[Corpora-List] Is a complete grammar possible (beyond thecorpus itself)?

[John F. Sowa]

Yorick and Rob,

YW> just for the record (because John cares about these things)...

Computability is relevant for any kind of processing, but the issues
of decidability and completeness are defined only for a formal system.

I do not believe that the latter issues can be meaningfully applied
to such a wildly informal system as a natural language -- perhaps to
certain specialized uses of NLs (such as mathematical sublanguages),
but not to the fundamental mechanisms of how people learn and use NLs.

RF> You'll confuse the issue with so many words.

OK.  I'll restate my point more succinctly:

Trying to get deep insights into how NLs work from research in
decidability and completeness is hopelessly misguided.

John


_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora