In the previous two chapters, I explored two major constraints (i.e., biological and cognitive) on adult L2 learning, in attempts to account for the general lack of success. I made the point that the bio-cognitive constraints control the ability to acquire an L2 and subsequently the ultimate degree of success. What also seemed clear from that discussion, however, was that these constraints were limited in their capacity to explain inter-learner and intra-learner differential failure.

This chapter is therefore purported to examine the latter through a microscopic analysis of failure. A major argument to be advanced is that within the general confine of bio-cognitive constraints, other factors – social, cognitive, environmental and psychological – may operate, independently or in tandem, leading to differential failure at both an inter-learner and an intra-learner level. First, by putting under scrutiny a number of empirical studies, I will (1) show evidence of local fossilization, (2) evaluate the major methodological approaches to researching fossilization; and (3) highlight the modular nature of fossilization. Then, I will present and discuss some linguistic structures that have been generally considered prone to fossilization. Finally, I will discuss the Multiple Effects Principle (MEP; Selinker & Lakshmanan, 1992) as a potential explanation for interand intra-learner differential failure.

Some Empirical Studies

Over the years, researchers who attempted empirical investigations of fossilization have generally adopted either of the two perspectives: a product perspective or a process perspective. In the product perspective, fossilization has been assumed, and the informants labeled, from the outset, ‘fossilized’, with the subsequent research effort being to confirm fossilization in such learners. Usually this is done through a ‘defossilization attempt’: if the attempt is unsuccessful, it is thought to provide evidence that the infor-mants are indeed fossilized. A process perspective, on the other hand, usually relies on a longitudinal, or sometimes, pseudo-longitudinal, study for establishing what is fossilizable.