News & Events

This talk claims that Contrastive Topic (=CT) implicatures are cross-linguistically more basic in representing scalar reasoning and conveying scalar implicatures than only and its exhaustivity operator, pervasively mobilized since Groenendijk et al (1984). CT is both scalar and conventional, contra Buring (2003). Scalarly, CT and even are systematically, semi-dually, related, with the former upward-denying, and the latter downward-denying in NPIs, unlike Chierchia (2004). Thus scalar implicatures, polarity and information structure are inseparably interwoven. RFR, or B accent in English, and C accent in French are all equivalent to CT morphemes in languages (Lee 1999) such as thi (exclusive CT) in Vietnamese, -nun in Korean and ?i]wa in Japanese both with a high pitch, and shi in Chinese. CT cannot occur discourse-initially, requiring a QUD with a Potential Topic (Lee 2000) explicit or assumed, as in Q: Did you meet (all) the (four) new-comers? A: I met [one]_CT. It generates the scalar implicature But I didn? meet more than one. The scalar [one]_CT in the answer is both topical and focal, evoking alternatives, a set of sets of propositions, and generating scalar implicatures, via Grice? maxims, as a D-linked partitive. CT is a ?onventional linguistic device.?? In contrast,if a weaker indefinite nonspecific scalar term occurs with the Concessive operator even/-to/-mo, it becomes an NPI, requiring a negative locally, I didn? meet [even one]. [i ge ren ye ??-neg]. In the relation between even and CT, if the NPI predicate is externally negated as It is not the case that she didn? (even) lift her finger, it becomes semi-dually equivalent to She lifted her FINGER_CT but not to *She lifted her finger. Ordinary predicates with ordered abstract degrees, as in [touch < kiss] and propositions (in even and CT) evoke scales in context. A recent revival of interest in CT fails to capture the relatedness of CT scalar implicatures to ?on-CT??ones. In only(p), p is presupposed (or conventionally implicated) and the denial of the entire alternatives is asserted. The only-like O/exh operator in this sense is not compatible with scalar implicatures that require the concessive discourse connector But. Even those utterances with no CT marking but simple focus marking must have a covert CT operator to generate a scalar implicature. The O(nly)/exh operator cannot explain why *I only met more than two. (Keshet 2009) and *He solved the problem (at 2:00) only elegantly (and not both elegantly and quickly and It is better to eat some of the cake than it is to eat all of it.)