Re: Yiklamu hits the big time
|From:||Mark P. Line <mark@...>|
|Date:||Monday, June 14, 2004, 16:48|
John Cowan said:
Yes, there are dozens of google hits for 'Yiklamu' which derive from a
blurb (presumably on the site you cite) that says it covers languages from
"Aboriginal" to "Yiklamu".
As long as we're on the subject of ostensibly compendious link-sinks for
I'm actually quite disappointed, frustrated and surprised at the fact that
there's not a major website out there that provides reasonably reliable,
reasonably theory-neutral (but terminologically controlled) sketches
(demographics, phonology, graphology, morphosyntax) of a *large*
collection of natlangs.
I guess there are a number of reasons why this hasn't happened. The most
important one might be that "theory-neutral (but terminologically
controlled)" really means using least common denominators, as it were, and
there's no coherent ontology (yet) that can support this kind of
Maybe this is a job for us ontologists...
Example for further cogitation: What would a linguistic ontology need to
specify for the term 'grammatical case' to allow somebody to decide, say,
exactly how many different cases there are in (some appropriately
specified variety of) Hungarian? Obviously, the ontology would have
exactly one definition for 'grammatical case', and that definition would
have to be applicable to all languages, not just Hungarian, in such a way
that case marking (or lack thereof) can be meaningfully compared across
the whole gamut of human languages. You can think of lots of similar
examples: parts of speech, TAM marking, semantic & pragmatic meanings of
word order (e.g. the topic/focus/subject fiasco), etc.
So-called language "universals" could be reformulated in terms of such an
ontology, thus hopefully alleviating some of the problems of "universals"
being formulated in such a theory-specific or language-specific way as to
be nearly useless. ("No human language is such that a rule moving alpha
inside a theta-role container has to be ordered prior to any rule
involving the square root of the sine of theta divided by alpha.")
("Finnish, Estonian, Hungarian and related languages all have vowel
harmony of the Uralic type.")
An ontology of this magnitude would also be pretty useful for conlang
development, too, with or without the bit about "universals".