The German collocation, Er ist nun sicher. Sein Nachbar hat eine Pistole gekauft. (He is now safe./He is now sure. His neighbor has bought a pistol.) is ambiguous, while the identical collocation, Er ist nun sicher, sein Nachbar hat eine Pistole gekauft (He is now sure his neighbor has bought a pistol.) is not. This difference in meaning clearly follows from the fact that the first collocation consists of two sentences, while the second consists of a single complex sentence. The difference is indicated by the punctuation, which reflects a salient difference in the pronunciation of the two collocations. In the first, the intonation falls on sicher and then rises again, while in the second, the intonation only falls completely at the end of the entire sequence. The degree of integration, or linkage, of the second clause with the first determines the semantic interpretation of the collocation, expressed in its prosody.
Clause integration can also be signaled lexically in German, for example, through the use of the complemetizer, dass (that) Er ist sich sicher, dass sein Nachbar eine Pistole gekauft hat. (He is sure that his neighbor has bought a pistol.) In German, a dependent clause containing dass can occur in positions that are not possible for the variant without dass: Er ist sich, dass sein Nachbar eine Pistole gekauft hat, inzwischen sicher is possible, but Er ist sich, sein Nachbar hat eine Pistole gekauft, inzwischen sicher is not. This pair of examples demonstrates that there are distinct degrees of integration. A dass-clause is more closely integrated into the main clause (and can therefore be positioned inside of it) than a dependent clause that lacks lexical marking of dependence. The contrast between Er hat gesehen, dass sein Nachbar eine Pistole gekauft hat (He saw that his neighbor had bought a pistol.) and Er hat seinen Nachbar eine Pistole kaufen sehen (He saw his neighbor buy a pistol) demonstrates further degrees of integration. The second example, but not the first, can only mean that he actually witnessed his neighbor buying the pistol: the seeing and the buying form a complex simultaneous event. This strong semantic integration is marked by the form of the verb and is also reflected, in German, in the limited word order possibilities of the non-finite dependent clause. It cannot occur outside the main clause core: Er hat gesehen, seinen Nachbar eine Pistole kaufen is ungrammatical.
Relative clauses show that integration is not always construction dependent, as different relative clause types show different degrees of integration. Non-restrictive relative clauses, German verb second relative clauses and restrictive relative clauses show increasing degrees of integration with respect to different levels of the grammar, including the discourse level. That the discourse level is also relevant can be easily observed in the presentative construction. For example, Es sang einmal ein Junge (There once sang a boy.), requires a follow-on clause which continues to speak about the boy, so that the two clauses form a discourse unit. In contrast, an expression like, so weit, so gut (so far, so good), marks a break with the following sentence.
Our project will investigate the formal marking of integration in selected constructions in Bantu, Creole, and Germanic languages. Our hypothesis is that we should find different degrees of integration, definable at the phonological, syntactic, semantic or discourse levels. Of particular interest for our project are cases of integration mismatch: linguistic constructions showing integration at one level of the grammar but dis-integration at some other level. Central research questions for our project include:
· How close is the correlation between phonological, syntactic and semantic integration in various relative clause types, including clefts?
· What is the degree of integration of various types of verb-second clauses and certain adverbial clauses in German , which are syntactically dependent but maintain some (relative) semantic and pragmatic independence, and of serial verb constructions in Creole languages, which pose the opposite problem?
· What effect do the different placements of complete argument and adjunct clauses have on integration at the prosodic, morphological and syntactic levels? How well integrated into the host clause are reduced clausal structures?
· What is the contribution of the expletive to the integration properties of constructions like presentatives, clefts or dependent clauses with correlatives?
Our research has the following theoretical implications. For phonology, we test whether prosodic phrasing algorithms refer only to constituent edges or whether they must also refer to degrees of integration. We will also investigate how well the levels of the prosodic hierarchy account for degrees of integration. For syntax, our research will certainly have implications for theories of the fine structure of the clause domain and for the theory of expletive elements. For semantics and discourse theory, we expect to contribute to the development of sustainable and empirically motivated explanations of the nature of clausal independence, dependence and of forms in between.