MLUw and MLUm

[Ann Peters]

A follow-up to Liz and Catherine:
   I agree that, especially in early morphosyntactic production there are
two things going on, and they probably need to be assessed differently.
On the one hand there is [A] the stringing together of content words/ideas
(MLUw?); on the other is [B] the increasing inclusion of grammatical
markers (MLUm?).
   The trouble with English is that so many of the grammatical markers
are free morphemes that one is tempted to think that counting *words*
is the way to go.  However, these capacities probably develop separately,
showing up as individual differences in early combination.  In fact, the
kids who go the [B] route are probably the "frame and slot" kids who
structure their early combinations around morphosyntactic frames.
I think one sees these patterns even more clearly when one looks at
languages (like Italian) with more bound morphology.
   I have found it useful to compute MLU in two ways:
1) just open-class lexical items (excluding free grammatical morphemes),
and 2) all morphemes, whether bound or free.
I believe these measures would equate better across languages than the
traditional MLUw that grew out of working with English.
I have tried to address some of these issues in my chapter in Slobin's vol.5:
A.M. Peters, 1997. "Language typology, prosody and the acquisition of
grammatical morphemes". In The Crosslinguistic Study of Language
Acquisition, vol.5, D.I. Slobin, ed. Hillsdale NJ: Lawrence Erlbaum
Associates, 136-197.
Ann Peters