Íslenskt mál og almenn málfræði - 01.01.2013, Qupperneq 56

54
Ásgrímur Angantýsson
Wiklund, Anna-Lena, Gunnar Hrafn Hrafnbjargarson, Kristine Bentzen og Þorbjörg
Hróarsdóttir. 2007. Rethinking Scandinavian Verb Movement. The Joumal ofCom-
parative Germanic Linguistics io(3):203-233-
Wiklund, Anna-Lena, Kristine Bentzen, Gunnar Hrafn Hrafnbjargarson og Þorbjörg
Hróarsdóttir. 2009. On the Distribution and Illocution of V2 in Scandinavian that-
Clauses. Lingua 119:1914—1938.
SUMMARY
Keywords: verb movement, embedded topicalization, transitive expletive constructions,
morpho-syntax, assertion, verbal inflection
The main topic of this paper is the status of Faroese among the Scandinavian languages,
in particular with respect to verb placement in embedded clauses. In Faroese, V3 is the
default word order in all types of embedded clauses. In this respect Faroese is very differ-
ent from Icelandic, where V2 (meaning simply ‘the finite verb in second position’) is
always the default order in subject-initial embedded clauses. But it is also different from
Danish, where V2 is always heavily marked in all types of embedded clauses.
There has been considerable controversy in the linguistic literature about the nature of
the V2-order in Faroese embedded clauses: Is it the result of a V-to-T movement (as is
typically assumed for Icelandic) or is it a root phenomenon, i.e. movement to the CP-
domain (commonly referred to as V-to-C). This paper argues that both kinds of V2-order
can be found in Faroese embedded clauses. On the one hand there is V-to-C and hence
there is very clear evidence that assertion plays a role in the distribution of V2-order in
subject-initial complement clauses in Faroese: If the complement proposition can be inter-
preted as the main assertion of the utterance then V2 is usually fine, but if the matrix pred-
icate expresses the main assertion then V2 is heavily degraded in most cases. Here there
are some similarities to Embedded topicalization (ET): ET gets positive judgments in the
assertive complements of siga ‘say’ and halda ‘believe’, as well as in the complement of the
semi-factive predicate finna útav ‘discover’, but almost nobody fully accepts it in other
types of embedded clauses. These results are even clearer with respect to Hooper &
Thompson’s (1973) classification of predicates than comparable results for ET in Icelandic,
where the acceptance rate of ET in complements of predicates of classes C and D was
much higher.
But embedded V2-orders in Faroese cannot all be attributed to V-to-C since they are
also accepted, by some speakers at least, in clauses where ET is completely impossible,
such as relative clauses and indirect questions (clauses that have no root properties). This
suggests that V-to-T is also a possibility in Faroese. Some versions of the so-called Rich
Agreement Hypothesis (e.g. Bobaljik and Thráinsson 1998) would then predict that tran-
sitive expletives should also be possible in Faroese and this prediction is borne out. It is
demonstrated, however, that speakers who distinguish tense and agreement morphemes
most clearly in their speech are neither more likely to accept transitive expletives nor V2-
orders in various types of embedded clauses than speakers who do not distinguish tense
and agreement morphemes as clearly. The relevance of this for different versions of the
Rich Agreement Hypothesis is discussed in the paper and a possible scenario for the his-