To: Scott B.
From: Geoffrey Klempner
Subject: Kant on the relationship between freedom and morality
Date: 6th April 2010 13:23

Dear Scott,

Thank you for your email of 26 March, with your essay for the University of London Ethics Historical Perspectives BA module, in response to the question, 'Does Kant offer a defensible account of the relationship between freedom and morality?'

You've had a good go at what is an extremely difficult question. Kant's account of freedom in the final part of the Groundwork for the Metaphysic of Morals is infuriatingly obscure. I remember reading this as a first year undergraduate, and feeling that there was something incredibly deep about it, yet I never actually managed to grasp what 'it' was.

According to Kant, all our actions including our moral actions are determined, insofar as we belong to the world of appearances or phenomena. However, we are simultaneously members of the noumenal world, or the world of things in themselves, and as such our actions are not governed by the law of causality.

This isn't some 'compatibilist' doctrine of how we can be 'free' even though our actions are determined (e.g. Hume) nor even a view like Spinoza's that freedom is the capacity for one's actions to be determined by reason.

To clarify, one thing you could have done is explain Kant's strategy in the Groundwork, or at least how he maps out the argument. He begins by analysing the notion of a 'good will', an analysis which leads him to the three (or four) formulations of the Categorical Imperative. But he recognizes that this is not a 'proof' that ethics is grounded in the Categorical Imperative, because he has yet to establish the 'possibility' of the Categorical Imperative. In order to do this (he thinks) he needs to invoke the metaphysics of phenomena and noumena.

Despite my best efforts, I can't make any coherent sense of this and I suspect that many students of Kant have similar doubts. You can say anything you like about the noumenal world and the objects in it because none of the constraints or laws that apply to the phenomenal world can be applied to the noumenal world. You can't even (as Schopenhauer observed) talk about noumenal objects (in the plural) because the very conditions for identity and difference are abrogated once we make the transition from the phenomenal to the noumenal world.

Take any moral man, who always acts according to the Categorical Imperative. His actions, as the actions of an entity in the phenomenal world, are fully determined by prior causes. To say that he is 'also' a noumenal entity whose actions are thereby 'uncaused' just seems incomprehensible.

However, the question asks about Kant's 'account of the relationship between freedom and morality'. Maybe there is something that can be salvaged, or is even worth preserving.

John McDowell, in his Harvard lectures published as 'Mind and World', talks about the distinction between the 'logical space of reasons' and the 'logical space of causes'. There does seem to be something very fundamental about the way we relate to other persons, by contrast with the way we relate to beings (such as non-human animals, or computers) which we do not regard as 'persons'. P.F. Strawson makes a similar point in his British Academy lecture 'Freedom and Resentment'.

If I say something to you, with the intention or expectation that it will result in some action by you, this can either be seen as engaging in discourse, or as merely controlling your behaviour, depending on my prior intentions. The former is persuasion, the latter is merely manipulation, or 'using' you. It is true that in both cases, the scenario can be analysed in terms of causes and effects -- sounds come out of my mouth, are detected by your ears, bring about changes in your brain etc. However, in the former case there is also available an alternative explanation which belongs to the 'logical space of reasons'.

It is possible to view Kant's enterprise as an attempt to derive rational principles of action which are in some sense analogous the laws of logic which apply to the things that we judge or believe. If I follow a logical inference and accept the conclusion, then there is in principle an explanation in terms of causes and effects, for what I come to believe as a result of performing that inference, but it is also true that I performed that inference simply because of the laws of logic.

In a similar way, when I knowingly and deliberately do an ethical action, the explanation is given by my application of the Categorical Imperative, even though it is also true that my action is an effect of a deterministic chain of causes and effects.

Corresponding to the distinction between hypothetical imperatives and categorical imperatives, one can make a distinction in logic between proofs from prior sets of contingent premisses (e.g. 'Socrates is a man, all men are mortal, therefore Socrates is mortal') and proofs of tautologies, or logically necessary propositions. The laws of ethics are analogous to logically necessary propositions, e.g. 'Do not tell a lie'. This is a principle governing actions, which does not depend on any prior assumptions (e.g. that being caught telling lies damages one's reputation).

In other words, just as there are things I must believe -- the theorems of maths and logic -- regardless of empirical evidence, so according to Kant there are things I must do, regardless of my prior desires or interests. So if Kant is right about the Categorical Imperative, then it follows that ethical actions are those which are determined purely by reason and not by prior desires or interests.

In the light of this, it is tempting to say that Kant didn't need to apply his theory of phenomena and noumena. If you think that there is a Categorical Imperative which governs our actions then in this sense you do believe that morality and 'freedom' are, in a sense, one and the same. This is a freedom we can have, even though every bodily movement we make is governed by laws of cause and effect.

All the best,

Geoffrey