Andersen's earlier work sought an explanation for how learners create and restructure their interlanguage systems as a product of participating in verbal interaction with more proficient speakers. It resulted in the 'Nativization Model,' according to which L2 acquisition consists of two general processes, nativization and denativization. In the former, learners make the input conform to their own internalized view of what constitutes the L2 system. That is, they simplify the learning task by forming hypotheses based on knowledge that they already possess (L1 knowledge and knowledge of the world). In Andersen's terms, they attend to an 'internal norm.' The result is the kind of pidginization evident in early language acquisition and documented in Schumann's work. In denativization, learners accommodate to an 'external norm'; that is, they adjust their interlanguage systems to make them fit with the input, making use of inferencing strategies. Denativization is apparent in depidginization (the elaboration of a pidgin language which occurs through the gradual incorporation of forms from an external source).
Subsequently, Andersen (1990) has recognized that nativization and denativization are not two separate 'forces' but aspects of the same overall process of acquisition. Andersen's later work is an attempt to develop the nativization model by further specifying the 'processes, cognitive operating principles, and communicative strategies' that fit with it.