Adjacency and discontinuity at crossroads: Learning verb-second word order in German
Complex constructions are nested constructions, i.e. they consist of subcomponents which are constructions of their own. In this presentation I will examine the acquisition of German verb inflection and word order from such a constructivist perspective. The acquisition of word order in German received much attention from the principles & parameters version of Generative Grammar. It was investigated how the acquisition of functional categories licenses verb movement, such that due to the subsequent acquisition of inflectional markers word order in early child German should be virtually error free. However, no definite of this theoretical proposal could be established. Also, these claims are mainly based on early utterance with simple verbs, but not with auxiliary constructions in which word order deviations become more obvious: In unmarked word order, objects and adverbial phrases separate auxiliary and main verb (i.e., Jane has the DGKL paper finished) and occupy the so-called middle field.
This study investigates early word order in the German high-density corpus of a monolingual boy. The period between age 2;0 and 2;6 (14,400 main verb constructions) marks the transition from a nonfinite to a finite system. At 2;0, auxiliaries are absent and only 10% of the main verbs are finite. At 2;6, 72% of verbs are finite, and auxiliary provision is up to 20% (the adult level). Contrary to earlier claims, non-canonical word order is frequent initially, and it takes about four months for word order to consolidate. In auxiliary constructions in utterances with five words or more, the child proceeds from marked and sometimes ungrammatical OVS word order in which heavy NPs are generally extraposed (1), to unmarked and templatic word order as in (2).
neue Trambahn hat Papa (ge)kauft . (age 2;2)
new streetcar has Daddy bought
'Daddy has bought a new streetcar'
haben Papa beim ICE nach Holland abgeholt . (age 2;4)
have Daddy at-the ICE to Holland picked-up
'(we) have picked up Daddy at the ICE-train from Holland'
Note that the early extrapositions are syntactically correct, but pragmatically odd. This suggests that other factors play a role in the acquisition of discontinuous word order. Initially, the adjacency of related information seems to be a major factor in the organization of the child's utterances: extraposition of the heavy NP keeps the organizing units simple, because the verbal elements are adjacent or in close proximity. Furthermore, the child acquires heavy NPs and complex VPs independently, because they do not co-occur initially. Quantitative analyses show that the boy has produced several tenthousand complex NPs and several thousand complex verb forms before these two constructions reliably occur together in unmarked word order. In light of the "starting small" hypothesis (Elman 1980) it is probable that children make use of small, contingent, and meaningful units to assemble building blocks, and that the syntactic integration of these components is yet another and independent step, which correlates with growing working memory (Freudenthal, Pine & Gobet 2006)