Vocabulary Building on the Move with the Prime Machine: Evaluation of a Cross-platform Corpus-based Game for English Vocabulary Learning

The development of corpus methods has led to a deepened understanding of the patterning of vocabulary, with positive impacts on learner dictionaries, vocabulary profiling and text analysis, as well as hands-on language learning with concordancers. While learning activities based on corpus data have been used in English language teaching for many years, these have tended to be generated by textbook writers and teachers manually selecting and organising corpus data. This paper evaluates a new corpus-based game which was developed based on the data-driven learning exercise “One Item, Multiple Contexts” (Johns, 1997) or “spot the missing word” (Hanks, 2013). This kind of exercise was gamified in a learner-friendly corpus tool – The Prime Machine. To evaluate the language learning potential of the automatically provided materials, a corpus-based analytical approach was employed, categorizing the patterning of the target items and distractors. The primings of lexico-grammatical features involving collocation, semantic association, and colligation were categorized and compared. The corpus-based analyses found reoccurring patterns and primings in the concordance lines across the learning phase and the gaming phase. Interviews with 5 Chinese university EFL students who had played the game were conducted. The interview questions focused on asking participants’ noticing behaviours, the compatibility of difficulty levels, and their perceptions towards the game. The responses showed that respondents noticed reoccurring language patterns and measured the difficulty level specific to their gaming experiences. A request was made for more explanation on missed patterns when the incorrect item was selected, and improvements on the reward system were suggested. All participants perceived the game to be beneficial for learning lexico-grammatical features of English vocabulary.