Detecting Article Errors Based on the Mass Count Distinction

Detecting Article Errors Based on the Mass Count Distinction Ryo Nagata1, Takahiro Wakana2, Fumito Masui2, Atsuo Kawai2, and Naoki Isu2 1 Hyogo University of Teacher Education, 942-1 Shimokume, Yashiro, 673-1494 Japan rnagata@info.hyogo-u.ac.jp 2 Mie University, 1577, Kurimamachiya, Tsu, 514-8507, Japan {wakana, masui, kawai, isu}@ai.info.mie-u.ac.jp Abstract. This paper proposes a method for detecting errors concern- ing article usage and singular/plural usage based on the mass count distinction. Although the mass count distinction is particularly impor- tant in detecting these errors, it has been pointed out that it is hard to make heuristic rules for distinguishing mass and count nouns. To solve the problem, first, instances of mass and count nouns are auto- matically collected from a corpus exploiting surface information in the proposed method. Then, words surrounding the mass (count) instances are weighted based on their frequencies. Finally, the weighted words are used for distinguishing mass and count nouns. After distinguishing mass and count nouns, the above errors can be detected by some heuristic rules. Experiments show that the proposed method distinguishes mass and count nouns in the writing of Japanese learners of English with an accuracy of 93% and that 65% of article errors are detected with a precision of 70%. 1 Introduction Although several researchers [1,2,3] have shown that heuristic rules are effective to detecting grammatical errors in the English writing of second language learn- ers, it has been pointed out that it is hard to write heuristic rules for detecting article errors [1]. To be precise, it is hard to write heuristic rules for distinguish- ing mass and count nouns which are particularly important in detecting article errors. The major reason for this is that whether a noun is a mass noun or a count noun greatly depends on its meaning or its surrounding context (Refer to Pelletier and Schubert [4] for detailed discussion on the mass count distinction). Article errors are very common among Japanese learners of English [1,5]. This is perhaps because the Japanese language does not have an article system similar to that of English. Thus, it is favorable for error detecting systems aiming at Japanese learners of English to be capable of detecting article errors. In other words, such systems need to somehow distinguish mass and count nouns in the writing of Japanese learners of English. R. Dale et al. (Eds.): IJCNLP 2005, LNAI 3651, pp. 815–826, 2005. c ⃝ Springer-Verlag Berlin Heidelberg 2005 816 R. Nagata et al. In view of this background, we propose a method for automatically dis- tinguishing mass and count nouns in context to complement the conventional heuristic rules for detecting grammatical errors. In this method, mass and count nouns are distinguished by words surrounding the target noun. Words surround- ing the target noun are collected from a corpus and weighted based on their occurrences. The weighted words are used for distinguishing mass and count nouns in detecting article errors. Given the mass count distinction, errors concerning singular/plural usage, which are also common in the writing of Japanese learners …

Content truncated to 500 words