Íslenskt mál og almenn málfræði - 01.01.2023, Qupperneq 188

et al. 2006, Schmitz 2006), with nominative on subjects and accusative on objects
(direct and indirect) being frequently overgeneralized. But interestingly, children
acquiring Icelandic also overgeneralize the dative (H.Þ. Sigurðardóttir 2002) in
subject and object position – unlike German children (Clahsen et al. 1996,
Schmitz 2006, Scherger 2018). In object case, dative arguments can appear instead
of benefactive prepositional phrases or accusative objects with verbs such as lesa
‘read’, kyssa ‘kiss’ and knúsa ‘hug’, and (seemingly a bit later) accusative motion
themes of verbs such as færa ‘move’, hrista ‘shake’ and grípa ‘catch’ can be replaced
with the dative. Additionally, children have occasionally been reported to over-
generalize dative subjects with predicates such as vera sveitt(ur) ‘be sweaty’ (see
experimental results from H.Þ. Sigurðardóttir 2002, based on an example report-
ed by Gunnarsdóttir 1996). Moreover, it has been established that children ‘over-
generalize’ the dative with psych verbs appearing with non-nominative subjects,
but this could be interpreted as the acquisition of existing variation rather than
overgeneralization (Nowenstein 2017). More research is clearly needed to
explain how the acquisition of Icelandic case marking takes place and what the
process can tell us about language acquisition on one hand and case marking on
the other. The following chapters present such research, delving into how chil-
dren understand case (Chapter 2), produce case (Chapter 3) and change case
(Chapter 4).

2. The meaning of case
In Chapter 2, I use two comprehension experiments, a minimal pair task (N =
48 preschoolers aged 2–6) and a forced-choice novel verb task (N = 146, children
aged 2–13), to show that children form associations between case and meaning
relatively early in acquisition, using case as a cue for verbal semantics. This is
interesting in the context of the syntactic bootstrapping theory (originally pro-
posed by Landau and Gleitman 1985), where it has been demonstrated that chil-
dren use syntactic information to learn verb meaning. Broadly speaking, the
results in Chapter 2 (Figure 1) show that datives (in contrast with the nominative
and accusative) are associated with experiencers in subject position and benefac-
tives/goals and motion themes in object position.
The strength of the association varies between age groups and contexts, with
the largest effects appearing with experiencers in subject position, despite the
low frequency and restricted semantics of non-nominative subjects in the input.
This is an example of how salient exceptions can constitute powerful cues and
opportunities for learning. The associations between the dative and goals and
motion themes are more elusive, yielding less clear interpretation contrasts based
on case. With novel verbs, the results show that in particular semantic contexts,
morphological case, a language-specific cue,6 can be as salient as the number of
arguments, a potentially universal cue which has been instrumental in work argu-
Iris Edda Nowenstein188
6 Language-specific in the sense that it is not universally (overtly) marked.