Skip to main content

Knowledge and Perception

Seminar discussions

To refresh your memories at the end of this course, and to provide some kind of continuity to the seminars that you have this term, what will follow below is a short series of summaries of the major topics we discuss. Remember, it won't be exhaustive of the issues and is no replacement for the live group learning of the seminars

Defining knowledge

The tripartite theory gives us three parts to knowledge: that it is i)justified ii)true iii)belief

It is one thing to propose these as necessary components of what it is to know something, it is quite another to claim that they are sufficient.

Gettier shows us the possibility that we might fulfil all three criteria and yet not have what we like to call knowledge. How? we might believe something, be justified in our doing so, and it 'happen' to be true without our really knowing that it is. We reach a state of justified true belief by pure coincidence.

What Nozick does, is build knowledge into the form of 'subjunctive conditionals'. IF it is raining, THEN I will believe it. The fact that I believe something is grounded in the truth of the world. Knowledge therefore only belongs to justified true beliefs that track/respond to, or are 'grounded' in, the truth of the world.

The question we ended with is whether we want to say that knowledge is a matter for 'objective' truths. That is, how do we know we are tracking the world correctly. Do we? If not can we ever KNOW we know...?


The question was whether the truth of one belief can depend on another without our risking a regress into infinite 'reasons for a reason' (Agrippas criticism). To solve this, we need to suggest some kind of 'foundation' for truth. That is, where we can say a belief a is valid in virtue of its being grounded in belief b, which, by its nature, is more stable or infallible.

However, the principle of foundationalism is perhaps more attractive than the actual spelling out. That is, we seem to be left with a stack of 'basic beliefs', pure sense data in our heads, as being the only thing we can 'infallibly' trust as true. This just doesn't seem good enough. The minute we try to describe the content of the experience we invoke terms that refer outside of it. Our experience as such can't both be given expression and retain its 'infallible truth'. On a more general level these beginnings force us to tell a story about how to link the mind with the world. Again, (i.e. return to days of Descartes...) it seems that by making the stuff of the mind infallible we've made whats outside of it seem unacceptably fallible.

In short, it is hard to see how we can find truth in a simple claim to knowledge such as 'it is raining'. Yes, the fact that its seems to be an experience of a given kind TO ME is infallible and true, but who cares, right?. We want to know if I'm right that it is actually raining. Any claim that I make about the actual rain 'out there' does not retain or even seem to be helped by the truth of experiential "basic beliefs".


Coherentists suggest that a belief is justified iff (if and only if) it coheres with the network of beliefs one already holds. The truth of an individual belief is therefore decided upon entirely in reference to the cognitive states of the believer. This makes it an internalist theory like foundationalism.

We asked what the nature of 'coherence' actually is; whether it means more than a simple lack of conflict between beliefs. Also, we questioned whether we can make sense of coherence as existing in degrees if it is what decides the truth or falsity of a belief. Does it make sense to say a belief can become more true with more coherence?

We also asked how we first came about this network of beliefs: what decided the truth of our first belief? without a network we have no notion of coherence, and therefore seemingly no idea of truth. Similarly troubling is the question we asked of how one would go about searching one's network of beliefs in order to judge the level of coherence of a new belief. We also wanted to know if and how coherentists explain our ability to change whole sets of beliefs on the basis of new incoming beliefs. If all that makes incoming beliefs true is coherence with what you know, it seems hard to see why you would want to allow a new belief to overthrow your old ones.

Also, coherentism leaves open the possibility that my belief network may be completely bizarre - that is, totally unrepresentative of what the world is actually like - in so far as it still has truth as long as it is coherent.

However much we may want to keep the idea that our beliefs need some kind of coherence, in so far as we understand it to be paradoxical to believe p and not-p simultaneously, the notion of coherence alone does not seem to provide a sufficient account of the truth of beliefs. A necessary part of the story perhaps, but not an exhaustive explanation.


Externalist theory says that the question of whether a given belief is justified is a matter of the way the world is (conditions 'external' to the subject). It is not a matter to be decided by the 'cognitive content' of the believer (against coherentism and foundationalism). This is in direct opposition to the internalist theories we have been working on.

Supposedly then, I acquire beliefs and knowledge about the world using what I know to be a 'reliable method'. We asked not only what 'reliable' was supposed to mean, but also what such a method would look like. Although the idea of the world deciding the content of our beliefs appealed, we felt it was not enough. It did not give an exhaustive story of what it is to have a 'justified' belief.


Arguably, there is a sense of elusiveness inherent to the notion of a 'reliable method'of acquiring knowledge. What contextualism says is that this is a function of the context in which the perception occurs. That is, we can only spell out what a reliable method is in any given scenario in relation to the context in which the perception occurs. On a dark night, with foggy vision, we are unlikely to trust our perceptions to the same extent that we would on a sunny day with clear vision. Although we appreciated the idea, and worked through the logical steps of some examples, we still didn't feel this gave us the sense of reliability that we needed to explain our acquisition of knowledge from belief.

Final Reflections

Arguably what is missing is some way to seal the gap of possibility of having coincidentally true beliefs. i.e. that the world is a certain way, I believe it to be this way, but my believing it is not caused by or based on the world being that way. (think back to the Gettier coins example with Smith having a 'coincidentally true belief')

In other words, we want to build into our description of a reliable method the fact that it is reliable just because we know it reflects the true state of the world. 'Knowing that we know for the right reasons' is important. It at least suggests that there may be some kind of necessary role for cognitive content of the believer in deciding what it is for a given belief to be justified. In short, that the idea solution may have to comprimise between internalism and externalism. Although this may seem easy at first, adopting two opposed positions may mean that the strengths of one are merely the weaknesses of the other.

What do you think?