Language-learning expertise

Landau, E. (2010). “From brain to language to accent.”  CNN Online. Retrieved on October 4, 2010:

Becoming a proficient speaker of at least one language is a hallmark of the typical human psychological development. When it comes to learning more than one language, however, our abilities seem much more widely dispersed. Why might some people display a greater “talent” for learning a second language (or more) than others?

By far the best known predictor of success at foreign language learning is the learner’s age.  An increasing number of children who grow up in bilingual environments from early on may well grow up to be fluent speakers of both their native languages. But you don’t have to be natively bilingual in order to master multiple languages at the native-speaker level. In a classic study of second-language acquisition by Johnson & Newport (1989), immigrants to the USA were tested for high-level mastery of English (including phonetic and grammatical nuances), and the results were examined as a function of age at initial immersion in the English-speaking environment. People who started learning English before the age of 7 tended to achieve native-like proficiency. From there on, the older one was at arrival, the less native one’s command of English is likely to be, but only through the age of 13. For those who arrived as teenagers or adults, age is no longer the most prominent factor. In other words, all of us are naturally gifted at language learning as children, and most of us loose most of this talent as we grow older, yet some adults retain greater residual ability than others.

One approach to interpreting this type of evidence is in terms of critical periods in brain development. According to this perspective, a child’s brain is in an earlier stage of biological maturation and therefore open to acquiring one or more native languages. Languages acquired during this period of sensitivity are represented in the brain differently than those acquired at a later developmental stage. The subject of the CNN piece, Dr. Ping Li, clearly represents this approach, with the major subject of inquiry being the neurobiological differences between native and non-native speakers’ processing of linguistic information.

An alternative approach, which I believe is particularly relevant to user-centered design, has to do with the social context in which languages are learned and personality differences that people display in negotiating this social situation. One major advantage that children have over adults as language learners (though probably not the only one) is that they are fully expected and culturally licensed to sound like children, i.e. to make grammatical mistakes, to get certain phonemes only approximately right, to talk about things that seem trivial to adults, etc. Aside from maturational or cognitive differences between children and adults, there are also differences in the type of social experience with language use that the two age groups typically have. One such difference is the psychological value of errors, i.e. of producing incorrect or partially correct forms. From a child’s perspective, saying something incorrectly is a valuable opportunity to elicit adult feedback, often in the form of reformulation (Choiunard & Clark, 2003). Such feedback provides children with a source of implicit negative evidence, which is crucial in making language possible to learn.

From an adult’s perspective, producing an incorrect linguistic form is an  invitation for the audience to draw any number of undesirable inferences about the speaker, e.g. that he or she is poorly educated, belongs to a low-status ethnic, cultural, or socio-economic group, or possibly cognitively taxed, stressed, anxious, etc. The psychological cost of making such a negative impression depends on a number of factors including certain personality characteristics (shyness, perfectionism, anxiety) and certain situational variables (social status, level of privacy, level of formality). It is safe to say, though, that adults as a group are more likely to avoid saying things incorrectly than children, which may be part of what makes them worse language-learners.

While no adult can be turned back into a child, and while altering people’s personality is beyond the scope of product design, creating learning situations that minimize the perceived social cost of using one’s incomplete knowledge of a language to communicate and elicit feedback is a major challenge for the design of language-teaching materials and methods.  Indeed, this pedagogical challenge goes beyond the specific domain of language and may be applied to acquisition of expertise in any sufficiently challenging domain of cultural knowledge. Consider, for example, on-the-job training. A typical novice employee feels a fair amount of pressure to blend in, or do what it takes to be accepted by the workplace social hierarchy. The first thing that you want to ensure in order blend in is that you don’t blunder, that you don’t appear incompetent, that you don’t reveal yourself to have an incomplete understanding of workplace culture. Yet these natural inclinations might actually work against the very goal of training, i.e. expertise acquisition, which may require a certain amount of experience making mistakes. What results is waste of cognitive effort at best and perhaps less-than-effective training efforts. Designing training environments that don’t discourage mistakes, but actually draw on them as learning opportunities, may go a long way towards making the most of people’s natural talent for learning.