more thoughts on MLU

[Joseph Stemberger]

A few thoughts about adult language and its relevance to MLU counts:

I agree with Ann Peters when she says:

> A basic issue for me has been: how many *separate pieces* of language can
> a particular child assemble at a given time? Asking such a question would
> lead me to count both 'fall' and 'fell' as single units.

But what about WALKED, in the speech of a child who uses -ED over 90% of
the time in obligatory environments and makes overregularization errors?
I think that many or even most would be tempted to count WALKED as two
morphemes.

But there are numerous papers on adult processing that suggest that at
least some regularly inflected forms are stored in the lexicon. Alegre &
Gordon suggest that any regular past tense form with a frequency over 6
tokens per million words of speech is stored. Ullman argues that forms
like BLINKED, with bases that rime with families of irregular forms, are
stored in the lexicon. Other researchers go even further.

Are these regular forms stored in such a way that they are processed
like single units, or is processing complex enough that they still count
as "separate pieces of language"?
I don't think that we know the answer to that question FOR ADULTS, much
less for children.

I think that we're going to have to accept that the facts are complex
enough that we aren't going to have a resolution of this anytime soon.



A question about comparing MLU across languages.
Presumably, if we calculate the MLU of the speech of adults to adults in
different languages, we'll get different mean MLU's for different languages.
Presumably, we'll find those same differences in speech by adults to
children.

Is there any way to take the differences in MLU in adult speech in
different languages, and create a way to adjust child MLU's to equate
for inherent differences between the languages?


---Joe Stemberger
UBC