Year: 2023
Type: dissertation
Abstract: The Icelandic case marking system has for decades proved an important testing ground \nfor theories on case and productivity. This is in part because Icelandic has the rare \nproperty of marking argument function redundantly, with both rich morphological case \nmarking and a relatively rigid word order (Kiparsky 1997, Fedzechkina et al. 2016). \nAdditionally, the Icelandic (non-default) dative is surprisingly robust and productive, both \nin object and subject case (e.g. Andrews 1976, Thráinsson 1979, Svavarsdóttir 1982, \nMaling 2002, Jónsson & Eythórsson 2005 and Barðdal 2008). The productivity of the \nIcelandic dative does appear in typologically expected semantic contexts, namely with \nexperiencer and recipient arguments, but also with e.g. themes of motion verbs (Maling \n2002, Jónsson 2003, Barðdal 2008). Despite these well-known associations between \nmorphological dative case and specific semantics, the relationship between case and \nmeaning in Icelandic is still a matter of debate. Although various patterns have been \nidentified (agents are always nominative, patients are typically accusative and indirect \nobjects (recipients) most frequently dative), linguists have pointed out that the correlations \nare not exact and exceptions can be found (e.g. Maling 2002, Thráinsson 2007, H.Á. \nSigurðsson 2012 and Wood 2015). Conflicting views can therefore be discerned in \nprevious work: Case productivity is assumed to be semantically conditioned while the \nrelationship between case and meaning is rejected on the basis of being too \napproximative. But do patterns (rules of language) always have to be absolute and \nwithout exceptions? \nIn this thesis I argue that from the standpoint of language acquisition, which has \nbeen lacking in research on Icelandic case marking, rules do not need to be exceptionless \nto be discovered by children and thus become part of the grammar. If the distributional \nevidence for case-semantics associations is present in children’s language environment, \nthey will learn these patterns as long as the number of exceptions is tolerable (Yang \n2016). This hypothesis is supported by the results of this thesis, which builds on \nexperimental data from 148 children aged 2–13 as well as incremental child language \ncorpus analyses. Broadly, the results show that children associate the dative more with \nexperiencers, recipients and motion themes in comparison to other thematic roles, \ndespite known exceptions. The association between the dative and applied arguments \n(experiencers and recipients) is formed earlier than the association with themes of motion \nverbs. The corpus analysis furthermore shows that these mappings of form and meaning \ncan be derived from the distributional properties of the input, which can also predict the \nwell-studied patterns of dative productivity and case marking variation in Icelandic. I argue \nthat the learnability perspective is crucial to the understanding of language variation and \nchange, as well as providing valuable insights into theoretical approaches to case. \nHowever, the main contributions of the thesis lie in the field of language acquisition, \nwhere the properties of the Icelandic case marking system also have the potential to \nbecome important testing grounds for various fundamental questions. The results of the \ndissertation have implications within three distinct but related topics in language \nacquisition research: (1) (Morpho)syntactic bootstrapping, (2) Rule formation and \nproductivity and (3) Variation and change in acquisition. \nFirst, I argue that the Icelandic data provides additional support for the claim that \nchildren can rely on language-specific, morphological cues to derive verb meaning \n(Göksun et al. 2008, Matsuo et al. 2012, Trueswell et al. 2012 and Leischner et al. 2016). \nThe results show that children acquiring Icelandic can, early on, use case to determine \nverb meaning when word order is uninformative. I furthermore show, by testing transitivity \nand case morphology within the same experiment, that case morphology can be as \nsalient as the number of arguments in specific contexts. Second, the data call for a model \nof rule formation which accounts for nested non-default productivity and does not hinge \non frequency to determine gradient overgeneralization, since patterns which are not \nstatistically dominant (the dative) can become productive and multiple rules can apply to \na context simultaneously. I argue this is most compatible with rule-based accounts which \nassume that rules are derived from distributional information in the input (Yang 2002, \nAlbright & Hayes 2003, O’Donnell 2015 and Yang 2016). Third, I show how such an \naccount, a combination of Yang’s Tolerance Principle (2016) and his Variational Model of \nlanguage acquisition (2002), can also account for the acquisition of the Icelandic case \nmarking system in well-known scenarios of language variation and change, and that a \ncase should be made for stochastic exceptions in addition to stochastic rules. I argue that \nthe underlying mechanisms in language acquisition, namely a constant search for \nproductive rules while tracking exceptions, and both convergence with and divergence \nfrom the language environment, shape the directionality and dynamics of language variation and change.
Source: Opin vísindi (Opin vísindi)
Author: Nowenstein, Iris Edda
Institution: University of Iceland
Language: English
Subfield: Language and Linguistics
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W6980349902
