Donohue, Cathryn2015-12-101871-5656http://hdl.handle.net/1885/59852Most work on case marking has focussed on the standard or 'default' case patterns, however more recent work has examined the marginal instances where non-standard cases are used to mark core grammatical relations. In this paper I investigate the use of logistic regression as a tool for validating competing analyses proposed to account for one such case pattern in Basque: transitive clauses where the object bears dative case. Several explanations for this dative marking have been proposed, appealing to notions such as telicity, animacy, and the 'person' of the subject/object. To evaluate these different proposals a database of naturally occurring sentences was created from existing corpora and coded for these different possible variables. Following Bresnan et al (In: Bouma G, Kraemer I, Zwarts J (eds) Cognitive foundations of interpretation. Royal Netherlands Academy of Science, Amsterdam, 2007) and others, a logistic regression model was fit to the data using these predictor variables to ascertain the most important factors determining the use of the dative case for Basque objects.Keywords: Basque; Case alternations; Dative object; Logistic regression201110.1007/s11525-010-9179-92016-02-24