Whamit!

This weekend, visiting Professor Jon Rawski gave invited talks at the CUNY Computational Linguistics Lecture Series, and the Workshop on Theoretical Phonology (WTPh) in Montreal.

Speaker: Charlie Yan (QMUL)
Title: On the locality profile of verb doubling in Mandarin Chinese
Time: Tuesday, May 7th, 1pm - 2pm
Location: 32-D461

Abstract: Within Mandarin Chinese root clauses, verb doubling effects are attested in either a clause-medial or clause-initial position (Cheng 2007; Meadows and Yan 2023; a.o.). I treat this contrast as VP-fronting to a lower or higher position in the clausal spine. Crucially, I show in this talk that cross-clausal VP-fronting to the lower, clause-medial position is only possible out of a limited class of complement clauses, unlike VP-fronting to the clausal periphery. I offer an approach to such movement restrictions based on the Williams Cycle (Williams 2003, 2013; Poole 2022), and predictions about VP-fronting across multiple levels of embedding.

Speaker: Xinyue Zhong (MIT)
Title: Tone sandhi in Mandarin-English bilingual speech
Time: Monday, May 6th, 5pm - 6:30pm
Location: 32-D831

Abstract: When bilingual speakers mix languages in their speech, two phonologies come into contact in real time, which raises the question of how potential incompatibilities between the two phonologies are resolved. This study investigates one such potential incompatibility concerning how Mandarin-English bilinguals handle Mandarin tone sandhi when the trigger is a word in English. Existing research (Cheng 1968, Gao and Wu 2022) suggests that tone sandhi can be triggered by English words based on pitch similarity. Nevertheless, my previous experiment (2021) additionally showed that English consonant clusters can block tone sandhi in places where sandhi is expected to apply based on pitch similarity.

Two hypotheses were proposed to explain this blocking effect: 1) an “illusory” vowel or low tone is associated with consonant clusters (cf. Dupoux et al. 1999), which interferes with the expected pitch patterns, or 2) the fact that consonant clusters are banned in Mandarin phonotactics prevents English syllables with clusters from participating in Mandarin phonological processes. A new production experiment was conducted this semester where two Mandarin tone sandhis, the Tone 3 sandhi and the bu sandhi, are tested with select English words embedded in otherwise Mandarin sentences. The results showed that neither of the hypotheses is fully consistent with the patterns in speakers’ sandhi production under the assumption that English syllables receive Mandarin tones in the tested sentences. The findings so far suggest that closer inspection of the phonetic properties of the English syllable triggers is needed, and that we might need to adjust our assumptions about what conditions tone sandhi in Mandarin.

Speaker: Viola Schmitt (MIT/Humboldt-Universität zu Berlin)
Title: Possible worlds and ontological symmetry
Time: Thursday, May 9th, 12:30pm - 2pm
Location: 32-D461

Abstract: The main point of this talk is a negative claim: I will argue that possible worlds differ from other objects via which construct meanings in terms (i) reference and (ii) plurality formation (i.e.., contra ontological symmetry, Schlenker 2006). Based on this claim, I will outline a puzzle, namely, why worlds should be constrained in this way and why — in the modal part of the language — quantification is seemingly dissociated from reference and plurality formation.

Speaker: Yosef Grodzinsky (The Hebrew University)
Title: High behavioral and neural selectivity in the processing of downward entailingness, and its theoretical implications
Time: Friday, May 10th, 3:30pm - 5pm
Location: 32-141

Abstract: (with I-An Tan)

A 50-year-old discovery, that Downward Entailing (DE) operators incur greater processing costs than their Upward Entailing (UE) counterparts (Just & Carpenter, 1971; henceforth the Monotonicity Effect) led to a series of psycho- and neuro-linguistic studies. These explored the manner by which DE operators license Negative Polarity Items (NPIs). We took several steps: first, we conducted behavioral studies that showed that the Monotonicity Effect is highly robust and replicable, demonstrably independent from generic psychological factors such as word frequency, length, on-the-fly learning, delayed verification, and more (Deschamps et al., 2015; Agmon et al., 2023, passim). Second, an fMRI study localized the Monotonicity Effect anatomically in a small and neuroanatomically coherent region that is adjacent to, but excluding, the left hemispheric Brodmann Areas 44,45 (famously known as Broca’s region, Grodzinsky et al., 2020).

These studies served as backdrop against which we conducted our next investigation: Once the Monotonicity Effect was indexed behaviorally and anatomically, we could use these indices to determine the role of DE-operators as NPI licensors. NPI licensing conditions have long been debated, with syntactic and pragmatic factors weighing in. We conducted a series of behavioral studies with English and Hebrew speakers, hoping to obtain evidence for a precise characterization of these licensing conditions. We tested sentences with 2 DE operators, hosted in Homer’s (2020) flip-flop environments, and in a series of controls. To enable a coherent interpretation of our results, we defined a direct mapping from DE-domains to processing costs. We found that (i) the Monotonicity Effect is determined not by the number of DE operators, but rather, by the monotonicity of the minimal constituent in which they reside; and (ii) that DE-ness is not a property of operators, but of environments (Gajewsky, 2005; Tan et al., 2024).

Finally, we tested the same materials in fMRI (Tan et al., in prep.). Despite difficulties, we not only replicated previous results, but also, provided a new perspective on the neural bases of DE-ness. We show how our results bear directly on the current debate about the nature of monotonicity, and conclude with some speculations about the cognitive functions of the new brain area our studies uncovered.