Íslenskt mál og almenn málfræði - 01.01.2011, Blaðsíða 163

The Topic ofthe Thesis and an Overview ofMethods and Results 161
tial complement clauses: If the complement proposition can be interpreted as the
main assertion of the utterance then V2 is usually fine, but if the matrix predicate
expresses the main assertion then V2 is heavily degraded in most cases. A possi-
ble explanation for this is the following. In Faroese, verb movement (or the posi-
tion of the finite verb in relation to VP-adverbs) can be observed by the language
learner in the primary linguistic data (PLD). However, the verbal inflection, being
somewhat simpler than its Icelandic counterpart, does not suffice to tell the learn-
er whether or not the IP is split (there is no clear evidence for separate tense and
agreement morphemes, cf. Bobaljik and Thráinsson 1998). Therefore, the learner
seeks other explanations and concludes that verb movement is largely restricted to
matrix-like embedded clauses, i.e. clauses that contain an assertion (V-to-C move-
ment). In Western-Jutlandic, the language learners are only exposed to very weak
remains of verb movement and ignore it for the most part when building their
grammar.
In table 2 we see a simplified overview of the questionnaire results regarding
Embedded Topicalization (ET):
Icel. Far. Övdal. W-Jutl.
Embedded Topicalization that-chuses with predicates of types A, B and E +/- + +/- +/-
that-chuses with predicates of types C and D -/+ - -/+ -/+
Indirect questions - - -
Adverbial clauses - - -
Relative clauses — — — -
Table2: An overview ofthe acceptability ofEmbedded Topicalization in different types
ofembedded clauses in Icelandic, Faroese, Övdalian and Western-Jutlandic
The four languages behave similarly with respect to Embedded Topicalization:
ET is only generally accepted in that-chuses that are complements of predicates
A, B and E. All of the examples of ET in the Icelandic corpora were also of this
kind. Faroese is the “best-behaved” language in terms of Hooper and Thompson’s
(1973) classification of predicates with respect to ET as it was with respect to
Vfin-Adv order.
There have been conflicting claims in the literature about the empirical situa-
tion regarding ET in the Scandinavian languages. Some linguists have assumed
that Topicalization is more easily or widely accepted in embedded clauses in
Icelandic than in the Mainland Scandinavian languages and others have assumed
that ET obeys similar restrictions in Icelandic to those in the Mainland Scandi-
navian languages. It has also been suggested that there are two varieties with
respect to ET in Icelandic, in the way that speakers of variety A allow Topicaliz-
ation quite freely in embedded clauses except for temporal dauses and embedded