Kant's Letter to Hume

A Pedagogical "Forgery" (© by J. Garrett)

Koenigsberg, Preussen
21 July 17761

Mr. David Hume
Edinburgh Scotland

Esteemed Mr. Hume,

I hope this finds you well.

As a devoted student and admirer of your philosophical works, which have exercised an important role in the development of my own thinking, I wanted to share with you some of my thoughts in the area of morals that have been stimulated by your writings. I do not assume that you will agree with all of them, but they may, in some small way, continue the conversation that is philosophy and to which you have contributed much and may perhaps stimulate a response from your eminent mind.

I have not as yet published any major works on these topics2, and it will take a few more years before I am sufficiently satisfied with the details to be bold to submit these to the press, but I can share the outlines that I expect my further inquiries will follow.

My own view is indebted to what you have written in the area of theory of knowledge, and it is there that I must start, before I can explain why I put emphasis in my philosophy of morals on something different from yourself.

Your theory of understanding and my study of what I call experience share a common starting point. You insist that all our knowledge, apart from the a priori knowledge that derives from relations between ideas alone, must be derived from contingent sense experience. Extremely important in your account is the fact that our sense impressions have relationships of contiguity and succession with each other. Continguity refers to spatial relationship. Of course there are degrees of nearness between objects. Succession refers to chronological order, what comes earlier and later from the perspective of time.

You boldly challenge the received notion that we have knowledge of cause and effect3 and the notion that we know that each of us have a substantial self or mental substance4, and so forth. We must set aside as unproven, you suggest, as we must all ideas that do not correspond to sense impressions.

Of course, as any careful student of your writings know, you do reinstate cause and effect by substituting for the supposed necessary link between them the habits that we form connecting earlier and later events, so that when we encounter an event of one type, we habitually expect an event of a second type to follow. In this way, you admit a notion of cause and effect even though, as you rightly state, we have no sense impression at all of a necessary link between them.

Similarly you point out that we have no sense impression of a continuously existing self, only a series of sense impressions, feelings, imaginations, and the like, none of which is permanent, that we group together for various purposes and call the ego or the self. But we have no sense impressions that are always present throughout the time that we assume our selves to persist.

I have to review these matters, of which you are most familiar, in being the most eminent proponent of such reasoning, so as to explain my innovations and departures from them.

You convinced me that our scientific knowledge must be closely tied to what we experience, and I do not depart from that. But I try to make some observations about what may be reasonably believed, though not scientifically known, and this lies beyond the realm of sense experience.

I also think I have discovered some interesting connections or analogies between experience and what may be reasonably believed.

A Priori Structures of the Mind5

My own account of contiguity and succession, as well as cause and effect, is somewhat different from yours.

I believe I have discovered two a priori forms of intuition, which are not rooted in sense data, but are added by the mind to unstructured sensation in order to create our knowledge that objects of sense are actually found at various points relative to space and time. In other words, space and time are not discovered in sense impressions themselves, but are added to the sense impressions by the mind. Without these a priori forms, we could not judge that this sensed object is closer to that sensed object, or that this sensed object came before that sensed object rather than their occurring simultaneously. What we experience would be more poorly ordered than the things about which we dream.

Besides a priori forms of intuition, there are also, I believe, a priori concepts of the understanding, which we also add to our sensations in order to produce our definite notions of the physical world. Among the a priori concepts of the understanding are:

Cause and Effect
Substance and Accident

I agree with you that these concepts are not found in sense impressions all by themselves. Where my own view differs from yours is that in my view the human mind imposes these a priori structures upon the sense impressions. Moreover, if the mind did not do this, we would have none of the knowledge that we do have of an organized world, with objects that persist over time, and events that are linked to one another as causes and effects.

My position is summed up in the slogan "concepts without percepts are empty, percepts without concepts are blind." Perhaps I should have said "percepts without a priori concepts are quite indefinite and disconnected."

Possible Beings Outside Experience

Only things within experience can be known, strictly speaking. Upon that I insist. But we can still believe in the existence of certain things that we cannot know. Things that lie outside experience that can be reasonably postulated include:

* The transcendental ego, or the unity of the a priori conditions of understanding with the a priori forms of intuition
* The Free Will
* The immortal soul
* God

But I am writing to you about the philosophy of morals. The transcendental ego is not directly relevant to that. God and the immortal soul are indirectly relevant to that. What is most important to morals is the Free Will.


Yet before I proceed to the philosophy of morals, I must refer to another point of agreement between us. You argue in Section VI of your Enquiry Concerning Human Understanding that liberty [moral freedom--J.G.] exists and is compatible with necessity [determinism--J.G.], once we understand determinism in terms of your philosophically refined notion of causal connection.

Now I understand your position on necessity to be similar to my own. In the realm of experience, which is firmly rooted on sense-impressions, events are organized according to necessity [i.e., deterministically].

Free Will: Postulate of Morality

Where I depart from you is that I maintain that we can make sense of the idea of the will as a free cause that is not located in the realm of experience. This is a cause that we cannot know, in the strict sense of "know," because we cannot have a perception of it. But we can believe in its existence. In fact, I claim that we must believe in it if we are to make sense of the moral practices in which we human beings operate and the responsibility human beings have for their own actions.

In other words, I believe that the a priori concepts, which I first discovered in my reflections on how we come to know objects scientifically, are also used by us in our ordinary reflections upon morals. The slightest reflection reveals the mind's application of the concept of cause in this area. In the area of theoretical understanding, that is, of experience, we relate cause and effect to the regular succession of events. But in the area of practical reason, of moral choice, we invoke the notion of an event that is not tied by necessity to preceding events. This is the notion of the free will.

I do not say that the free will has no relationship to laws, however. It has a relationship to a different kind of law, what I call the moral law.

The Unqualified Good6

It is my view that the only thing that can be conceived as good without qualification is the good will. Notice I did not say that I, or anybody else, know that there are wills or that I am one. However, this is something that everybody, even those who verbally deny it, must believe if he is to share in the practice of morality.

An unqualified good is a good that is always admirable, worthy of respect, and it is contrasted with qualified goods, which are sometimes admirable, sometimes not.

Fearlessness is admirable in a good person, but not in a scoundrel.

Fame is admirable, when it is found in a good person, but not in a despot.

Likewise power is admirable when, say, it is conjoined with the idea of perfect benevolence, but not when it is conjoined with tremendous malice, as in the idea of a malicious prince of demons.

Even happiness is not admirable, when we imagine it conjoined with a ruthless despot, who is not in the least haunted by his malicious deeds.

Action from Duty

What distinguishes the good will from other wills? My answer is that the good will is a will that acts from duty, that does as he or she should.

I distinguish action from duty from two other sorts of action:

Action contrary to duty. An example is that of a shopkeeper who cheats his customers. It is likely that he does so because he wants to avoid costs to which he is morally committed. He might conceivably do it because -out of excessive indulgence toward his wife--he wants to buy her pleasant things he cannot otherwise afford.

Action conforming to duty but from inclination. An example is that of a shopkeeper who keeps his contracts because he fears getting a bad reputation and losing customers or he fears being arrested as violating the law governing businesses like his own.

Action from duty (which also always conforms to duty). An example is that of a shopkeeper who keeps his contracts not because of fear of ill repute or loss of money or jail time, but because keeping his contracts conforms to duty. He acts out of respect for his duty, which is given by the moral law.

In these simple examples I have assumed that the moral duty coincides with the legal duty, although I am well aware that it does not always coincide.

Therefore, one who acts from duty acts from respect for the moral law.

Inclination in Hume's Account of Morals

The important contrast is the contrast between action from inclination and action from respect for the moral law.

If I properly understand your important Enquiry Concerning the Principles of Morals7, you trace all our moral judgments to our inclinations, which you group under two primary headings, self-love and benevolence. You associate the latter with what you call the sentiment of humanity, the capacity to place ourselves sympathetically in the position of other persons, not only those who are close to us and with whom we interact in our personal lives, but also those who are very remote from us.

If you were to say to me, "Well and good, Kant, but I do not have any impressions of any pure motive of respect for the moral law as such. Even such respect as I have for the positive law, the law established by lawgivers and monarchs, derives from my respect for justice, which I am convinced, as is everyone who reflects upon it, is of the greatest social utility. My identification with social utility rests in turn primarily upon my capacity for benevolence or my sense of humanity, which you classify with inclinations."

I would respond to you that you are exactly right: We cannot know with scientific certainty that we have such a motive as respect for the law as such. Yet, I claim that we do have the notion of a conceivable good will, and that linked unavoidably with that conceivable good will is the notion of acting not from inclination but from respect for the moral law.

The Moral Law and Reason

This will be clearer when I explain what I take the moral law to be. It cannot be the satisfaction of inclinations. It cannot be unqualifiedly good to promote the pleasure of others for the sake of this pleasure itself, because pleasure is not an unqualified good; nor can it be to promote the possession of power for others, for similar reasons.

We must find some other content for the moral law. And since it cannot be found in inclination, in our empirical selves, it must be found in that other part of our selves, our reason, in particular, our practical reason.

Hypothetical Imperatives

What, you ask, remains when all the empirical content of our practical reason is removed from it? Solely the idea of universality, say I. And in the area of moral conduct or praxis, the key idea is what we are commanded to do or if you like, what we ought to do. When we consider what we might be moved to do by inclinations, we are faced with a hypothetical imperative:

If you want to be rich, then you ought to invest in this or that kind of stock.

If you want to be happy, then you ought to act in accordance with justice and avoid offending your neighbors.

These are all hypothetical imperatives, they are not truly universal. They presuppose a inclinations that a person, as a purely rational being, need not have.

By contrast, an imperative that expresses pure practical reason, and is not tied to inclinations, would be a non-hypothetical imperative, addressed to all and sundry rational beings without qualification. It says "Do X," not "if you want Y, do X."

Such an imperative is not only addressed to every rational being but also takes universality as its content.

Maxims of conduct and the Categorical Imperative

Since it must do so by informing our conduct, there must be a way in which it can test the rules which we adopt in our daily lives. I am speaking here of maxims of conduct, such as "I shall always tell the truth" or "I shall always tell the truth unless it's in my interest not to do so."

As you can see from these examples of maxims, they are not universal. But our categorical imperative is universal: It states

Act only on those maxims thou couldst will to be a universal law.
I believe that in this formula I have discovered the moral law itself, which is addressed by pure reason to every embodied rational being, a rule which is addressed universally and whose very content is universality.

What's more, it supplies a test for the maxims that we use to govern our lives. If a maxim can actually be conceived as a universal rule, and if I can will-decide that I would wish --that it become a universal law, that is, a rule that everyone would live by, then it is my duty to live by it myself.

If I make my choices only by maxims that have passed this test, then my will is a good one.

Another Formulation of the Categorical Imperative

I do not have space to explain my further reasoning, and you are doubtless a bit fatigued by what has become a dissertation instead of a letter, but suffice it to say that there is another formulation for this categorical imperative, which I believe is logically equivalent to the first, but illuminating in a different light. It goes

Always act so as to treat humanity (by which I mean the practical reason of human beings) as an end, never as a means only.

This means that it is our supreme duty to act in ways that respect the practical rationality or free rational will of our fellow human beings, not only our own. We must never use others as a mere means to our ends.

Although we do not know that there is a free rational will in each and every one of us, the very practice of moral evaluation requires the postulate of such a capacity. If we are to judge one another as responsible for our actions-praising some and blaming others, we must posit that we all belong to the kingdom of ends.

Inclinations are of no moral worth, they have no dignity and do not merit our respect. The goodness of the will alone merits our unqualified respect, and the mere presence of the free will, even if it is only potentially good and not actually so, calls for the respect that is due rational beings as distinct from beasts.

Should your occupations permit you to craft responses to any of these reflections, I should be grateful to receive them.

Your humble fellow inquirer,

I. Kant

1. David Hume died August 25, 1776. So, even if he had received such a letter from Kant, it is unlikely that he would have had opportunity for a response. Which does not prevent us from imagining how he might have responded had he regained his health and lived another decade.

2. Immanuel Kant published his Critique of Pure Reason in 1781, his Fundamental Principles of the Metaphysics of Morals in 1785, and his Critique of Practical Judgment in 1787.

3. In Hume's Enquiry Concerning Human Understanding, published in 1748.

4. In Hume's Treatise of Human Nature, published in 1739 and 1740.

5. This section and the next are based on ideas in Kant's Critique of Pure Reason (1781).

6. This section and the next and the final four sections are based on Kant's Fundamental Principles of the Metaphysics of Morals.

7. This work was published in 1751.