Solf workshop (ScanDiaSyn 2006 Grand Meeting)

The CP-domain in Icelandic

Ásgrímur Angantýsson, University of Iceland

It is a puzzling fact that conditions for topicalization, stylistic fronting (SF) and so-called Scandinavian word order in Icelandic (‘real’ V3) depend on the type of the embedded clause. In this paper I discuss which structural properties of embedded clauses might give rise to these different conditions and propose new ideas about the CP structure in Icelandic.

I adopt the idea that the IP is split in Icelandic but simple in the Mainland Scandinavian languages (cf. Þráinsson 1996 and Bobaljik and Þráinsson 1998). According to this analysis, the verb moves but still follows the adverb in the case of ‘real’ V3 in Icelandic. I point out that SF does not always have the same semantic/pragmatic effect which is a problem for all the previous analyses of the phenomenon. It has been argued that SF has a focusing effect (Hrafnbjargarson 2004) which is, in my view, true for some cases of so-called SF but not others. Along the lines of Rizzi (1997) I propose that in Icelandic, FocP can be merged above TP in case certain type of CP (OP/wh-CP) is merged to AgrSP. Similarly, I suggest that TopP can be merged above AgrSP in case certain type of CP (that-CP) is merged to AgrSP. According to my analysis, the Mainland Scandinavian languages have no articulated CP domain but following Vikner (1995) I assume that an additional CP (CP2) can be merged above IP in case the CP(1) is a complement of a bridge verb.

I explore various examples of word order variation in different types of embedded clauses and show how they can be accounted for according to this analysis. The acceptability of embedded topicalization depends on the type of embedded clause in question in that it doesn’t work in relative clauses and clauses with a wh-pronoun. I show that the fronting process – traditionally called SF – has focusing effects in some cases (e.g. in relative clauses) but not in others (e.g. that-clauses). This difference has not been pointed out in the literature before, as far as I know. Finally, I show that Scandinavian word order is fine either with an unstressed pronoun or a full NP in the subject position in indirect questions with a wh-pronoun and in relative clauses. In temporal clauses it is fine if the subject is an unstressed pronoun but in most other embedded clauses it is difficult both with a full NP and an unstressed pronoun.

One of the most interesting conclusions is that so-called stylistic fronting is a result of two different processes: topicalization (e.g. in that-clauses) and focusing (e.g. in relative clauses). Some of the fronting processes in embedded clauses that traditionally have been called SF in the literature (see Þráinsson 2006 for critical discussion) are thus analyzed as topicalization. This analysis also emphasizes the similarities between ‘real’ stylistic fronting (as in relative clauses) and the V3 phenomenon in Icelandic, called Scandinavian word order; both are focusing processes that occur in similar constructions.

Selected references
- Bobaljik, Jonathan David and Höskuldur Þráinsson. 1998. Two Heads Aren’t Always Better than One. Syntax 1:37-71.
- Hrafnbjargarson, Gunnar Hrafn. 2004. Stylistic Fronting. Studia Linguistica 58:88-134.
- Rizzi, Luigi. 1997. The Fine Structure of the Left Periphery. In Liliane Haegeman (ed.): Elements of Grammar. Handbook in Generative Syntax. Kluwer Academic Publishers, Dordrecht/Boston/London, 281-337.
- Vikner, Sten. 1995. Verb Movement and Expletive Subjects in the Germanic Languages. Oxford University Press.
- Þráinsson, Höskuldur. 2006 [To appear]. The Syntax of Icelandic. Cambridge University Press, Cambridge.