Abstracts of the Plenary Talks
Plenary talk 1: "On Reducing Prosodic Domains to Phases: Successes and Challenges"
by Michael Barrie (Sogang University)
This talk will address recent advances in the syntax-phonology interface. We start by covering basic aspects of Match Theory and Phase Theory and discuss how the two can be melded into a single, minimalist framework. The core idea presented is that prosodic categories correspond to phases. Mismatches may arise and are dealt with either by more highly ranked constraints or by post-syntactic prosodic restructuring. Recent research of Mongolian, Korean, and Blackfoot will be presented. We will end with some remaining challenges to this minimalist proposal.
Plenary talk 2: "Linguistic experience in our theories of language: an attitude verb perspective"
by Nick Huang (National University of Singapore)
How does our experience with language shape how we learn and process it? Recent research into this classic question, especially within computational psycholinguistics and natural language processing, suggests that many important linguistic distinctions can be reasonably captured using just the statistics of one’s linguistic experience. For instance, various proposals have sought to model a sentence’s acceptability as a function of the probability of observing the sequence of words in that sentence; similarly, an influential approach models the meanings of words based on what other words they co-occur with.
Despite these empirical successes, I argue that a more nuanced view of the role of linguistic experience is still necessary. I present two case studies centred around the syntax and semantics of attitude verbs, such as “think”, “want”, “shout”. The first case study concerns what I will refer to as “bridge effects”: the variation in how acceptable it is for wh-dependencies to cross into the finite complement clauses of certain attitude verbs (e.g. What did Jo think/??shout that Sam saw _?). The second case study concerns how children learning English and Mandarin Chinese might learn the meaning of these verbs, many of which describe highly abstract, hard-to-observe mental states. Drawing from acceptability judgment studies, corpora, and computational models, I argue that both cases cannot be straightforwardly accounted for solely based on the statistics of easily-observed linguistic features. Rather, the case studies suggest a role for more abstract constraints and/or learning biases that involve non-trivial abstraction of one’s linguistic experience.
Plenary talk 3: "Beatboxing Phonology" by Reed Blaylock (University of Southern California)
Beatboxing is a type of non-linguistic vocal percussion that can be performed as an accompaniment to singing or rapping, or as a standalone performance. But despite having many sounds that are unattested in any language and communicating no linguistic message, beatboxing has phonology: its own phonological units and organization that are surprisingly similar to the phonological units and organization of speech systems from around the world.
The nascent research field of beatboxing science has begun to develop theories about the speechlike cognitive units of beatboxing, their organization, and their relationship to speech. This presentation highlights some key properties of beatboxing phonology using evidence from videos of expert beatboxing acquired with real-time MRI. Examples are also provided from beatrhyming, a type of music performance in which a person sings and beatboxes simultaneously. The evidence points to the possibility that speech and beatboxing share a mutual phonological foundation.
The real-time MRI videos used in this analysis were collected by Nimisha Patil (@nimitzbeatbox) and the Speech Production and Articulation kNowledge group (SPAN, https://sail.usc.edu/span) led by Shri Narayanan at the University of Southern California. Many of the videos can be watched in advance at https://sail.usc.edu/span/beatboxingproject. The project was supported by NIH R01DC007124 and NSF IIS 1514544.