Íslenskt mál og almenn málfræði - 01.01.2023, Side 209

direct objects, accusative case, and patient role are associated with one another, as
are indirect objects, dative case, and goal/benefactive role. But patterns that don’t
conform to these basic associations are many, and individual variation in applica-
tion of Dative Substitution in adult speakers complicates the picture of the input
to children. That is, there are exceptions within the system (e.g. verbs that take
dative or accusative subjects, or nominative objects) and variation among adult
speakers in use of the system (Dative Substitution involves speakers sometimes
producing dative subjects with verbs that historically have taken accusative or
nominative experiencer subjects). Both types of variation or exceptions introduce
complications for theoretical accounts of learning, if we believe that children con-
struct their mental grammar through processes that rely on inductive bias and
bootstrapping from universal tendencies to associate certain cases with certain
syntactic roles, and syntactic roles with certain semantic roles, rather than piecing
their grammar together one lexeme at a time.
On the comprehension side, Nowenstein shows (chapter 2), through picture
selection tasks using both existing verbs and with novel verbs that label (presum-
ably familiar) actions and states, that children preferentially interpret dative sub-
jects as experiencers, even though most experiencer subjects are marked with
nominative in adult Icelandic. Thus, dative in subject position is a strong cue to
the thematic role of that argument. Interpretation of objects is more weakly asso-
ciated with particular cases and roles, but at least in older children dative case on
objects is associated with goals/benefactives and, less strongly, theme objects that
undergo motion. The readiness to associate dative case with experiencer subjects
may have to do with the fact that you can have nominative-accusative syncretism
in some cases, but never nominative-dative syncretism without also including
accusative in the syncretic form. Thus, dative stands out morphologically in con-
trast to both nominative and accusative. It could also have to do with a widespread
preference for subjects to be associated with nominative case in nominative-
accusative languages (and in Icelandic).
On the production side, Nowenstein shows (chapter 3) that children more
readily produce dative case on objects than on subjects. This reveals an interest-
ing asymmetry between comprehension and production. Namely, children more
readily interpret dative on subjects as indicating an experiencer role than they do
dative on objects with any particular role, but they more readily apply dative case
to objects than to subjects when thematic roles make the dative an appropriate
choice. But there are also asymmetries in the types of objects that children apply
dative to in production. With existing verbs, children use dative with the objects
of those verbs that historically take dative objects, mostly with the verb meaning
‘give’ although this seems to be the indirect object, not the direct object argument.
There is some overgeneralization of accusative and even nominative on these
objects, but that is mainly limited to the youngest children. There is even some
overgeneralization of dative on objects of verbs that have historically taken geni-
tive objects, particularly the verb meaning ‘ask’. With novel verbs children pro-
duce dative on subjects of the novel verb meaning ‘like’ and the adjective meaning
Review and commentary on Nowenstein’s dissertation 209