Assessing Speech Perception and Comprehension in Bilinguals
An interview with Lu-Feng Shi
Posted May 25, 2016
Interview conducted by François Grosjean
Speech-language-hearing professionals often have to assess the speech perception and comprehension of bilinguals who suffer from a hearing or processing impairment. How they do so is an interesting topic that we know little about. One researcher who has done a lot of work in the field is Dr. Lu-Feng Shi, Associate Professor of Communication Sciences and Disorders at Long Island University. He has kindly accepted to answer a few of our questions and we wish to thank him wholeheartedly.
What do speech-language-hearing professionals do in a clinic and who are those who seek their help?
In North America and some European nations, speech-language pathologists deal with speech and language disorders, whereas audiologists take care of hearing disorders. We work with a wide age range of patients, from infants to older adults, more specifically with practically anybody for whom there is a breakdown in communication that causes functional limitation and/or participation restriction. Our primary goal is to enhance effective communication.
Can you give us some examples of patients that you assess?
Our clientele includes children who have speech, language, and hearing impairment (e.g., fluency and articulation disorders, language development issues, congenital hearing loss). As you can easily imagine, many are also bilingual. Adults typically have developed speech and language skills, so their communication difficulty is usually due to hearing impairment. When evaluating their hearing, we use tones as well as speech stimuli. As for our elderly patients, in addition to hearing loss, some are recovering from a stroke and may have motor speech issues. Here too, the patients' language background and in some cases their bilingualism have to be considered.
Are bilinguals assessed like monolinguals?
With monolingual clients, we present test items in their native language. For a bilingual client, things are more complicated. We know that a bilingual is not two monolinguals in one person (see here), and therefore it is not adequate to compare the result of a test done in one language to the norm based on monolinguals in that language. Doing that has the potential of misdiagnosing the bilingual as having hearing or processing issues when that might not be the case.
Some might suggest that you allow for a few more errors during the assessment of bilinguals.
It's fine to accept that a normal-hearing bilingual client misrecognizes a few more words than the monolingual. We know his/her compromised performance is due to a language proficiency difference rather than a hearing/processing disorder. But most cases are not as clearly cut. What if a bilingual individual has both a hearing/processing disorder and a language proficiency issue? How much of the poor performance can be attributed to the disorder versus the difference in language proficiency? How can we then make a recommendation or set the goal for rehabilitation?
Should bilinguals be tested in all their languages then?
Ideally bilingual clients should be evaluated in every language they use on a daily basis. After all, the goal is to enhance communication. If they speak English at work but Spanish at home, then we should try to test them in both English and Spanish. But this is very hard in practice. To test a bilingual in two languages, we need a clinician who is equally proficient in both languages and who also understands the phenomenon of bilingualism. We also need two tests, one in English and the other in Spanish, for example, with comparable psychometric properties. We then need to administer each test in listening conditions that best simulate the situations in which these languages are used.
Couldn't an interpreter be used?
Yes, an interpreter can help us communicate with a client, but cannot help us judge the client’s responses and score the test. Audiologists write down and score the responses to words or sentences in real time, so proficiency in the test language is essential. Moreover, bilingual clinicians should understand the culture each language is related to so as to be aware of the client’s concerns and know how to best address them. There is also a trust issue here; as Leo Morales and his colleagues have shown, clients respond better to clinicians who share their cultural background.
Tell us more about test materials when assessing bilinguals.
I’m going to use English and Spanish again as examples. The most widely used test material in English includes a set of monosyllabic words which are prevalent in the language. By contrast, the most common words in Spanish have two syllables. These words have more phonotactic (syllable structure) and lexical cues and are therefore easier to recognize than those with one syllable. As such, I would expect poorer performance on an English word recognition test than a Spanish test in a perfectly balanced English-Spanish bilingual. If one is not aware of how these languages are structured, one might think the bilingual has a problem with English when that is not the case.
What about listening conditions?
In most clinics, speech recognition is only carried out in quiet, but ironically, even individuals with moderate hearing loss do not complain of difficulty with speech in a quiet room. To most, hearing becomes difficult in noisy situations such as in crowded places. Therefore, many scholars advocate testing clients in noise so as to better appraise their difficulty in real-life listening, in addition to testing them in quiet to get the baseline. The same goes for bilinguals; only this time we are talking about four test conditions (quiet and noise for each language). That requires a lot of time not only for the clinician but also for the client. Whether a third-party insurer will pay for all this is anyone’s guess.
If the ideal practice is not practical, what may be the second best practice?
This is a very difficult question for which we don't yet have an answer. We have shown with Diana Sánchez that the performance of bilinguals in one language does not correlate with performance in the other, so it may be necessary to test in both languages if we are interested in getting the full picture. On the other hand, if for some clients from a given group, testing in both languages yields the same diagnosis and leads to the same set of recommendations, then perhaps it is adequate to just test in one language for diagnostic purposes.
What about rehabilitation?
We cannot assume that improving listening skills in one language automatically improves skills in another language. Each language has a unique set of phonemes and they are subject to the effects of hearing loss in a different way. English is notorious for highly confusable high-frequency, low-amplitude sounds such as /f/, /s/, and /θ/. English also allows clusters of up to four consonants, making correct recognition of each consonant in the cluster very difficult (e.g. the final consonants in glimpsed). Romance languages such as Spanish and Italian are more listener-friendly, in a way, as they have prominent vowels and limited cases of diphthongs and consonant clusters.
My guess therefore is that English-Spanish bilinguals may have more difficulty with English than Spanish speech as a result of hearing loss and thus require more rehabilitative effort in English, even if English is the stronger language of the two. This is an area in which more investigation is clearly needed.
So what are your conclusions on assessing bilinguals?
Without more research, I'll refrain from being too assertive. However, good clinical work is based on a sound rationale. As such, I would suggest that we test a client in one language if that alone can lead to a clear diagnosis. When rehabilitation is called for, then we may want to evaluate and follow up with both languages.
For a full list of "Life as a bilingual" blog posts by content area, see here.
Photo of a woman with headphones from Shutterstock.
Morales, L. S., Cunningham, W. E., Brown, J. A., Liu, H., & Hays, R. D. (1999). Are Latinos less satisfied with communication by health care providers? Journal of General Internal Medicine, 14, 409-417.
Shi, L.-F., & Sánchez, D. (2010). Spanish/English bilingual listeners on clinical word recognition tests: What to expect and how to predict. Journal of Speech, Language, and Hearing Research, 53, 1096-1110.
François Grosjean's website.