46 Steven Franks
Regardless of the precise mechanism assigning the dative, the question identified above
of course remains: Why is it that, in identical contexts, the dative can be assigned to a
semipredicative but not to a subject NP? Also remaining under the direct case assign-
ment scenario are the questions listed at the end of Section 5.1, and of course the
overgeneration problem: How do we prevent the SD from applying in OC contexts? I
return to these and other issues in Section 6, which draws upon the collective insights
in the SD literature in an attempt to devise a unified workable approach.- Components of a solution
This final section seeks to bring together as many strands of reasoning about the SD
as possible.6.1 Some leading ideas
I begin by briefly cataloguing some leading ideas which will guide the account to be
laid out below.6.1.1 Agreement versus assignment
Whereas adjectives agree, nouns are assigned case. The reason that ordinary adjec-
tives do not enter into the SD construction is because this dative is the result of direct
assignment rather than agreement (e.g. with a PROdat). As argued above, the semi-
predicatives can be directly assigned case, but ordinary adjectives cannot. The fact that
adjectives are not a target for probe-goal case feature valuation can also be seen in the
contrast in case behaviour between predicate adjectives and nouns in Polish. Whereas
in Russian there is a choice between agreement and instrumental for both nouns and
ordinary adjectives, in this language in copular sentences nouns must be in the instru-
mental and adjectives must agree; see Bondaruk (2013) for references and details:
(51) a. Jan był / chce być szczery/*szczerym.
Jan.nom was / wants be.inf sincere.nom/*inst
‘Jan was/wants to be sincere.’
b. Jan był / chce być *szczery człowiek
Jan.nom was/ wants be.inf *sincere.nom person.nom
/szczerym człowiekiem.
/sincere.inst person.inst
‘Jan was/wants to be a sincere person.’
Polish thus wears the agreement versus assignment contrast on its sleeve.