- New insights into the syntax and semantics of complementation: Introduction to the special issue*
Dependent clauses are one of the prominent examples illustrating the ability to generate recursive structures by a computational mechanism. Three types of dependent clauses have usually been distinguished in a broader sense: (i) complement clauses, (ii) adverbial clauses, and (iii) correlative or relative clauses. The question of whether these three types can be reduced to a single abstract structure has inspired several fruitful lines of research, resulting in a mass of new empirical findings and leading to novel results. For example, many authors have proposed to analyze different types of dependent clauses as correlative/relative clauses (see Arsenijević 2009, Bhatt & Pancheva 2006, Caponigro & Polinsky 2011, Geis 1970, Haegeman 2012, Kayne 2014, Krapova 2010, Kratzer 2006, Moulton 2009, 2015, among many others). This special issue of the Historical Syntax section of Language, entitled New insights into the syntax and semantics of complementation, focuses in particular on the diachronic syntax and semantics of dependent clauses and shows to what extent complement, adverbial, and relative/correlative clauses may be related to each other.
In what follows, I briefly outline the most important findings of the contributions collected in this special issue and highlight how they contribute to the diachronic discussion on clause-linkage in general.
Katrin Axel-Tober’s article, ‘The development of the declarative complementizer in German’, is concerned with the issue of how the German subordinate complementizer dass ‘that’, which introduces declarative complement clauses and triggers verb-final position, evolved in the history of German.1
In the example given in 1, the matrix predicate wissen ‘know’ embeds a finite CP-clause introduced by the declarative complementizer dass ‘that’. The assumption has been that it developed from the nominative/accusative form of the neuter demonstrative pronoun das ‘that’. Accordingly, the paratactic structure given in 2 is supposed to have given rise [End Page e23] to the hypotactic structure exemplified in 1, shifting the clause boundary (cf. Behaghel 1928, Ebert 1978, Horacek 1964, Lehmann 1988, Lenerz 1984, to name but a few).
The example in 2 contains the cataphoric demonstrative pronoun das ‘that’, occurring at the end of the first main clause and pointing forward to the content of the second main clause. Similar examples exemplifying the structures given in 1 and 2 are attested as early as Old High German (750–1050) (examples adapted from Axel-Tober, this issue).
Axel-Tober challenges the standard grammaticalization analysis of how dass ‘that’ emerged, offers a novel reanalysis scenario that is also applicable to other West Germanic languages, and provides abundant crosslinguistic evidence supporting the view that dass did not evolve directly from the cataphoric demonstrative pronoun das ‘that’ (see also Axel 2009 and Axel-Tober 2012:Ch. 2 for more details). Accordingly, the declarative complement clause developed from an explicative relative clause, as given in 5 (example adapted from Axel-Tober, this issue).
Following this line of reasoning, the dependent thaz-clause given in 5 should be analyzed as an explicative relative clause. What is crucial about its status is that (i) it is headed by the relative particle thaz, which occupies the C-position, (ii) it is attached as an adjunct clause to a functional projection in the matrix clause, and (iii) its content refers to the correlative element tház ‘that’ placed in the matrix clause. According to Axel-Tober, such a structure gave rise to the development of complement clauses; see the Old High German example in 4 with a silent correlative element, in which the thaz-clause is taken to function as a complement clause and thaz ‘that’ is the declarative complementizer triggering the verb-final position. In other words, there are no differences between the source and the target structure on the surface. Several advantages follow from this kind of reanalysis. It not only shows that there is no need to assume a radical change of the sentence boundary (main clause → embedded clause); it also indicates that thaz was a C-head before and after the reanalysis, regardless of whether it introduced an adjunct or a complement clause.
The issue of relative and...