Íslenskt mál og almenn málfræði - 01.01.2011, Side 165

The Topic ofthe Thesis and an Overview ofMethods and Results 163
Icel. Far. Övdal.
Stylistic Fronting (of past participles):
that-chuses (impersonal passives) +/- +/- -
Indirect questions (impersonal passives) + -
Relative clauses + +/- -
Expledve Insertion:
Temporal clauses (weather predicates) +/- + +
Relative clauses + +
TECs (main clauses) +/- -
Subject gaps:
Temporal clauses (weather predicates) +/- -
Relative clauses +/- -/+ +/-
Table 3: A comparison ofthe acceptability ofSF and related constructions in different
types of embedded clauses in Icelandic, Faroese and Övdalian
In Icelandic and Faroese, SF was more widely accepted in relative clauses than in
ffctff-clauses. The Övdalian speakers completely rejected fronting of past partici-
ples in both clause types. In all languages, Expletive Insertion received a high
score in temporal clauses with weather predicates. In Faroese and Övdalian,
Expletive Insertion was also accepted in relative clauses, which was very different
from the situation in Icelandic, where such insertion is bad (this was not tested in
the IceDiaSyn project though). Leaving the subject position empty (subject gaps)
in relative clauses was generally acceptable in Icelandic and, to a certain extent in
Övdalian, while most speakers rejected it in Faroese. Most of the older speakers
of Icelandic also accepted subject gaps in temporal clauses with weather predicates
while most of the Faroese speakers and many of the younger speakers of Icelandic
rejected such examples.
At least certain sets of the SF-data can be properly treated under the so-called
head movement approach. The motivation for analyzing SF as an adjunction to I
(or AgrS) rests on the prohibition of head movement to a specifier position. This
analysis also explains the absence of focus effects and it accounts for the relation
between verb movement and SF (Icelandic vs. Mainland Scandinavian), i.e. that
V-to-I movement is a necessary condition for SF although it is presumably not a
sufficient condition as we have seen (cf. the situation in Övdalian and Faroese).
Under Jóhannes Gísli Jónsson’s (1991) analysis of SF, the subject gap condition is
accounted for in terms of feature checking. The SF-element is head-adjoined to
the finite verb and moves along with it to I. As a result, the finite verb is “too low”
in the structure to check the relevant features with a lexical subject. Therefore,