Topics are Conditionals: A Case Study from Exhaustification Over Questions

PACLIC 28 !522 Topics are conditionals: A case study from exhaustification over questions Yurie Hara Department of Linguistics and Translation, City University of Hong Kong 83 Tat Chee Avenue Kowloon, Hong Kong SAR y.hara@cityu.edu.hk Abstract This paper argues in favor of Haiman’s (1978) idea that conditionals and topics are analo- gous. The evidence comes from exhaustifi- cation over topicalized questions, which have the same semantics as conditional questions (Isaacs & Rawlins, 2008). 1 Introduction Similarities between conditionals and topics are identified by many linguists (Haiman, 1978, 1993; Collins, 1998; Bittner, 2001; Bhatt & Pancheva, 2006; Ebert et al., 2008). Some languages use an identical morpheme to mark topics and condition- als. In Japanese, for instance, a conditional suffix nara is used for both conditional and topic construc- tions. When nara follows a clause as in (1-a), the clause serves as an antecedent of a conditional sen- tence. When nara attaches to a NP as in (1-b), the NP is the topic of the sentence. (1) a. Taro-ga Taro-nom kuru come nara, if paatii-wa party-top tanosiku fun naru. become ‘If Taro comes, the party will be fun.’ b. Taro-nara Taro-if kaeri-mas-ita. go.home-pol-past ‘As for Taro, he went home.’ This paper offers another piece of evidence for the virtual identity of topics and conditionals. In partic- ular, I argue that topics have the same semantics as conditional antecedents in that both serve as context- shifters. In dynamic semantics, conditionals are de- fined in terms of a two-step (Stalnaker, 1968; Kart- tunen, 1974; Heim, 1982) or three-step (Kaufmann, 2000; Isaacs & Rawlins, 2008) update procedure: (2) c+ ‘if P, Q’ = (c ∩ P ∩ Q) ∪ (c ∩ P), where a context c and propositions P and Q are sets of possible worlds. (3) 1. A derived context is created by updating the speech context with the antecedent of the conditional (c ∩ P). 2. The derived context is updated with the consequent (c ∩ P ∩ Q). 3. The original context learns the effects of the second step. To illustrate briefly, in (4-a), the initial context is assertively updated by the antecedent ‘Max comes’, that is, the worlds that make the proposition false are deleted. The derived context is then assertively up- dated by the consequent ‘we’ll play poker’. Finally, the worlds removed in the second step are also re- moved from the original context. (4) a. If Max comes, we’ll play poker. b. There’s food in the fridge, if you’re hun- gry. (Haiman, 1978, 564) The idea of context-shifting nature of conditionals might be clearer with so-called biscuit conditionals like (4-b). In (4-b), the antecedent ‘if you’re hun- gry’ shifts the context so that the assertive update of the consequent ‘There’s food in the fridge’ becomes relevant or optimal (Franke, 2007, 2009). Copyright 2014 by Yurie Hara 28th Pacific Asia Conference on Language, Information and Computation pages 522–531 PACLIC 28 !523 Just like English if-clauses, the Japanese Topic- marking wa serves to shift the context. Let us take the following …

Content truncated to 500 words