To: Scott B.
From: Geoffrey Klempner
Subject: Rationality and mechanisms of belief formation
Date: 8th December 2009 12:58

Dear Scott,

Thank you for your email of 30 November, with your essay for the University of London Epistemology module, in response to the question, 'What if any connection is there between the rationality of a belief and the reliability of the mechanisms or processes responsible for forming the belief?'

You commented that you found this topic 'confusing'. I can see here that you have been doing plenty of reading around the topic of epistemic rationality (by the way, I would appreciate references at the end of your essay!) but were uncertain exactly which parts of the discussion were relevant to the question. The result is that you go into a number of issues that aren't directly relevant: for example, the question whether, if the aim of epistemic rationality is the formation of true beliefs, whether it is more rational to have more beliefs regardless of their content or importance. That would be relevant to a question along the lines of, 'What are the criteria for epistemic rationality?' or, 'How would you define epistemic rationality?'

What the question asks for is an account of the connection between two notions: rationality of belief, and reliability of belief-forming mechanisms. A lot of what you say is relevant to this question, but this is made less obvious by the confusing structure of your essay.

It is always a good idea, when considering a question about 'connections', to look for plausible claims about necessity and sufficiency. Later, one can complicate this picture by considering, as you go on to do, different notions of 'rationality' such as 'deontic' rationality.

If I go about forming my beliefs in a rational way, is that sufficient to guarantee that my methods of investigation will lead reliably to truth? What assumptions are built in to the idea that thinking rationally is a good way to get to the truth?

Descartes considers this problem in Mediations. You are familiar with the use of the evil demon scenario to cast doubt on the existence of an external world. But there is also an equally important application to the reliability of methods of investigation. Imagine an evil demon who has created a world where coin tossing (your example) is a better way of discovering the truth than evaluating evidence. What this shows is that there seems to be an assumption that the world itself is 'rational', that things happen in a rational, predictable, orderly way, rather than randomly or by magic.

Pursuing this further would take us off-topic, into the justification of induction. But you can see how it is a substantial question why it is 'better', for someone seeking truth, to go about truth seeking in a 'rational' way.

Is rationality of belief formation necessary for the reliable formation of true beliefs? You give the example of a child forming the belief, 'There are horses in that field'. (Actually, you get this wrong: If there are no horses in the field but only advertisement hoardings that look like horses from a distance then the child's belief is false, and so can't be 'knowledge'. I guess what you meant to do was construct a Gettier-type example. But why?) You also give the example of seeing a dog in the garden and immediately forming the belief, 'There is a dog in the garden.'

This raises two issues: The first issue concerns the beliefs of young children, which have not been formed by rational deliberation because the child is incapable of anything but the most rudimentary rational deliberation. If I hold up a red rattle to a three year old and ask, 'What colour is it?' and the child says, 'red', or ask, 'What is it?' and the child says, 'rattle', is that knowledge? What is the minimal requirement? (A computer image recognition program capable of identifying red objects or rattles doesn't have 'knowledge'.)

A second issue concerns all the ways in which we form beliefs (like the one about the dog) which don't explicitly involve rational deliberation. What is the connection with rationality? Arguably, our *unreflective* beliefs are no less rational for not being formed as the result of reflection, because we are capable of bringing rational considerations to bear at any time. E.g. what I saw in the garden can't be a horse (better example than a dog) because there is no way a horse could have got into the garden. So that's a good 'reason' for looking again, more closely.

This still leaves the issue that you address, concerning 'deontic' rationality, where one follows the norms (right or wrong) accepted in your society, and notions of rationality according to which it is (in some sense, to be explored) a priori true that processes of rational belief formation reliably lead to truth, of which there are both 'internalist' and 'externalist' varieties.

All the best,

Geoffrey