Contemporary work in sociolinguistics and linguistic anthropology is increasingly taking up social justice as an explicit framework, building on historical scholarship that laid the foundation for understanding how language use is tied to social inequality. In such efforts, educational contexts are crucial settings for intervention, sites where sharing scholarly insights about language can provoke effective public conversations about the roots and contemporary manifestations of social inequality linked to language. Inspired by this work, we suggest that conducting such language-related educational efforts within our institutions of higher education can make crucial contributions to the advancement of social justice.

In this talk, we describe our efforts to offer workshops on language and social justice for students involved in service learning, as part of the Swearer Center co-curriculum. We discuss in detail a workshop that focused on the topic of the “word gap”: research on disparities in how children perform in school that is based on the notion that the vocabulary of children from low-income communities lags behind that of their more affluent peers. This research claims that by the time they enter kindergarten, low-income children display a 30 million word deficit when compared to higher-income students. Narratives about this so-called “word gap” are among the most pervasive discourses about language and social inequality in the United States today, and have inspired programs around the country designed to illuminate this disparity—including, here in Providence, the 5 million dollar initiative Providence Talks. However, such interventions, and the research on which they are based, have also been the target of substantive critiques by linguistic anthropologists and sociolinguists. We suggest that teaching students and others about the word gap” debate can help them gain a deeper understanding of how language works to uphold and continually reproduce social inequalities, while also inspiring them to think about how language can be used as a tool for challenging these injustices.

Two-dimensionalism is no doubt the most interesting development to follow Kripke’s Naming and Necessity. It is interesting with respect to proper names and names of natural kind terms since it combines Frege’s distinction between meaning and reference with Kripke’s views. Chalmers has written extensively on two-dimensionalism and has recently given a detailed account of propositional attitude reports. In this paper I will raise a problem for Chalmers’ account of belief reports. I will argue that empty names are a challenge to Chalmers’ analysis of belief reports.

Granted that there are belief reports with empty names in the subject position of the that-clause, Chalmers’ analysis should assign those belief reports truth values in accordance with our intuitions. Since empty names are names that appear to fail to refer, two-dimensionalism has to be combined with some view about the nature of the semantic values of empty names. I will consider three different views: the semantic value is the empty set, the semantic value is a created abstract object, and the semantic value is assigned in possible worlds considered counterfactually. On the first two views, I think, Chalmers’ analysis of belief reports is false. The third view is, I think, partially successful.

Language production and comprehension involve rapid integration of diverse information sources. Stereotypes—implicit mental associations among concepts—can influence event expectations and thus bias preferred linguistic descriptions. Above and beyond the effects of stereotypes on event expectations, however, can the mappings between event expectations and preferred linguistic descriptions themselves be biased? Here we provide an affirmative answer in one test case: in English speakers’ preferences regarding pronominal references to individuals whose gender is not known or determined, expectations that the individual might be female manifest in “she” pronoun preferences at a lower rate than expectations that the individual might be male manifest in “he” pronoun preferences. In large-scale experiments conducted during the 2016 US and 2017 UK electoral campaigns, where major-party candidates and fluctuating electoral prospects provided natural experiments in changing event expectations, we found that “she” references to the next head of state were consistently disadvantaged relative to election-outcome expectations. In further experiments we find that this bias also generalizes more broadly to pronoun production preferences in a wider variety of contexts. Finally, we investigate the influence of pronoun gender on event interpretation. For a rational comprehender calibrated to the production preferences described above, the signal provided by “she” that the referent may be female should be stronger than the signal provided by “he” that the referent may be male. In text memory experiments, however, we find the opposite to be the case. Our combined experimental results imply that, in linguistic descriptions of events involving a individual whose gender may not be known, speaker expectations regarding that individual’s likely gender are not faithfully transmitted — rather, expectations that the individual may be female are systematically under-conveyed. These findings pose challenges for rational theories of pragmatics, and exemplify how the tools of experimental psycholinguistics can contribute to our understanding of implicit cognition.

This talk reports ongoing work in collaboration with Veronica Boyce, Till Poppels, and Titus von der Malsburg. An initial paper can be found here.