A Classifier-Based Approach to Preposition and Determiner Error Correction in L2 English

Proceedings of the 22nd International Conference on Computational Linguistics (Coling 2008), pages 169–176 Manchester, August 2008 A classifier-based approach to preposition and determiner error correction in L2 English Rachele De Felice and Stephen G. Pulman Oxford University Computing Laboratory Wolfson Building, Parks Road, Oxford OX1 3QD, UK {rachele.defelice|stephen.pulman}@comlab.ox.ac.uk Abstract In this paper, we present an approach to the automatic identification and correction of preposition and determiner errors in non- native (L2) English writing. We show that models of use for these parts of speech can be learned with an accuracy of 70.06% and 92.15% respectively on L1 text, and present first results in an error detection task for L2 writing. 1 Introduction The field of research in natural language process- ing (NLP) applications for L2 language is con- stantly growing. This is largely driven by the ex- panding population of L2 English speakers, whose varying levels of ability may require different types of NLP tools from those designed primarily for native speakers of the language. These include applications for use by the individual and within instructional contexts. Among the key tools are error-checking applications, focusing particularly on areas which learners find the most challenging. Prepositions and determiners are known to be one of the most frequent sources of error for L2 En- glish speakers, a finding supported by our analysis of a small error-tagged corpus we created (deter- miners 17% of errors, prepositions 12%). There- fore, in developing a system for automatic error detection in L2 writing, it seems desirable to focus on these problematic, and very common, parts of speech (POS). This paper gives a brief overview of the prob- lems posed by these POS and of related work. We c⃝ 2008. Licensed under the Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported li- cense (http://creativecommons.org/licenses/by-nc-sa/3.0/). Some rights reserved. then present our proposed approach on both L1 and L2 data and discuss the results obtained so far. 2 The problem 2.1 Prepositions Prepositions are challenging for learners because they can appear to have an idiosyncratic behaviour which does not follow any predictable pattern even across nearly identical contexts. For example, we say I study in Boston but I study at MIT; or He is independent of his parents, but dependent on his son. As it is hard even for L1 speakers to articulate the reasons for these differences, it is not surpris- ing that learners find it difficult to master preposi- tions. 2.2 Determiners Determiners pose a somewhat different problem from prepositions as, unlike them, their choice is more dependent on the wider discourse context than on individual lexical items. The relation be- tween a noun and a determiner is less strict than that between a verb or noun and a preposition, the main factor in determiner choice being the specific properties of the noun’s context. For example, we can say boys like sport or the boys like sport, de- pending on whether we are making a general state- ment about all boys or referring to a specific group. Equally, both …

Content truncated to 500 words