It’s hard not to think of the acquisition of non-native phonological contrasts when reading about the research on London cab-drivers described here (well, if you’re a linguist, that is). I found the blog post via a discussion of the study here.
Essentially, London cab-drivers are better at learning new routes than a control group when the new route is located in an unfamiliar city, but they have a harder time learning a new route than a control group when it is located in London, and thus competes with their previous knowledege. This latter effect seems similar to the observation that it’s harder to learn a new phonological contrast after having already learned a language, compared to when acquiring a first language.
Various people have indeed argued that our ability to learn new types of phonological contrasts might deteriorate not (just) because we are past a critical period of brain development, but (also) because after having learned our first language(s), it’s become harder to fit new contrasts into our existing ‘map.’ By having learned a language we might have developed “an inability to inhibit access to existing (and now competing) memory representations,” to quote from the blog post on cab drivers. This idea is advocated here:
Perhaps one could test whether there is really a parallel between the results found for cab-drivers and language learners by comparing how well we can learn a particular new contrast when it’s presented as part of a very different language as opposed to as part of a different dialect of our native language. The latter should actually be more difficult, if the parallel would hold up, although one would somehow have to control for the familiarity with the lexical material. And I guess there should be less of an issue in learning a new gestural contrast in a sign language if you’ve only learned a spoken languages before, compared to having previous knowledge in a sign language. Maybe someone has already tested this?