Kashiwagi-Wood, AkikoPickwick, Austin2019-05-062019-05-06http://hdl.handle.net/10323/6741In this paper, I hypothesize that information structure is responsible (at least in part) for the distribution of the Japanese conditionals –ba, –tara, nara, and to. I seek to show how such a hypothesis is supported by various findings in the previous literature on the Japanese conditionals, and that an information-structural analysis of the conditionals should be pursued in future research. In connection with the latter goal, I also review the approach outlined in Riester et al. (2018) for the analysis of information structure in natural language, and discuss its limitations as a tool for testing the above hypothesis. I make three specific hypotheses with regards to the information-structural distribution of the conditionals: (1) that –ba cannot be a delimitator, whereas –tara, nara, and to may be, (2) that antecedents in sentences using –ba are subsumed under either wide or narrow focus, and that (3) antecedents marked with to are more salient in discourse than those marked with –tara. I also show that the approach of Riester et al. (2018) (1) underpredicts the information structure of delimitators and (2) cannot measure the salience of memory-activatedness of linguistic items.JapaneseConditionalsInformation structureDiscourse analysisPragmaticsThesis