Re: Some questions on phonology

On Tue, 14 Oct 2008 21:28:11 +0200, Philip Newton <philip.newton@...>
wrote:

>Have a look at http://www.xibalba.demon.co.uk/jbr/pleisto.html , which
>attempts to create a Neanderthal-era language.
Sweet, a new piece by JBR!  Great stuff.

>It doesn't even go the
>route of specific, individual phonemes but applies features to
>syllables as whole. (So, for example, "bot" would not be a possible
>syllable because the "voiced" or "voiceless" feature would have to
>apply to the entire syllable, not just a phoneme at the onset or
>coda.)
Yes, but I think this glances off the point.  The reason it doesn't have
phonemes is because it cuts out the whole middle layer of representation:
where modern languages have
  *words*  composed of several  *phonemes*  each with values of several
*features*
this paleolang has simply
  *words*  consisting of values of  *features* .
So words are monosyllabic, and it's inconceivable that they could be
otherwise, since that would mean giving a feature two different values for
the different points in time and there's just no room for that in the
representation.  Thus for instance there aren't even coda consonants (beyond
those that can be thought of as manifestations of supersegments), given that
there are onset consonants.  On the other hand there could be (and in JBR's
there is) a value of the voicing feature which is "starts off voiceless,
becomes voiced", so a word like /ka/ is not ruled out.

Alex