ABSTRACT
Context of learning became an important second language acquisition (SLA) construct around the year 2000, largely in response to concerns that second language (L2) research overwhelmingly focused on understanding the internal changes that occur during development. SLA researchers can better characterize and predict the effects of learning contexts by studying and comparing corpora representing both learners’ input and output in a given setting/mode. As the L2 socialization research has shown, different contexts of learning impose different communicative demands on learners, which in turn will encourage them to use particular discourse types and sets of linguistic features. Corpus research indicates that the input that learners receive, which the setting/mode features establish, influences the nature of L2 output. A valid learner corpus can be constructed with simulated or authentic data. Simulated data entails learners participating in a range of tasks that reflect how the learner uses the L2 in a given context.
