Synthetic a priori

We can break Kant’s approach down into two parts. Firstly, he accepted Hume’s distinction between conceptual knowledge and empirical knowledge, but he added a third dimension that could account for metaphysical cognition.

Kant distinguished between a priori and a posteriori knowledge - that is propositions we know to be true without checking (e.g. ‘a triangle is a shape with three sides’) and those we learn through experience (e.g. ‘the mug is on the table’) but he also distinguishes between analytic and synthetic knowledge. A proposition is analytic if the predicate is ‘contained within’ the subject. So for example the predicate ‘has three sides’ is part of, constitutive of, the concept of a triangle. The description ‘has a length’ is part of the concept of ‘a stick’. A Synthetic proposition on the other hand is one where the predicate is not contained within the subject. ‘Is green’ is not part of the meaning of table - it can be part of the meaning of a particular table, but not of the concept table in general. Kant now has a four-way distinction:

The analytic a priori is basically just Hume’s Relation of Ideas and the synthetic a posteriori is his Matters of Fact. Kant dismissed ab initio the idea that anything might be analytic a posteriori. But the key concept was synthetic a priori: ideas which were known without needing to be experienced, but where the predicate was not contained within the subject. This, Kant thought, was the nature of Metaphysical Cognition.

The copernican revolution

Secondly, Kant needed to show how such synthetic a priori knowledge was possible. Kant’s key idea was that this knowledge was not knowledge about the world, but a precondition of knowledge about the world.

What stops this image being nothing but a random collection of pixels?

There are patterns in this photo, shapes that we recognise as a woman and her children, forms that we know to be ragged clothes and a concerned expression on her face. Where do these patterns, these forms exist? Are they out there in the world? Or are they in our minds?

John Locke, for example, believed that our mind conformed to the objects we experience - that the pattern and order in our experience was a feature of the world. Kant rejected that view and instead argued that the objects we experience conform to our mind. These patterns were features of our mind, not the objects in themselves. Kant called this shift a ‘Copernican Revolution’. In the same way that Copernicus recognised that the earth only appeared to go round the sun, Kant thought we must reject the idea that the order in our experience corresponds to anything other than the way things appear to us.

For Kant, our minds had a kind of default factory operating system - a collection of synthetic a priori ideas.

And this operating system was a prerequisite for us making sense of our experiences. To look at something without the appropriate synthetic a priori ideas was like trying to use a webcam without the appropriate drivers. Our minds will simply be unable to read the information. 

Just as the correct driver/program is a precondition of a computer being able to use a webcam, the synthetic a priori concepts are the precondition for our experience.