Íslenskt mál og almenn málfræði - 01.01.2009, Síða 67

65
Tilbrigði í fallmörkun aukafallsfrumlaga
SUMMARY
‘Variation in the case-marking of oblique subjects:
Dative substitution in Old Icelandic?’
Keywords: dative substitution (dative ‘sickness'), accusative substitution, alternating
verbs, Old Icelandic
In Modern Icelandic and at least since the mid I9th century, there has been a tendency to
substitute (older) accusative case on experiencers for dative, often termed dative substitu-
úon or dative ‘sickness’. However, Halldór Halldórsson’s 1982 diachronic study indicated
that in older Icelandic, nominative substitution rather than dative substitution was more
prevalent in contexts where accusative is substituted for dative later in the history of
Icelandic. The present study complements the Old Norse part of Halldór Halldórsson’s
work, based on a systematic investigation of verbs with oblique experiencer subjects in
Old Norse texts, corpora, concordances and dictionaries, which were available electroni-
cally. Halldór Halldórsson (1982) found Old Norse examples of a dative experiencer
where an accusative experiencer is expected only with skorta ‘lack’ (Old Icelandic, 2
instances) and lysta ‘want, please’ (Old Norwegian, 1 example).
A new study brings the number of verbs with varying accusative/dative case marking
°f oblique experiencer subjects from two to ten—although with five of those verbs (shown
in parentheses), dative is older, as well as more common, than accusative (accusative sub-
stitution’): angra ‘bother’, (anagja ‘please’), bila ‘fail’, (bíbaga ‘suit’), (bresta ‘lack, fail), (byrja
behove’), harma ‘grieve’, (höfga ‘nap, drowse’), lysta, skipta ‘change, be of importance, and
skorta. This fact suggests that accusative (subject) case was at least semi-productive, which
is rather unexpected if accusative on experiencer subjects is entirely idiosyncratic as is
often assumed. The scarcity of data poses a problem in analysing the examples with dative
cxperiencers in accusative contexts, as it is not always clear whether the dative uses involve
an underlying dat-nom or dat-acc case frame (see below). In the case of bresta, it seems
that a change from the pattern dat-nom to acc-acc has taken place in later Icelandic—the
first (philologically sound) instance of an accusative experiencer is attested in a i6th-cen-
tury text. These facts certainly do cast some doubt on the legitimacy of an analysis as dative
substitution, thus making it tempting to look for other explanations.
A different analysis is offered where dative in accusative contexts with the verbs in
question is taken to resemble the behaviour of alternating verbs. The verbs angra and bi-
haga have already been analysed as alternating verbs, angra in Old Swedish, bíhaga in Old
Icelandic, and based on the alternating behaviour of líka ‘líka’ in Old Norse, the group of
such verbs is argued to have been larger than in Modern Icelandic (cf. Jóhanna Barðdal
^998,1999a). Such an analysis is more compatible with the dat-nom pattern and the fact
^hat nominative/accusative substitution rather than dative substitution occurs in Icelandic
until the mid I9th century. In (standard) Modern Icelandic, for example, bresta, klœja ‘itch’
and þverra ‘lose, recede’ generally occur with the pattern acc-acc, klœja with acc-pp, the
I^tter two of which are only attested with dat-nom until the I9th century. In addition,
skorta could be used personally in the same source where a dative experiencer has been
^ound with the verb. This indicates that the underlying argument structure of the dative
examples of skorta may not be the dat-acc pattern characteristic of dative substitution.