What the Way We Form Our Words Reveals about Us

In August 2020, Pew Research predicted that more than half of elementary and high-school students would attend class virtually in the fall due to the pandemic, and this is only part of the picture. With workplaces, classrooms, and health care visits all pushed online, our voices have quickly become more important than ever. This rapid shift to remote communication has reduced attention to nonvocal communicative cues while simultaneously increasing the salience of the vocal elements of conversation. Poor audio quality caused by unstable internet connections or lack of webcam access can compound this issue because people must then rely more on the sound of their voice to carry their input. Since the digital divide disproportionately affects minority and rural communities, these groups now have more difficulty communicating and being heard. The result is that we find ourselves in a pivotal moment to recognize voice-dependent discrimination, known as “linguistic profiling,” and to reflect on the harm that existing discrepancies in internet access produce when coupled with underlying prejudices that perpetuate inequalities in everyday life.

In any language community, speakers’ accent, pronunciation, word choice, and other linguistic features (e.g., hesitations, intonation, volume) all convey social information about them, including their geographic origin, race, ethnicity, gender, and/or socioeconomic background. These quick judgments can be innocuous, such as inferring that your coworker is from the Midwest when they ask for a “pop” instead of “soda.” But what happens when these linguistic clues link to stereotypes and prejudices? In such situations, the assumptions made about the way a person speaks can translate into negative assumptions about the person themselves—this is linguistic profiling. Page 345 →As we navigate the continuing waves of COVID-19, it is unlikely that we will be back to face-to-face meetings anytime soon. We must thus ask ourselves: how are biases toward minority communities emphasized differently when professional communication is pushed onto virtual platforms, and are we able to delink internal bias from the content of what a speaker is telling us?

In this essay, I will provide an overview of the vast topic of linguistic profiling using current examples from the United States and will show how the pandemic may make instances of such discrimination more prevalent. In highlighting this, I aim to contextualize linguistic profiling within the long histories of inequity against minorities in the United States. Then, based on my own scholarship on the also-profiled Andalusian Spanish dialect, I demonstrate how the present crisis makes these deeply entrenched biases more obvious and offers an opportunity to create more inclusive practices.

What Is Linguistic Profiling?

Making assumptions based on speech is an integral and often positive part of communication, but when this leads to stereotyping and discrimination, those assumptions can create harm and make communication difficult for many speakers. Just as racial profiling uses people’s appearance to assume their race and activate social stereotypes, linguistic profiling uses language to assume characteristics about a speaker.1 By using auditory cues that might identify which race, ethnicity, or linguistic subgroup a speaker belongs to, listeners unintentionally activate implicit biases they hold regarding these groups, thereby leading to inequitable treatment.

Beyond this, the effects of linguistic profiling intersect with other types of bias relating to social identity. In a study by Kelly Wright,2 housing applicants speaking African American Language (AAL)3 or Southern American English (SAE) were less likely to be offered access to housing in the United States than speakers of mainstream US English, even when the applicants had only communicated via phone. Notably, the effect was greater for AAL speakers than for speakers of SAE, who are predominantly white.4 A recent publication by a research team from the University of Michigan5 reviews empirical outcomes documenting such inequalities due to language perceptions across multiple social spheres—education, employment, media, justice systems, housing markets, and health care institutions.

Of course, these issues are further complicated due to the complex history and social contexts surrounding the use of minority languages. In the case of Page 346 →the United States, the shadow of slavery, widespread Jim Crow policies, and centuries of structural racism provide a backdrop for many of the current race- and class-based struggles manifesting in the present time of unrest and exacerbated inequality. The resulting power dynamics, which taint our relationship with language and which position language itself as a structural element maintaining racial logics, allow for implicit bias to discredit the voices of linguistic subgroups and deny equal treatment. In the trial following the murder of Trayvon Martin in Florida, for example, Martin’s friend and key witness, Rachel Jeantel, testified using AAL. Following the trial, one of the jurors reported that Jeantel was “hard to understand,” with another stating that “her testimony played no role whatsoever in their decision.”6 Though she provided crucial details to the jury, having been on the phone with Martin at the time of his death, Jeantel’s testimony was largely ignored during the jury’s deliberations due to their unfamiliarity with Jeantel’s speech patterns.7 This is not a one-off: throughout judicial and nonjudicial environments, marginalized language varieties are often misheard, misjudged, or ignored altogether, creating an unjust power structure of language that promotes discrimination. These situations create opportunities for implicit bias to manifest and ignore underrepresented voices.

To avoid the dangers of such inequality, many speakers of marginalized dialects “code-switch” to sound like the group with more social power. Code switching describes the process of alternating between two or more languages or dialects in day-to-day life. This phenomenon is so commonplace and well understood that it is actually the premise for a popular film: in the hit movie Sorry to Bother You, a Black telemarketer has trouble connecting with his clients until an older operator advises him to “use his white voice.” Code switching is one example of many unfair adjustments minorities make to better fit into a white-dominated culture. These problems, and the choice of whether to code-switch (if able), are intensified by COVID-19 for students and professionals alike who now rely more on their voices than ever for daily communication. However, code switching is a double-edged sword, as doing so ends up strengthening the positive associations of the prestigious dialect without legitimizing the use of the marginalized dialect in such a setting. These associations are also strengthened by the expectations of those in power. Some teachers demand the reproduction of accent-less speech from second-language students, including many Latinx youth in the United States. As a result, learners are internalizing the idea that accented English is not acceptable in any situation.8

In effect, linguistic profiling both reinforces and perpetuates existing prejudicesPage 347 → and social inequalities, often hurting ethnic minorities, immigrants, and other cultural subgroups such as younger speakers or members of the LGBTQ+ community. In a period of widening social inequality, the existence of and potential increase in linguistic profiling can be especially harmful.

Linguistic Bias and Social Justice in 2020

Coinciding with the intensified struggle with linguistic prejudice and the litany of other issues brought on by the pandemic has been a moment of cultural reckoning about anti-Black racism. The Black Lives Matter movement has origins tied to the history of anti-Black prejudice, including linguistic prejudice. Founded in 2013 following the acquittal of Trayvon Martin’s killer, Black Lives Matter aims to acknowledge the presence of systemic racism and protest the ways that anti-Black bias continues to undermine human rights, especially in the form of extrajudicial killings. Through this combined context, discussions about linguistic profiling are especially relevant and help us better understand the many ways the pandemic will change our means of communication and the judgments we make about one other.

Yet it remains impossible to determine how the intersection of the pandemic and the Black Lives Matter movement will change the way that marginalized voices are treated in the United States and abroad. The movement has greatly emphasized the role of implicit bias in widening social inequalities. Will the ongoing campaigns usher in new civil rights laws that pay legal attention to linguistic profiling? How might the role of language in the judicial system change as legal proceedings are conducted remotely, when the very existence of linguistic profiling is often denied by lawyers and police? With schools adjusting to remote and hybrid setups, how might students be treated differently when they rely more heavily on their voices to relay information or ask questions?

As we move forward in our new world of remote interaction, it is imperative that we answer such questions by understanding how linguistic profiling occurs and acknowledging implicit biases. Specifically, it will become necessary to make linguistic profiling a sustained part of the ongoing conversation about race. We must also deconstruct harmful patterns during communication, both by listening more closely to the content of what our interlocutors say rather than how they are saying it, and by normalizing other’s speech through exposure and deliberate learning efforts. For example, a recent publication on bias against AAL in the transcriptions of white and some Black court reporters Page 348 →speaks to the importance of deliberate inclusion of marginalized speech varieties in professional training.9 This type of training fosters familiarity for non-AAL speakers as well as destigmatization both at the internal (among AAL speakers) and external (among non-AAL speakers) levels.

Unmuting Voices in a Pandemic

Early in the pandemic, the Pew Research Center found that 53% of Americans believed that the internet was essential to them during the viral outbreak.10 Their results also reveal deeper inequities: across both urban and rural areas of the United States, one-third of parents say their children rely on their cell phones to finish schoolwork due to a lack of computer access at home, a statistic that rises to 43% for lower-income parents. The same survey found that 40% of low-income parents believe their children will need to seek out public Wi-Fi to finish assignments, compared to 22% overall.¹⁰ With both disease spread and unemployment disproportionately affecting lower-income and minority communities, reliance on strong internet access is only further weakening these communities’ abilities to support themselves. The compounding effects of inadequate internet access or devices with the higher digital literacy demanded by online schools create difficult learning environments for thousands of students, especially children who do not natively speak English but are expected to learn the language with limited in-person interaction.11 The same is true for minorities seeking medical treatment—not only are African Americans already less likely to receive quality treatment from doctors,12 but those with poor internet access are now less able to access telemedicine services that make health care safer during the pandemic.

These technological inequities represent another intersection of linguistic prejudice with existing disparities. Therefore, conducting research into the speech cues that prime stereotypes will contribute to honest discussions of the systemic biases that are inextricably linked with nonstandard speech. Scholars in linguistics are well positioned to combine empirical findings and humanistic insights in ways that can influence social and policy change. Craft et al. (2020) highlight the importance of linguistics especially in driving some of these shifts:

Because it is not widely believed that language is a signal of who we are, where our roots lie, or who the people are that we think of as ours, it also is not common to believe that discrimination based on language is akin to Page 349 →discrimination based on identity, history, regionality, or community. This is one area in which linguists have a special role to play in the societies in which we live because we are the professionals who know these orientations to be inaccurate, and we have the skills and tools to explain how and why. (402)

It will be difficult, nonetheless, to find direct evidence for whether linguistic profiling has increased during the pandemic, especially because many people fail to acknowledge their biases or otherwise mask them. However, with the pandemic already exacerbating existing inequalities and the transition online making the sound of our voices more relevant, the conditions are ripe for more instances of linguistic profiling. I therefore encourage speakers of all language varieties to consider the situations where such bias may occur and do the work to address these forms of discrimination.

My scholarship on Andalusian Spanish offers one example of how this can happen. Although the distinction is regional rather than racial, Andalusia offers parallel examples to the United States in its demonstration of how language plays a role in propagating social inequalities. For many Andalusians, the most salient feature that sets them apart from other Spaniards is their dialect, in a similar vein that linguistic profiling in the United States often involves ethnic or regional biases deduced from judgments about speech. By understanding which linguistic elements comprise the Andalusian dialect and contribute to Andalusian identity, we can better identify instances of linguistic profiling and give legitimacy to the accent. As our digital future pushes us to rely more heavily on voices in our communications, now is the time to reflect seriously on this point.

An Ocean Away, Similar Sentiments Prevail

Andalusia is an autonomous region in southern Spain that includes well-known cities such as Granada, Málaga, and its capital, Seville (see Figures 1 and 2). It is the most populous autonomous region of the country, governing over eight million people, and represents a contemporary vacation spot for travelers from Northern Europe. Many signature cultural icons come from Andalusia, including visual artists such as Pablo Picasso and Diego Velázquez, and poets and playwrights such as Luis de Góngora, Gustavo Becquer, and Federico García Lorca. Prior to the fifteenth century, large areas within Andalusia were ruled by North African Moors, whose legacy continues to influence local culture, particularly in some of the regional vocabulary Page 350 →and surviving architecture. Economically, Andalusia’s rural landscape yields large exports of wine, pork, and olive oil. However, this agriculture-based commerce results in a lack of major industrial urban centers that are more common to northern Spain. Additionally, after decades-long repression and population decimation during Francisco Franco’s dictatorship (Andalusia was a bastion of many anti-Franco communist and anarchist parties), weak upward mobility continues to hold back southern households from achieving higher incomes.13

Andalusia is not a linguistically homogeneous region. The southern provinces have variable use of seseo versus ceceo (i.e., the use of either the “s” or “th” sounds to pronounce the letters “s,” “c,” and “z”), in addition to various ways of producing the “ch” and “sh” sounds in words like muchacho ‘boy’ or chorizo ‘sausage.’ Furthermore, the core elements of Andalusian Page 351 →speech, such as eliminating the “s” at the end of words, actually represent an advanced development of Castilian Spanish, along with other linguistic innovations. Yet, due to its unconformity with Castilian Spanish, which is the prestige variety, Andalusian Spanish speakers enjoy less social value than normative speakers and are often treated as speaking an improper, poorly educated form of Spanish, which can negatively impact speakers’ self-esteem and self-efficacy.14 This is in part because the norm for what Castilian Spanish should look and sound like is established by the government agency Real Academia Española. Foreign books are typically translated into normative Castilian Spanish, even in places where a regional dialect has enough common prestige to merit a local translation of unique words or phrases. Overall, the lack of recognition of Andalusian Spanish as a valid accent continues to propagate stereotypes and harm the self-esteem of its speakers.

In addition, Andalusian speakers commonly face a variety of stereotypes attributed to the region; they can be stereotyped as lazy, unproductive, unprofessional, and self-interested.15 While some depictions of the region show a lively and expressive populace, the imagery is also of a Spain that is rural, local, traditional, and religious, with the impression that it is less developed than the northern half of the country.16 In many cases, readers of books or viewers of films are not explicitly told a scene is set in Andalusia; rather, the social context is made obvious through the use of the Andalusian accents by local characters.

The economic consequences of the pandemic have only heightened the social inequalities felt by Andalusians. Tourism accounts for 14% of the region’s GDP, and the slowing of travel has contributed to layoffs that have affected over a quarter of the region’s workers, the worst in mainland Spain.17 Thousands of jobs in Andalusia are at risk, as myriad European countries issue stern travel warnings or require visitors returning from Spain to quarantine. In a direct parallel, the economic impact of the shutdown in the United States has also disproportionately affected minority communities: people of color are less likely to be able to telework, more likely to work in essential positions that leave them vulnerable to infection, and also more likely to work in low-wage industries that are experiencing record layoffs. Amid these challenges faced by the speakers of minority dialects, language-based stereotypes may seem minor in comparison; however, by judging speakers preemptively, their learning, employment, health care, and other needs are affected during a time of crisis. As we navigate the continuing waves of COVID-19 and remain in Page 352 →a remote environment, we need to give ourselves extra checks to ensure that we are not relying on bias when talking with interlocutors of different dialects or accents.

The Role of Academics in Addressing Linguistic Bias

Now more than ever, the pandemic and the Black Lives Matter movement show how language prejudice further widens social gaps. This sentiment is echoed in Craft et al. (2020): “Linguists bear the greatest responsibility to examine not only how language functions as an object of study but also how that object interfaces with the systems our societies create” (402).18 Since moving to remote research and teaching in March 2020, I have reflected on how I can translate my research and teaching into a form of activism in the context of current events. Essentially, across professions and specializations, there is room for everyone to meaningfully consider the implications of our work and to open space for innovation.

For my students, class discussions surrounding Andalusian speech open the door to specific conversations about what speech elements prompt stereotypes. These conversations, in my view, are essential for deconstructing the stigmas attached to marginalized varieties. To offer one example of how this has happened, in June 2020 my students and I collaborated on a project to explore the depth of linguistic stereotypes in Andalusia across Spain. In a seminar entitled “Do You Speak Andalusian?,” We developed a survey about language attitudes toward Andalusian Spanish. There were two parts to our survey: first, dialect identification, and second, attribute ratings. In the first phase of the survey, we asked the participants to identify the region of Spain most associated with the pronunciation that they heard in the stimulus. The respondents overwhelmingly labeled the Andalusian pronunciations as belonging to Andalusia, showing that the dialect is easily and immediately identifiable.

In the second phase of the survey, we assessed the attitudes the listeners had toward the pronunciation of each stimulus. Our preliminary findings reveal that after hearing just one word or phrase from a speaker, Spanish listeners were more likely to rank the Andalusian pronunciations as less serious, less educated, less productive, and less urban, while ranking the Castilian pronunciations the highest on the same scales. These attitudes were shared by respondents from both within and outside of Andalusia. Somewhat unexpectedly, for some of the phonetic features that we studied, the AndalusianPage 353 → respondents ranked themselves more harshly than the Castilian respondents. Put differently, Andalusians seem to have heavily internalized the negative stigmas associated with their accent. Linguistic profiling is therefore not just a manifestation of implicit bias between groups but also an internal process affecting people’s self-judgments.

I remind my students that you can’t understand a language without understanding its culture—this includes recognizing and mitigating implicit bias. By acknowledging the implications of our research for understanding the human experience, linguists offer legitimacy to claims of negative linguistic profiling and encourage conversation around topics of bias and systemic prejudice. I emulated these goals through the aforementioned course and am compelled to advocate for more contextual education in my courses that prepare students for the social and historical background of the population groups they are studying.19 To this end, I recently mentored a student’s undergraduate thesis titled “Sound Change in Western Andalusian Spanish: An Acoustic Analysis of Phonemic Voiceless Stops,” which carries out a full-scale linguistic analysis on some of the most stigmatized features of the Andalusian dialect, approaching them using the same framework that applies to any standardized dialect.

As discussed earlier in the research by Kelly Wright on AAL,³ results like the ones from my seminar on Andalusian Spanish are not unique to a single country or area. Whereas the Spanish paradigm relates to the stereotypes assigned to a particular region, linguistic profiling in the United States can play on stereotypes rooted in its long history of racism and ethnic prejudice. As a global phenomenon, its impacts vary, but by drawing attention to findings on the Andalusian accent, I hope to build on previous research on marginalized varieties.

Other linguists at the University of Michigan similarly urge awareness of social issues through public engagement. The College of LSA dean, Anne Curzan, hosts a weekly NPR podcast called That’s What They Say, which analyzes how the English language is changing, including dialectal pronunciation differences. As exemplified by her popular TED talk,20 Curzan aims to make linguistics accessible to a wide audience by validating and explaining different forms of language. For example, Curzan is a strong advocate for accepting the use of the singular “they” pronoun in English.21 Another professor, Teresa Satterfield, researches bilingualism and language variation; her public engagement occurs through local, community-based work. Specifically, Satterfield spearheads a Saturday-school education program, called En Nuestra Lengua, for children of Spanish-speaking immigrants who live in southeast Michigan. Page 354 →This program helps legitimize the source accent of the Latinx community by developing academic and literacy skills in Spanish for bilingual children. In addition to fueling their academic passions, colleagues like Curzan and Satterfield are developing innovative ways to promote awareness and create positive change.

Raising the Stakes

There is no perfect time to initiate conversations aimed at confronting our individual biases; we need to use our voices, power, and privilege to unapologetically advocate for equality in every communicative circumstance. Through my own scholarship and teaching on Andalusian Spanish, I aim to contribute to the legal, cultural, and social legitimacy of claims of linguistic bias, both in Andalusia and elsewhere. Trauma, bias, and stigma are deep and intergenerational, but we have power as individuals to reject these destructive cycles and choose understanding and acceptance. Examining the automatic, split-second judgments we make about a speaker may seem like an inconsequential act of protest, but as we further depend on audio-reliant communication platforms, this represents a crucial step toward continuing our evolution into a less-prejudiced society.