This study explored the effects of meaning-focused and form-focused instruction on aspects of foreign-language learner’s spoken English, argument production, and attitude towards English use. Four freshman English classes (N=136) from a university in Taiwan participated in this quasi-experimental study. Two classes were randomly assigned to either the Collaborative Reasoning (meaning-focused instruction) condition or the Dictogloss (form-focused instruction) condition. Students in the two conditions discussed the same four topics, one topic per week, and each topic for a 60-minute class session. A battery of pre and posttests was administered, including an open-ended speaking task, attitude surveys, a standardized English test, and a reading fluency test. The major finding of the study was that students who had meaning-focused Collaborative Reasoning discussions increased their fluency in the posttest speaking task, while learners who participated in form-focused Dictogloss discussions had reduced fluency, after controlling for pretest fluency and other language proficiency measures. A corroborating finding is that Collaborative Reasoning students had higher posttest reading fluency than Dictogloss students, controlling for pretest reading fluency and other factors. The speaking fluency and reading fluency results suggest that Collaborative Reasoning students may have automatized core linguistic processes more than Dictogloss students. Overall, analyses of students’ talk suggested that meaning-focused instruction may nurture language performance that is more communicative, and form-focused instruction may foster language performance that is more restricted but accurate on target forms. The findings support the idea that conversational interaction is a foundation for language development. Evidence from this study suggests that incorporating meaning-focused instruction in the classroom could be one way of nurturing mastery of English as a communicative tool in an English-as-a-foreign-language context.