It was pure good fortune that I encountered Complexity Theory (CT) some years ago. It originates in the physical sciences, and I am a mere dilettante in them. However, from the beginning of my acquaintance with CT, I quickly realized that it had much to offer applied linguists. It challenged my concept of language as a static rule-governed system, maintaining instead that “the act of playing the game has a way of changing the rules” (Gleick 1987, p. 24) , and doing away with the need to posit an innate LAD and preformationism (“the assumption that in order to build a complex structure you need to begin with a detailed plan or template”) (Deacon 2012, p. 50). Besides, I could see how CT had the potential to unite important ontogenetic and phylogenetic processes: Language development, language learning, language evolution, language use, language change. However, it was not only that CT was theoretically satisfying; it also resonated with my personal experience. For one thing, adopting a systems perspective on issues of interest in applied linguistics, rather than a piecemeal approach, made a great deal of sense to me. I had grown somewhat discouraged by common research approaches to second language acquisition (SLA) as I understood them. I found them to be reductionist, atomizing the object of concern and then studying one atom at a time, often through single treatment, pre-testpost-test designs. Controlling for other factors and overly deterministic, it sought to identify the causal factor in SLA, e.g., comprehensible input at an i+1 level. Such approaches too readily dismissed variability as noise or measurement error or attributed it to “outliers.” They treated context as a backdrop, removed from the main action. They failed to capture the dynamicity of processes, leading Elman (2003, p. 430) to remark that “This