Second Language Acquisition

Speaking a second language (L2) means having acquired a new inventory of phonemes (speech sounds), new words, and new inflectional and syntactic rules. Because this new knowledge must be stored somewhere in the brain, it does not require neuroimaging experiments to conclude that a bilingual speaker’s brain must be in some way different from that of a monolingual speaker. The question is rather: Can the difference between the brain of a bilingual speaker and a monolingual speaker tell us something about the functional organization of a bilingual language system? Quite different answers are possible, ranging from ‘more of the same,’ meaning that at all processing levels the neural substrates for first language (L1) and L2 processing completely overlap, to ‘radically different,’ as suggested by early hypotheses that an L2 relies on the right rather than the left hemisphere of the brain. In between, there are hypotheses suggesting that L2s rely predominantly on a subset of L1 processes, such as storage and declarative processing, but not rule or procedural processing. Such hypotheses predict that a subset of the brain regions involved in L1 processing, namely those subserving rule or procedural processing, should not be recruited for L2 processing. Neuroimaging studies on bilingual speakers are thus potentially of high relevance for our understanding of the functional architecture of a bilingual language system. At the same time, they face considerable methodological problems. In experiments involving a task, differences in neural activity patterns may be explained by general differences in task performance rather than the components of the task that the experiment was designed to assess. Given that most speakers are less proficient in their L2 than in their L1, however, equal task performance is rarely achieved. A second problem is that experimental variables such as ‘age of L2 acquisition’ are almost inevitably confounded with other variables such as ‘L2 proficiency’ or vice versa. Often, it becomes a matter of interpretation which of the observed neuronal activation differences are due to experimental variables and which are due to possible confounds. In this situation, comparisons across studies can be helpful because they may indicate which activation patterns or electrophysiological signatures are reliably found in certain tasks when performed in the L2, irrespective of particular L2 speaker characteristics. They may also show brain activity that is reliably found for subsets of studies testing particular types of L2 speakers. This article therefore focuses on experimental tasks that have been used in a number of hemodynamic (i.e., positron emission tomography (PET) and functional magnetic resonance imaging (fMRI)) and electrophysiological studies of L2 processing, such that an assessment of the reliability of findings is possible. Where possible, we attempt to identify L2 speaker characteristics that seem to influence observed brain activation patterns.