We are like sailors who have to rebuild their ship on the open sea, without ever being able to dismount it in dry-dock and reconstruct it from the best components. (Otto Neurath)
Knowledge isn’t built on a foundation. Deducibility is a transitive relation; the conclusion of a valid argument cannot entail anything that is not also entailed by the premises. Nothing is built, since all that is entailed is the original set of premises. Since the foundation cannot be logically weaker than the set of all it entails, the foundation is the entire building.
The metaphor of construction in epistemology is a cognitive trap. The notion of “building from a foundation” cuts off creativity and diversity, reducing epistemology to unpacking the logical content of a few propositions. What does this achieve? Very little, almost nothing is produced but restatements of common beliefs – ‘I exist,’ ‘there is thinking,’ and so on – and little can come from them without some additional, and quite dubious, assumptions about phenomenal states.
I’m willing to grant that phenomenal states are indubitable to the justificationist out of a sense of leniency. Unfortunately for the justificationist, this does nothing to further their program. One cannot have access the noumenal world that easy. It is as if a man sets out to cross the Grand Canyon with little more than a running start.
No theory is ever proven beyond doubt, or made more probable by experience, or even had positive reasons to support. Stop that kind of talk, for it goes nowhere: it’s the mythical philosopher’s stone. It doesn’t exist. It never existed.
Such ends should be rejected. It’s ridiculous to respond to an explanation of why certainty is something we can never have with offering a theory of knowledge that provides it. That be to miss the point: anything that offers certainty must be taken to be false until it addresses all the available criticisms of its position. If the alchemist objected to the chemist’s program on the grounds that the chemist could not transmute lead into gold, it would be a laughable objection. Those that seek justification are, so I conjecture, alchemists.
Critical Rationalism is such an alternative solution. It cuts the Gordian knot by dispensing with any kind of support or certainty. Instead of worrying about finding reasons for belief, we should attempt to find problems with our current ideas, propose solutions and then criticise them according to some criteria, like whether they solve the problems they are meant to solve, whether they conflict with other theories, whether they conflict with experimental results, and so on.
In a sense, I’m a fallible or conjectural apriorist. I make the best of my imaginative and critical faculties to eliminate error, identify new (and hopefully deeper) problems and invent better solutions to my problems. For instance, I often reject a theory (or a proposal for action) before it has ever been tested. Why? Sometimes, theories are just bad explanations.
First, I need to stress that there’s no such thing as an established a posteriori fact. All experience is that of complex phenomena linked together in sequence by a fallible and prejudicial brain. Experimental testing to establish facts, even if they are designed to refute a theory, is then not the main point of human action. It’s subsidiary, a part of all possible criticisms one may subject a theory. Experience doesn’t arbitrate between theories — humans arbitrate.
Judgement is important, the decision that one will adopt the results of an experiment or not, that one will adopt the results of a critical discussion or not. One possible judgment is that a theory can be rejected even though a test is not necessary in order to decide against it.
An infinite number of theories are rejected as bad explanations long before they could be tested. Why should I consider the claims of people that assert they have ESP, or are the reincarnated spirits of Atlantians, or that the moon is made of green cheese? They do not stand up under the slightest scrutiny.
Empirical tests should be thought of a particular kind of judgment one makes, of settling the most difficult scientific disagreements. Testing is then not the primary method of science — it is a critical discussion. Crucial tests, then, aren’t so crucial in making most of our judgments. All knowledge is created by conjecture/criticism. The rules governing how to settle critical preferences between theories are then a matter of preference.
What I’m getting at is that the demarcation problem should be superseded by a larger problem — that of rationality. (This was Bartley’s problem.) What demarcates between rational and nonrational or irrational arbitration? How do we determine the comparative merits of rival theories? Which theory are we to reject and which one are we to (tentatively) prefer? This, rather than Popper’s solution to the demarcation problem, is the revolutionary part of Critical Rationalism.