A Preliminary study on computer-aided acquisition of Cantonese vocabulary by Mandarin speakers using a parallel corpus

The increasing trend of people from mainland and Taiwan coming to Hong Kong to study and work features the domain of teaching Cantonese as a foreign dialect. Owing to Cantonese’s unofficial status and the tiny size of the market, research in this field is still at the beginning stage and is thus unfavourable to the learners. Nevertheless, the need for this domain of study is greater than ever as Mandarin speakers are increasingly integrated into the Hong Kong society. Previous studies in this field mainly focused on pronunciation (e.g. Shī 2002, Ráo 2003, Ki 2006, author 2010). While there have been studies on computer-aided learning of Mandarin vocabulary and idiom (e.g. Yang & Xie 2013; Lewis, Luk & Ng 1998), less attention has been paid to Cantonese. In this paper, we report our experience in teaching Cantonese vocabulary to Mandarin-speakers with the aid of the first parallel corpus that aligns transcribed Cantonese speech and the written Mandarin counterpart (author 2011).In general, a polysemous word in one language often has different counterparts in another language for each meaning. In our case, for instance, Mandarin “le” can be functioned as both a perfect aspect particle and a mood particle. However, in Cantonese, the functional counterpart for perfect aspect particle is “jó” while that for the mood particle is “la”. This often poses obstacles to learners. We have thus conducted an experiment on a group of Mandarin-speaking undergraduate students at the time when they have learnt Cantonese for two months from a Cantonese class in university. They were invited to look for the Cantonese counterparts of some polysemous words in Mandarin in our corpus as well as to analyze their meanings or functions.Pre-test, post-test and delay post-test were offered to test assess the knowledge towards of the master of these Cantonese counterparts by our subjects. We have administered a pre-test to collect a set of Cantonese words that each student did not understand. In personalized computer-assisted vocabulary learning session, given the Mandarin counterpart, each student was asked to search for half of these words in the corpus. Our experimental results show that when contrasting the words that the subjects have checked, analyzed from the corpus and those that the subjects have not, significant differences were exhibited in both post-test and delay post-test but not in pre-test. It can be seen that our corpus is an effective device of computer-aided acquisition of Cantonese vocabulary.