Íslenskt mál og almenn málfræði - 01.01.2023, Side 213

iris edda nowenstein
Response to Butt’s and Becker’s reviews

I would like to begin by expressing my deep gratitude to Miriam Butt and Misha
Becker for accepting to serve on the external committee and generously engaging
with the work presented in my thesis. It is invaluable for an early career
researcher to receive the detailed and constructive feedback they provided before,
during and after the defense, and I believe that their perspectives will very much
shape the future directions of this work. Although the reviews published in the
present volume of Íslenskt mál only present a fraction of their valuable input, I
would like to engage with two points which also constituted an important part
of our discussions during the defense itself: (1) The need for a more fine-grained
or decompositional analysis of the semantic features at play in the Icelandic case
marking system and its acquisition and (2) The possible generalizability of the
findings cross-linguistically, including within an evolutionary perspective on the
Tolerance Principle (Yang 2016).
Both Butt and Becker pointed out the dissertation’s lack of a fine-grained
analysis of the semantic features which determine, at least in part, the case mark-
ing of arguments in Icelandic. Becker points this out in the context of the exper-
iments, where animacy effects should have been controlled for and tested system-
atically, and Butt discussed this regarding the annotation of the corpus (Ice -
CASE), where a coarse annotation using classic thematic roles (agent, patient,
goal, theme, experiencer) was used instead of a decompositional feature-based
approach (with e.g. animacy, sentience, volition, motion and instigation for
agents). I want to stress that I agree with their criticism and think this should be
the way this research is taken forward. In terms of systematically controlling for
and testing animacy in the experiments, the lack of such distinctions is in part
due to the fact that the novel verb production experiments were designed to
reproduce the contexts which most clearly elicit dative regardless of animacy, but
this is also the consequence of the data being gathered early in the process of the
project and therefore containing various flaws which have become apparent in
hindsight. Regarding the coarse thematic annotation of IceCASE, the original
plan was to conduct a much more fine-grained annotation in addition to the cur-
rent one, and I did indeed set out to annotate the possible argument structures of
the 410 verbs of the corpus not only with the thematic roles mentioned above,
but also the criteria of Grimm’s (2011) lattice, Dowty’s (1991) proto-role features,
Ramchand’s (2008) first phase syntax event roles, Svenonius’ (2002) aspectual dis-
tinctions as well as Vendler’s (1957) aspectual classes and Barðdal’s (2008),
Maling’s (2002) and even Levin’s (1993) lexical semantic classes. The goal of this