Classification of Lexical Collocation Errors in the Writings of Learners of Spanish

29379

Authors

nan

Abstract

It is generally acknowledged that colloca- tions in the sense of idiosyncratic word co- occurrences are a challenge in the context of second language learning. Advanced miscollocation correction is thus highly desirable. However, state-of-the-art “col- location checkers” are merely able to de- tect a possible miscollocation and then of- fer as correction suggestion a list of collo- cations of the given keyword retrieved au- tomatically from a corpus. No more tar- geted correction is possible since state-of- the-art collocation checkers are not able to identify the type of the miscollocation. We suggest a classification of the main types of lexical miscollocations by US Ameri- can learners of Spanish and demonstrate its performance. 1

Keywords

nan

Full Content

Proceedings of Recent Advances in Natural Language Processing, pages 529–536, Hissar, Bulgaria, Sep 7–9 2015. Classification of Lexical Collocation Errors in the Writings of Learners of Spanish Sara Rodr´ıguez-Fern´andez Universitat Pompeu Fabra Roberto Carlini Universitat Pompeu Fabra {sara.rodriguez.fernandez, roberto.carlini, leo.wanner}@upf.edu Leo Wanner ICREA and Universitat Pompeu Fabra Abstract It is generally acknowledged that colloca- tions in the sense of idiosyncratic word co- occurrences are a challenge in the context of second language learning. Advanced miscollocation correction is thus highly desirable. However, state-of-the-art “col- location checkers” are merely able to de- tect a possible miscollocation and then of- fer as correction suggestion a list of collo- cations of the given keyword retrieved au- tomatically from a corpus. No more tar- geted correction is possible since state-of- the-art collocation checkers are not able to identify the type of the miscollocation. We suggest a classification of the main types of lexical miscollocations by US Ameri- can learners of Spanish and demonstrate its performance. 1 Introduction In the second language learning literature, it is generally acknowledged that it is in particular id- iosyncratic word co-occurrences of the kind take [a] walk, make [a] proposal, pass [an] exam, weak performance, hard blow, etc. that make language learning a challenge (Granger, 1998; Lewis, 2000; Nesselhauf, 2004; Nesselhauf, 2005; Lesniewska, 2006; Alonso Ramos et al., 2010). Such co- occurrences (in lexicography known as “colloca- tions”) are language-specific. For instance, in Spanish, you ‘give a walk’ (dar [un] paseo), while in French and German you ‘make’ it (faire [une] promenade / [einen] Spaziergang machen). In En- glish you take a step, while in German you ‘make’ it ([einen] Schritt machen) and in Spanish you ‘give’ it (dar [un] paso). In English, you can hold or give [a] lecture, in Spanish you ‘give’ (dar [una] clase), but you do not ‘hold’ it, and in Ger- man you ‘hold’ it ([eine] Vorlesung halten), but do not ‘give’ it. And so on. Several proposals have been put forward for how to verify automatically whether a collocation as used by a language learner is correct or not and, in the case that it is not, display a list of poten- tial collocations of the keyword (walk, step, and lecture above) of the assumingly incorrect collo- cation. For instance, a Spanish learner of En- glish may use *approve [an] exam instead of pass [an] exam. When this miscollocation is entered, e.g., into the MUST collocation checker1 for veri- fication, the program suggests (in this order) pass exam, sit exam, take exam, fail exam, and do exam as possible corrections. That is, the checker offers all possible <verb> + exam collocations found in a reference corpus or dictionary. However, the display of a mere list of correct collocations of a given keyword is unsatisfactory for learners since they are left alone with the problem of picking the right one among several (potentially rather sim- ilar) choices. On the other hand, no further re- striction of the list of correction candidates or any meaningful reordering …

Content truncated to 500 words

Labeled Text

It is generally acknowledged that collocations in the sense of idiosyncratic word cooccurrences are a challenge in the context of second language learning. Advanced miscollocation correction is thus highly desirable. However, state-of-the-art “collocation checkers” are merely able to detect a possible miscollocation and then offer as correction suggestion a list of collocations of the given keyword retrieved automatically from a corpus. No more targeted correction is possible since state-ofthe-art collocation checkers are not able to identify the type of the miscollocation. We suggest a classification of the main types of lexical miscollocations by US American learners of Spanish and demonstrate its performance. 1

Publication Details

Paper ID:
29379
Year:
2015
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan