Animacy Acquisition Using Morphological Case

12364

Authors

nan

Abstract

Animacy is an inherent property of en- tities that nominals refer to in the phys- ical world. This semantic property of a nominal has received much attention in both linguistics and computational linguis- tics. In this paper, we present a robust unsupervised technique to infer the ani- macy of nominals in languages with rich morphological case. The intuition behind our method is that the control/agency of a noun depicted by case marking can ap- proximate its animacy. A higher control over an action implies higher animacy. Our experiments on Hindi show promising re- sults with Fβ and Purity scores of 89 and 86 respectively. 1

Keywords

nan

Full Content

International Joint Conference on Natural Language Processing, pages 64–72, Nagoya, Japan, 14-18 October 2013. Animacy Acquisition Using Morphological Case Riyaz Ahmad Bhat LTRC, IIIT-Hyderabad, India riyaz.bhat@research.iiit.ac.in Dipti Misra Sharma LTRC, IIIT-Hyderabad, India dipti@iiit.ac.in Abstract Animacy is an inherent property of en- tities that nominals refer to in the phys- ical world. This semantic property of a nominal has received much attention in both linguistics and computational linguis- tics. In this paper, we present a robust unsupervised technique to infer the ani- macy of nominals in languages with rich morphological case. The intuition behind our method is that the control/agency of a noun depicted by case marking can ap- proximate its animacy. A higher control over an action implies higher animacy. Our experiments on Hindi show promising re- sults with Fβ and Purity scores of 89 and 86 respectively. 1 Introduction Animacy can either be defined as a biological property or a grammatical category of nouns. In a strictly biological sense, living entities are ani- mate, while all non living entities are inanimate. However, in its linguistic sense, the term is syn- onymous with a referent’s ability to act or instigate events volitionally (Kittil¨a et al., 2011). Although seemingly different, linguistic animacy can be im- plied from biological animacy. In linguistics, the manifestation of animacy and its relevance to lin- guistic phenomena have been studied quite exten- sively. Animacy has been shown, cross linguisti- cally, to control a number of linguistic phenomena. Case marking, argument realization, topicality or discourse salience are some phenomena highly correlated with the property of animacy (Aissen, 2003; Bresnan et al., 2007; De Swart et al., 2008; Branigan et al., 2008). In linguistic theory, how- ever, animacy is not seen as a dichotomous vari- able, rather a range capturing finer distinctions of linguistic relevance. Animacy hierarchy proposed in Silverstein’s influential article on “animacy hi- erarchy” (Silverstein, 1986) ranks nominals on a scale of the following gradience: 1st pers > 2nd pers > 3rd anim > 3rd inanim. Several such hierarchies of animacy have been proposed following (Silver- stein, 1986). One basic scale taken from (Aissen, 2003) makes a three-way distinction as humans > animates > inanimates. These hierarchies can be said to be based on the likelihood of a referent of a nominal to act as an agent in an event (Kittil¨a et al., 2011). Thus higher a nominal on these hier- archies higher the degree of agency/control it has over an action. In morphologically rich languages, the degree of agency/control is expressed by case marking. Case markers capture the degree of con- trol a nominal has in a given context (Hopper and Thompson, 1980; Butt, 2006). They rank nomi- nals on the continuum of control as shown in (1)1. Nominals marked with Ergative case have highest control and the ones marked with Locative have lowest. Erg > Gen > Inst > Dat > Acc > Loc (1) In this work, we demonstrate that the correla- tion between the aforementioned linguistic phe- nomena is highly systematic, therefore …

Content truncated to 500 words

Labeled Text

Animacy is an inherent property of entities that nominals refer to in the physical world. This semantic property of a nominal has received much attention in both linguistics and computational linguistics. In this paper, we present a robust unsupervised technique to infer the animacy of nominals in languages with rich morphological case. The intuition behind our method is that the control/agency of a noun depicted by case marking can approximate its animacy. A higher control over an action implies higher animacy. Our experiments on Hindi show promising results with Fβ and Purity scores of 89 and 86 respectively. 1

Publication Details

Paper ID:
12364
Year:
2013
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan