Semantic predictability of implicit causality can affect referential form choice

The role of grammatical role and thematic role predictability in reference form production in Mandarin Chinese

“…In contrast to grammatical role, thematic role predictability did not affect Mandarin speakers’ choice of referential form. This finding is consistent with the literature that suggests that thematic role predictability is unrelated to referential form production (e.g., Kehler et al, 2008 ; Fukumura and van Gompel, 2010 ; Kaiser et al, 2011 ; Rohde and Kehler, 2014 ; Lam and Hwang, accepted ), but incompatible with the prior work that suggests that thematic role predictability affects pronoun use (e.g., Arnold, 2001 ; Rosa and Arnold, 2017 ; Weatherford and Arnold, 2021 ). Note, however, that the predictability effect, if attested, was relatively weak compared to the grammatical role effect.…”

“… Rosa and Arnold (2017) found that the predictability effect did not overturn the subjecthood effect and was detected only when participants showed variation in their referential form choice. Furthermore, predictability did not influence referential form choice for implicit causality (IC) verbs in numerous studies ( Kehler et al, 2008 ; Fukumura and van Gompel, 2010 ; Rohde and Kehler, 2014 ; but see Weatherford and Arnold, 2021 ). For example, Fukumura and van Gompel (2010) found that English speakers were more likely to refer to the stimulus (Ana) when asked to complete sentence fragments such as (2).…”

“…Critically, Arnold’s (1998 , 2001 ) Expectancy Hypothesis suggests that entities that are likely to be mentioned again are accessible and tend to be referred to with pronouns (see also Givón, 1983 , 1988 ). For example, Rosa and Arnold (2017) found that English speakers used more pronouns when referring to the goal than the source following sentences like (1) (see also Arnold, 2001 ; Weatherford and Arnold, 2021 ). Kaiser et al (2011) , however, provide evidence against the Expectancy Hypothesis.…”

A Bayesian Approach to German Personal and Demonstrative Pronouns

“…Perhaps in the light of the strong, counterintuitive dissociation it posits, it has been the predictions of the strong form of the Bayesian hypothesis that have received the most attention in the literature. Whereas early studies have provided evidence to support it ( Rohde, 2008 ; Fukumura and van Gompel, 2010 ; Rohde and Kehler, 2014 ; inter alia), some more recent studies, primarily by Arnold and colleagues, have found limited effects of semantic factors (thematic roles) on production ( Rosa and Arnold, 2017 ; Zerkle and Arnold, 2019 ; Weatherford and Arnold, 2021 ; see also Arnold, 2001 ). These contradictory findings leave us with the looming questions of what the source of the disparities are, and of what type of model can explain the extant data as an ensemble, especially given that the identified effects of semantic factors on production are typically more limited or otherwise inconsistent than theories that rely on a singular notion of entity prominence would predict.…”

Individual Differences in Print Exposure Predict Use of Implicit Causality in Pronoun Comprehension and Referential Prediction

“…We test this idea explicitly in Experiment 3 with the use of a metalinguistic prediction task. We know that verb bias affects predictions: when people read a fragment like Ana admired Liz , they judge that Liz should be more likely to be mentioned than Ana ( Guan and Arnold, 2021 ; Weatherford and Arnold, 2021 ). Our question here is whether this bias varies across individuals.…”