Kant's Critique of Pure Reason

Presented by Andrew in February 2018

This is a stream of consciousness blurb about Immanuel Kant's major work – The Critique of Pure Reason (CPR). I expect many Kant scholars would violently disagree with it and say it's a complete misrepresentation.

Kant says in the long intro – entitled the Prolegomena – 'I was awoken out of my dogmatic slumber' by David Hume's sceptical writings, which Kant saw as an attack on metaphysics. Kant thinks metaphysics is very important, and that everybody does it whether they realise it or not. But he thinks most people do it wrong.

Do you agree that most people 'do metaphysics', that they form metaphysical views about what the world is, what it's made of, how it came to be etc?

Two key ideas of Hume's particularly goaded Kant:

Hume said that he could not see how 'causality' could be a feature of the universe.
Hume said that 'we could not know any synthetic truths a priori'

A 'synthetic truth' is a true proposition that is not 'analytic'.

An analytic proposition is one whose predicate concept is contained with the subject concept. Here subject and predicate have the meaning you learned at high school when you studied grammar. The most common example used is that the proposition 'All bachelors are unmarried' is analytic because the concept of being unmarried is contained within the concept of being a bachelor.

We say we can know something a priori if we can figure it out simply by logic without having to look. Mathematical theorems, like '5+7=12' (Kant's chosen example) can be known a priori. We can figure it out without having to first do lots of experiments with fives and sevens of pebbles, beans and tin cans.

Kant claimed that '5+7=12', like most mathematical theorems, is a synthetic truth that can be known a priori. He said it's synthetic because the concepts of 5 and 7 are not contained in the concept of 12.

Do you think Kant is right about that? What does it mean for one concept to be contained in another?

This series of eight lectures at Oxford Uni by Dr Dan Robinson presents a droll, opinionated, sympathetic version of Kant's Critique of Pure Reason. If, like me, you don't like to sit at a computer watching a long video, you can download them as podcasts to listen to while walking, driving or in bed. Dr Robinson is almost as good as Garrison Keillor at putting one to sleep, but he does then tend to barge his way into dreams in an infuriating fashion, refusing to stop talking despite repeated, increasingly hysterical requests.

Written summaries of CPR are available:

on wikipedia
on spark notes
on internet encyclopaedia of philosophy (sections 1-3 only. The remainder of the article covers Kant's other works. CPR is covered in sections 2b-g and 3, with the sections before that covering Kant's life and the background to CPR)
stanford encyclopedia of philosophy has this root article on CPR, and has a number of others. As is common with SEP, it gives a long, deep treatment aimed at the real enthusiast.

You may also find useful the first part of this Oxford Uni lecture by Canadian Dr James Grant. He's talking about a different work by Kant, but the opening provides an interesting and useful description of Kant's peculiar use of vocabulary (including his inconsistency in using it!), the usual meaning of some of Kant's favourite terms, and the many challenges of trying to read and understand Kant.

The above summaries differ quite considerably from one another as to what they bring out as the main points. That is evidence of how long, rambling and unclear CPR is (a criticism that Kant even makes himself!), that different people can make very different interpretations of what its main ideas are. The enormous volume of literature discussing what it all means is probably only exceeded by that about The Bible, Shakespeare and I Am the Walrus. The fun thing about that is that there's always room to disagree, always something to discuss.

Kant finished up CPR by criticising how philosophers often try to apply reason outside the scope within which it is valid. This leads to what he calls the Antinomies and Paralogisms, which are ridiculous conclusions obtained when trying to do that. Although Kant was a convinced theist, he insisted that attempts to prove the existence of God fell into those categories of error, and that the only ground for believing was Faith.

Do you agree with Kant that there are limits to what reason can address? What are things it can't address? What should we do about those areas: ignore them? Or is there some other way to approach them?

Kant identified what he saw as the three important questions of philosophy, towards which people tried to apply reason:

What can I know?
What must I do?
What may I hope?

Do you think those questions cover what philosophy is about? Are these ones important? Are there other important ones omitted?

The Critique was aimed primarily at the first of the three questions. He addresses the second in a later work (Groundwork of the Metaphysic of Morals). The third is covered by the section of CPR on Antinomies and Paralogisms.

Here are what I see as key ideas in CPR:

Rationalists vs Empiricists. In the two centuries before Kant the major division in philosophy was between Rationalists, who thought we could to some extent reason about how the universe was, and Empiricists, who asserted that knowledge could only come from empirical observation. Famous rationalists were Descartes, Spinoza and Leibniz. Famous empiricists were Locke, Berkeley and Hume. Kant is seen by some as bridging the divide, or at least trying to.
Intuitions, percepts, concepts. For Kant, 'intuition' usually refers to something like a raw sensory input, like a single pixel of colour from a rod or cone in our retina. 'Sensibility' is the receiving of intuitions, or the capability to do so. A 'percept' is one or more intuitions that have been processed to be something like a shape. It is more sophisticated than an intuition. A 'concept' is more sophisticated still and means something that fits into a coherent mental framework. My no doubt erroneous interpretation is that the Transcendental Aesthetic (see below) allows us to convert Intuitions into Percepts, and our Transcendentally Deduced Categories (see below) allow us to convert those into Concepts. The 'Understanding' is the conversion of percepts into concepts, or the ability to do so.
Transcendental Aesthetic (see text of Manchester Uni lecture dedicated to this here). This is the in-built process of interpreting sensory inputs in a framework shaped by space and time. Without this, our Intuitions would be just a Mess of shapeless data.
1. Do you think space and time are a feature of the universe or rather a feature of the way our brains interpret the universe? Do you think this question even makes sense? Can you imagine there being other ways of interpreting the universe?
Transcendental deduction of the categories. This is the in-built ability to allocate percepts to categories – eg to identify that a tree-shaped object is a tree. But Kant's 'transcendentally deduced' categories are much less specific than things like trees. Wikipedia presents them as a 3 x 4 table for which, if Kant is true to form, each cell will expand out into another table, and so on. Make of it what you will.

Table of Judgements

5. Synthetic unity of apperception. This is the ability to knit together different Intuitions to form knowledge of a phenomenon. Dr Robinson gives the example of a cup of coffee, saying that we get visual, smell, taste, feel and even audible inputs about the coffee, not all at the same time, which we somehow manage to synthesise into knowledge that there's a cup of coffee there. The way he describes it, it does seem astonishing, given the inputs occur at different times and go to different parts of the brain. He adduces this as another example of a priori knowledge.

Are abilities like the last three really knowledge, or are they more like capabilities? Do you think the use of the word 'knowledge' is appropriate, and sufficient for Kant to suggest he has rebutted Hume's claim that a priori synthetic knowledge is impossible

6. Kant says that in order for us to be able to unify the Intuitions to synthesise the phenomenon of a cup of coffee, there must be a single, continuing perceiver – an experiencing subject – which he calls the Transcendental Ego. This conflicts with Hume's observation that, when he went looking for his 'self', all he could find was 'bundles of perceptions'. Hume could see no reason for insisting there is a perceiver.

Do you think a perceiver is necessary in order for there to be perceptions? Do you think this question means anything?

7. Phenomena vs (unknowable) noumena. A phenomenon is an experience of an aspect of the world, like seeing a table and maybe feeling its hardness and hearing one's knuckles rap on it. The phenomenon is presumably made up of information that has been processed into concepts and percepts. The original 'source' of each phenomenon – the 'actual table' – Kant calls a noumenon, a thing-in-itself (ding-an-sich in German). It seems to follow from Kant's worldview that the actual nature of noumena is unknowable. All we can know is the phenomena to which they give rise. If time and space are frameworks added by our brains then maybe noumena don't 'really' exist in a spacetime framework, but are rather in some other, unknowable, unimaginable framework (if any).

8. Refutation of idealism. Kant's phenomenon-noumenon distinction led some to 'accuse' him of being an Idealist like Berkeley, an Idealist being someone that claims only ideas are real, and that 'matter' is just an idea rather than having existence separate from people's ideas of it. Kant strenuously rejected this and rebutting the accusation was a major reason for his writing the 1787 second edition. He decided to label his view about the Phenomena-Noumena distinction 'Transcendental Idealism' in the hope of persuading people that he wasn't a ridiculous (in the eyes of the accusers) real Idealist like Berkeley. You've probably notice by now that Kant absolutely loves the word 'Transcendental' ('transzendental' in German)!

Do you think Kant was an Idealist? Do you see any difference between saying that things in themselves are unknowable (unimaginable) and saying that they don't even exist?

9. What Kant calls the Empirical Ego is different from the Transcendental Ego. It is the self we perceive when we reflect upon our consciousness. It's like what we see when we look at ourself.

Items 3-5 are about ways we process and arrange information, which I would say are hard-coded into our brains (apparently this interpretation is called 'psychological nativism'), but Robinson says that Kant insists they are not psychological. The Manchester Uni lecture linked above discusses this at some length.

What does it mean to say that these capabilities are not psychological? What else could they be?

* * *

Finally: all my knowledge of Kant's philosophy is from secondary, tertiary and even quarternary sources. I've barely read a whole paragraph that he actually wrote himself. For those with more determination than I have, a free translation of the German original can read or downloaded (in pdf or e-book form) here.

References to places in CPR are often in the form Annn / Bmmm where nnn and mmm are numbers. The nnn refers to the paragraph number in Kant's original edition of 1781, while the mmm refers to the para number in his major rewrite of 1787.

Bear in mind that any English version is a translation, which magnifies the considerable uncertainty that already existed in the original German as to the meaning of Kant's words and sentences.