bees , undergenlzn

[Joyce Tang Boyland]

I think that children's regularization of "be" is more profitably
thought of as a case of *under*generalization, where pieces of the
paradigm are being split off to serve particular different functions.

That is, there's the "be" that means "act in such a way"
(which is consistent with the input to children: be good, be nice, etc.),
and then there's the copula (which, in the input, is often just a clitic
'm, 're, or 's).  It's pretty reasonable for kids to undergeneralize "be"
to mean just the "act thus" sense.

It's comparable, say, to the splitting off of "haf to" from the rest of "have".
My husband, for example, says things like "They were hafing to shout all day"
where "haf to" is just the "must" meaning.

Children undergeneralize all the time, and adults do it too.
It makes sense to me to see this as a specific case of a general phenomenon.

Incidentally, this isn't a case of grammaticization (the opposite, really),
but i don't think it's far-fetched to say that processes like this
often contribute to grammaticization. (for further discussion of general
cognitive processes in grammaticization, ask for my dissertation).

Joyce Tang Boyland