EAI OnKnowledge

We saw above that an essential part of the scientific method for understanding the world, induction, was heavily dependent on the existing basis of "knowledge", facts accepted as being true.

There are theorems according to which people gathering information about the same world in mutually "fair" ways will eventually agree on any general subject; these theorems somehow explain Science. Now, there are subjects for which no satisfying notion of fairness is possible: for instance, because observation of certains facts can only be done once, by few witnesses who can't always be trusted; or because that observation costs too much, and isn't worth it (at least, so we expect), because there are so many observations, that we don't know which to focus upon and try to integrate with other data into the knowledge base; because we don't know for sure which part of the base can be simplified by some explanation; etc. Our problematics is thus to understand how to manage this knowledge base; how to choose how to expand

fortune: "My opinions may have changed, but not the fact that I am right". Our knowledge base needs not be increasing, without ever questioning its own content; it needs not be even consistent.

This page is linked from: Ethics and Information