Raciolinguistics: Dialects, Race, and People

The first parts of this module established two important points:

• In many daily situations, we do not notice the many accents and dialectal differences that surround us almost constantly. Differences in accents and dialects do not necessarily cause communication breakdowns; in fact, we usually do not notice other people’s accents or dialects at all.

• In experimental situations, listeners have been repeatedly demonstrated to perceive accents or dialectal differences that do not exist in the speech signal, and to make negative judgments about both the speech and the speaker on the basis of their incorrect perceptions, in ways that have been experimentally shown to be connected to the listener’s beliefs about the race and/or national background of the speaker.

We need to consider one final and common situation: Listeners often do notice the real accent or dialect of a real person in a real-life situation. We can usually both see and hear the speaker, and we are almost always correct in assuming that the speech we are hearing is being produced by the speaker we are seeing. We also often have at least some correct information about what we might describe as the speaker’s racial, ethnic, cultural, national, or familial background. What happens in these situations? When we notice that someone speaks differently from the way we speak, what are we noticing, what are we judging, and why?

Yes, We Are Talking About Race.

One of the things we notice, in many situations, is what we believe to be the speaker’s race, ethnicity, national background, or family background, so let’s go ahead and talk specifically about these variables and about their relationships to language. To do so, we need to start by accepting and normalizing, first of all, that the constructs many people tend to refer to as race, ethnicity, and national or familial background can reasonably enter our thinking about speech and language.

(This discussion avoids the question of whether race “exists,” because most people think they know that it does and because some of the constructs that people tend to be refering to when they use this word affect our practice. If you are not familiar with the question or with its complexities, you might want to explore Rogow, 2003, and the other resources available at https://www.racepowerofanillusion.org/ . )

It is not wrong, racist, trendy, progressive, political, politicized, “woke,” or otherwise problematic to be aware that people live in different countries, speak different languages, speak different dialects, have cultural and personal backgrounds that often relate in part to where they and previous generations of their families have lived, and can look different from each other in some ways. Similarly, it is not wrong to assume that the members of a reasonably defined group of people may have some characteristics in common. In fact, as we addressed in Module 2, the members of any group must have something in common, if they are to function as a group, and identifiable groups of people do tend to cluster along subsections of multiple descriptive dimensions – including which dialects of which languages they tend to use to communicate (Hofstede, 2011). Languages and countries and people’s backgrounds are imperfectly related, of course, but they are undeniably related; indeed, we defined languages in Modules 6 and 7 in part by referring to nations or political divisions.

Bringing race, ethnicity, and family or national background into a conversation about dialects and languages, in other words, is necessary, not problematic in itself.

The problems that tend to connect dialect and race begin to arise, especially in a country as multiethnic, multiracial, multilingual, and multidialectal as the U.S., when listeners’ beliefs about other people, and their resulting judgments, decisions, and actions, go beyond a simple, accurate, descriptive awareness of true group tendencies. It is one thing to believe, as one example, that many people in Japan have dark hair and speak Japanese. It is another thing entirely to believe that people in the U.S. who look as if they or previous members of their family might have come from Japan are bad instructors.

As extreme as that example might sound, real-life examples of such incorrect negative judgments about complete people are widespread. You might be able to think of examples related to recent events in your own life, the lives of your family and friends, or in your town. You might also be able to remember examples that made national headlines, or you might be thinking of patterns of human behavior that you know about from around the world or throughout history.

We can also gather many examples from many types of real-world and observational research. In legal research in the U.S., for instance, jurors have repeatedly been demonstrated to perceive witnesses, defendants, and attorneys who speak African American English more negatively (i.e., as less professional, less educated, less trustworthy, and more likely be guilty), as compared with how they perceive witnesses who use a Western or Northern American English dialect (see Kurinec & Weaver, 2019). In a completely different realm, Chicano English, Russian-influenced English, and Italian-influenced English are routinely used intentionally for television and movie characters because these accents and dialects are reliably interpreted by audiences as signaling gang membership or the character’s tendency toward violence (Lawless, 2014). In one example from management research that I find particularly striking, Post et al. (2009) studied analyses of “managerial potential” made by higher-level administrators about engineers from India who were working in the U.S. After examining multiple variables from the speakers and from the administrators, Post et al. concluded, among other findings, that “the absence of insecurity in Indians’ self-reported English fluency [was] detrimental to the evaluation of their managerial potential” (Post et al., 2009, p. 241).

Can you work out what that last sentence means? Engineers from India, working in the U.S. and with any “managerial potential” at all, would have come primarily from relatively privileged socio-economic groups. They would have been natively bilingual (in Hindi and English, for example), or they would have learned English in school starting at a relatively young age and used English extensively in their schooling and their profession. They probably reported no insecurity about their ability to use English for the same reason you might report exactly the same thing: You are, correctly, secure about your ability to use English.

The managers evaluating these engineers, however, assumed that the engineers’ English would be poorer because they were from India or judged the engineers’ Indian English to be in some way worse than their own (the managers’) American English. As a relatively predictable next step, the managers then interpreted the engineers’ self-report of feeling secure about their ability to use English as if it reflected a problematic lack of self-awareness, even though the managers had not measured the engineers’ self-awareness and had no information about the engineers’ self-awareness. The managers had leapt all the way from their stereotype that people from India do not speak English well, to assuming that the engineers who reported feeling secure in their ability to use English must lack the self-awareness that a good manager needs, to rating the engineers as less prepared or less qualified than other people for managerial positions because of their lack of self-awareness.

Post et al.’s (2009) study, and the many other examples we have mentioned, all demonstrate some core aspects of a field that has recently been named raciolinguistics (see Alim et al., 2016; Rosa & Flores, 2020). As Rubin and colleagues had previously noted, raciolinguistic assumptions exist and can be primed by either piece of information; that is, knowing (or believing we know) either the speaker’s race or the speaker’s dialect, or both, tends to activate listeners’ assumptions about both race and dialect. As a result, in real-life situations, listeners’ perceptions, beliefs, and ensuing decisions about speech and also about speakers are often influenced by their initial judgments about the speaker, including an initial judgment that the speaker looks to be a member of what the listener believes to be a particular racial, ethnic, social, or other group. Some of the prejudices demonstrated in these situations have also been described as implicit biases, if the judge is unaware of holding the belief that their decisions or actions demonstrate (see Blair et al., 2011, for a discussion about implicit and explicit biases in healthcare), and they are related to the constructs also described as microaggressions (see Williams, 2020, and Williams et al., 2021). Rosa (2019) described the many problematic raciolinguistic assumptions that people hold by noting that sometimes people act as if a person can “look like a language” or “sound like a race” – something that is distinctly problematic when the judge’s views of the dialect, language, or race in question are negative and wrong.

Considering Some Implications

How are you doing with this discussion? Are you cheering, dozing, or feeling annoyed again about the social-justicey feel of it all?

Once again, let’s gently remind ourselves and each other that we are each on our own journeys; we all encounter information through the lenses of our previous experiences and current beliefs.

Wherever you might be in your journey, though, do you see the important implications here for your practice or for our entire profession?

Race has been intertwined with language throughout human history in many ways, some of which many of us now judge, from our current perspectives, to have been problematic. Sometimes race, ethnicity, or national or familial background ended up related to language because of military or other takeovers of lands and peoples and the resulting imposition of languages (as Kahra’s three-circle model for World Englishes recognized, from our Module 7; see also Von Esch et al., 2020, for one recent review focused on language teaching around the world). Sometimes, race, ethnicity, and familial background ended up related to language because of one country’s social history, including immigration patterns or a history of certain groups being enslaved by other groups. Because of the sensitive nature of these topics, and because they can be interpreted in many ways, daring to mention race at all is often seen as problematic, for a wide range of reasons — as has emerged in many political discussions in the U.S. during the mid 2020s, and as Von Esch et al. (2020) addressed for language teaching (see also Lilienfeld’s 2017 objections to the construct of microaggressions and, especially, Williams’ 2020 response).

To be effective in client-centered, culturally and individually appropriate speech-language pathology, however, we must be willing to discuss race, because we must recognize who we are, who our clients are, and how the people in our communities have interacted over time. The basics of clinical service delivery in speech-language pathology start with interacting respectfully and compassionately with real people, developing a complete understanding of a real person’s speech and language, and then making decisions with them about which next steps make sense for them in their contexts. We simply cannot achieve those goals if we are at the same time somehow trying to completely ignore large parts of people’s contexts.

We also must think about race in speech-language pathology because the matched guise research, other research from raciolinguistics, and substantial other information also shows us that when we think we are noticing a speaker’s dialect, and when we believe we are making clinical or professional decisions about a speaker’s dialect, something much more complex might be happening. Remember: Kang and Rubin’s listeners decided that the speaker they were hearing was hard to understand, Hay et al.’s (2006) listeners decided they heard specific phonemes that the speaker objectively had not produced, and the administrators in Post et al.’s (2009) study decided that the engineers from India they worked with had limited self-awareness in ways that would limit their ability to serve as managers. What if our own clinical and professional decisions are being similarly influenced, at some level, by our underlying stereotypes or biases about our clients’ and colleagues’ dialects or about their racial, ethnic, national, regional, or religious backgrounds?

As Lippi-Green (2012, p. 335) concluded, biases, stereotypes, and discrimination based on dialect or language are not limited to decisions about specific speech or language features but are related, instead, to the negative “social circumstances and identities” that listeners incorrectly associate with that dialect or language. Our judgments about dialects, in other words, often are not judgments about speech or language patterns at all; they are judgments that stem from our underlying, and often unlearned, beliefs, stereotypes, or prejudices about groups of people.

The implications for us as speech-language pathologists are complicated but important: As we think about accents and dialects, we probably need to focus our energies less on phonological patterns or specific morphemes and more on understanding and reducing the continuing influence of race-based and other social stereotypes in our society, in our profession (see Holt, 2022; Yu, Nair, et al., 2022), and possibly even in our own decision-making. The information provided by raciolinguistic research serves as important fundamental information and can help each of us, as the unique individuals that we are, to think about how we might make our best contributions to our profession and to our society.