I know this question has been widely asked, and that the answer may not be as straightforward as the question, which is partly why I'm asking. It's been a long time since I sat in a philosophy class, but I'm seeing Kant's name pop up in relation to human-AI value alignment, which is one of my preoccupations.

  • What are some examples of categorical imperatives?

By "categorical imperatives" one can mean universalized maxims, but I'm also interested in alternative formulations in the context of modern ethics.

  • 2
    In the context of Kant, there is only one Categorical Imperative, although it has at least six equivalent forms. All other imperatives are contingent and all other categorical notions are indicative. Are you intending to ask for alternative statements of the Categorical Imperative, or are you asking for examples of universalized maxims to which the Categorical Imperative would apply?
    – user9166
    Commented Nov 29, 2017 at 21:52
  • 1
    The parenthesized part is really another question, and it is much easier to answer, since someone with the right book at hand (not me) can just pull them out of the table of contents. Kant performs a big circular proof that four of them are equivalent using two extra ones as lemmas. You might want to just split that off and ask it as a matter of reference.
    – user9166
    Commented Nov 29, 2017 at 22:16
  • 1
    On Kant's definition, hypothetical imperatives are instructions for achieving a goal, categorical imperatives are absolute commands (Kant has a convoluted argument that there can be only one categorical imperative, but it is not very convincing today, just as his argument for synthetic a priori). In Christianity, say, the ten commandments will be categorical imperatives, and similarly in other religious ethics. Outside of religion modern ethics moved away from Kant's absolute imperatives just as modern epistemology moved away from his absolute a priori.
    – Conifold
    Commented Nov 29, 2017 at 22:22
  • 1
    They certainly are, even if the categorical imperative is a synthetic a priori. There are reasons for rejecting the imperative aside from the general reasons for relativizing a priori, or for rejecting its relativized form.
    – Conifold
    Commented Nov 30, 2017 at 0:38
  • 1
    @Conifold. Thanks for the clarification. I was asking the question from Kant's standpoint. You are taking an external view, which is fine. Of course, from Kant's standpoint the 'even if' hardly applies since it is plain that the CI is synthetic a priori : 'this categorical ought represents a synthetic proposition a priori' (Grundlegung : III, 454. Excuse my quoting in English from Korsgaard's Cambridge ed.; I do not have the German text immediately to hand.
    – Geoffrey Thomas
    Commented Nov 30, 2017 at 13:07

2 Answers 2


The terminology of hypothetical and categorical imperatives is rather specific to Kant. Roughly, hypothetical imperatives give commands conditioned on one’s purposes (if you wish to succeed in life study hard, etc.), while categorical imperatives are unconditional, absolute. The problem with authentic examples is that according to Kant “There is therefore only a single categorical imperative, and it is this: act only according to that maxim through which you can at the same time will that it become a universal law”. But Kant clearly does not distinguish between ethics and meta-ethics (principles for selecting ethical principles), a distinction to which modern philosophy became sensitivized by developments in logic and semantics in the early 20th century, and his “single categorical imperative” is a meta-ethical principle. Using the term loosely, all irreducible absolute commands of ethics proper are “categorical imperatives”. For example, so are the ten commandments of Judaism and Christianity, and their analogs in other religions. Kant’s implication is that all such absolute commands reduce to his universalizable maxims, and only they do. His reasoning to that effect is long-winded, our Philip Klöcking does his best to explain it here.

The ideas of absolute commands and of objective morality, as well as their identification with universalizable maxims, came under heavy fire in modern philosophy. The former point is obvious in the context of postmodernist relativization of epistemological, ethical, etc., principles to history and culture, so I will focus on the latter. MacIntyre argued that many dubious maxims are very well universalizable: "keep all your promises except one", "let everyone except me be treated as a means" (he denies that Kant’s derivation of the “formula of humanity” from “the” categorical imperative is valid). Anscombe and Mackie pointed out that the idea of objective “oughts” is implicitly based on the idea of God as the lawgiver, and does not make much sense without it, as Hume’s critique of is-ought derivations shows, see virtue ethics.

Moreover, the “categorical imperatives” in ethics, including “do not kill”, invariably lead to “wrong” conclusions resulting in growing lists of exceptions or sophistic distinctions like the Thomist doctrine of double effect. The latter attempts to distinguish between foreseeable harmful side effects of bringing about a good end, and harmful means of bringing about the same good end. The distinction turns on whether one merely foresees killing 1 person in the process of saving 5, or intends to kill the said 1 to save the said 5. Many see this as a distinction without a difference, along with its sister distinction between acts and omissions. Mackie in Ethics characterizes both as artificial, and explains how moral dilemmas undermine the idea of absolute commands in general:

The main reason why it has been thought important by Catholic moralists is that it seems necessary if there are to be absolute moral rules, for example ones which forbid murder, adultery, or apostasy in any circumstances. If an agent is equally responsible for all the foreseen consequences of an immediate action (or failure to act) as well as for that action itself, there will be conflict cases where different rules, or even different applications of the' same rule, require incompatible actions, so that the rule (or rules) cannot (both) be absolute.

[…] There is, indeed, one other way in which absolute prohibitions could be maintained… we could distinguish positive acts from omissions and frame absolute rules only about positive acts. No conflict cases could then arise, because in any conceivable set of circumstances all prohibitions of kinds of positive act (including the bringing about of certain evils as known second effects) could be obeyed at once by complete inaction. But do we want to maintain absolutism in either of these ways?

The problem of causal responsibility is complex even aside from the issues of double effect, see Is it a logical flaw to blame someone for an event if they were simply its causal factor?, and is often approached by analyzing “intentions”. But Anscombe offered a general objection to ethical theories, which, like Kant’s, judge not merely actions, but actions combined with intentions. It is known as the problem of relevant descriptions. The same action may be intentional under one description but not under another (in Anscombe’s example under “pumping water” but not under “contracting muscles”), this is not unlike the interpretational sophistry of double effect. For Kant’s universalization scheme to work at all we should be able to ascribe intentions to others, and he apparently did not appreciate how problematic that is. Here is from Shumski’s The Problem of Relevant Descriptions and the Scope of Moral Principles:

In her seminal attack on modern moral philosophy, G. E. M. Anscombe claims that Kant's ‘rule about universalizable maxims is useless without stipulations as to what shall count as a relevant description of an action with a view to constructing a maxim about it’. Although this so-called problem of relevant descriptions has received considerable attention in the literature, there is little agreement on how it should be understood or solved… I argue that the problem consists in the fact that Kant's formula of universal law seems to stand in need of an account of moral sensibility that does not render the formula superfluous. But, as my discussion of existing solutions reveals, there can be no such account.

This is not to say that universalizability, at least in a weak sense of abstracting from individual traits, do not have a place in modern ethics, see Do “if everybody did it” arguments commit a fallacy? Mackie in Ethics calls universalizability an "in some sense indisputable" condition on moral systems. And he endorses "categorical imperatives", as unconditional rules of behavior imposed by a society or institution for pragmatic reasons, e.g. facilitating decision making under predictive uncertainty or ensuring better outcomes in the prisoner's dilemma situations.

  • I'm wondering why "You must worship only God and no one else" commandment (which is among 10) is CI then. It does not follow from CI, because the opposite commandment also is universalizable.
    – rus9384
    Commented Sep 3, 2018 at 15:49

Since it has an 'only' in it, the normal form of the Categorical Imperative lends itself most convincingly to proofs by contradiction, so he mostly proposes a number of maxims that do not universalize.

Most famously "Lie." fails to universalize because it is not possible for everyone to lie all the time. If you told only lies, you would not be able to establish the truth, in order to make sure you were actually lying. So if you made this truly universal, people would forget how to do it.

Others of these include "Kill people." "Make threats." "Undermine other cultures."

The example he chooses to open with is one of the few positive cases: "Increase your holdings through any safe means."

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .