I think that it does, to the extent that developing a mental
model depends on communication :-) To a very large extent,
your vocabulary limits your ability to communicate. Whether
this 'vocabulary' consists of words in a natural language or
symbols in a constructed language (mathematics or logic or
data modeling), communication depends largely on shared
understanding of those symbols.
I suppose that you _could_ develop mental models entirely
without communicating with someone else -- I haven't really
given that much thought. But a large part of the power of
our mental models comes from shared understanding. So perhaps
we have a chicken-and-egg situation here. Does language come
before mental model development, or mental model development
come before language? I suggest that you can only build very
primitive models in a pre-language environment, without
involving someone else in their refinement and validation.
And I would further suggest that _which_ natural language you
learn and use further affects the kinds of mental models you
can build -- or at least the ease with which you can build them.
Michael Ayers
mbayers@mmm.com (612) 733-5690 FAX (612) 737-7718