Complexity Theory (CT) provides a useful framework for understanding Second Language Acquisition (SLA). Using an ecological model, CT studies the dynamic processes of change and emergent outcomes over time, tracing back how trajectories may have been affected by changes in and interactions among multiple variables and subsystems. Language learners do not follow linear learning paths, but rather their progress depends on a variety of interlocking variables. Developmental trajectories will look different for each learner. That is particularly the case for informal language learning, in which learners co-adapt from a wide variety of resources, leading to divergent outcomes (Godwin-Jones, 2018a). A CT approach emphasizes the dynamics of the interplay between learner variables and the people, artifacts, and services constituting the L2 learning system. It aligns well with usage-based linguistics. Viewing language learning from a CT perspective has a number of consequences for Computer Assisted Language Learning (CALL) research, explored here. [For the complete proceedings, see ED600837.]