Research Article: Abstract social categories facilitate access to socially skewed words

Date Published: February 4, 2019

Publisher: Public Library of Science

Author(s): Jennifer Hay, Abby Walker, Kauyumari Sanchez, Kirsty Thompson, Shane Lindsay.


Recent work has shown that listeners process words faster if said by a member of the group that typically uses the word. This paper further explores how the social distributions of words affect lexical access by exploring whether access is facilitated by invoking more abstract social categories. We conduct four experiments, all of which combine an Implicit Association Task with a Lexical Decision Task. Participants sorted real and nonsense words while at the same time sorting older and younger faces (exp. 1), male and female faces (exp. 2), stereotypically male and female objects (exp. 3), and framed and unframed objects, which were always stereotypically male or female (exp. 4). Across the experiments, lexical decision to socially skewed words is facilitated when the socially congruent category is sorted with the same hand. This suggests that the lexicon contains social detail from which individuals make social abstractions that can influence lexical access.

Partial Text

The distribution of word usage across different speaker groups is not even. Some groups may talk about certain topics more than others or use different words to talk about the same topics. This leads to certain words being used more by some types of speakers than others. For example, the word children is more likely to be said by women than men [1]. Therefore each word has a unique social distribution in which it is encountered, or–in the words of Bakhtin–“[e]ach word tastes of the context and contexts in which it has lived its socially charged life” [2](p.293). Recent results suggest that speakers and listeners are sensitive to the differing social distributions of individual words. For example, in speech perception, words are accessed faster when produced in experience-congruent voices [3, 4], and in speech production, these words are more likely to be produced with experience-congruent phonetic variants [5, 6].

This experiment tests for an implicit association between word age and the categories OLD and YOUNG. In this experiment, participants engaged in a combined IAT and Lexical Decision Task where they sorted words and non-words while also sorting old and young faces. The words came from Walker & Hay [3] and were words that were skewed in their usage across speakers of different ages, such that older speakers were more likely to use some of these words than younger speakers, and vice versa. We predict that participants will be faster at accessing real, older words if they are using the same hand to sort real words and older faces. If we find a significant facilitation effect of face and word congruence, this would provide support that words used more by older or younger speakers are associated with older and younger people respectively.

In this experiment, we test for an implicit association between words used more often by men/women and male/female faces. The design is similar to experiment 1, except we used words that are biased in production frequencies such that they are used more often by men than women, or vice versa (as opposed to older and younger speakers). The purpose of this experiment is to conceptually replicate experiment 1, using a different social category of speakers.

Experiment 2 (and the associated lexical decision task), indicate that congruence between a person’s perceived gender (through voice or face) and the gendered distribution of word use facilitates lexical access. Both voices and faces are involved in the speech experience, however, so a congruency effect might arise in the absence of any abstracted generalization regarding gender. In experiment 3, we examine the relationship between words and the gender (female, male) of people who use those words most often by testing for an implicit association between words and a more secondary gender association via objects. That is, we get more abstract by using objects that have themselves been associated more with females or males, but aren’t necessarily directly associated with memories for the particular words we are interested in. If this effect is observed, then this would provide strong evidence that our observed associative effects are being driven by higher-order contextual labelling of individual speech events, where words and items are being categorized in (and abstracted to) the larger category of either FEMALE or MALE. We set out to test this in experiment 3, by replacing the faces used in experiment 2 with photos of gendered objects.

Experiments 2 and 3 demonstrate that people have associations between MALENESS and FEMALENESS and words that are used more by men and women. In both studies, the IAT task involved sorting gendered images (faces or objects) into MALE and FEMALE categories, making gender an explicit, conscious part of the task.

This project aimed to investigate the implicit associations people have between word usage and social categories. Across four experiments, we found that participants associated words that they have experienced more from certain groups of speakers with abstracted social categories. In experiment 1, reaction times suggested that younger listeners had an association between words encountered by older/younger speakers and older/younger faces respectively. In experiment 2, accuracy rates suggested that listeners had an association between words encountered by female/male speakers and female/male faces, respectively. In experiment 3, we showed that listeners had an association between words encountered by male/female speakers and male/female objects respectively (in both response times and accuracy rates), and in experiment 4, we showed that we did not have to ask participants to sort things explicitly by gender to see this effect emerge in accuracy and in response times.

In this study we presented the results of four combined Lexical Decision and IAT experiments, which all investigated whether lexical access was facilitated when words that participants have experienced more from certain types of speakers were paired with faces of people in that social group, or with objects stereotypically associated with the social group. We find evidence to support this hypothesis in all experiments. Critically, the results suggest that people’s socially-skewed experiences with words result in not just an association between the word and certain types of speakers, but with the word and more abstract social categories. We argue that this provides further support for hybrid models of lexical representation, in which both linguistic and social categories are associated with phonetically detailed lexical storage.




Leave a Reply

Your email address will not be published.