On Accent Bias and Intelligibility in Accent Training

By Jennie Parker

Accents can be correctly identified in utterances as short as "hello."

Several years ago, when I taught in an intensive English program at an American university, I was assigned a special section of the freshman experience (FE) course designed for international freshmen. Like most universities, ours required freshmen to enroll in a one-credit course that introduced them to various aspects of college life: campus resources, safety, study skills, etc. The section I taught was designed exclusively for international freshmen and was co-taught by the intensive English program and the office of international studies. (Enrollment in this section was optional for international students; alternatively, they could enroll in a general FE course). Part of the curriculum of this special section involved cross-cultural reflections and English language support.

Mid-semester, my students, who came to the U.S. from many countries, including Colombia, Mexico, Italy, China, South Korea, and Ivory Coast, gave oral presentations. The main learning objective was to deliver a well-prepared, organized, coherent presentation. A secondary objective was to address English language skills. As part of this, I designed a special section of the rubric where I listed mispronounced words and their corrections.

I will never forget a student from China who spoke clear English with a Mandarin accent. Though she had presented beautifully and earned an A, when I handed back her feedback, her face got red, and her hand shot up in the air. She scolded me in front of the class for implying that her accent was “wrong” and somehow inferior to mine.
​
I was taken aback. That was certainly not what I had intended to do, and the only reference to accent at all on the rubric was implicit: the list of mispronounced words and their corrections. I thought I was helping her and her classmates to communicate more clearly with their North American peers and professors.

Was I wrong to include pronunciation? I felt ashamed for having offended her, but also confused. Wasn’t part of my job as an ESL instructor to address pronunciation? If so, how could I have done so in a way that would have been better received? The search for these answers led me to a wealth of sociolinguistics literature on accent.

First, let’s be clear, we all speak with an accent. The way we pronounce vowels and consonants and apply word stress, rhythm, and intonation all contribute to what is our accent. In the United States, the majority of people speak English with an accent referred to as General American English; most Americans would say this accent lacks any distinctive characteristics, and it may be erroneously described as “unaccented.” In addition to this standard accent, some estimate there are roughly 24 additional nonstandard accents in the United States, which may point to a speaker’s regional, ethnic, or socioeconomic characteristics. But when you think of English as a global language, for it is spoken by nearly 1.5 billion people around the world, the number of accents increases to roughly 160.

So, which accent is correct? No governmental or royal society exists to regulate the English language, and most linguists would agree that there is no one correct, valid English accent. Well, then, which accent is most desirable? That depends on the listener. Own-accent bias suggests that people prefer to hear their own accent. Accent helps us to distinguish our in-group from out-groups. The logic goes, those who speak like us are the most similar to us. But is that true?

Our perceptions of others’ accents depend on our social and cultural associations with that accent. In other words, accent triggers our stereotypes of a speaker. According to accent prestige theory, accents are used as cues to judge characteristics of speakers. More specifically, these characteristics can be divided into two categories: status (including intelligence, education level, social class, competence, and level of success) and solidarity (including friendliness, trustworthiness, and kindness). Studies have shown that the standard accent is rated higher in both categories by people who speak the standard accent (such as General American in the U.S.). However, those who speak with a nonstandard accent also give higher ratings to those who speak with a standard accent in the status category, but rate those who speak with an accent similar to their own higher in the solidarity category.

Since moving from the Midwest to the South over 20 years ago, I’ve become more sensitive to how some of these stereotypes tend to play out in movies. It seems that oftentimes, characters with a Southern accent in film and animation are friendly but dim-witted. Think Mater in Cars. In contrast, British accents in American theaters are often correlated with prestige or intelligence. Think James Bond.

Unfortunately, in the real world, these accent biases create opportunities for discrimination. John Baugh, a linguistics professor at Washington University, coined the phrase, “linguistic profiling,” which refers to discriminating against a person based on auditory cues, such as dialect and accent. His studies have shown that ethnicity can often accurately be determined from short speech samples – as short as “hello” – and that those who speak with a dialect or accent devalued where they live may fall victim to discrimination in areas such as housing, banking, and medicine. I remember a friend’s British husband joking that in the process of becoming an American citizen, whenever he had to go through security checkpoints in government buildings, he just had to say “hello” and was waved through, avoiding the scrutiny that others endured. Of course, more than auditory cues may have contributed to the guards’ prejudices, but I thought of Baugh’s studies.

Perpetuating accent bias is a real danger of accent training. Awareness should be raised as to the concept of accent bias: of course, a speaker’s accent is not actually an indicator of his intelligence, success, or trustworthiness, for example, but it can elicit social evaluations, perhaps even unconsciously, that put him at an unfair advantage or disadvantage. And completely changing one’s accent to suit the environment is neither easy nor desirable for most. Accent training should, therefore, be approached sensitively with these issues in the minds of both the accent coach and the client.

Certainly, the concept of accent is heavy with meaning: identity, stereotypes, a source of pride or stigma. Yet, another aspect of accent is often overlooked in discussions on it: intelligibility. One’s accent consists of features that trigger far less of an emotional response: pronunciation, word stress, rhythm, and intonation.  To a large extent, these features need to be addressed in language education to help learners be understood. It’s been nearly a decade since I taught that FE class.  In that time, I’ve taught pronunciation (or accent) to thousands of English language learners living in the U.S. who fall up and down the socioeconomic scale, with education levels ranging from little to no formal education to post-graduate degrees. Their feelings on accent have ranged from pride in their native accent to pride in cultivating a General American accent, as well as learners who have given accent identity little to no thought. While anecdotal, I feel comfortable drawing the following conclusion with a degree of certainty: despite their differences, all of these learners want to be understood.

With that in mind and with all I’ve learned about accent and its weight, I’ve devised a simple two-pronged test when it comes to providing feedback on pronunciation. The first question I ask myself is if the feedback is solicited. In other words, has the learner come to me specifically for accent or pronunciation work, or has she come to me with another primary goal, such as fluency or grammar improvement? Next, if the feedback is solicited, I ask myself whether it is necessary. I can only know if it is necessary if I know learners’ goals. Over the years, most have told me that they want to be better understood in English; others have said that they want listeners to focus more on the content of their message and less on the delivery. Far fewer have said that they want to speak English like a North American. These goals determine whether my feedback is necessary. That is, am I providing a correction that improves intelligibility or am I correcting every phoneme that is inconsistent with a General American accent?

When I think back to the FE section that I taught, my pronunciation feedback failed the two-pronged test. The student who was rightly upset with me hadn’t solicited my feedback on pronunciation. Registering for a freshman experience course is very different from registering for an English pronunciation class. I imagine pronunciation feedback was nowhere on the published course description. Second, I hadn’t asked her about her learning goals as they relate to pronunciation, and, furthermore, I would have readily admitted that her English was already highly intelligible. But I gave her feedback, anyway, exposing my unconscious bias that my standard accent was the correct one.
​
I cringe when I think back on that experience now (and if you’re reading this, dear student, please accept my sincere apologies). Over the years, I’ve come to regard accent training as moving along a continuum that stretches from unintelligible to intelligible speech for a target audience. The continuum looks different for each English speaker because the audiences may be different and what lies at the very end of the intelligibility scale is the speaker’s target accent. That target may be speaking intelligible English with a Spanish accent, for example, or it may be speaking English with a General American accent, or Received Pronunciation (i.e., the standard British accent), or any of the other over 150 English accents around the globe.  Thus, an effective accent training program must be learner-centered and highly customizable.