Whamit!

Speaker: Haoming Li (MIT)

Title: Assertion and presupposition of change-of-state verbs across different aspects (Part 2)

Time: Wednesday, March 18th, 1pm - 2pm

Location: 32-D461

 

Abstract: Sudo (2012); Zehr & Schwarz (2018), among others, have drawn attention to the different discourse and projection behaviors of presuppositions that are entailed by the assertion (e.g., the prior negative state presupposition of stop) versus those that are non-entailed by the assertion (e.g., the gender presupposition of herself). Doron, Fox, & Wehbe (2025) note that two dimensional systems of assertion and presupposition specification are too unconstrained and propose to retrieve the assertion of sentences from trivalent propositions via an algorithm that takes into account the deletability of the presupposition triggers (herself represents the deletable pattern, while stop represents the non-deletable pattern). This talk presents new data on the non-uniform presuppositional behavior of certain change-of-state verbs like arrive in different aspects (arrive behaves deletably in the present perfect but non-deletably in the simple past) which calls for the system in DF&W, where the same presupposition trigger can in principle be variably deletable or non-deletable depending on the semantic context, and which motivates lexical decomposition approaches to change-of-state verbs. In the second installment of the talk, I will recapitulate the empirical picture and the main analytical ingredients, and then explain in more detail how the analysis works. In addition, I will address an issue raised against the DF&W and my own application thereof, i.e., why stop itself cannot be decomposed in the same manner as arrive. I will show that stop might have two different decompositions depending on whether it is used in a pure change-of-state manner or in a habitual manner. In the former, stop shows the same aspectual alternation in projection like arrive and should indeed be decomposed similarly. In the latter, stop remains the poster child of the non-deletable pattern, and should be decomposed in a differently way.

 

 

 

Abstract: We present new evidence supporting the cycle as a grammatical mechanism in phonology. The evidence comes from the distribution of stress and vowel length in Nazarene Arabic, an understudied variety of Palestinian Arabic spoken in Nazareth. We show that the Nazarene Arabic pattern can be accounted for by cyclic versions of both rule-based phonology and Optimality Theory. The same pattern poses a challenge to Base-Derivative Correspondence – an alternative mechanism within Optimality Theory, according to which cyclic effects result from transderivational constraints that enforce similarity between related surface forms. We show that Base-Derivative Correspondence cannot account for the data, because the cyclic application of processes in Nazarene Arabic decreases rather than increases similarity between surface forms. Overall, this study highlights a divergent prediction of the cycle and Base-Derivative Correspondence, suggesting that phonological theory should include the former mechanism.