Dr. Sandra Kübler and Dr. Patrícia Amaral co-authored the chapter “Disambiguating altro: A View from Another Language,” with Jeremy Dickinson, in the volume edited by Dr. Amaral, Other: Ambiguity, Constraints, and Change (Syntax and Semantics, Vol. 45).
Abstract:
This chapter investigates the ambiguity of Italian altro and its cross-linguistic counterparts through the lens of computational corpus linguistics. Specifically, it examines the disambiguation of altro between the non-identity and additivity meanings through a corpus of translations between Italian, our target language that displays ambiguity, and German, a language in which the two meanings are lexicalized. We use a word aligner, which automatically learns which word in a source sentence corresponds to which word in the translation. We investigate how a machine learning algorithm (a Support Vector Machine) given the task to distinguish between these meanings, resolves the ambiguity. We compare the results obtained with features that we manually produced, based on the previous literature on the semantics of altro, with part-of-speech features identified by the algorithm as contributing to the disambiguation task. Our results show that the non-identity meaning is easier to identify than the additive meaning and that it is broadly associated with contexts displaying definiteness. We hypothesize that the additive meaning is harder to identify by the algorithm because this meaning may be expressed by a wider range of lexical resources in German. The chapter concludes with a discussion of the implications of these results for an analysis of the semantics of altro.
The College of Arts