Some current projects:

Language production of sequences (with Gary Dell)

Episodic memory for multi-word sequences (with Aaron Benjamin, Colin Bannard, Lili Sahakyan)

To what extent does our memory for alcoholic beverages or handsome wizard depend on how often we have experienced alcoholic or wizard? Do we encode phrases in the same way that we do for words? I have found that people know how common a phrase is, which influences their responses in recognition memory. However, accurate discrimination of studied and unstudied phrases depends on the components (the words) and the meaning (Jacobs et al., 2016, JML). This suggests that while phrases are more than the sum of their parts, words are better cues to recognition memory.

More recently, I have been exploring the same effects in recall. In recall, common and uncommon words that are just as fluently understood concepts are equally well-recalled. However, common phrases are better recalled than uncommon ones. We are currently tying these two results into theories of language production.

Reduction and production-internal processes (also with Duane Watson and Torrey Loucks)

In recent work we have found that practice in inner speech does not count as repetition for duration reduction purposes. However, producing a word’s homophone does count as a repeated sequence (Jacobs et al., 2015, JML). On our account, acoustic feedback leads to shortened transitions between phonemes, which leads to reduction for identical phonological sequences.

Currently I am testing how much acoustic feedback matters for repetition reduction by manipulating the amount of feedback speakers receive from their own productions.

Probabilistic processes in language production (also with Kay Bock)

Speakers and listeners converge on the statistics of their language, and are influenced by the language they hear and the people they talk to. However, syntactic priming studies to date have not looked at participants’ biases prior to exposure to primes. In a series of studies I have been looking at whether participants change proportionately with the magnitude of the difference between the experimental statistics and their initial linguistic biases.

Machine learning techniques for dialogue and discourse analysis

Human language processing and language modeling

At Texas

I also used to do work in the Cognition and Communication Lab with Zenzi Griffin and in the Child Language Lab with Colin Bannard.