Multiword Expressions in Child Language

Multiword Expressions in Child Language Rodrigo Wilkens1, Marco Idiart2, Aline Villavicencio1 1Institute of Informatics, 2Institute of Physics Federal University of Rio Grande do Sul (Brazil) rodrigo.wilkens@inf.ufrgs.br, marco.idiart@gmail.com, avillavicencio@inf.ufrgs.br Abstract The goal of this work is to introduce CHILDES-MWE, which contains English CHILDES corpora automatically annotated with Multiword Expressions (MWEs) information. The result is a resource with almost 350,000 sentences annotated with more than 70,000 distinct MWEs of various types from both longitudinal and latitudinal corpora. This resource can be used for large scale language acquisition studies of how MWEs feature in child language. Focusing on compound nouns (CN), we then verify in a longitudinal study if there are differences in the distribution and compositionality of CNs in child-directed and child-produced sentences across ages. Moreover, using additional latitudinal data, we investigate if there are further differences in CN usage and in compositionality preferences. The results obtained for the child-produced sentences reflect CN distribution and compositionality in child-directed sentences. Keywords: Multiword Expressions, Compound nouns, compositionality, Language Acquisition 1. Introduction The increasing availability of psycholinguistic, lexical and ontological resources and of more precise and ro- bust natural language processing tools has enabled the au- tomatic annotation of language acquisition corpora with additional sources of information. In particular, among them resources like WordNet (Miller, 1995) and specialised datasets contain information about Multiword Expressions (MWEs) such as noun compounds (police car) (Kim and Baldwin, 2006; Nakov, 2008a), phrasal verbs (break down) (McCarthy et al., 2003) and collocations (e.g. salt and pep- per) (Seretan, 2011; Eryi˘git et al., 2011). These resources can be used as basis for annotating MWE occurrence in corpora such as the English portion of the Child Language Data Exchange System (CHILDES) (MacWhinney, 1995), which contains transcriptions of child language data. In this paper we introduce the resulting resource, CHILDES- MWE, which contains almost 350,000 English sentences annotated with more than 70,000 distinct MWEs of various types, including compound nouns and phrasal verbs.1 The resulting annotated resource can be used as basis for language acquisition studies. In this paper we use it for examining how MWEs feature in child language, focusing on compound nouns (CNs) in English. Firstly, using lon- gitudinal data, which follow specific children across time, we want to examine the link between the input and out- put of the children in terms of CNs, verifying if there are differences in the distribution of CNs in child-directed and child-produced sentences. Secondly, as in terms of seman- tics CNs range from compositional to idiomatic combina- tions, we want to determine if there is any effect of compo- sitionality in the sentences to which children are exposed and in those they produce. This paper is structured as follows: in §2 we describe some related work on MWEs and child language; and in §3 we present CHILDES-MWE and describe the annotation pro- 1CHILDES-MWE is freely available at html://www.inf. ufrgs.br/pln/resource/CHILDESMWE.zip. cess. The materials and methods used for this work are in §4, along with an analysis of the results obtained. We finish with conclusions and …

Content truncated to 500 words