter the
domain of pure speculation. It doesn't mean that it's wrong, nor that it's not
done carefully. It just means that when you compare uncertain data with
uncertain data, the uncertainty doesn't add up. It multiplies. So of course the
reconstruction can be carried on further. But the result cannot be considered
to have the same strength as the reconstructions of PIE or Proto-Semitic for
instance. They can be considered like plausible hypotheses maybe, but without
any other data (for instance archeological), they will only remain so.

Christophe.

http://rainbow.conlang.free.fr

Take your life as a movie: do not let anybody else play the leading role.