This study uses text mining tools with R to investigate the correlates in English as a foreign language (EFL) student teacher classroom talk with an attempt to address the learningpromotion features in foreign language classrooms. The teaching materials and the transcripts of two parallel classes are analyzed through term frequency (tf) and inverse document frequency (tf-idf) to demonstrate the degrees of linguistic alignment in different foreign language classes and how the learning promoting mechanism works based on the contents and forms of classroom discourse. It is found that the discourse of both classes is significantly correlated with the baseline discourse (i.e., the teaching materials) at the level of word frequency. However, zooming in on the term weight of the discourse using tf-idf, the teacher and student talk of both classes show a low degree of relevance to the teaching materials. It should nevertheless be noted that the teacher talk in both classes demonstrates some learning promoting features such as vigorous student-teacher interaction and instant feedback conveyed by various response tokens in the teacher talk. This study shows that text mining tools can provide new ways of quantifying the dynamics of classroom discourse.