Íslenskt mál og almenn málfræði - 01.01.2008, Side 65

Becoming Perfect: Observations on Icelandic vera búinn að 63
resultative reading, on the other hand, asserts that the target state of a
telic event (e.g. lose the keys or arrive) holds at the reference time;
this is the only reading available for the resultative construction with
vera ‘be’ + participle of an unaccusative verb in Icelandic.12
Experiential perfects place the entire event time before the reference
time, and therefore allow the event to be repeated; in Icelandic, the
úo/a-perfect typically has an experiential reading.
Now, as shown by Jóhannes Gísli Jónsson (1992) and Wide
(2002), the distribution of hafa and vera búinn að is not at all as clear
cut as the distinction between the different readings of the perfect, as
these are understood here. Consider the example in (20) below, where
vera búinn að co-occurs with an adverbial of iteration and where the
time of the individual events is completely anterior to the reference
time. With the present terminology, this is an example of an unam-
biguous experiential perfect; yet, vera búinn að is not ungrammatical:
(20) Ég er búinn að týna lyklunum fimm sinnum,
I am fmished to lose the keys five times
en ég er sem betur fer með þá núna.
but I am luckily enough with them now
‘I have lost the keys five times, but luckily enough I have
them now.’ (cf. Jóhannes Gísli Jónsson 1992:139, note 10)
In order to account for examples as in (20), Jóhannes Gísli Jónsson
assumes a considerably wider defmition of resultative perfect than one
where the result state is specified by the underlying predicate.
According to Jóhannes Gísli Jónsson, (20) is a resultative perfect, but
the result state is “something like the state of being sick and tired of
losing the keys” (1992:139, note 10). Along similar lines, Wide
argues that in (21) below, where a child has repeatedly been sliding
°ff a piece of fumiture, the perfect with vera búinn að is used by the
adult to comment on an activity in the physical setting, and to change
12 Cf. e.g. McCawley (1981:84) who assumes that the resultative reading is an
experiential perfect with an implication about the present situation. Under his
account, a construction that has one of the readings but not the other would be impos-
sible.