For decades educational psychologists have bemoaned the black box approach of much research on learning, that is, the focus on product rather than process, and the absence of fine-grained analysis of the learning process in the individual. One way that progress has been made on this point in the last couple of decades is through cognitive neuroscience, but even there what is documented is mostly the product of learning, possibly at different stages, rather than the process. An alternative way of trying to understand processes that are hard or impossible to observe is to infer them from the way individual difference variables interact with linguistic and contextual variables. In this article I draw from a wide variety of studies to illustrate how the interactions between aptitudes and treatments, age and treatments, age and aptitudes, age and structures, and aptitudes and structures point to different learning processes. I stress how such interaction research is much more than the sum of its parts, making suggestions for a research agenda to sharpen the focus on the process in second language learning.