Computer Aided Colloquial Language Learning

In this senior design project, we solve part of the problem of teaching common speech to second language English learners by focusing on idioms. We chose to focus on idioms because their figurative nature make them difficult to understand to second language learners, as their meanings can not be deciphered though a literal translation. Thus in order to understand idioms, the user must have prior knowledge of their meanings and their appropriate usage. It is difficult to learn more idioms through existing tools, such as an idiom dictionary. Dictionaries are only useful if the user knows what idiom they are looking for, as they are only searchable by idiom. However, this does nothing to help the user learn new idioms. We plan to incorporate this idea of an idiom dictionary, but alter it to allow users to expand their knowledge of idioms and appropriate usage. There are two main aspects to our project. The first is a browser of idioms and definitions, much like existing dictionaries. However, the system also returns example paragraphs found in the New York Times. By reading the idioms in context, the user will learn more about appropriate idiom usage. This is much better than existing dictionaries that only show one example sentence, where one sentence may not be sufficient for complete comprehension. The second aspect is the main differentiating factor in our approach the focus on keywords in definitions. Through various algorithms, we derived a list of keywords for each idiom based on the definition. When a second language learner has an idea on what he would like to say, this sort of approach is very helpful in finding creative ways of saying it. Using keywords to organize the idioms allows us to present the data in a way that is more suitable for learning. By viewing sets of related idioms at a time, the student would be able to understand the concepts and learn more at once. The system eliminates the need for a teacher and allows this to be a stand alone tool to teach this specific aspect of English. Our system is a teacher and has a better dictionary than current solutions, which makes it the perfect teaching aid for a second language English learner. 1. Related Work Most English learning aides focus on proper grammar and syntax. However, a few attempt to teach colloquial style English. 1.1 Collocations One system automatically assists second language learners with correct collocations (Chang et al., 2008), which helps improve colloquial English. The system takes in a written sample, checks a reference corpus for any matching collocation patterns, and gives suggestions of collocation usage. Although the system's process is similar to our design, it only focuses on verb-noun collocations. Additionally, the system was specifically designed to match between Chinese and English only. Our system differs in that it is designed to improve proper usage of idiomatic expressions and designed for second language learners of any first language background. 1.2 Error correction and examples Another system uses the web for error correction (Gamon et al., 2008). This system is based on transcribed speech. The suggestion provider captures the potential error in the input sentence, while the example provider queries the web for example sentences. However, there has been no study on the usefulness or correctness of these example sentences. Our system queries articles from the New York Times, a valid source of data that is grammatically correct and of appropriate difficulty level, rather than random data from the internet in order to provide useful example sentences to the user. 1.3 Idiom dictionaries Online idiom dictionaries provide idioms and their definition and example sentence, such as English Club Idioms Dictionary and The Free Idioms Dictionary. These are only searchable by idioms. If a user wants to learn new idioms, typing in a keyword will only return an idiom containing the keyword, which most likely has no relation to its definition. Our system makes idioms searchable by meaning, which is much more helpful in teaching the user exactly what he is intending to say. Another online tool, English Daily, attempts to teach users idioms. This site embeds the idiom and the definition in a 3-5 sentence example. This site has a similar approach as ours in terms of providing examples to teach proper usage and assist in comprehension. However, there are only a select number of idioms provided (44). Our system covers more than 700 idioms and teaches users a much wider range of idioms. 2. Technical Approach