Episteme
knowledge proper · epistemic knowledge
Greek for knowledge in the strict sense — what is known rather than merely believed. Plato required belief to be tethered by an account; Aristotle reserved the term for demonstrative knowledge of what cannot be otherwise, distinguishing it from craft and practical wisdom.
In practice
A consultant who can predict what a change will do, and say why, has episteme. One who reliably gets good results and cannot say how has techne. The second is worth more per hour and is much harder to transfer, teach or sell.
The common mistake
Treating knowledge as justified true belief and assuming the definition is settled. Gettier (1963) broke it in three pages, and sixty years of repair attempts have produced no agreed replacement.
Someone who guesses correctly and someone who knows have both got the right answer. That the second has something more is obvious; saying what it is has resisted every attempt for two and a half thousand years.
The Greek accounts
Plato poses the problem in the Meno: true beliefs are as useful as knowledge but they run away, and are tied down by working out the reasonPlato, Meno 97e–98a. The image is of the statues of Daedalus, which run away unless fastened. True opinion tied by an account of the reason — aitias logismos — becomes knowledge and stays put.. The Theaetetus then tries to define knowledge as true belief with an account and abandons each version of 'account' it considers, closing without an answer. Aristotle narrowed the term instead of defining it: episteme is demonstrative knowledge of what cannot be otherwise, and it sits alongside four other intellectual virtues in Nicomachean Ethics VI — techne, craft knowledge of how to make; phronesis, practical wisdom about what to do; nous, grasp of first principles; and sophia, which combines nous and episteme. Most modern confusion about the word comes from collapsing those five into one.
Gettier
The justified-true-belief account survived until Edmund Gettier's three-page paper (1963). His cases produce justified true belief that is plainly not knowledge, because the justification is disconnected from what makes the belief trueGettier, E. (1963). 'Is Justified True Belief Knowledge?' Analysis 23(6). In the standard form, a subject infers a true conclusion from a justified but false premise. The belief is true, and justified, and true for reasons having nothing to do with the justification.. The repairs came quickly and none has held: no-false-lemmas conditions face further counterexamples; Goldman's reliabilism (1979) makes knowledge a matter of the process rather than the reasons; Nozick (1981) required the belief to track the truth across nearby possibilities. Timothy Williamson (2000) concluded the programme was misconceived — knowledge is not analysable into belief plus conditions but is the primitive, and belief should be explained in terms of it.
A different sense entirely
Michel Foucault's épistémè (1966, 1969) names something else: the historical a priori of a period — the unnoticed conditions that determine what can count as a serious claim at all. It is not knowledge but the substrate that makes particular knowledge possible, and it shifts discontinuously. The two senses share a word and not a subject, and running them together produces a great deal of bad writing.
The objections
Sixty years of Gettier literature has produced counterexamples to every proposal and agreement on none, which is itself evidence. Two readings are available: that the concept is not the kind of thing that admits necessary and sufficient conditions, or — Williamson's — that the analytic direction was backwards from the start.
There is also a standing objection to Aristotle's restriction. Episteme as demonstrative knowledge of the necessary excludes almost everything anyone actually knows, since empirical claims about contingent matters do not qualify. That is a coherent position and an inconvenient one, and much of the modern history of epistemology is the attempt to loosen it without losing the distinction from mere opinion.
What it rules out
It rules out treating a correct answer as evidence of knowledge — Plato's point about true belief. It rules out confidence as a criterion. And it rules out the assumption that 'justified true belief' is a settled definition; it is a refuted one, still taught.
It does not rule out acting on true belief. Plato's own observation is that for guiding action it does just as well; what it lacks is stability, which matters over time and not in the instance.
Sources
Aristotle. Nicomachean Ethics VI; Posterior Analytics I. · Foucault, M. (1966). Les Mots et les choses; (1969). L'Archéologie du savoir. Gallimard. · Gettier, E. (1963). 'Is Justified True Belief Knowledge?' Analysis 23(6). · Goldman, A. (1979). 'What Is Justified Belief?' In Justification and Knowledge. · Nozick, R. (1981). Philosophical Explanations. Harvard University Press. · Plato. Meno 97–98; Theaetetus. · Williamson, T. (2000). Knowledge and Its Limits. Oxford University Press.
Concept web
Open the full webQuestions
What is episteme?
Greek for knowledge in the strict sense, as against mere belief. Plato required true belief to be tied down by an account of the reason; Aristotle reserved it for demonstrative knowledge of what cannot be otherwise.
What is the difference between episteme, techne and phronesis?
In Nicomachean Ethics VI, episteme is demonstrative knowledge of necessary things, techne is craft knowledge of how to make something, and phronesis is practical wisdom about what to do in particular circumstances. They are distinct intellectual virtues, not degrees of one.
What is the Gettier problem?
Gettier (1963) produced cases of justified true belief that are clearly not knowledge, because the justification is disconnected from what makes the belief true. Sixty years of proposed repairs have produced no agreed replacement definition.
What did Foucault mean by episteme?
Something different: the historical a priori of a period — the unnoticed conditions that determine what can count as a serious claim at all. It is the substrate that makes particular knowledge possible, and it shifts discontinuously between periods.