The Bottlenec k H ypothesis in L 2 A c q uisition : A S tudy of L 1

The Bottleneck Hypothesis (Slabakova, 2008; 2013) aims to partially answer the question of why certain properties of language are harder or easier to acquire in a second language (L2). Based on a comparison of different linguistic modules and how they are acquired, the hypothesis holds that functional morphology is the bottleneck and consequently the most challenging part of L2 acquisition. It is argued that the reasons for this are 1) that differences between languages are located in functional morphology (cf. the Borer–Chomsky Conjecture (Baker, 2008; Chomsky, 1995)); and 2) that functional morphology bundles a variety of semantic, syntactic and morphophonological features which affect the acceptability and the meaning of a sentence. Feature Reassembly (Lardiere, 2009) and mismatches in feature realization are responsible for most of the acquisition challenges. The study presented in this article tests the predictions of the Bottleneck Hypothesis by investigating L1 Norwegian speakers’ knowledge of core syntax and functional morphology in L2 English. It is predicted that Norwegian learners make fewer errors with syntactic operations than with functional morphology, and furthermore, that knowledge of core syntax improves faster than knowledge of functional morphology as the speakers become more advanced. We test two constructions that do not match in English and Norwegian: subject–verb agreement, obligatory in the L2, and Verb-Second (V2) word order, obligatory in the L1. The former represents knowledge of functional morphology and the latter knowledge of syntax. This is the first experimental study designed to directly test the hypothesis. Our main findings show that participants experienced more challenges with agreement than with verb movement, as they had considerable problems identifying ungrammatical agreement in an acceptability judgment test. This lack of sensitivity was established not only at the lower proficiency levels, but also among the more advanced learners. In addition, there was a stronger correlation between non-V2 word order and proficiency