p. 17
"...are pairs of form and meaning"
This seems to be very loosely speaking. The AVMs look more like triples
of phonology/syntactic form/semantic representation, but the statement
might subsume phonology and syntax under form, which however leaves
"meaning" questionable, because at best it is again only part of the
picture as some variant of logical form (but not meaning in anything
like a Montagovian sense). There's maybe a few too many open
questions about this formulation.