To: Matthew M.
From: Geoffrey Klempner
Subject: Kant on freedom and lying
Date: 10th February 2011 11:28

Dear Matthew,

Thank you for your email of 2 February, with your essay for the University of London Diploma Ethics Historical Perspectives module, in response to the question, 'Does Kant have a defensible account of freedom?' and your email of 5 February resending an essay which you originally attempted to send on 15 January, in response to the question, 'Is Kant right in thinking that appeal to the categorical imperative is enough to show lying to be morally wrong?'

Kant and lying

Kant actually wrote an essay devoted to this question, 'On the Supposed Right to Lie from Philanthropic Concerns', which is included in some but not all editions of the 'Groundwork for the Metaphysics of Morals'. In the essay Kant is adamant that lying can never be morally right even in a case where you need to lie to protect the life of an innocent person.

Although it is OK to spend some portion of the essay going into the background of Kant's moral philosophy and the various formulations of the categorical imperative, the examiner expects you to concentrate on Kant's treatment of lying. Your only contribution is to make the obvious point that the claim seems counterintuitive: there are cases where bad things would happen (e.g. criminals discover the whereabouts of a witness you are protecting) if you are not prepared to lie.

But, then, Kant isn't concerned with consequences as such. You make a good point that if all we have as the touchstone for moral decision making is the categorical imperative, then the goodness or badness of consequences must be judged by this principle alone. We can't appeal to our evaluation of consequences to bolster the principle or fill in the gaps where it doesn't seem to give the answer we want.

Another point you make, which deserves to be expanded upon, is the contrast between cases where a maxim fails the test of the categorical imperative because the agent's intention would be self-contradictory, and one which fails the test because there is something incoherent or self-contradictory about the consequences. Is Kant slyly appealing to consequences here? Or is this just another way to refer back to the question of the agent's intentions? If I intend an action, and at the same time foresee that the consequences, or wider consequences of my action would frustrate my intention, then that is another way for an intention to be 'incoherent or self-contradictory'?

I think that Kant has a argument in respect of the question of lying which deserves to be reckoned with (although I obviously don't agree that it follows that one can never lie under any circumstances whatsoever). There is in fact a paradox here. Any attempt to state the 'exceptions' to the law that one must never lie is self-frustrating. I wrote something about this in my lectures on business ethics. Here is an extract from unit 5:

Consider the seemingly innocuous statement, 'I sometimes tell a lie when I'm in a tight spot.' What exactly does this mean? If I say this to you, then you are to understand that if I am ever in a tight spot, my words are not to be believed. But in that case, I cannot lie to you when I am in a tight spot, because a necessary condition for successfully lying is that one's words are believed by the person lied to. The only time I can successfully lie to you is when I am not in a tight spot, because then you are not expecting me to tell a lie. Let's consider now a variation on this scenario. 'I nearly always tell the truth, and only lie when I am in a very tight spot.' Even in a tight spot, I will tell the truth and face the adverse consequences, but in a very tight spot I will lie. (For the purposes of this argument we can assume a common understanding concerning what constitutes the difference between a 'tight spot' or a 'very tight spot', and when either situation obtains.)

By making this statement, I have effectively devalued my word. I might still think that I can still get away with lying in a tight spot. However, you are now justified in reasoning as follows: as I can no longer get away with lying when I am in a very tight spot, the tightest spot where I can successfully lie is a tight spot. So you will not believe me in this case either. Nor, repeating the same reasoning, will you have any reason to believe me when I am in a slightly tight spot. Generalizing from this example, to admit that one sometimes lies, to gain any advantage whatsoever, is logically self-defeating. This argument shows that it is self-defeating to own up to lying. You might think that this is still different from saying that it is wrong to tell a lie. Obviously, if you're going to gain an advantage by lying, the best strategy is to present oneself as someone who is totally honest and trustworthy.

That may very well be so. But the point remains that whenever we discover that someone has lied, we have no choice but to condemn the action. If I discover that X lied to get out of a tight spot, and I excuse the action on that account, then that is tantamount to my saying that I believe that it is acceptable to lie in those particular circumstances. That's what I would do in that person's shoes. And now I am in exactly the same situation that I was in before, when I admitted that I sometimes lie to get out of a tight spot. An action which we will never freely admit to and always condemn, is by definition always wrong. (See http://ethicaldilemmas.co.uk.)

Kant on Freedom

In the 'Groundwork for the Metaphysics of Morals' Kant presents the question of freedom as the search for a 'third term' which would provide the necessary justification of the categorical imperative. It is not sufficient merely to analyse our ordinary moral notions, specifically the notion of a 'good will' in order to uncover the underlying assumption that the sole basis for moral decision making is the categorical imperative. Why not? Because our moral notions, the language of morality itself, could be fatally flawed. We could be wrong. We could all be living under the illusion that there is such a thing as a categorical imperative when in fact there is not.

Kant is adamant that the only way that the categorical imperative can be show to be justified -- at least to be 'possible' -- is if the coherence of the assumption of human freedom is demonstrated. We can't prove that we are free, in the necessary sense, but we can at least believe in the possibility of freedom. In the final section of the Groundwork, Kant presents the theory of phenomena and noumena as the framework which would allow for the possibility of human freedom.

As you argue in your essay, causality relates to the phenomenal world, while our metaphysical freedom derives from the fact that we exist as noumenal subjects and not merely as phenomenal selves.

This is deeply puzzling. It's bad enough that we are asked to believe in the 'existence' of a noumenal reality concerning which nothing can be known, to which none of our empirical concepts apply, not even the a priori notions of substance or cause, not even (as Schopenhauer observed) the concept of number (talk of 'things in themselves' or 'noumena' presupposes that there is more than one noumenon or thing in itself, but if we can't apply concepts to them how can these 'things' be counted?).

However, there is a more sympathetic reading of Kant's views on freedom, which may or may not have been in the back of his mind when he put forward the phenomena-noumena theory.

In contemporary philosophy, there has been quite a lot of discussion of the contrast between the 'logical space of reasons' and the 'logical space of causes'. You will find this, e.g. in John McDowell's Harvard lectures 'Mind and World'. The idea here involves a strong resistance to the kind of reductionism which would seek a 'translation' of talk of 'reasons for action' into physicalist language. The finesse here involves recognizing that reasons can (and indeed must) be 'causes' as Davidson argued. So it would be wrong to think that these two 'logical spaces' have no point of overlap. My decision to write this email today because it was overdue and I didn't want to keep you waiting any longer (sorry!) caused me to open this file and start typing.

Reason isn't a 'first cause' in the sense of the belief that God is the first cause, the uncaused cause, in the series of causes and effects (something you mention). Rather, the logical space of reasons is self-enclosed, in a way which prevents us from looking beyond the reasons we could give for our actions to the physical machinery working underneath. (There is some scope for 'looking beyond', e.g. in depth psychology.)

Is this observation enough to rescue Kant? Consider the laws of logic. These laws, such as the law of non-contradiction are recognized as being absolute. They are principles which govern rational discourse. Any apparent exceptions must be accounted for. It is much less controversial, to say the least, that human thought 'obeys' the laws of logic, than to say that human reasoning about morals 'obeys' the categorical imperative. And yet, the two claims are alike in that according to Kant both define what it is to be 'rational'.

Consider someone following a chain of logical inference. The success or failure to reach the correct conclusion depends on the machinery of the brain, a long chain of causes and effects. But that doesn't mean that the laws of logic as such are not absolutely necessary. In the same sense, one might argue, the fact that we are empirical beings subject to the laws of causality like everything else in the universe does not show that the categorical imperative is not absolutely necessary, as being definitive of what it is to be a being whose actions are governed by 'reason'.

All the best,

Geoffrey