Pragmatic and Idealized Models of Knowledge and Ignorance

It is a pervasive assumption in recent analytical philosophy that knowledge can be defined as a modality representing a rational agent's true and consistent beliefs. Such views are based on rationality assumptions. One is that knowledge can only consist of true propositions. This way of speaking is sharply at odds with the way we speak about knowledge, for example, in computing, where a so-called knowledge base can be a database, that is, a set of data that has been collected and is thought to consist of true propositions, even though, realistically speaking, many of them might later be shown to be false or untenable. In this study, opposed to the truth-implying idealized sense of "knowledge" there is postulated a pragmatic sense of the term in which knowledge is sometime defeated, or shown to be false as new data comes in.I The idealized conception of knowledge has worked its way, through the developments of epistemic modal logic based on the abstract notion of possible worlds, into analytical philosophy. Possible worlds, introduced as a technical device in the work of Kripke (1963) on modal logic, dominated the vocabulary and methods of analytical philosophy from that point on. This device purported to provide philosophers with a powerful explanatory apparatus for explaining the notion of knowledge. Hintikka (1962) started by arguing that his epistemic modal logics represented what is meant by the concept of knowledge, but later shifted from a descriptive to a normative approach in which he argued that epistemic logic represents an idealized model of how a rational agent knows (Girle 2003: 121). Girle's comment (121) is whether such normative models are real enough or too ideal, suggesting that, through such idealization, philosophers may have been led rather uncritically down a path that has confused more than intelligently guided the study of everyday arguments, based on the notion of knowledge that philosophical analysis should to be able to analyze and understand as part of their methodology. This paper puts forward a contrast between this idealized model of knowledge and a pragmatic model. On the latter approach, the term "knowledge" is defined in terms of a search through a database that may be more or less complete. As this dynamic process continues, more and more propositions are collected, and may be verified and falsified. Two factors are shown to be very important in judging knowledge in this framework. One is the question of whether the database is complete, in the sense that all the true propositions are known. The other is the contrast between what is known and what is not known. On this model, knowledge is defeasible, meaning that