those adjectives

[Avery.Andrews at anu.edu.au]

I'd suggest that the phenomena of `long distance case agreement' in Icelandic
provides reasonable evidence that predicate adjectives have SUBJ in that
language.  Note that copulas are optional in adjectival functional control
constructions, so it wouldn't do to have complex predicate formation
of the adjective and the copula, with the latter being responsible for
managing the SUBJ attribute (my paper in the 1982 volume contains more
about the details of these constructions than a sensible person would
really want to know).

On the other hand, there are many situations where the agreement phenomena
within NPs work differently than predicate adjective agreement (the latter
typically being richer).  So Germanic languages (including Icelandic) often
have a `weak/strong' distinction that seems to be keyed to the determiner
for attributive adjectives, but this never happens for predicate adjectives,
even in languages like Icelandic where predicate adjectives agree.  Likewise
Russian seems to basically have case agreement for attributive adjectives
but not for predicate adjectives (at a minimum, case agreeemnt is obligatory
for attributives but not for predicatives).

So I'd go for a bifurcated treatment (and find no problem with Miriam's
rule suppressing the SUBJ argument in the course of deverbalization of
`bellende').

  Avery Andrews