Some Problems with Falsification: An Illustration from SLA Research

Abstract

This paper reviews some philosophical literature which argues that falsification is difficult, if not impossible to achieve. In the applied linguistics literature, there is the claim that successful SLA is the result of social, psychological, and affective factors. The argument that this claim has been falsified is examined and is found to be untenable. Various problems with falsification as a strategy in SLA theory construction and evaluation are discussed, and an argument is made that exploration of constructs to generate new formulations is a more productive approach to SLA theory building. Beretta (1991) points out in his recent paper on theory construction that second language acquisition (SLA) researchers are in virtual agreement that theories and hypotheses must be falsifiable. Following the philosophy of Popper (1972), SLA researchers maintain the conviction that truth cannot be achieved by accumulating confirmatory evidence because it is always possible that some future observation or experiment will provide negative results. Therefore, a hypothesis can only be maintained if, after repeated testing, it survives disconfirmation or falsification. Thus, a criterion which our field uses in judging theories and hypotheses is that they must be falsifiable; that is, they must be amenable to disconfirmati on by some empirical test or observation. The Popperian view appears to be the default position in the research community and in society in general. Children become familiar with it when learning about the 'scientific method' in elementary school. It is reinforced in high school science classes and is explicitly taught in university research design and statistics courses.