Abstract Haase
Modelling learner conceptualization of morphological causatives in L2
The grammaticalization of cause-effect relationships has been studied in a number of languages but few studies employ a typological-comparative method (notable exceptions are Comrie, 1989; Song, 1996. Even fewer employ a cognitive approach as suggested by Langacker’s case hierarchy for the parameter of control in a cause-effect relationship. Further, little is known about the acquisition of causation for second language learners.
In this study the conceptualization of learners of causation in English in its different lexicalizations is studied against the backdrop of a typologically close L1, German. The study will focus on morphological causative verbs. Data evaluation will lead us to propose L2 learner theories about the formation and productivity of these causatives. The findings also suggests consecutive refinement of learner theories which enables Bayesian modeling.
It is hypothesized that these models are complex and under influence of multiple cognitive triggers. These are parameters like L1 interference, frequency, productivity and syntactic variance concerning transitivity (i.e. causative-inchoative alternation). Therefore, insight can be gained through learner judgment elicitation tests on category membership. Test subjects were asked to rate acceptability of existent (like soften) and nonexistent but plausible (like freshen) causatives. In a follow-up test subjects rated acceptability of lexical causatives in clauses (as in The soup cools) including erroneously used morphological causatives (*The price cheaps). Scores were used to calculate a general scheme of acceptability of causatives in a second language. Finally, the identified hypothetical parameters are used to model learner conceptualization of the likelihood of a morphological causative in the second language.
It will be proposed that theory formation of learners about these causatives receives further refinement on basis of parameters that are at the learner's disposal. As these parameters are complex, learners probably come to their decisions unconsciously but in a similar way they reach at other conclusions as well. The judgment on category membership can be modelled on the grounds of Bayes' theorem as a series of cues. Learners will then adapt their conceptualization on basis of gradual knowledge and therefore re-negotiate probability of category membership.
The evaluation of the data and the theoretical considerations allow for establishing some tentative parameters. Attribution of category membership of morphological causatives is influenced by at least four different cues:
cue 1: absolute frequency, a learning effect/parameter due to higher exposition of learners to more frequent items
cue 2: productivity, a highly productive deadjectival item gains "weight" against its root
cue 3: material/shape, as morphological Vcaus often refer to object properties
cue 4: a syntactic parameter that recognizes causative/inchoative alternation but is further influenced by world knowledge parameters, e.g. about the status of psych verbs.
These cues are hypothesized to determine learner conceptualization. As will be shown, Bayesian decision making can be a method to quantitatively capture the parameters discussed above and hence model learner conceptualization.
Selected references
Brooks, Patricia J. and Michael Tomasello (1999) "How children constrain their argument structure conditions". Language 75, 4, 720-38.
Comrie, Bernard (1989) Language universals and linguistic typology. Second edition. Chicago: The University of Chicago Press.
Gelman, Susan A. and Melissa A. Koenig (2001) "The role of animacy in children's understanding of 'move'". Journal of Child Language, 28, 683-701.
Song, Jae Jung (1996) Causatives and causation. London: Longman.