Language plays a crucial role in the formation and maintenance of stereotypes about social categories. When communicating about other people and their behaviors, language reflects the categories we have created, and what we stereotypically expect of these groups. In turn, through our linguistic expressions we share our stereotypic views with message recipients.

In the present project, we aim to gain a better understanding of how and when stereotypes are expressed in everyday language use. Thereby we build on our framework (Beukeboom & Burgers, 2019) in which we integrated earlier research on stereotyping and language use. We distinguish two categories of relevant linguistic biases: (1) biases in linguistic labeling, and (2) biases in describing the behaviors and characteristics of categorized individuals.

First, we develop an annotation guideline in which we bring together the different linguistic means though which people reveal their stereotypic expectancies in labels and behavior descriptions.

Second, we will use this guideline to annotate the language produced by participants in our experiments. In these experiments we test various factors that can influence whether or not people start stereotyping when spontaneously describing the behavior of others (i.e., generalize across individuals and situations). We also test the effects of variations in language use on stereotype formation in message recipients.

Finally, we will develop computational methods to automatically detect stereotypes in texts to uncover implicit biases in a variety of real-life contexts.

Taken together, this project creates a more encompassing understanding of the role of language in perpetuating stereotypes and discrimination in society, and facilitates future scientific and applied research by creating the ability for stereotype detection in a variety of contexts.