Íslenskt mál og almenn málfræði - 01.01.2023, Side 205

subjects in modern Icelandic (Schätzle 2018) and is the middle formation of an
originally locational predicate (‘find at’).
As shown by Deo (2003) for Marathi, such changes do not occur all at once
and they do exhibit variation and some exceptions. The class of intransitive verbs
with an oblique goal that Deo identifies were the first to be reanalyzed as expe-
riencer predicates and are found almost exceptionlessly with dative subjects in
Old Marathi. This pattern can be accounted for by the historical reanalysis posit-
ed in Schätzle (2018) and Beck and Butt (2023). The psychological predicates with
an originally nominative-accusative pattern show variation between nominative
and dative subjects in Old Marathi, much of which continues into Modern
Marathi, with a subset of verbs (‘reveal’, ‘think’, ‘observe’) settling on dative sub-
jects and accompanying this with a concomitant change in semantics (‘occur’,
‘appear’, ‘realize’, respectively). This development can be seen as a regularization
of the system in which experiencers are aligned with the dative to express an
abstract goal semantics (roughly as in (2a)) and which then also extends to the
class of transitive verbs that do not have prototypical sentient agents. These are
still firmly nominative-accusative in Old Marathi, but have changed to a dative-
nominative pattern in Modern Marathi (again, often with a concomitant change
of lexical semantics).
Even though the Marathi data compiled by Deo (2003) remains to be ana-
lyzed in full, it suggests a diachronic path of change which is driven by a seman-
tically motivated regularization of the system in that datives first express both
abstract and concrete goal/endpoint semantics and then get extended to sentient
goals/locations, which are in turn reinterpreted as experiencer arguments, even-
tually leading to a regularization by which all experiencer arguments are consis-
tently marked with the dative. Nowenstein’s results similarly point to the changes
in case marking patterns as involving a semantically driven regularization of the
system so that datives are being more closely aligned with experiencer semantics,
while more agentive subjects are becoming to be realized more consistently via
the nominative.
There are thus extremely interesting parallels to be found in the diachronic
and synchronic case systems of Icelandic and Indo-Aryan and in the paths of
change exhibited by the languages. An intensive and deep exploration of these
crosslinguistic parallels is very likely to yield interesting insights that will seri-
ously impact the current state of the art. For example, the Marathi data has not
as yet been analyzed fully, with one puzzling aspect having been a number of
exceptions Deo (2003) also identified as part of her work. However, given the
development of the Tolerance Principle in the intervening years and its success-
ful application to various types of diachronic data, one very interesting new path
of research to pursue would thus be the application of the Tolerance Principle to
the Marathi data.
Another fruitful point of comparison would be the patterns of case acquisi-
tion. There is some work on the acquisition of case in Urdu/Hindi which also
suggests a reliable use of semantic generalizations as a guide to acquiring case
Icelandic datives: A view from South Asia 205