Automatic Extraction of Rules Governing Morphological Agreement

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 5212–5236, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 5212 Automatic Extraction of Rules Governing Morphological Agreement Aditi Chaudhary1, Antonios Anastasopoulos2,†, Adithya Pratapa1, David R. Mortensen1, Zaid Sheikh1, Yulia Tsvetkov1, Graham Neubig1 1Language Technologies Institute, Carnegie Mellon University 2Department of Computer Science, George Mason University {aschaudh,vpratapa,dmortens,zsheikh,ytsvetko,gneubig}@cs.cmu.edu antonis@gmu.edu Abstract Creating a descriptive grammar of a language is an indispensable step for language docu- mentation and preservation. However, at the same time it is a tedious, time-consuming task. In this paper, we take steps towards automating this process by devising an auto- mated framework for extracting a first-pass grammatical specification from raw text in a concise, human- and machine-readable format. We focus on extracting rules describing agree- ment, a morphosyntactic phenomenon at the core of the grammars of many of the world’s languages. We apply our framework to all languages included in the Universal Depen- dencies project, with promising results. Us- ing cross-lingual transfer, even with no expert annotations in the language of interest, our framework extracts a grammatical specifica- tion which is nearly equivalent to those cre- ated with large amounts of gold-standard an- notated data. We confirm this finding with hu- man expert evaluations of the rules that our framework produces, which have an average accuracy of 78%. We release an interface demonstrating the extracted rules at https: //neulab.github.io/lase/. The code is publicly available here.1 1 Introduction While the languages of the world are amazingly diverse, one thing they share in common is their ad- herence to grammars — sets of morpho-syntactic rules specifying how to create sentences in the lan- guage. Hence, an important step in the understand- ing and documentation of languages is the creation of a grammar sketch, a concise and human-readable description of the unique characteristics of that par- ticular language (e.g. Huddleston (2002) for En- 1https://github.com/Aditi138/ LASE-Agreement †: Work done at Carnegie Mellon University. glish, or Brown and Ogilvie (2010) for the world’s languages). One aspect of morphosyntax that is widely de- scribed in such grammatical specifications is agree- ment, the process wherein a word or morpheme selects morphemes in correspondence with another word or phrase in the sentence (Corbett, 2009). Languages have varying degrees of agreement rang- ing from none (e.g. Japanese, Malay) to a large amount (e.g. Hindi, Russian, Chichewa). Patterns of agreement also vary across syntactic subcate- gories. For instance, regular verbs in English agree with their subject in number and person but modal verbs such as “will” show no agreement. Having a concise description of these rules is of obvious use not only to linguists but also language teachers and learners. Furthermore, having such de- scriptions in machine-readable format will further enable applications in natural language process- ing (NLP) such as identifying and mitigating gen- der stereotypes in morphologically rich languages (Zmigrod et al., 2019). The notion of describing a language “in its own terms” based solely on raw data has an established tradition in descriptive linguistics (e.g. …

Content truncated to 500 words