Íslenskt mál og almenn málfræði - 2020, Page 298

distribution of Adv−Vfin, I propose that Adv−Vfin was accommodated under
the same or similar conditions that give rise to SF, perhaps by adjunction. There
are then two types of contexts in particular that I argue facilitate such a reanaly-
sis, namely relatives where there is a gap due to movement out of the embedded
clause, resulting in a potential ‘subject’ in the matrix clause, and embedded claus-
es with a controlled subject pro gap where the verb either raises or is left in situ.
These two structures are exemplified below (see section 3.7.3 in the dissertation
for details and references):
(1)a. Haraldur var eini nemandinn sem ekki hafði lesið bókina.
Haraldur was only student-the that not had read book-the
b. Haraldur var eini nemandinn sem hafði ekki lesið bókina.
‘Haraldur was the only student that had not read the book.’
(2)a. Sistur þinar biðia baðar skiælandi að heilsa þier og grata
sisters your ask both weeping to greet you and cry
af þvi að ei gatu skrifað
of it that not could write
‘Both your sisters give you their greetings weeping and cry because they
could not write to you.’ (MalJen-1821-01-1X.xml)
b. Hrædd er jeg um að komi ekki svo fljótt norður
afraid am I about that come not so soon north
‘I am afraid I will not come that soon to the north.’ (JakJon-1853-01-04.xml)
Null subjects as in (2) were lost during the 19th century (Þóra Björk Hjartar -
dóttir 1993, Kinn et al. 2016) and while SF is widely accepted, its frequency has
been going down (see Halldór Ármann Sigurðsson 2017). For structures like
(2a,b), I assume that the representation must contain a gap for the subject, the
structural position of which is uncertain (see my dissertation p. 123). Unlike the
structure in (1a,b), the null subject of (2a,b) can be made overt, which arguably
makes Adv−Vfin configurations, even where no subject is ever overtly present
as in (1), more opaque or elusive to speakers and facilitates reanalysis.
SF has gradually been becoming more and more a written language phenom-
enon and it is mainly kept alive by the fact that it is often needed as an alterna-
tive for the expletive það and even maður constructions in formal contexts. It
thus comes to serve a stylistic purpose in the written language, being stylistically
marked and judged more acceptable by older speakers than by younger speakers
(cf. Ásgrímur Angantýsson 2011). An adaptation of the Adv−Vfin variant could
be modelled after these two environments and that would give some kind of
insight into why the restrictions on Adv−Vfin in Modern Icelandic are the way
they appear to be. This part I take to be “change from below”.
Sundquist (2003) only considered the role of SF in the reanalysis of verb rais-
ing as V-to-C or else V-in-situ and no V-to-I. That this is a plausible scenario is
Heimir Freyr van der Feest Viðarsson298