Two broad perspectives have been advanced to account for the differences observed between native (L1) and non-native (L2) speakers in ultimate attainment and processing. In one view, such differences are fundamental and possibly selective, with particular parts of the language system becoming hard or impossible for late learners to acquire in a native-like way. In another view, L1-L2 contrasts are attributable to general factors such as slower processing speed or amount of exposure, and can be expected to be more gradient in nature, as well as more general in scope. In this talk, I will present results from experiments and meta-analyses examining L2 morphological processing, as a test case for these larger perspectives.

Our results indicate that differences between L1 and L2 speakers show remarkable selectivity and are restricted to specific parts of the morphological processing system (e.g., inflection, conjugation classes). In contrast, other sub-domains of morphology (e.g., word formation) can be processed in a native-like way, even when the L2 was acquired later in life. At the same time, the observed L1-L2 differences were often found to be graded rather than all-or-none, for example, becoming more pronounced at later ages of acquisition. This suggests that a full account of L2 grammar and processing will require models that can accommodate gradient levels of non-native-likeness and of morphological constituency, while nevertheless capturing the internal differentiation of the language system.