University of Pennsylvania Working Papers in Linguistics


This paper explores transfer of parasitic gap (p-gap) constructions from English into German by heritage speakers in Wisconsin. Kathol (2001) argues that German lacks ‘true’ p-gap constructions compared to English. Engdahl (1983:73/2001) introduces an accessibility hierarchy of domains in which p-gaps are accepted:

(1) Engdahl’s accessibility hierarchy for occurrence of MGCs (partial)

most accessible least accessible

manner adv. > temp. adv. > purpose clauses > that, than > when, because > relative clause

[untensed domains] [tensed domains]

The licensing of p-gaps may thus be variable in several regards, including across complement vs. relative vs. adjunct clauses, and more basically between tensed and untensed domains. We probe whether the licensing strategies for p-gaps of a dominant L2 (English) can affect an incompletely-acquired L1 (German) that does not license such gaps and, if so, whether such strategies follow Engdahl’s hierarchy. The presence of p-gaps would support the work of Grosjean (2008), whose view predicts that English syntax may surface (i.e., ‘seep through’) in spoken German if English has become the dominant language for an individual.

Our results support the theory of ‘grammatical seeping’, and our speakers in general behave in accordance with the predictions of Engdahl’s hierarchy. They produce p-gaps in English-to-German translations relatively frequently in manner clauses, often in temporal clauses and rarely in relative clauses. In temporal clauses we find considerable syntactic restructuring. In the least accessible context, relative clauses, speakers restructure more fundamentally, in order to eliminate the gapping environment altogether.