Íslenskt mál og almenn málfræði - 01.01.1980, Blaðsíða 178

176
Joan Maling
generalization of V/2 to clauses that would otherwise begin with the
finite verb. In Section 3, this Subject-Gap Condition on Stylistic In-
version is shown to provide an argument for the existence of oblique
subjects. Finally, in Section 4, various conditions on what can precede
the finite verb are discussed, including conditions on the use of the
dummy það, and on the occurence of unfilled subject gaps.
1. Defining „second position“
Topicalization in main clauses in Icelandic is fairly common. One
study of Icelandic word order, Kossuth(1978), reported that on the
average, 20% of declarative sentences is modem Icelandic prose begin
with something other than the subject. Topicalization in embedded
clauses is much less frequent. However, it is by no means tme that only
subjects can precede the finite verb in embedded clauses as has often
been asserted.2 An example of Topicalization in an embedded clause is
given in (1); the topicalized object NP has been italicized:
1 2 3
(1) Sigga Ólína mundi vel eftir því, að Brynjólf hafði hún oft og
mörgum sinnum getað fengið
(Gestur Pálsson, Tilhugalíf, 5. k.)
Note that the finite verb is in second position in the embedded clause,
as indicated by the numbering. V/2 applies equally to both main and
embedded clauses in Icelandic.3 First position for the finite verb (V/l)
is typically reserved for main clauses, namely, for direct questions, im-
peratives, and the so-called „narrative“ style characteristic of Icelandic.
In embedded clauses, the finite verb comes first in conditional sentences
without the conjunction ef. Ignoring such conditionals, let us explore
the hypothesis that V/2 holds without exception in embedded clauses
in Icelandic.4
2 For example, by Haugen(1976:84).
3 See Maling & Zaenen(1978a,b) for discussion of the consequences of this fact
for the derived constituent structure of questions and topicalizations.
4 Although V/2 is the general rule in embedded clauses, there are some ex-
amples of the finite verb in third position as the result of an adverb in second
position, a possibility not noted by Bergsveinsson(1969) in his otherwise exhaustive