In this paper we examine the topicalization paradigm for ten different verbal constructions in German. We argue that a uniform explanation for the observed behaviors follows from the interpretation of the relevant expressions as (parts of) lexical representations. To this end we motivate a revision of Functional Uncertainty as proposed in Kaplan and Zaenen 1989 to account for filler/gap relations in long-distance dependencies. We assume with the original formulation of this principle that topicalized elements share values with the (grammatical) function status of an entity an indeterminate distance away. We appeal to the inventory of functions posited within LEXICAL-FUNCTIONAL GRAMMAR (LFG), inclusive of the frequently neglected PREDICATE function, which, we argue, is associated with both simple and complex predicates. In addition we show that topicalization, given this function-based proposal, should not be limited to maximal categories. We argue that the need to posit a PREDICATE function for German topicalization is supported by an independent line of research within LFG concerning the analysis of complex predicates. For this purpose we employ the proposals of T. Mohanan (1990/1994), which argue for the independence of the construct PREDICATE from its categorial realization. We show that this type of proposal extends to provide a uniform account of the German topicalization paradigm. This permits us to explain the similarities and differences in the behaviors of various sorts of predicators as well as certain idiomatic expressions interpreted as complex predicates.

This paper investigates the influence of Dutch, Swedish, and English on the syntax of Pidgin Delaware, a contact language used in the Middle Atlantic region in the seventeenth century. Arguments are presented against Thomason's (1980) view that the pidgin predated European contact; instead, the structures of the pidgin are viewed from the perspective of Dutch speakers attempting to learn the Delaware language. The theoretical framework of Van Coetsem 1988 is used to explain which Algonquian features were successfully acquired by the Dutch and where the Dutch imposed features from their native language in the early, formational stage of the pidgin. In addition, subsequent changes in Pidgin Delaware are attributed to its use by Swedish and English speakers.