The combinatorial variability hypothesis in the second language

Variability in the production and comprehension of inflectional morphology is widely attested and well documented in L2 acquisition. The sources of this variability are, however, disputed. In this paper, the authors look at the potential of Adger's (2006) Combinatorial Variability Hypothesis (CVH) to explain variable performance in this respect. The CVH explains intra-personal morphosyntactic variation as arising from the combinatorial mechanisms of language itself. It offers an evaluation metric of uninterpretable feature combinations, predicting which forms are going to surface more, and possibly even become defaults, in the acquisition of inflection. The experimental study involves English native speakers who are beginning or intermediate learners of German. They are asked to choose missing subjects among four options in copula sentences. Results indicate that, when choosing pronominal subjects, learners exhibit the pattern of errors predicted by the CVH. Frequency of the inflectional morphemes does not explain the results. The authors argue that there may be different underlying sources for variable L2 morphosyntactic performance: feature re-assembly, perceptual salience, phonological regularity, and semantic complexity may all be relevant. When all of these factors are held constant, though, and errors still persist, only language-internal explanations, such as combinatorial variability of features, are really credible