International Encyclopedia of Systems and Cybernetics

"The result of learning and cognition, tested in practice (ideas, conceptions)" (UNESCO – UNEP Glossary, 1983, p.17).

Of course, many other definitions have been proposed. The UNESCO Glossary however adds significant comments: "Includes the cognition of real objects and phenomena, laws of nature and society, environmental problems and the correct reflection of all this cognized information in the process of thinking" (Ibid). Taking in account the following reflexions, bearing on the general conditions of knowledge, this is as good a general definition as any other.

H.von FOERSTER observes that there is ".. an almost universal confusion in which "knowledge" is seen as a commodity, i.e. is identified with substance rather than with process… An educational system that confuses learning with the dispensing of goods called "knowledge" may cause some disappointment in the hypothetical receivers, for the goods are just not coming: there are no goods" (1974, p.21).

This view about knowledge thus enounced is notwithstanding generally unheeded by A.I. specialists who, in D. GREGORY's words, believe that:

"Knowledge is a commodity. We can trade it, teach it, learn it, remember it, represent it, discover it…

"Knowledge is distinct from its knower – just like data are distinct from computer disks;

"Knowledge is sets of true facts together with the rules for combining them;

"Knowledge can be reduced to sets of primitives – just like matter is ultimately reducible to fundamental particles" (1993, p.67).

For the practical application of these disputable views, which seem to confuse knowledge with simple information, or even mere unstructured collections of data, see hereafter T. ÖREN's Knowledge's General Taxonomy.

As an antithesis, D. GREGORY states "PASK's idealist axioms":

"Knowledge is a process, not a state. It is not "that which is known", but the process of knowing (alias observing, alias distinguishing).

"Knowledge consists in coherent systems of relationships that connect knowers with their world, not facts with each other, independent of their knowers… (note that the sky is not blue to insect eyes – But then who is to say that any concept of "sky" exists for an insect, to begin with).

"Primitivity is with respect to the knower, not the world: knowing that food satiates hunger is elementary knowledge for a small child. But very elaborate for a biologist" (1993, p.68).

In the same vein as von FOERSTER, D. BOHM and F.D. PEAT write: "Knowledge of reality does not therefore lie in the subject, nor in the object, but in the dynamic flow between them" (1987, p.67). In J.L. VULLIERME words: "The object of knowledge is the knowledge of the object by a local, and localized observer" (1990, p.151).

Knowledge is thus the result of a recursive transaction between an observer, already endowed with a basic cerebral organization, and reality and, consequently, supposes a perpetual readjustment of the internal state of the subject. This internal readjustment tends however to become more and more stabilized through time, due to the formation of learning algorithms (individually and socially conditioned).

Finally, in W. REEVES words, knowledge is: "… neurally incorporated and organized information (which) supplies us with a usable and lasting sense of the world" (of course, in our own terms) (1992, p.1102).

It is noteworthy to emphasize this opinion, based on the neurology of communication, in relation to some of the following entries which precisely accept in an implicit way the quite different notion of knowledge as something tangible that can be stored, located, transformed, etc… (This viewpoint proper to computer scientists should not be uncritically accepted outside of computer science).

As a collective cultural endeavor, I. PRIGOGINE observes that: "… the ideal of complete knowledge, which has haunted Western science for three centuries… is only meaningful for the limiting case of stable dynamic systems. The message of the 2nd law is precisely that we are living in a world of highly unstable motions, to which the classical ideal does not apply" (1986, p.2).

In this sense, the complete and absolute knowledge ideal should be replaced by noncontradictory coherence in knowledge, submitted to constant validation through a falseability process "a la POPPER".