To: Scott B.
From: Geoffrey Klempner
Subject: Is a belief formed by a reliable mechanism justified?
Date: 19th November 2009 13:57

Dear Scott,

Thank you for your email of 16 November, with your essay for the University of London Epistemology module, in response to the question, 'Critically assess the claim that any belief formed by using a reliable mechanism is justified.'

This is an excellent answer to the question, which offers a credible critique of the reliabilist theory, in favour of a view of knowledge as an achievement brought about by the exercise of the knower's 'epistemic virtues or cognitive facilities'.

At the core of this critique is a plausible counterexample to the reliabilist model. A broken thermometer which gives fluctuating readings unconnected with the actual temperature serves as a 'reliable means' of determining the temperature of the room, because someone is secretly altering the temperature of the room to match the thermometer.

As reliabilism has been considered as a strategy for overcoming 'Gettier counterexamples' to the thesis that knowledge is justified true belief, it is certainly a blow to the reliabilist that a Gettier-type of counterexample can be found to reliabilism.

I would like to discuss this point further, as it is the heart of your argument. First, is your example of the broken thermometer correctly describably as a Gettier case? I have doubts about this. In each of Gettier's counterexamples, and also the many other examples offered by epistemologists making the same point, the intuition being appealed to is that a justified true belief cannot be knowledge, if it turns out to have been true 'by accident'.

However, in the case of the broken thermometer, you admit that it is no accident that the temperature of the room matches the reading on the thermometer. Someone (or some thing, such as a thermostat controlled by the thermometer) is reliably adjusting the temperature of the room to match the thermometer reading.

You say that this contradicts *another* plausible principle, that the intention of a belief is to reflect how things are in the world. If things are set up so that the belief is rendered true by changing the world (=temperature of the room) then this violates a condition for knowledge.

In that case, it isn't a Gettier-type counterexample. At best, we can say that it has a family resemblance to Gettier cases.

But I'm not convinced anyway. Let's take the variation which I proposed, that the temperature of the room is controlled by a central heating thermostat, which is itself controlled by the thermometer readings. The advantage of this is that we have eliminated the aspect of 'trickery' (your mischievous daughter). Our intuitions tell us that something can't be knowledge if it is the result of trickery. But that's not really what is at issue here. We are assuming, for the sake of argument, that the process (whatever it is) is reliable, and therefore discounting, e.g. the possibility that your daughter will eventually get bored and fail to make the appropriate adjustments.

Let's say I move into a new house which has 'thermometers' in every room. Every time I care to look, I am reliably informed about the temperature of that room. What I don't know (and which would annoy me greatly if I did know it) is that the 'thermometers' are not acting as thermometers. They are set up to give random 'temperature readings' which arise from an internal computer chip generating random numbers. However, I don't see that this false belief invalidates my knowledge. I get on the phone to the central heating service engineer to complain that the central heating system is running way too hot. 'What temperature is the room now?', he asks. 'It's 28 degrees Centigrade.' This is knowledge for the central heating engineer, so surely it is for me also.

What I do think your example counts against is the response to Gettier along the lines of, 'Knowledge is justified true belief, which does not involve any false assumptions'. This is surely too strong. In fact, I suspect that if followed through consistently, this definition of knowledge would lead to total scepticism. Any belief, if examined in sufficiently meticulous detail, will reveal false assumptions somewhere or other.

As for the view of knowledge as an achievement of someone who exercises 'epistemic virtues or cognitive abilities', people differ in how cognitively 'virtuous' they are. Maybe I should have been more suspicious when the temperature which I set on (what I thought to be) the central heating 'thermostat' had no discernible affect on the temperature of the house. But knowledge is a sufficiently robust notion to survive inefficient or unvirtuous approaches to knowledge gathering. Undoubtedly, being less cognitively virtuous than, say, my neighbour, it is not surprising that my neighbour knows a lot more than me and has far fewer false beliefs. But what I know, I do *know*.

All the best,

Geoffrey