ALAN BERETTA and GRAHAM CROOKES

This paper addresses an important part oftheory construction, the production ofnew ideas, that is, the process ofdiscovery, in order to determine what value insights from the philosophy, history, and sociology ofscience might.have for the emerging discipline of SLA. We recognize the current conflict between those who espouse the rationality of science and those who point to social forces and personal motives as causal agents in the process of discovery. We present a case that endorses the role of reasoning in discovery and which accepts the need for social mechanisms appealing to the interests ofindividual scientists in order to explain how rationality flourishes. We give particular attention to a set of reasoning strategies for generating hypotheses, identified by Darden (1991), and illustrate them with examples from SLA, where possible, and neighboring fields. We proceed to argue that a plausible social explanation for the centrality ofreasoning is that the institution ofscience has evolved in such a way that its interests coincide with the interests of individual scientists, an account which is based primarily on the work of Hull (1988). Again, we indicate possible implications for the new field ofsecond language acquisition. 1.2 Discovery versusjustification oftheories Before we can address the question of the relevance of philosophers' accounts to SLA theory construction, we must first of all clarify which aspects of theory construction we are talking about. Are we interested in generating hypotheses (the context of discovery) or assessing their value (the context of justification) or both? The dichotomy, usually associated with Reichenbach (1938), makes a very broad distinction, one that is too broad for Laudan (1980). Laudan takes justification to be concerned only with the evidence relevant to the finished research report. Fully formed theories are assessed after the data are in. He finds that the allocation of the rest ofthe enterprise to discovery-'every stage in its history prior to its ultimate ratification' (Laudan 1980: 181)-ascribes to it a very wide scope which he prefers to further divide: he posits a context of 'pursuit' between discovery and justification, restricting discovery to whatever is involved in generating a new idea. There are problems in ignoring the possibility that justificatory principles might be invoked even in the context of discovery, but it seems reasonable to assume, as Laudan does, that much of the overlap within the trichotomy will occur in the context of pursuit. Therefore, henceforth we will respect Laudan's division and construe discovery as a concern limited to the generation of new ideas. The idea of formulating rules that would lead to discoveries had a long and illustrious history going back to antiquity and flourishing in the works of Bacon, shelved just for the moment; perhaps any given ad hoc hypothesis is really a legitimate elaboration ofthe theory; and so on. Moreover, would scientists judge in the same wayson each of these issues if the theory in question were theirs or that ofa rival? Again, it seems not: Mulkay (1991) provides interview data concerning a highly successful biochemist who professes to observe Popperian principles; according to his rivals, he simply refuses to see that a conjecture of his own has been refuted. As Mulkay suggests, there is a gap between the skeleton of logical rules and complex, often confusing practice. The gap provides room for maneuver. However, the greatest difficulty with observing Popperian principles is that they are prescriptive. They explain how science should be done, which is rather like the tail wagging the dog. The phenomenon in need of explanation is how science could have achieved what it has, that is, how science is done.' Much of philosophy of science for the last 30 years has been addressing this issue and many interesting accounts have emerged. In view of this, and bearing in mind that we need to make a clear distinction between prescriptivists and those who investigate actual scientific practice, there is good reason to consider the insights offered by the philosophy, history, and sociology of science. And that is the basic issue that this paper addresses: can SLA benefit from serious attempts to understand how science works? But first the question needs to be refined somewhat by considering a related question that has been treated dichotomously: the 'discovery' versus the 'justification' of theories. Universityof Hawaiiat Manoa Michigan State University