Universal Dependencies and Quantitative Typological Trends. A Case Study on Word Order

Universal Dependencies and Quantitative Typological Trends. A Case Study on Word Order Chiara Alzetta+, Felice Dell’Orletta*, Simonetta Montemagni*, Giulia Venturi* + Universit`a degli Studi di Genova, Via Balbi 5, Genova, Italy *Istituto di Linguistica Computazionale “Antonio Zampolli” (ILC–CNR) – ItaliaNLP Lab, www.italianlp.it via G. Moruzzi, 1- Pisa, Italy chiara.alzetta@edu.unige.it, {felice.dellorletta,simonetta.montemagni,giulia.venturi}@ilc.cnr.it Abstract The paper presents a new methodology aimed at acquiring typological evidence from “gold” treebanks for different languages. In particular, it investigates whether and to what extent algorithms developed for assessing the plausibility of automatically produced syntactic annotations could contribute to shed light on key issues of the linguistic typological literature. It reports the first and promising results of a case study focusing on word order patterns carried out on three different languages (English, Italian and Spanish). Keywords: Linguistic Knowledge Extraction, Dependency Treebanks, Linguistic Typology 1. Introduction The interaction between linguistics and computational lin- guistics has a long history dating back to the 60’s. In Kuˇcera (1982), it is explicitly stated that “computational linguistics provides important potential tools for the test- ing of theoretical linguistic constructs and of their power to predict actual language use”. This still appears to rep- resent a key objective, as claimed e.g. by Martin Kay in his ACL Lifetime Award speech in 2005 (Kay, 2005), or by the more recent papers gathered in the Special Issue of the journal “Linguistic Issues in Language Technology” (LiLT) focusing on the relationship between language technology and linguistic insights (Baldwin and Kordoni, 2011). Af- ter more than 40 years from the first pioneering studies, the growing availability of linguistic resources such as anno- tated corpora for many languages combined with the in- creasing reliability of Natural Language Processing (NLP) methods and tools enables the acquisition of quantitative evidence ranging across different levels of linguistic de- scription which can significantly contribute to the study of open issues of the theoretical linguistic literature. This holds particularly true for the area of typological stud- ies which can benefit a lot from this synergy, making it possible to acquire quantitative evidence shedding light on how, why and to what extent languages vary with respect to key features covering major areas of language structure. By exploring collections of linguistically annotated corpora for different languages, complex and articulated frequency dis- tributions of language constructions can be extracted. In- formation acquired from available corpora can significantly enrich typological descriptions of languages such as the World Atlas of Language Structures (WALS) (M. S. Dryer and M. Haspelmath, 2013) 1, the most commonly-used and broadest database of structural (phonological, grammatical, lexical) properties of languages whose data are based on primary sources such as grammars, dictionaries and scien- tific papers. But impact and role of this information type cannot be limited to the descriptive level. Typological evi- dence inferred from linguistically annotated corpora for dif- 1Available online at http://wals.info ferent languages can significantly contribute to model lin- guistic variation within and across languages. Word order variation represents a widely investigated topic of the typo- logical literature whose recent developments …

Content truncated to 500 words