Semantic predictability of implicit causality can affect referential form choice

The production of referring expressions is influenced by the likelihood of next mention

The present study provides converging evidence across three next-mention biases that likelihood of coreference influences the choice of referring expression: implicit causality (IC), the goal bias of transfer-of-possession (ToP) verbs, and implicit consequentiality (I-Cons). A pilot study and four experiments investigated coreference production in German using a forced-reference paradigm. The pilot study used object- and subject-biased IC verbs, showing a statistically marginal influence of next-mention bias on referential expressions, albeit mediated by grammatical function and feature overlap between antecedents. Experiment 1 focused on these features for object reference with ToP verbs, showing effects of coreference bias. In a within-participants comparison, Experiment 2 showed comparable effects for two classes of IC verbs, stimulus-experiencer and experiencer-stimulus predicates. Experiment 3 replicated and extended the IC form effects to another verb class, agent-evocator verbs. Finally, Experiment 4 revealed effects on anaphoric form also for I-Cons, while simultaneously replicating the effect observed for IC.

A systematic evaluation of factors affecting referring expression choice in passage completion tasks

There is a long-standing controversy around the question of whether referent predictability affects pronominalization: while there are good theoretical reasons for this prediction (e.g., Arnold, 2008), the experimental evidence has been rather mixed. We here report on three highly powered studies that manipulate a range of factors that have differed between previous studies, in order to determine more exactly under which conditions a predictability effect on pronominalization can be found. We use a constrained as well as a free reference task, and manipulate verb type, antecedent ambiguity, length of NP and whether the stimuli are presented within a story context or not. Our results find the story context to be the single important factor that allows to elicit an effect of predictability on pronoun choice, in line with (Rosa and Arnold, 2017; Weatherford and Arnold, 2021). We also propose a parametrization for a rational speech act model, that reconciles the findings between many of the experiments in the literature.

Choice of nominative and topic markers in Korean discourse

In two written sentence completion experiments, I tested whether the choice of nominative and topic markers is subserved by the same mechanisms as anaphor choice and referent choice in Korean. I found that morphological marker choice was sensitive to various factors including connective, grammatical role, and verb semantics, whereas anaphor choice was sensitive only to grammatical role. I also found that morphological marker choice did not systematically vary with verb semantics unlike referent choice. These findings reveal two important properties of referential form production. One is that morphological marker choice relies on different mechanisms than underlie anaphor choice and referent choice. Another is that connective plays a significant role in determining referential form. Despite its role in referent choice and pronoun comprehension, connective has received little attention in current models of referential form production. The current study is the first to systematically test the effect of connective on referential form choice and to demonstrate its effect on morphological marker choice. By revealing how morphological marker choice is related to anaphor choice and referent choice, and identifying the factors that influence morphological marker choice, the results of the study improve our theoretical understanding of morphological marker production and reference production.

A chimpanzee by any other name: The contributions of utterance context and information density on word choice

An important feature of language production is the flexibility of lexical selection; producers could refer to an animal as chimpanzee, chimp, ape, she, and so on. Thus, a key question for psycholinguistic research is how and why producers make the lexical selections that they do. Information theoretic approaches have argued that producers regulate the uncertainty of the utterance for comprehenders, for example using longer words like chimpanzee if their messages are likely to be misunderstood, and shorter ones like chimp when the message is easy to understand. In this work, we test for the relative contributions of the information theoretic approach and an approach more aligned with psycholinguistic models of language production. We examine the effect on lexical selection of whole utterance-level factors that we take as a proxy for register or style in message-driven production accounts. Using a modern machine learning-oriented approach, we show that for both naturalistic stimuli and real-world corpora, producers prefer words to be longer in systematically different contexts, independent of the specific message they are trying to convey. We do not find evidence for regulation of uncertainty, as in information theoretic approaches. We offer suggestions for modification of the standard psycholinguistic production approach that emphasizes the need for the field to specify how message formulation influences lexical choice in multiword utterances.

A cognitive modeling approach to learning and using reference biases in language

During real-time language processing, people rely on linguistic and non-linguistic biases to anticipate upcoming linguistic input. One of these linguistic biases is known as the implicit causality bias, wherein language users anticipate that certain entities will be rementioned in the discourse based on the entity's particular role in an expressed causal event. For example, when language users encounter a sentence like “Elizabeth congratulated Tina…” during real-time language processing, they seemingly anticipate that the discourse will continue about Tina, the object referent, rather than Elizabeth, the subject referent. However, it is often unclear how these reference biases are acquired and how exactly they get used during real-time language processing. In order to investigate these questions, we developed a reference learning model within the PRIMs cognitive architecture that simulated the process of predicting upcoming discourse referents and their linguistic forms. Crucially, across the linguistic input the model was presented with, there were asymmetries with respect to how the discourse continued. By utilizing the learning mechanisms of the PRIMs architecture, the model was able to optimize its predictions, ultimately leading to biased model behavior. More specifically, following subject-biased implicit causality verbs the model was more likely to predict that the discourse would continue about the subject referent, whereas following object-biased implicit causality verbs the model was more likely to predict that the discourse would continue about the object referent. In a similar fashion, the model was more likely to predict that subject referent continuations would be in the form of a pronoun, whereas object referent continuations would be in the form of a proper name. These learned biases were also shown to generalize to novel contexts in which either the verb or the subject and object referents were new. The results of the present study demonstrate that seemingly complex linguistic behavior can be explained by cognitively plausible domain-general learning mechanisms. This study has implications for psycholinguistic accounts of predictive language processing and language learning, as well as for theories of implicit causality and reference processing.

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

Evidence suggests that English speakers use pronouns when referring to the grammatical subject and predictable thematic role. We tested how grammatical role and thematic role predictability affect different types of referential forms, namely, overt pronouns and null pronouns in Mandarin Chinese. We found that both overt and null pronouns were sensitive to grammatical role. However, we did not find any evidence that overt and null pronouns were sensitive to thematic role predictability. Although null pronouns were influenced by grammatical role, the rate of null pronouns for subject reference was very low compared to that of overt pronouns. Given the frequent occurrence of null pronouns in Mandarin, our results suggest that the use of null pronouns may not be explained by a simple grammatical role mechanism.

Greater entropy leads to more explicit referential forms during language production

Speakers can refer to previously-mentioned words (e.g., the actor) using attenuated referring expressions such as pronouns (he), or elaborated referential forms such as repeating the original word (the actor). Predictability is theorized to influence form of reference during language production: More attenuated forms may be used to refer to more predictable words, presumably because predictable words are already active in memory, and therefore require less linguistic signal during subsequent reference. However, the reported results are mixed. The current study examines the effect of entropy, an information-theoretic metric that captures the predictability of all, not just one, referential candidate, on the production of referential forms. A meta-analysis combining data from multiple experiments (492 participants, 405 items) revealed that greater entropy leads to more explicit referential forms, suggesting that entropy might intensify the competition between referential candidates during language production, reducing total memory activation.

A Bayesian Approach to German Personal and Demonstrative Pronouns

When faced with an ambiguous pronoun, an addressee must interpret it by identifying a suitable referent. It has been proposed that the interpretation of pronouns can be captured using Bayes' Rule: P(referent|pronoun) ∝ P(pronoun|referent)P(referent). This approach has been successful in English and Mandarin Chinese. In this study, we further the cross-linguistic evidence for the Bayesian model by applying it to German personal and demonstrative pronouns, and provide novel quantitative support for the model by assessing model performance in a Bayesian statistical framework that allows implementation of a fully hierarchical structure, providing the most conservative estimates of uncertainty. Data from two story-continuation experiments showed that the Bayesian model overall made more accurate predictions for pronoun interpretation than production and next-mention biases separately. Furthermore, the model accounts for the demonstrative pronoun dieser as well as the personal pronoun, despite the demonstrative having different, and more rigid, resolution preferences.

The Effects of Semantic Role Predictability on the Production of Overt Pronouns in Spanish

In order to refer in any language, speakers must choose between explicit forms of expression, such as names or descriptions, or more ambiguous forms like pronouns. Current models suggest that reference form is driven by subjecthood, where speakers in English choose pronouns for the subject, and speakers of null pronoun languages like Spanish or Italian use null pronouns. We test this generalization by examining the effect of a different factor, thematic role predictability, on reference production in Spanish. In stories about transfer events (e.g., Ana gave a ball to Liz), speakers prefer to use pronouns more for reference to goals (Liz) than sources (Rosa and Arnold, Journal of Memory and Language 94:43-60, 2017). However, this has not been examined for null pronoun languages. In two experiments, we demonstrate that Spanish speakers are also sensitive to thematic role, but it primarily affects the rate of overt pronouns (ella, el) rather than null pronouns. These results highlight the need to include semantic constraints in models of reference production for null-pronoun languages.

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

In three experiments, we measured individual patterns of pronoun comprehension (Experiments 1 and 2) and referential prediction (Experiment 3) in implicit causality (IC) contexts and compared these with a measure of participants’ print exposure (Author Recognition Task; ART). Across all three experiments, we found that ART interacted with verb bias, such that participants with higher scores demonstrated a stronger semantic bias, i.e., they tended to select the pronoun or predict the re-mention of the character that was congruent with an implicit cause interpretation. This suggests that print exposure changes the way language is processed at the discourse level, and in particular, that it is related to implicit cause sensitivity.