Time: 11:00am - 12:00pm Venue: Scape 0.13
Amelia Kimball (LLF, CNRS, Uni. Paris Diderot, Paris 7; https://ameliakimball.com/) will be giving two short talks on two related but different projects:1. Prosody + Memory Collaborators: Jennifer Cole (Northwestern), Stefanie Shattuck-Hufnagel (MIT), Gary Dell (University of Illinois)"Is prosody understood with abstract categories? What elements of intonation are recorded in memory?"Prosody in American English has been modelled as abstract pitch accents, analogous to phoneme categories in generative phonology. My work uses experiments to test the representation of pitch accents in memory, and directly compare this representation to phonemes. I find that listeners are able to remember fine details of pitch accents, and that compared to phonemes, there is little evidence of category boundaries even in a task that is highly sensitive to boundary effects.2. Meter + Regularity + MemoryCollaborators: Duane Watson (Vanderbilt), Loretta Yiu (University of Washington), Jennifer Cole (Northwestern)"Are patterns of word stress regular in English? How do word stress patterns affect memory?"It is known that if word stress creates a pattern (meter/rhythm), listeners have some sensitivity to this pattern in the moment. If regular patterns extend, this could allow listeners to entrain to the signal and predict upcoming speech . But everyday speech is quickly unfolding, and acoustic measures indicate that conversational speech is not strictly regular. In an experiment with Jennifer Cole I use a transcription task (RPT, Cole, Mo, & Baek 2010, see also Cole & Shattuck-Hufnagel 2016) to test whether everyday speech shows regular patterns.We find that regular patterns of alternating stressed and unstressed syllables are rare in samples from a corpus of conversational American English. In another study with Duane Watson and Loretta Yiu, we use serial recall to test whether online sensitivity to metrical patterns has effects on downstream memory. We find that irregular patterns show an advantage in memory.