In addition to the material world there is an immaterial world which acts on the sense organs and, through these, produces ‘ideas’ in the human brain: these are of two kinds, ‘ideas of sensation’ – the physical world as perceived – and ‘ideas of reflexion’ – the operation of the mind upon the ideas of sensation and in fact the ‘world of thought’. Locke freely admits he does not know how the stimulation of sense organs produces ideas, but what he does claim to know is that knowledge of the material world consists in the ‘ideas’ of it we have in our minds. Like Descartes, Locke lacks mystery: instead of mystification there is confession of ignorance. Enter upon the scene Bishop Berkeley, his soul full of distaste for the ‘materialism’ of Locke. If, says Berkeley, our knowledge of the material world consists in the ideas of it we have in our heads, what reason is there for supposing that anything other than the ideas exists? Locke, like Descartes, concludes that the world is what it seems to be: a material thing and a mental thing, a collection of pieces of material and a collection of mental ideas. But Locke goes on to say that all the mind can know is the mental ideas and thus, says Berkeley, deprives us of any ground for supposing that the pieces of material exist. Suppose the material world did not exist: would anything about our perception of the world be changed? If all we know is the mental ideas nothing would be changed. The material world is an unnecessary and undemonstrable hypothesis: so away with it. Berkeley abolishes matter and declares the world to be a creation of mind. Everything is ‘really’ mind.

Now this conclusion is to the commonsense mind fantastic. Johnson kicking a stone and declaring ‘I thus refute him!’ is likely to represent the normal reaction to it. But I would ask the reader to observe, firstly that it is no more ‘fantastic’ than any other hypothesis which asserts the superior reality of some world other than the physical, and secondly that it does actually follow from the conclusions of Locke, which nobody in the eighteenth century regarded as anything but the plainest sense. The reason Berkeley's conclusion seems so very odd is that it is insufficiently radical: it assumes too much of what is in fact oldstyle commonsense, so that the arguments Berkeley offers in support of it are merely ingenious rather than convincing. Locke points not towards the simple substitution of one kind of knowledge for another, but towards a radical critique of knowledge itself, a thorough-going analysis and assessment of what it is we really know. This critique is undertaken by David Hume. Hume asserts that our mind consists of ‘impressions’ and ‘ideas'. Impressions are what Locke had called ‘ideas of sensation’, our perception of the physical world; ideas are images of impressions, formed in thinking and reasoning: ergo we can have no idea of anything of which we have received no impression. What has not been perceived cannot be known. You think you know fire causes heat, says Hume, but you do not know that, because causation is not something perceived. You can perceive that A is bigger than B or on top of B or always accompanied by B, but you cannot perceive that it causes B. ‘Causation’ is the expectation that what has always been conjoined in the past will always be conjoined in the future: but you cannot know that the future resembles the past because what does not yet exist cannot be perceived: you suppose it will, but that is a matter of habit. The consequence is that real knowledge of the world is fragmentary; that we have connected knowledge of it is an illusion born of habit and laziness.

Kant read Hume and it changed his life. He wanted to contradict him but couldn't: he found him irrefutable. And yet he must somehow have gone wrong, since it was impossible permanently to acquiesce in the notion that the reasoning powers of mankind were illusory. When he finally detected how Hume had gone wrong his solution was the foundation of ‘German metaphysics’ in the modern sense and provided the immediate background to the metaphysics of Schopenhauer. Kant's undertaking is a new description of the human mind. It falls, he says, into two parts: the part which perceives and the part which thinks (vide Locke's ‘ideas of sensation’ and ‘ideas of reflexion’ and Hume's ‘impressions’ and ‘ideas’). The perceiving part of the mind receives the impressions conveyed by the senses, and Kant calls these impressions ‘particulars’; the thinking part is the organ of the understanding and the objects of the understanding he calls ‘concepts’. The application of concepts to particulars constitutes ‘synthetic judgements’. What is a synthetic judgement? ‘Judgement’ is Kant's term for ‘proposition’, and he first classifies judgements as being either analytic or synthetic.