NSF project: Does discourse priming change reference comprehension strategies? (#1917840)
NSF project #1917840: Does discourse priming change reference comprehension strategies? (2019-2022)
Abstract
Understanding pronouns requires comprehenders to use the context to infer the intended referent. For example, in a story like “Ana cleaned up with Liz. She needed the broom,” people tend to assume that “she” refers Ana, who was recently mentioned and in the prominent position of subject. This subject bias is one example of several well-known biases that stem from the discourse context, yet little is known about how these biases are learned. We test the exposure hypothesis, namely that people learn discourse conventions through exposure to patterns in the linguistic input. For example, if I frequently encounter discourse where the subject is re-mentioned (e.g., “Ana liked Liz. Ana also liked Matt. She invited the two of them to her party,”) I will expect to encounter that same subject-reference pattern in the future, and this expectation biases my interpretation of ambiguous pronouns (Arnold, 1998, 2001). Consistent with this idea, people with higher print exposure exhibit a stronger subject bias for pronoun comprehension (Arnold et al., 2018), but the mechanisms behind this effect have received little study.
It is important to understand the role of exposure to pronoun comprehension for several reasons. First, several current models account for reference comprehension in terms of reference probability, namely the likelihood that the speaker will mention referent x (Arnold, 1998; Frank & Goodman, 2012; Hartshorne et al., 2015; Kehler & Rohde, 2013). This implies that people calculate these probabilities. Second, recent findings suggest that people use the same contextual information to both understand pronouns and make predictions about who will be mentioned (Zerkle & Arnold, 2019; Guan & Arnold, 2019), which supports this class of models. However, almost nothing is known about how people estimate referential probability. The proposed project fills this gap by testing whether the frequency of referential patterns in recent experience can change the way people understand pronouns, under the assumption that if a type of reference is highly frequent, it will be judged as probable in new situations. We use a paradigm developed by Williams and Arnold (2019). They primed adult participants with stories that either always re-mentioned the subject character (“Will hiked with Liz. He looked at the map,”) or always re-mentioned the nonsubject character (“Will hiked with Liz. She looked at the map.”) Priming changed how adults interpreted ambiguous pronouns in the experiment, where subject primes boosted the subject effect and nonsubject primes increased nonsubject interpretations. This shows that short-term exposure to discourse patterns can change pronoun interpretation strategies. Yet numerous questions remain about how people learn about referential probability from discourse exposure. The proposed project tests how short-term exposure to discourse exposure modulates pronoun comprehension mechanisms. This project examines adults; future work will test how exposure also shapes development in children. Aim 1 is to examine what kinds of generalizations people draw from exposure. Does exposure to primes in one structure (e.g., one verb class) affect other structures? Does priming affect both semantic and syntactic biases? Can people learn contingent, verb-specific patterns through priming? Aim 2 examines how much input is necessary to learn a referential pattern, and how long it lasts. How is learning influenced by the concentration of a prime type in the context, both overall and relative to competing primes? Do longer, narrative texts have the same priming effects as our simple two-sentence stories? Does the learning from short-term exposure extend beyond the experimental session, and if so for how many days?
Journal Publications
Johnson, E., & Arnold, J. E. (2022). The Frequency of Referential Patterns Guides Pronoun Comprehension. JEP:LMC. http://dx.doi.org/10.1037/xlm0001137 [co-funded by NSF #1651000]
Arnold, J. E., Mayo, H., & Dong, L. (2021). My pronouns are they/them: Talking about pronouns promotes singular they. Psychonomic Bulletin and Review. DOI: https://doi.org/10.3758/s13423-021-01905-0 [co-funded by NSF #1651000]
Ye, Y., & Arnold, J. E. (2023). Discourse-Level Adaptation in Pronoun Comprehension. Language and Linguistics Compass, e12481. https://doi.org/10.1111/lnc3.12481.
Ye, Y. & Arnold, J. E. (2023). Learning the statistics of pronoun reference: by word or by category? Cognition.
Arnold, J. E., Venkatesh, R., & Vig, Z. (In press). Gender Competition in the Production of Nonbinary ‘They’. Glossa Psycholinguistics.
Technical reports
Arnold, J. E., Mayo, H., & Dong, L. (2020). Individual differences (or the lack of them) in comprehension of singular they. Technical Report #3. UNC Language Processing Lab, Department of Psychology & Neuroscience, University of North Carolina – Chapel Hill, Chapel Hill, North Carolina.
Papers in prep or under revision / review
Arnold, J. E., Venkatesh, R., & Vig, Z. (under review). Gender Competition in the Production of Nonbinary ‘They’
Ye, Y. & Arnold, J. E. (under review). Implicit causality affects pronoun use in interactive fragment completion tasks. Ms., University of North Carolina at Chapel Hill. [SUPPORTING MATERIALS] (email for copy of paper) [co-funded by NSF #1651000]
Arnold, J. E. (under review). Hearing pronouns primes speaker to use pronouns. Ms., UNC Chapel Hill. (email for copy of paper)
Arnold, J. E. (under review). What categories matter for tracking discourse statistics? Ms., UNC Chapel Hill. [SUPPORTING MATERIALS] (email for copy of paper)
Theses
Johnson, E. (2022). The effect of multi-structure contexts on pronoun comprehension. Ph.D. Dissertation, University of North Carolina at Chapel Hill.
Ye, Y. (2023). Tracking Referential Adaptation Online in A Visual World Paradigm. Masters Thesis, University of North Carolina at Chapel Hill.
Conference Presentations
Arnold, J. E. (2024). Learning discourse patterns through exposure: Mixed input helps identify informative categories. Poster (expected); Experimental approaches to Linguistic Meaning, June 2024, U. Pennsylvania.
Arnold, J. E. (2024). Learning discourse patterns through exposure: Mixed input helps identify informative categories. Poster (expected); Human Sentence Processing Conference, May 2024, U. Michigan.
Roy, S., Johnson, E., & Arnold, J. E. (2024). Poster (expected); Human Sentence Processing Conference, May 2024, U. Michigan.
Ye, Y. & Arnold, J. E. (2024). Implicit Causality Affects Pronoun Use in Interactive Fragment Completion Tasks. Poster, Human Sentence Processing Conference, U. Michigan. (June 2024).
Arnold, J. E. (2023). Hearing pronouns primes speakers to use pronouns. Talk, Human Sentence Processing Conference, Pittsburgh, PA.
Arnold, J. E., Payst, N., Kakoki, E., & Wang, L. (2023). Online comprehension of nonbinary “they” is easier with supporting linguistic context. Poster, Human Sentence Processing Conference, Pittsburgh, PA.
Ye, Y., & Arnold, J. E. (2023). Implicit Causality Can Affect Pronoun Use in Fragment Completion Tasks. Poster, Human Sentence Processing Conference, Pittsburgh, PA.
Ye, Y., & Arnold, J. E. (2022). Learning the statistics of pronoun reference: by word or by category? Talk, Architectures and Mechanisms of Language Processing, York, England, Sept. 2022.
Ye, Y., & Arnold, J. E. (2022). Implicit Causality Can Affect Pronoun Use in Fragment Completion Tasks. Poster, Architectures and Mechanisms of Language Processing, York, England, Sept. 2022.
Arnold, J. E. (2022). Can pronoun production be primed? Poster (expected), Architectures and Mechanisms of Language Processing, York, England, Sept. 2022.
Ye, Y., & Arnold, J. E. (2022). Learning the statistics of pronoun reference: by word or by category? Poster, Human Sentence Processing conference, Santa Cruz, CA (virtual), March 2022.
Arnold, J. E., Venkatesh, R., & Vig, Z. (2022). University students produce nonbinary singular “they” almost as well as “he” or “she”. Poster, Human Sentence Processing conference, Santa Cruz, CA (virtual).
Arnold, J. E., Wall, A., & Steele, T. Are both syntactically and semantically-based pronoun dependencies stored in memory? Brief Talk, CUNY conference on human sentence processing, University of Pennsylvania, virtual (March 2021).
Langlois, V., & Arnold, J. E. Adaptation to discourse patterns depends on the relative frequency of competing structures. Brief Talk, CUNY conference on human sentence processing, University of Pennsylvania, virtual (March 2021).
Arnold, J. E., Mayo, H. & Dong, L. Singular they understood better after explicit introduction. Talk, CUNY conference on human sentence processing, University of Massachusetts, Amherst (March 2020).
Ye, Y., Weatherford, K., & Arnold, J. E. Implicit Causality Can Affect Pronoun Use in Fragment Completion Tasks. Brief Talk, CUNY conference on human sentence processing, University of Pennsylvania, virtual (March 2021).