About the workshop

Venue for the workshop is updated: Himalaya 105, IIIT Hyderabad

Registration for the workshop can be done on ICON's main conference page: see this link.

This workshop is focused on assessing the current status of research on verbal multiword expressions (MWE) in Indian languages and creating a strategy for future directions, particularly in terms of lexical resources. This workshop will be organized as part of ICON 2019 in Hyderabad, India.

Motivation

Verbal multiword expresssions, which include phenomena such as complex predicates, support verb/light verb constructions and causatives among others are prevalent among South Asian languages (Masica, 1993). A range of linguistic research across theoretical frameworks has examined such expressions more closely over the years e.g (Jayaseelan, 1988; Abbi and Gopalakrishnan, 1991; Butt, 1997 ;Mohanan, 1997; Davison, 2005; Ramchand, 2008) (among others).

These expressions have also received some attention in the computational linguistics literature with the development of automatic multiword detection systems for Indian language MWE (Sinha, 2009; Das et al., 2010; Chakrabarti et al., 2008; Begum et al., 2011; Vaidya et al., 2016) and computational grammar representations (Poornima and Koenig, 2008; Ahmed et al., 2012; Vaidya et al., 2019).

Over the years, psycholinguistic research on the processing of these expressions has begun to emerge, using a variety of methodologies from in the psycholinguistic literature (Vaidya and Wittenberg (forthcoming), Dasgupta et al., 2015).

Despite this rich and growing body of work on these expressions, only a few lexical resources have been developed for the multiwords in South Asian languages (Although they are annotated in the Indian language treebanks, (Bharati et al., 2002; Palmer et al., 2009)). Lexical resources that cater specifically to verbal multiwords have been developed for several European languages with the PARSEME project (Savary et al., 2017) and corpora such as STREUSLE (Schneider and Smith, 2015).

Lexical Resources

This workshop will aim at having a discussion around imagining/designing such as resource or shared task for verbal multiwords in South Asian languages, that would be of interest to theoretical linguists, psycholinguists and computational linguists alike. The discussion would be centered around whether existing annotation designs can be adapted, or changed with respect to the particular challenges of South Asian languages, where such expressions are highly productive.

Accordingly, the program will consist of three talks that examine the problem of complex predicates from different points of view: theory, typology, psycholinguistics and computational linguistics. The panel discussion will focus on some of the questions about the design of lexical resources. We hope to use this workshop as a launchpad for bringing together researchers in this area and to use it as a platform to discuss future developments.

References