real proto-lang

[Jens Elmegaard Rasmussen]

On Sat, 7 Jul 2001, petegray wrote:
> [...]
> So [...] I think there are
> advantages in treating our reconstructions as if they were actual spoken
> human languages - even while admitting the limits of our knowledge.

You're so right. We would not understand developments like {kw} > p or
{gw} > b if we did not care about the phonetic substance; these changes
are natural, i.e. small, because the elements concerned remain labialized
stop at all stages. We wouldn't even understand a conditioned change like
that of word-final *-m going to -n if we were not permitted to draw on our
knowledge about the phonetic substance that tells us that the element is
still a nasal consonant, even after the change. We would no doubt have to
deal with suggestions like *deiwos > Eng. God for which there may just
conceivable be parallels which we just don't bother to look for because we
know it would be silly. If actual pronunciation counted for nothing such
suggestions would have to be taken just as seriously as *nu: > Eng. now,
at least until they had been properly checked. Since we do not normally do
that, we do indeed work on the assumption that PIE and all the
intermediate stages between it and the attested daughter languages were
real and actually spoken languages. And that's what has taken us where we
are.

Jens