Íslenskt mál og almenn málfræði - 01.01.2023, Blaðsíða 208

In its focus on language acquisition, Nowenstein’s dissertation does not shy
away from grappling with core questions about what elements of the language
system can be reasonably presumed to have innate biases and which can’t. As
Nowenstein observes, “Children are expected to both diverge from and match
their caregivers’ speech” (p. 184). The question of why children match what they
match and diverge where they diverge is, in fact, the crux of virtually all work on
child language acquisition – and it serves as a potent reminder that children in fact
create their own grammar. On the one hand, it’s not an accident that children
speak the language of their parents and/or community. On the other hand, chil-
dren don’t speak exactly that language, but rather their own invention that over-
laps to a very significant degree with the grammars of the people around them.
But digging into the details of what patterns in the input they recreate and where
they diverge (and, more importantly, the why of each of those) is a huge challenge.
In this work, because of the focus on morphology, there is a language-specific puz-
zle of how children interpret individual case morphemes and how they draw the
mappings between these morphemes and argument roles, both syntactic and the-
matic. Thus, a major contribution of this dissertation is to drill down into assump-
tions about universal argument structure mapping patterns (syntactic bootstrap-
ping), and to explicate how language learners can both use, and overcome chal-
lenges presented by, language-specific morphology in the Icelandic case system.
Having spelled out the main strengths of the dissertation, let me now turn to
a finer-grained commentary on how this dissertation moves the field of language
acquisition forward. The syntactic bootstrapping hypothesis (Gleitman 1990;
Gleitman et al. 2005; Landau and Gleitman 1985, i.a.) provides a framework for
understanding how children can begin to discover the meanings of verbs by
exploiting robust linkages between argument structure (how many and what cat-
egory of arguments are found within a clause) and the semantics of the verb
(whether it denotes a solo action or state, a causative event, an event of transfer,
a mental state, etc.). Support for this approach comes from studies using a variety
of experimental methodologies, conducted with speakers of a variety of lan-
guages, including languages that have null subjects (e.g. Mandarin; Lee and Naigles
2008) and, critically, even in languages where morphology provides an arguably
more reliable cue to causativity (Kannada; Lidz, Gleitman and Gleitman 2003).
Overwhelmingly, these studies provide support for the basic insight that argu-
ment structure constrains verb meaning. What Nowenstein argues is that mor-
phology still matters, and in fact can play an important role alongside argument
structure, in children’s learning of verb meanings. On the basis of evidence from
Icelandic, a language with strict word order such that case morphology seems
almost redundant, she makes the case for morphosyntactic bootstrapping.
Nowenstein shows that there are some clear, general associations among case
marking, syntactic role and thematic role of arguments of intransitive, monotran-
sitive and ditransitive verbs in Icelandic. These associations mostly conform to
the widespread and familiar associations found in nominative-accusative lan-
guages. Namely, subjects are associated with nominative case and agent role, and
Misha Becker208