Reliablism is the position that beliefs are justified to the degree to which they are the product of reliable belief-forming processes. It differs from justification and coherentism in that unlike both of these positions, reliabilism denies that beliefs need to be justified by other beliefs. Instead, reliabilism holds that states of affairs in the world are what justify beliefs, specifically states of affairs relating to the causes of our beliefs. Under this model, if particular processes lead to beliefs that are generally correct, those processes are reliable and therefore beliefs that result from them are justified. Beliefs that result from unreliable processes, even if they happen to be right on occasion, are not justified.

As an example, it might be said that beliefs formed on the basis of careful scientific investigation are reliable, and therefore justified. Occasionally these beliefs may turn out to be wrong, but in general science is a realiable process, and so beliefs formed on the basis of scientific investigation are justified. By contrast, beliefs formed on the basis of consulting duck entrails are not justified (even if they happen to be correct), because these beliefs were formed by a process which in general does not produce reliable beliefs. It is important to understand that reliablists do not argue that beliefs are justified by the belief that they are formed by reliable processes (e.g. “I am justified in believing the cell is there because I believe the microscope is a reliable instrument”). Rather, it is the reliable processes themselves which are held to do the justifying (e.g. “I am justified in believing the cell is there because the microscope is a reliable instrument”). This is considered to be a weakness by some who believe that only beliefs can justify other beliefs, and who argue that it doesn’t make sense that a belief could be justified by a reliable process even if one is unaware that the process is reliable.

Another objection to reliabilism is that reliability can only be determined relative to some specific domain. For example, consulting duck entrails is pretty useless for most things, but may be a reliable way of diagnosing intestinal diseases in ducks. Similarly, human vision is often reliable, but there are some cases in which it is not. The problem for the reliabilist then lies in specifying when particular methods grant justification and when they don’t. Obviously one cannot simply say ‘a method is reliable only in the cases where it gives the correct answer’, for that amounts to saying that a belief is justified if and only if it is true, a result that few philosophers would be happy with. Reliabilism thus remains an active but highly contested response to the problem of justification.

Further Reading

Reliabilism: introductory piece from

Reliabilism: more detailed analysis from the Internet Encyclopedia of Philosophy

Reliabilist epistemology: a detailed overview of classical and modern approaches to foundationalism from the Stanford Encyclopedia of Philosophy