Linguistic Discrimination

Linguistic discrimination is a learned behavior that people exploit for discerning difference that also has the insidious ability to disadvantage individuals (Craft et al., Reference Craft, Wright, Weissler and Queen2020). To discriminate categorically is a survival skill. For example, animals can tell which young belong to them by looks and smell. As humans, we use this skill to help us move throughout the world (e.g., discerning a stop sign from a yield sign by shape and color). However, when we bring in the social dimensionality of the human experience, the natural aspects of discrimination can predispose us to prejudice. Notwithstanding this, people can have this discerning function without invoking prejudice. This is a socially motivated choice but can mistakenly be seen as innate given innovations such as the Implicit Association Test (Greenwald et al., Reference Greenwald, McGhee and Schwartz1998). However, because bias is learned, it can be unlearned. I attempt to show how we can avoid the slippery slope from discerning difference to enacting discrimination by diving into the psychological components of prediction and processing while being informed by the sociolinguistic literature and methodologies.

Linguistic Inquiry: A History of Siloed Camps

Within the past seventy years, the field of cognitive science has brought a new lens to the study of linguistics, where an interest in structure and variation within languages has expanded to research on how language is processed and perceived (Miller, Reference Miller2003). Despite this new perspective, there is relatively little knowledge of cognitive processing within multidialectal societies and minority language varieties. For instance, while AAE is the most well-studied minoritized variety of English in the U.S., there is still so much we do not know about cognitive processing of the variety. Sociolinguistic methods have historically provided a nuanced view of the social dimensions of speech variation (Bucholtz, Reference Bucholtz1999; Eckert, Reference Eckert2012; Norris et al., Reference Norris, McQueen and Cutler2003; Sumner & Samuel, Reference Sumner and Samuel2009). However, the field has not broadly drawn on methods that account for language processing happening live as it is used in context. Psycholinguistics offers an array of methods to measure processing, from brain regions to neural and behavioral responses and beyond. At the same time, psycholinguistics has historically focused on standard linguistic varieties (Hanulíková et al., Reference Hanulíková, Van Alphen, Van Goch and Weber2012; Van Berkum et al., Reference Van Berkum, Van den Brink, Tesink, Kos and & Hagoort2008). Consequently, psycholinguists have not considered how languages with variable social statuses are cognitively processed. Having access to more data nowadays, we can make robust arguments regarding linguistic phenomena, but at the same time if we are not interrogating the perspectives we come from—that is, lifting the curtain to recognize biases brought to bear in research—we will inherit the same problems we currently are dealing with.

Perceptions of African American English

This article focuses on AAE as a lens through which to better understand the necessity of a sociocognitive approach to analyzing linguistic discrimination. Americans have a commonsense notion of people sounding Black (Purnell et al, Reference Purnell, Idsardi and Baugh1999). Despite longstanding interest in research on AAE, there is still a question about the differences between “sounding Black,” “sounding like an AAE speaker,” and use or nonuse of AAE grammar. There are also conflicting notions regarding who speaks AAE, evidenced in the ever-evolving names used to refer to the variety (i.e., Ebonics, Black English, African American English, African American Vernacular English, and African American Language to name a few; Lanehart, Reference Lanehart2009; Purnell et al., Reference Purnell, Idsardi and Baugh1999; Wolfram, Reference Wolfram2007). Even more, there are also African and Caribbean Black people in America who may or may not use AAE (Blake & Shousterman, Reference Blake and Shousterman2010), giving further insight into how race is implicated in perceived speaker identity. All of the precise elements that characterize “sounding Black” remain inconclusive, though it is clear that there are some elements of Black Americans’ speech that greatly affect individuals and communities, including most blatantly, cases of racial profiling (Purnell et al., Reference Purnell, Idsardi and Baugh1999; Rickford & King, Reference Rickford and King2016; Wright, Reference Wright2020).

Among laypeople, AAE has typically been associated with individuals having lower socioeconomic status (SES), as evidenced not only from common assumptions but also as the foci of much research on AAE is on lower SES speakers. However, these prejudices also transcend SES, as seen in the literature on Middle Class African American English (Weldon, Reference Weldon2021). Salient examples include hostile treatment of high-profile individuals such as Anita Hill, when called to speak against Clarence Thomas’ Supreme Court appointment (Mendoza-Denton, Reference Mendoza-Denton1995), and harsh and unwarranted criticisms of Vice President Kamala Harris in the 2020 political debates (Holliday, Reference Holliday2020). Negative perceptions of AAE can have legal consequences, such as those extensively discussed regarding Rachel Jeantel during George Zimmerman's murder trial (Rickford & King, Reference Rickford and King2016; Slobe, Reference Slobe2016). Jeantel, as a speaker of AAE and Haitian Creole, was subject to both her grammatical usage and the sounds of her language being put on trial, despite the fact that she was the key witness in the murder of Trayvon Martin. All of these instances provide examples of speech—the percept—being judged based on discrimination and bias.

As evidenced here, Black Americans have a particular sound that leads to a particular “percept,” or “object of perception,” which can have discriminatory outcomes. As AAE researchers have expanded their conceptualizations of the variety to account for different social factors across gender, region, and socioeconomic class (Holliday & Villarreal, Reference Holliday and Villarreal2020; King, Reference King2021; Lanehart, Reference Lanehart2009), it has become clear that there is a hierarchy of AAE features ideologically associated with the variety—that is, some features are strongly enregistered (Agha, Reference Agha2005; e.g., “axe” for “ask”), while others are not, like auxiliary flexibility (Weissler, Reference Weissler2021). Even in the absence of AAE morphosyntactic features, people can still discern when speech is coming from a Black individual, as evidenced by the growing body of research on the prosodic and intonational features of AAE (Holliday, Reference Holliday2021; Holliday & Villarreal, Reference Holliday and Villarreal2020; Thomas & Reaser, Reference Thomas and Reaser2004; Weldon, Reference Weldon2021). So, while the racist outcomes are clear, the specific linguistic mechanisms are not, which is where inquiry into cognition comes into play in the pursuit of linguistic justice for racialized varieties and speakers, such as looking at how people's experience with AAE and usage of the variety modulates their interaction with it in real-time experimental paradigms (Weissler, Reference Weissler2021).

Sociocognitive Linguistic Analysis for Social Justice

While linguistics as a field is no stranger to the benefits of experimentation, it is crucial to incorporate the effects of Anti-Blackness into the experimental frame (Baker-Bell, Reference Baker-Bell2020, Charity-Hudley et al., Reference Charity-Hudley, Mallinson and Bucholtz2020). As linguists and cognitive psychologists become more attuned to talking about social justice issues and human behavior, we need to be discussing these behaviors through a sociocognitive window. Few researchers have used sociocognitive techniques to tap into people's percepts concerning foreignness and differences, gleaned from speaker behavior. Van berkum et al. (Reference Van Berkum, Van den Brink, Tesink, Kos and & Hagoort2008), showed that speaker identity modulates processing for listeners when sentences are introduced from typical and atypical actors (e.g., sentences more likely to be spoken by a man than a woman, or by an adult than a child) using electroencephalography (EEG) techniques. Their results indicated that Dutch listeners showed semantic neural responses of surprisal (called N400 event-related potential effects) for speaker inconsistencies (e.g., sentences like, “I like wine before bed” spoken by a child). This research supports the hypothesis that we take in perceived speaker information rapidly when processing sentences. Hanulíková et al. (Reference Hanulíková, Van Alphen, Van Goch and Weber2012) evidenced that listeners show neural responses of expectation violation when Dutch natives made grammatical errors, but not when L2 speakers made errors (such as a Turkish-accented Dutch speaker). These results indicate that listeners may modulate their expectations based on speaker identity, such that they might expect mistakes from a foreign-accented-sounding speaker but not from the Dutch native-sounding speaker.

Regarding American Englishes, EEG research shows that listeners alter their linguistic expectations when hearing AAE and Standardized American English (SdAE) (Weissler & Brennan, Reference Weissler and Brennan2020). Listeners show differential processing depending on whether they heard speech coming from a Black man speaking either AAE or SdAE and a white man speaking SdAE. Specifically, they showed neural responses of expectation violation to the respective varieties. There were no neural responses at all to grammatical errors (i.e., “The clown, he'll blowing up balloons”) in AAE speech. Like in Hanulíková et al. (Reference Hanulíková, Van Alphen, Van Goch and Weber2012), this research indicates that speaker identity and language variety may both be taken into account during processing. The results also indicate that listeners are not necessarily hearing SdAE when they hear a Black person use that variety, as evidenced by variable neural responses when the Black man used AAE versus SdAE. Though standard features may be evoked in the SdAE context, this speaker still “sounds Black,” contributing to further understanding of how perceived identity interfaces with cognitive processing (Weissler, Reference Weissler2021).

Additionally, visual processing studies also suggest that perceived speaker identity affects linguistic processing. Eberhardt et al. (Reference Eberhardt, Goff, Purdie and Davies2004) show that presenting faces of Black people increases participants’ speech of detection of crime-relevant objects. Additionally, their work shows that presenting objects and abstract concepts such as “crime” and “basketball” result in increased looks toward Black male faces in a Visual World Paradigm. Similarly, Liu et al. (Reference Liu, Rigoulot and Pell2014) hypothesized that listeners make judgments about speakers rapidly, given linguistic data and perceived identity information relating to emotion within American and Chinese speakers and faces. Their results suggest that culture-specific social norms that regulate how emotions are expressed in socially appropriate ways (i.e., display rules; Liu et al., Reference Liu, Rigoulot and Pell2014), play a crucial role in how people from different cultures process emotional speech.

Approaches such as EEG and eye-tracking provide a window into cognitive linguistic processing of individuals. However, it is important to mention the challenges alongside the benefits. EEG specifically has limitations in that researchers can only look at the processing of one structure at a time, since the event-related potential measure focuses on single events, which can be at odds with looking at multiple features in tandem. Innovations such as virtual eye-tracking (e.g., Webgazer, described in Papoutsaki et al., Reference Papoutsaki, Sangkloy, Laskey, Daskalova, Huang and Hays2016), are beneficial as they open up possibilities for researchers to work with populations that are not easily accessible in traditional laboratory settings. Researchers bringing laptops to the field as well can mitigate issues where access to computing technologies are limited in lower resourced communities, to minimize the use of bulky and costly equipment. The aforementioned study (Weissler, Reference Weissler2021), looking at how people's experience with AAE via their usage of the variety increases their accuracy in identifying AAE emotional speech with Black women's faces was facilitated via virtual eye-tracking. These considerations are worth keeping in mind when incorporating psycholinguistic methods into sociolinguistic experimental conceptualizations. All in all, interdisciplinary work allows more nuanced investigation of language, and being aware of the costs and benefits of each decision during inquiry is important to bring understanding of the feasibility and scope of the research.