Joe Pater's (2019) target article calls for greater interaction between neural network research and linguistics. I expand on this call and show how such interaction can benefit both fields. Linguists can contribute to research on neural networks for language technologies by clearly delineating the linguistic capabilities that can be expected of such systems, and by constructing controlled experimental paradigms that can determine whether those desiderata have been met. In the other direction, neural networks can benefit the scientific study of language by providing infrastructure for modeling human sentence processing and for evaluating the necessity of particular innate constraints on language acquisition.

The English pronoun they is currently undergoing a rapid change, in that they is increasingly being used to refer to specific (named) individuals as a singular personal pronoun. Although it has been used with a singular, indefinite antecedent for centuries, singular specific they is relatively new and coincides with rising recognition of the fluidity of gender identity and expression. For many individuals, they/them pronouns fit their gender identity best. However, such individuals are at a high risk of being misgendered because this new usage of they is neither well established grammatically nor part of prescribed use. In two experiments, adults from across the United States created short written narratives about individuals of different gender presentations. We varied whether participants saw a pronoun in the stimuli and, if so, whether they saw they, he, or she. We found that singular specific they was used less than she/he and that they-usage increased for those who reported being more familiar with it and with the LGBTQ+ community more generally. We further found that images that appeared androgynous or nonbinary were more likely to elicit singular specific they than were images that appeared binary. Finally, we varied whether participants received brief information about the person that included singular specific they. This type of modeling led to dramatic increases in they-production overall, and increases were most robust for participants who reported higher familiarity. Overall, this research illustrates that characteristics tied to social experience, modeling, and visual cues to an individual's gender identity are highly informative for the production of singular specific they. More broadly, we illustrate that language-processing costs related to language production can be boosted for users and therefore can intervene in the likelihood of misgendering.

Whether lexical representations are stored as abstract forms or exemplar tokens is the focus of much debate in both the phonological and word-recognition literature. This research report examines the recognition of words that have undergone Canadian raising and/or intervocalic flapping. Two eye-tracking experiments suggest that listeners are slower to fixate words that have undergone one or more phonological processes within their own raising dialect, supporting the idea that they must calculate a mapping from surface word forms to more abstract representations. Implications for representational and phonological theories are discussed.

We report an experiment eliciting ordering preferences for BINOMIAL EXPRESSIONS (e.g. bread and butter vs. butter and bread) in order to investigate the respective influences of productive and item-specific knowledge in language processing. Binomial ordering preferences reflect both (i) productive constraints involving phonological, semantic, and lexical properties, and (ii) item-specific relative frequencies. Bayesian and exemplar-based computational models of acquisition and use predict influences of both productive and item-specific knowledge on ordering preferences, with item-specific knowledge playing a smaller role the lower the expression's overall frequency. Our results confirm this prediction, but also reveal a role of item-specific knowledge even for binomials with overall frequency less than one in ten million. These findings bring a quantitative perspective to the debate over the roles of productive and item-specific knowledge in language.

We investigate a noncanonical agreement pattern in American English in which a fronted WH-phrase appears to control agreement on an inflected auxiliary, as in Which flowers are the gardener planting? (Kimball & Aissen 1971). We explore this phenomenon with five acceptability- judgment experiments and interpret the resulting data with the aid of a quantitative model of the judgment process. Our study suggests that fronted WH-phrases interfere with agreement primarily as a function of their linear and structural position, and that this effect is not significantly modulated by overt case or thematic cues in off-line judgments. We suggest that our findings support a model of agreement processing in which syntactic phrases compete to control agreement on the basis of their structural and linear position with respect to the inflected verb.

In grammaticalization studies, reanalysis is understood as the assignment of new meaning to formally unchanged elements, supported by bridging contexts compatible with the old and the reanalyzed meaning. The source determination hypothesis (SDH) predicts that parallel grammaticalization trajectories occur crosslinguistically, as similar source meanings give rise to similar inferences. One such pattern is the development of recent past markers from FINISH constructions. While grammaticalization pathways are well-documented crosslinguistically, the SDH has never been tested experimentally. In this study, we examine whether modern English speakers are sensitive to inferences arising from a bridging context identified as relevant to the grammaticalization of Old Spanish FINISH into a recent past marker. In a temporal distance judgment task, we examined whether the bridging context identified for Old Spanish facilitates an inference of temporal immediacy in contemporary English, where the construction has not been grammaticalized. In line with the SDH, the bridging context enhanced an inference of immediacy in contemporary English (Exp. 1), with specific grammatical features of the source determining its strength (Exp. 2). These results not only demonstrate the viability of testing hypotheses about language change using experimental pragmatics but also call for empirically refining the concept of source determination.

Although web-based data collection has become increasingly popular in (linguistic) research over the past years, many researchers are still cautious about collecting data via the internet. Thus, this study aims at comparing web-based and lab-based testing of linguistic manipulations that have resulted in robust findings in previous lab-based research on bilingual language processing. A total of 134 L1 German students of L2 English participated in two experiments in a web-based (n = 78) or lab-based setting (n = 56). The study examined potential language co-activation through cognates in an English Lexical Decision Task (Experiment 1) and the use of L2 lexical and syntactic information in English relative clause processing in a Self-paced Reading Task (Experiment 2). We found comparable evidence of lexical and syntactic processing in both groups in both experiments. Critically, this paper provides important methodological implications for web-based data collections with second language learners.

Various theories have been proposed in the field of second language (L2) sentence processing research and have significantly advanced our understanding of the mechanisms underlying L2 sentence interpretation processes. However, many existing theories have only been formulated verbally, and little progress has been made towards formal modelling. Formal modelling offers several advantages, including enhancing the clarity and verifiability of theoretical claims. This paper aims to address this gap in the literature by introducing formal computational modelling and demonstrating its application in L2 sentence processing research. Through practical demonstrations, the paper also emphasises the importance of formal modelling in the formulation and development of theory.

This study extends the line of linguistic relativity research by assessing the effect of the French grammatical gender system on French speakers' and learners' perception of objects. Four groups of 140 adults (English monolinguals, French monolinguals, English–French bilinguals and French–English bilinguals; N = 35 each) rated 32 selected objects' gender by assigning them a masculine/feminine voice on a slider. We also assessed the participants' second-language (L2) proficiency. Multilevel modelling results revealed that French monolinguals and English–French bilinguals rated objects' gender in line with the French grammatical gender system. The effect of French on perception was not reduced by acquiring English, as French–English bilinguals performed on par with French monolinguals. Moreover, the effect was independent of L2 proficiency. These findings suggest that learning a gendered L2 affects the perception of objects – thus supporting the linguistic relativity hypothesis.