Unclogging the Bottleneck: The role of case morphology in L2 acquisition at the syntax-discourse interface

ABSTRACT

The Bottleneck Hypothesis argues that properties of inflectional morphology explain why second-language learners may face persistent difficulties in articulating meaning in target-language forms. In particular, the acquisition task proves even harder when first and second languages differ in the way they organize the mapping of functional features onto inflectional morphemes. Against this backdrop, the present study focuses on the interplay between case morphology and acquisition of syntax-discourse interface phenomena in the second language German of French native speakers. Whereas the Bottleneck Hypothesis predicts that discourse-driven syntactic alternations come for free when case morphology is in place, a concurrent prediction is made by the Interface Hypothesis, which suggests that the syntax-discourse interface is vulnerable per se, irrespective of L1-L2 differences at the interface between morphology and syntax. We tested 45 L2 learners in (i) a semicontrolled production task in context (element rearrangement task) targeting argument reordering (object fronting) as a function of the information status of discourse referents; (ii) a fill-in-the-blanks task involving the use of case endings; and (iii) a general proficiency test. Contra the Interface Hypothesis, results showed that argument reordering was affected by discourse properties of referents. Turning to the Bottleneck Hypothesis, learners’ command of case was identified as a reliable predictor of convergence in discourse-to-syntax mappings, which confirms the key role of inflectional morphology in the acquisition of other grammar areas, including interfaces between grammar and grammar-external domains such as discourse.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Focusing on French and German, the World Atlas of Language Structures reports: (i) one measure of morphological distance indicating that German has four cases, whereas French exhibits no case marking (Iggesen 2005, 202f.); (ii) one measure of syntactic distance, which stipulates that French has a dominant SVO order, whereas no such dominant pattern can be identified in German (Dryer 2005, 330f.).

2 Frey (2004b) suggests that German has a designated position for topics at the left periphery of the middlefield.

3 In other cases, inversion of subject and object leads to ungrammaticality, with incorrect agreement between verb and preverbal argument.

4 The constructs backward-looking and forward-looking derive from Centering Theory. Backward-looking centers denote given referents; forward-looking centers refer to the elements introduced in the current utterance, which, as such, provide potential links to the subsequent discourse (see, e.g., Walker, Joshi & Prince 1998).

5 Weber & Müller (2004) specify, however, that whereas the assumption that pronouns precede full NPs is verified for SVO sentences, the effect of constituent weight is more disputable in the case of OVS sentences. Furthermore, Chiarcos (2009) demonstrates that the tendency to position short expressions before longer ones applies primarily to the ordering of constituents within the middlefield; prefield expressions are on average longer than middlefield expressions.

6 Regular case assignment designates the functional use of case in a clause to indicate the function of arguments, i.e., the relationships that exist between arguments and predicational elements.

7 http://www.wortschatz.uni-leipzig.de/Papers/top10000de.txt.

8 Regarding automatization, the question whether implicit knowledge is the exclusive locus of automatic processes or whether explicit knowledge can also move toward automatized knowledge is a matter of debate. In this respect, DeKeyser (1997, 2003) argues for a distinction between implicit knowledge and automatized explicit knowledge, which are both functionally equivalent, i.e., characterized by rapid access to representations, the difference being due to the awareness criterion. Contrary to implicit knowledge, automatized explicit knowledge is assumed to involve consciousness about representations and their use.