Discussion of consequentialism has tended to be dominated by the popularity of utilitarianism, so that if you find a consequentialist, they will almost always be either a utilitarian or someone who is heavily influenced by utilitarians. But deontology has several very popular, and very different versions, and we need a basic way to classify the most important varieties. Since deontology is by definition based on a theory of obligations, the best classification will focus on this. There are several ways to classify theories of obligations, but one of the more useful starts with a basic recognition: some obligations are made, and only exist because they are made. For instance, promising, making a contract, and legislating, are all ways in which we sometimes create obligations that did not exist before, and that only exist because we made them. An unmade promise doesn't obligate; you are bound only to contracts that have actually been made; laws have no force if they are not passed. If some obligations exist only because they are made, this leads to a natural question: Are all obligations like this? There are two obvious answers you could give to this question: Yes and No. And deontologists do in fact divide on this question, and in ways that make their different answers to it important for their approaches as a whole. There is no standardized terminology for these two groups, but I will call the Yes-sayers 'moral positivists' and the No-sayers 'moral naturalists'.
Moral positivism has at least one very obvious advantage, namely, that moral positivists can give you a unified theory of obligation. All obligations can have the same basic explanation,and it will always have the same basic structure: you get the obligation when the obligation-maker makes it. Different moral positivisms will propose different criteria for who counts as an obligation-maker, or different accounts of how the obligation gets made, but if every obligation is made, every obligation is explained by its obligation-maker. If you have questions about whether a particular obligation exists, you just need to look to see whether there is reason to think someone made it; if you want to understand any obligation better, look at what its maker was doing in making it. That's very nice, and makes reasoning about obligations in some ways very straightforward.
We do run into a problem, however, that a moral positivist certainly must address. It's obvious that rules governing the state (state laws) are made by the state legislature; it's also obvious that the state legislature can only make laws for its own state. It's obvious that city ordinances are made by the city, and don't extend beyond the city. It's obvious that some obligations derive from contracts, and also that the contract only binds those who are parties to it. But where is the obligation-maker for moral laws, especially given that moral laws are generally taken to apply very widely and to trump other laws? Murder is morally wrong even if no state or federal law is passed against it; laws passed by legislatures can be morally wrong and corrupt; we expect people to follow basic moral rules quite generally. Where is the obligation-maker capable of making obligations that apply to human beings generally?
Two answers to this question are especially popular. The first is that society is in some way the general obligation-maker. Society as a whole could certainly give moral obligation a broader basis of authority than other laws. An obvious question is how society would do this; there are a number of possible answers, of which custom and some kind of contractual process are probably the two most popular. There is no standardized name for this group of ethical approaches; we could perhaps call them something like 'social positivism'. One obvious advantage social positivism has is that, in principle, you could discover moral obligations just by investigating the society you are considering. An obvious disadvantage is that human beings are often members of several societies simultaneously, and it is unclear how you would take this into account. A further question is what to do about societies that we usually think of as very bad in some way -- we usually take some societies to be wrong in some of the obligations they seem to be making (e.g., slavery), so we'd need to know how to handle such cases. (This is one reason why some social positivists prefer a contract-based approach rather than a custom-based approach; making a contract seems a more structured process, so can perhaps help by limiting the kinds of things that are allowed.)
But people often have the idea that moral obligations are much stronger than social positivism can provide. So what possible obligation-maker can give you a stronger foundation for authority than society as a whole? Here enters the other popular moral positivism: divine command deontology (often more loosely called 'divine command theory', although strictly speaking you could have a divine command theory of obligation without being a deontologist). If you want an obligation that is as strong as possible, it seems it would have to be universal, inescapable, and in some way prior to all other obligations. If that's what you think the fundamental moral obligations should be like, then the obvious candidate for an obligation-maker who could provide this is God. An authority with absolutely universal jurisdiction, who has knowledge and power such that obligations coming from it cannot be evaded, and who is prior to and independent of every other authority certainly sounds like a divine authority. There is a lot to recommend this as a theory of obligation -- as a moral positivism, divine command deontology gives us a completely unified theory of obligation in which all obligations have a definite cause, which is useful, and by taking God to be the cause, we manage to keep this advantage while establishing moral obligations as vastly more authoritative than obligations we human beings make, thus lending the weight to morality that many people think it has. It is a very elegant solution.
There are many very bad objections to divine command deontology. (For instance, a common one is that it's repulsive to say that God could make bad things obligations. But nothing about divine command deontology requires that its supporters have to accept the claim that God could make anything an obligation, and most don't; all it requires is that if anything is an obligation, it is so because God made it so.) But there are obvious questions. One disadvantage that divine command deontology has in comparison to social positivism is that while social positivism in principle always gives you a way to find out what your obligations are (by looking at what societies actually do), almost everyone faced with a divine command deontology immediately asks the question, "How do we know what God has made an obligation?" It doesn't seem that you can just look and see. Different divine command deontologists will have different answers to the question, but it's a good example of how an elegant solution to a problem can at the same time raise new questions that have to be answered.
While divine command deontologists and social positivists do say some interesting things about ethical reasoning occasionally, it tends not to be a major component of their approach -- on moral positivist views in general, if you want to find out what our obligations are, it's usually better to go to the authority that makes the obligations rather than to try to reason them out on your own. So it's inevitable that any approach to ethics that focuses on ethical reasoning is going to spend a bit more time with moral naturalism.
Moral naturalists reject the view that all obligations are made; some fundamental obligations are natural. There are different ways you could do this. You could, for instance, hold that our basic moral obligations are biological instincts; we all have to accept them because it's part of our nature to accept them, although this is due wholly to our biological history. There are probably other answers you could give. But far and away a more commonly accepted answer is that moral obligations are fundamental rational principles, that they are based in reasoning and are necessary in a way somewhat analogous to basic mathematical principles. And far and away the most influential moral naturalism is what we call Kantianism, which brings us to Immanuel Kant.
Engraving of Kant by Johann Friedrich Bause; for discussion of various contemporary likenesses of Kant, see Steve Naragon's web resources on the subject.
Immanuel Kant (1724-1804) spent most of his life in Königsberg, Prussia (modern-day Kaliningrad, Russia); his family were hard-working Lutheran Pietists, Pietism being a movement that emphasized personal conversion, devout study of the Bible, and practical pursuit of holiness. At the University of Königsberg, he studied a wide range of philosophers, and then afterward became a tutor for a few years. He eventually returned to the University of Königsberg to teach as a professor for forty years. For the first fifteen years, Kant did not receive a salary from a University; he was paid directly by students, so he spent a very large amount of time each week lecturing on different subjects in order to get enough students to pay his bills (he also at one point began to supplement his salary with a job as sublibrarian). Fortunately for him, he turned out to have considerable talent for lecturing and was very popular. He was eventually appointed chair of logic and metaphysics, and therefore entered a more stable period. It wasn't until the 1780s, though, that he began publishing the works for which he is best known, including his major works on ethics, Groundwork for the Metaphysics of Morals and Critique of Practical Reason.
Given Kant's background, it's not surprising that people first coming to Kant find a very academic style of writing, and Kant is famously difficult for students just starting to look at questions of ethical reasoning. But it's easy to misdiagnose why he is difficult. Our first instinct is to say that it must be because he is very complicated. I want to suggest that the first instinct here is wrong. Kant is possessed by an idea of morality that is not complicated by simple. Everything in Kant's ethics is about this idea: preparing to understand it, considering it in several different lights so as to understand it correctly, applying it. Kant's ethics is just the unfolding of a simple idea. If we have difficulty with it, it is because we are complicated. Faced with Kantian simplicity, our natural tendency is to add qualifications, exceptions, room for excuses, things that can change or that depend on circumstances. But precisely Kant's point is that if you do this, you are trying to make morality be something it is not.
Kant thinks that even if we just look at our ordinary, everyday view of morality, it becomes clear that it has to be in some way necessary and unconditional. Most of the things we call good are only conditionally good. For instance, intelligence is good if you use it properly. But the only thing we usually call good that can never be misused at all is a will that is good in the willing itself. That is, in ordinary terms, the good that can't be misused is doing one's duty because it is one's duty. (We can do our duty for other reasons, of course, but we recognize that this is defective. For instance, when hypocrites do what is right because they get some benefit from it, we still criticize them; it is doing the right thing for the wrong reason.) And the purer the case, the more obvious this is: we recognize moral action most clearly when the person does what is right even if they are going to be hurt or even killed because of it. To do one's duty because it is one's duty is, in a particular area of life, to will in conformity with a standard that makes the will unconditionally good; this standard can only be an unconditional standard. It has to be a universal law. But this means that it cannot depend on any circumstances, any consequences, anything changeable or uncertain like feelings or experience. It cannot have exceptions, cannot allow excuses, and cannot depend on our own preference, convenience, or expectations.
This is, of course, what I meant by saying that while Kant is simple, people keep trying to add complications. We like to make excuses for ourself. We like to find reasons why we, special little people, do not need to act according to some moral principle even though other people do. But in moral law there are no excuses and no exceptions. If moral life is too hard, we like to try to redefine it so it's easier; indeed, we have a bad habit of trying to define ourselves into being good people. But moral law just is what it is; you don't get to put that chin-up bar right under your chin. Moral law doesn't have to fit your standards; you have to fit the standard of moral law. You can't get it from your feelings, which change from time to time and from person to person; you can't get it from your experiences; you can't get it from examples; you can't get it from society. What you can get from these things may be good if used the right way; but everything that comes from these things can be misused. There are conditions attached. But moral obligation is unconditional. The way Kant puts it is that morality must be a categorical imperative, not a hypothetical imperative. A hypothetical imperative is a rule that comes with conditions (even if they are not specifically stated): Do this if X. It only identifies something as good for some possible or actual purpose. A categorical imperative is unconditional: Do this. There can't even be any assumed or implicit conditions. No further purpose needs to be considered.
What kind of law can be unconditional? What kind of imperative can be categorical? It can't have any content that depends on particular circumstances. It has to be the kind of law that is universal, applying to any and all possible rational beings, without exception, in every situation, without exception. So the content can't include anything that is not always available to every rational being making a choice. So what is going on when a rational being is making a choice? Kant thinks this involves making a rule for oneself. This rule that we make for ourselves in a choice is what Kant calls a maxim. Every maxim has a means-end structure. It identifies something valuable (the end) that provides a reason to act, and it also identifies a way of acting (the means) appropriate to the value of that thing that provides a reason to act. The maxim, then, captures what kind of thing we are intending to do and why we are intending to do it. The moral law is going to have be a standard for maxims, so it will have to tell us the kind of maxim that we are allowed to have; and in particular, the only maxims that we are allowed to have are those that can be willed unconditionally. And this gives us the categorical imperative, which effectively says that the moral law is to will in accordance with the moral law. As Kant says in the Groundwork for the Metaphysics of Morals (pp. 51-52 [4:420-421]):
If I conceive, however, a categorical imperative, then I know at once what it contains. For since the imperative contains besides the law only the necessity of the maxim*) to be in conformity with this law, the law, however, contains no condition to which it was limited, in this way nothing but the universality of a law in general remains over to which the maxim of the action is to be in conformance, and which conformity alone the imperative properly represents as necessary.
The categorical imperative is thus only a single and indeed this: act only according to that maxim, through which you at the same time can will, that it becomes a universal law.
Kant's view has the implication that whenever you are doing something wrong, you are in some way contradicting yourself, either directly or indirectly: you are, as a rational being, recognizing a universal standard, but you are also trying to excuse yourself from that standard. As he puts it (pp. 57-58 [4:424]):
If we now pay attention to ourselves during each transgression of a duty, then we find that we actually do not will that our maxim should become a universal law, for that is for us impossible, but the opposite of it should instead generally remain a law; only we ourselves take the liberty to make for ourselves or (even only for this time) to the advantage of our inclination an exception to it. Consequently, if we weighed everything from one and the same point of view, namely of reason, then we would find a contradiction in our own will, namely, that a certain principle be objectively necessary as a universal law and yet subjectively not hold universally, but should permit exceptions.
A moral maxim will have no excuses; it just is holding yourself to a standard that can be a standard for every rational being. If you want to know whether your maxim is a good one, you just take it and 'universalize' it; if you get a contradiction, either directly (what you are trying to do is itself incoherent) or indirectly (what you are trying to do is not the sort of thing you can always will as a standard), then it is bad. If it is 'universalizable', it is good.
In a sense, that's it. That's Kantian ethics: Act only according to that maxim through which you can at the same time will that it becomes universal law. It is the only possible categorical imperative, so all of moral life is based on this rule, this one moral precept, that can be put in one sentence. In practice, Kant recognizes that this is quite abstract and not always very easy to apply to real situations. It's also not very clear on its own how it relates to the ways we usually talk about moral choices. We now know what we are talking about, in abstract terms, when we talk about right and wrong, but how does it actually work in practice?
There is only one categorical imperative, but that doesn't meant that we can't rephrase it so that it might be easier to use. So Kant gives us three specific formulations of the categorical imperative. These formulations are the categorical imperative put into a different vocabulary, emphasizing different things in the imperative by using different kinds of analogy, so that it can more easily be related to certain aspects of our experience and to our usual ways of talking about moral choices. Let's take the above statement of the categorical imperative and mark the most important words with an asterisk:
Act only according to that *maxim, through which you at the same time can will, that it becomes a *universal *law.
These three seem to be the most important: a moral maxim is a universal-law maxim. So let's take each in turn, and find a way of stating the categorical imperative that emphasizes that word or idea.
(1) The Law of Nature Formulation. Let's start with universality. When we look around for other things that we generally consider to be universal or conditional, we find an obvious case in laws of nature. The law of gravity is not conditional on who you are, nor does it care what excuses you make. So if we wanted to get a better grasp on how a maxim can be willed as universal, we can make an an analogy to laws of nature; in a sense, just as we blow up a picture if we want to see more details, so also, if we want to see whether our maxim can be consistently willed as universal, we can blow it up to the size of the universe. That will certainly capture universality. Thus we get a restatement of the categorical imperative:
Act as if the maxim of your action were to become through your will a universal law of nature.
You can think of your moral choices as making the kind of world you want to live in. So imagine that this is literally true: when you make a choice, that becomes a law of nature. If you break a promise in order to get something you want, that maxim, 'Break a promise in order to get something you want', becomes a law of nature, so that every time anyone, anywhere, wants something that they can get by breaking a promise, they do. This seems to be an incoherent world: we're describing it as if there were real promises (otherwise they wouldn't need to be broken), but in reality if everyone automatically breaks a promise when it's convenient for getting something, just like everyone falls if they walk off a ledge, that's as good as to say that promises don't mean anything, and thus that there are no real promises. At the size of the universe see the problem: you are treating a promise as something that matters morally, as something that should be kept; but you are also treating promises as if you aren't really held to that standard. You're trying directly to will a contradiction.
On the other hand, sometimes things are more indirect. If you refuse to help someone in need because you want to stay comfortable, it doesn't seem like we have a direct contradiction. But we do have an indirect one. Maybe a universe can exist in which it's a law of nature that nobody helps anyone if they can stay comfortable that way, but given the limitations of human nature, no one can guarantee that they will never need help, so you can't always will that this be the way the world works. What you're willing is not inconsistent, but you aren't able to will it consistently. This also shows that you are trying to make yourself a special exception from universal law.
(2) The End in Itself Formulation. A maxim, I said, has a means-end structure: it identifies something of value (the end), and tells you what to do to treat as valuable (means). But a moral maxim, to fit the unconditional law, is going to have to take as its end something that is of absolute worth or value (something that is an end not because it is the means to some other, more important end, but because it is an end in itself. But that tells us that it's going to have to be an end that is always available, so to speak, something that can be treated as valuable no matter what. It can't be something to be used; and it can't be something that we would have to work to get the value from. There is nothing that can fulfill this function, Kant thinks, except rational nature itself -- which of course is always available to a rational being as something to be valued. Rational nature ('humanity', as we often call it) marks out rational beings as persons not things, as ends in themselves. So we get the next way of stating the categorical imperative:
Act in this way, that you use humanity in your own person, as well as in the person of every other, always at the same time as an end, never merely as a means.
As with the Law of Nature formulation, this can show up a direct contradiction (you are treating people as if they were merely a means to get what you want) or an indirect contradiction (you are not treating someone in a way that is appropriate to treating them as an end).
(3) The Kingdom of Ends Formulation. The third important idea, law, requires a bit more maneuvering. Our will must act so that something can be law. We might call this 'legislation', although it's not quite the same as what we usually think of as legislation. Thus the categorical imperative tells us that morality is about our will legislating universally through its own maxims. This Kant calls autonomy; the opposite is heteronomy, when the will treats as its standard something it gets from somewhere other than itself. Now, the place we most often think about legislating is in discussions of politics or society. If we use the word 'kingdom' to describe a system of rational beings united by a common law, then in a society that worked on moral laws everyone would be treated as ends in themselves. This is the kingdom of ends. It doesn't exist, but in a sense our will makes it possible: you can choose to make laws like a legislator in the kingdom of ends, in the moral society. (And of course, given what we just said, everyone in the kingdom of ends is the legislator of the kingdom of ends, and everyone takes themselves to be subject to the laws, without question. It is literally a society governed by rule of law and not men.) If the categorical imperative gives us the idea of autonomy, we can picture this idea in terms of the kingdom of ends. It is a society in which every member has dignity, a worth beyond all price, because no one is treated as a mere means to something else. Thus we have the third formulation:
Act according to the maxim of a universally legislating member of a merely possible kingdom of ends.
If the Law of Nature formulation investigates the maxim by expanding it to the size of the world, the Kingdom of Ends formulation investigates it by expanding it to the size of a society.
It's easy enough to see how these formulations might be useful in different ways at different times, depending on whether the situation is more abstract (Law of Nature), concerned with our relationships with another (End in Itself), or concerned with society as a whole (Kingdom of Ends). It's also the case that they connect to different parts of our moral vocabulary: making a moral world (Law of Nature), treating people as not just to be used (End in Itself), building a moral society (Kingdom of Ends). It's probably not accidental that each of these could be seen as a secularized analogue of a theological concept, whether that be providence, or the image of God in people, or the messianic community or New Jerusalem.
One way Kant puts the relationship among the three is that the Law of Nature formulation focuses on the form of morality (its structure), the End in Itself formulation focuses on the matter of morality (its content), and the Kingdom of Ends formulation focuses on the totality of morality (how its structure organizes its content, since law is what links universality and the maxim). Probably related to this is that Kant talks about the Law of Nature formulation as if it gave us the possibility of morality and the End in Itself formulation as if it gave us the actuality of morality. But they are all one categorical imperative, since in every case you are to act according to the maxim that can be willed as universal law; it's just the emphases that are different.
The formulations help bring the categorical imperative down to earth a bit. But it's still a strict standard. Can we even act according to it? Obviously Kant would say that moral law doesn't depend on whether we can act according to it -- it doesn't depend on us at all, because that would make it conditional. But it would be very unfortunate if we were morally obligated and had little chance of fulfilling our obligation. There's a sense in which the categorical imperative is the whole of Kant's ethics. But if we stop there, we can also be easily misled. Kant is, I think, very aware of how high a standard it establishes, and does explore, to a certain extent resources that we have available for being moral. Moral law doesn't depend on human nature, but human nature is a nature that takes into account moral law in a number of ways, and in practical terms, even if not in absolute terms, these can be quite important. But that is another set of topics for another post.