Over the course of my lectures on ethics, I’ve been emphasizing moral dilemmas that contemporary philosophers have discussed, using consequentialism as a guiding theme, mentioning what consequentialism would say about those cases. Now I want to step back and talk about competing moral theories, and the alternatives to consequentialism.

The most commonly cited alternative to consequentialism is “deontological ethics,” which you can just remember as being “what Kant said.” However, Kant said several different things about ethics, and its not clear they’re all in harmony with each other. Kant’s moral ideas are called “categorial imperatives.” The most often cited one is probably “Always act according to that maxim whose universality as a law you can at the same time will” (That’s the Wikipedia translation).

Now, when I studied Kant back as an undergrad, the teacher was a philosopher named Keith Yandell. He really liked Kant. But what was his position on that particular “categorical imperative”? He insisted that Kant didn’t really mean it, the reason being that if you were really devious, you could come up with “maxims” that sanctioned all sorts of immoral behavior. Let me see if I can explain this problem.

First, what is a “maxim”? For Kant, it’s a sort of second-order moral principle, below the categorical imperative. Beyond that, not so clear. On the subject of lying, it seems Kant thought there were two possible maxims: “never lie,” and “it’s okay to lie when you feel like it,” or maybe even just “lie.” Kant claimed that the second option leads to contradiction, because it would destroy the institution of communication on which lying depends. But you could think up other maxims–maxims that let you lie in your particular case, but don’t lead to problems overall. Some of these tricks might be a bit cheesy–referencing you yourself, or anyone who happens to have your particular combination of first and last name, or someone dressed as you are. But if you thought carefully, you might be able to come up with much more subtle sanctions for your own lying. And this problem is the core of Keith Yandell’s objection to this part of Kant’s ethics.

Another trouble comes from the fact that Kant seems to have thought that the only way to identify bad rules is that they lead to actual contradictions when universalized according to the categorical imperative. John Stuart Mill, in his famous book /Utilitarianism/, stated the problem with this point better than I could:

This remarkable man [Kant], whose system of thought will long remain one of the landmarks in the history of philosophical speculation, does, in the treatise in question, lay down an universal first principle as the origin and ground of moral obligation; it is this:–’So act, that the rule on which thou actest would admit of being adopted as a law by all rational beings.’ But when he begins to deduce from this precept any of the actual duties of morality, he fails, almost grotesquely, to show that there would be any contradiction, any logical (not to say physical) impossibility, in the adoption by all rational beings of the most outrageously immoral rules of conduct. All he shows is that the consequences of their universal adoption would be such as no one would choose to incur.

The other major moral principle Kant put forth is that you always have to act with respect for persons–also, never treat another persons as a mere means, rather than an end. It’s not clear that the first part, about respecting persons, has any anti-consequentialist implications at all. It could be argued that to let five people die just because you don’t want to kill one person isn’t respectful to the five people. The part about not using people as a mere means does point vaguely towards an anti-consequentialist conclusion, though it runs into the problems that we saw with similar ideas in my first lecture on consequentialism. It also may or may not rule out cases of getting one person to benefit another that we would normally think harmless–again, the principle is a bit unclear.

So Kant’s views have some problems. There is, however, a theory which tries to incorporate the best parts of Kant and consequentialism, called “rule consequentialism.” The idea is, first, that traditional consequentialists have carelessly assumed we should focus on actions, rather than rules. Rule consequentialists will talk about traditional consequentialists as “act consequentialists.” In some ways, rule consequentialism is surprisingly similar to Kant’s deontological ethics, except that rule consequentialism explicitly says that the consequences matter, with everyone treated equally. That’s a barrier to gaming the rules for your own benefit, and addresses Mill’s criticism that Kant claimed not to care about consequences, but couldn’t derive any actual moral rules without reference to consequences.

J.J.C. Smart, in a defense of consequentialism written around the middle of the 20th century, argued that rule consequentialism amounted to rule-worship. Why follow a rule that doesn’t have the best consequences in your case, just because it would have good consequences in other cases? Rule consequentialists have a response to the criticism: a rule might have good consequences only when followed by a solid majority of people, but at the margins breaking the rule isn’t likely to have bad consequences.

Perhaps the most plausible instance of this claim is voting: it’s good to have a solid majority of the population voting, but one person’s vote is unlikely to change anything, no matter what the political slogans and chain e-mails tell you. The clearest example of this was in the 2000 Bush vs. Gore election, when I was 13. Presidential elections, as you’re all hopefully aware by this point in your academic careers, aren’t directly decided by your votes, but rather are directly decided by the electoral college and only indirectly decided by your votes for electors. The situation outside of Florida was such that whoever got Florida’s electors would win in the electoral college. And Florida’s system is that whoever got the most votes there, even by one vote, got all the electors. Thus, it became very important to know exactly how many votes each side had. *Exactly how many.* Tensions ran high, but when you cut through all the partisan rhetoric, what we learned was that our vote counting systems just aren’t that accurate. The margin of error was in the hundreds of votes.

In other words, we learned that a single vote doesn’t count.

Looking at this situation from the point of view of an act consequentialist, you’d see that you shouldn’t vote. Since a single vote doesn’t count, and not voting saves you time, and doesn’t really affect anyone in any other way, not voting will have the best consequences overall. Therefore, act consequentialists shouldn’t vote. Rule consequentialism, however, accounts for why we should vote, because it would be good if everyone voted.

Act consequentialists, though, have a response to this argument: it commits the mistake of thinking that by acting in accordance with a rule, we can magically get everyone else to follow it as well. To the rule consequentialist’s question of “what if everyone did this?” the act consequentialist can reply “they won’t.” Voting won’t cause everyone else to vote; not voting won’t cause everyone else to not vote.

This becomes even starker when you consider the problem of resisting tyrants. If everyone always resisted tyrants, without regard for their own safety, tyranny would quickly become impossible. That doesn’t mean you, individually should run out and get yourself killed whenever you find yourself living under a tyrant. That won’t do anyone any good. Randomly resisting won’t magically cause everyone else to do the same, if you want to be really effective you have to be willing to plot with greater discretion than you’d have to if a random act of resistance could automatically spark a general rebellion.

Actually, the problem is worse than that, because if we try to ask what would happen if everyone followed a proposed rule, the rule need say nothing about how to respond to rule breakers, including tyrants, because we’re asking what would happen if such people didn’t exist. So in its strictest application, rule consequentialism fails to tell us what to do in situations we regularly find ourselves in. As a historical aside, it should be noted that people who’ve tried to follow through rule-based reasoning have gotten some bizarre results. Kant, for example, claimed that you must absolutely never lie, even to someone who’s trying to kill someone else and wants to know where their intended victim is hiding. So according to Kant, if you were hiding Jews from the Nazis during WWII, you would actually have to confess this to them readily.

At the end of the day, I think it’s sensible to treat act consequentialism as the most coherent form of consequentialism, because our acts are ultimately what we have control over. Trying to avoid this leads to the problems I’ve discussed: rule-worship, imagining we can set the principles by which everyone acts, and so on. Further, I think a lot of what theories like rule consequentialism are supposed to accomplish.

For example, Brad Hooker has suggested we include a clause in the formulation of rule consequentialism saying that you should take into account the cost of establishing a rule in society. The problem here is that trying to get a rule generally accepted is a type of action, and is therefore something that traditional, or act consequentialism will recommend in some circumstances. This may seem odd, but it’s a confusion to think that there’s some logical necessity requiring us to always inculcate people with correct moral principles, write them into law, etc. Claiming sound morality requires that is a substantial moral claim, as open to challenge as any other moral claim.

There are some cases where it’s very plausible that we shouldn’t try to inculcate a perfect moral code, as given human fallibility people are likely to misapply it. Rather, we should promote whatever moral principles will do the most good. Remember the case of torture and the ticking time bomb? It’s plausible to think there are hypothetical situations where torture would be the right thing to do, but as a matter of fact, we can’t actually trust the government with the power to torture people. This kind of reasoning can be applied even to our own mental habits.

And as for voting: it can make sense to encourage others to vote, to provide them with incentives to do so and give them grief for not voting, even though without any such incentives there would be no good reason to vote.

I’m going to try to summarize what I think are the strengths and weaknesses of consequentialism, as they’ve played out over these last four lectures. First, it seems that the consequences of our actions matter. Second, it seems that what we have the most immediate control over is our actions, not the rules that all people must operate on. Third, some of the distinctions that anti-consequentialists want to make run into difficulty on close examination. Fourth, some of the things that non-consequentialist theories are supposed to provide can be provided by consequentialism, once you recognize the range of things that can be considered actions–such as promoting rules.

None of this, though, changes the fact that consequentialism has some highly counter-intuitive implications. Normally, we don’t think we are required to kill one person to save five, or to be as concerned for the welfare of strangers as we are for our own. It’s not clear exactly how we should deal with these situations in philosophy–where an otherwise attractive theory tells us our commonsense beliefs are wrong. We want to be able to find errors in our moral thinking, but we don’t want to deny obvious moral truths.

One response philosophers have made to this debate is to give up the search for all-encompasing moral principles, like consequentialism and Kant’s deontology, and do ethics piecemeal. This is sometimes turned into a theory onto itself, and called “moral particularism” or some such. My own thought, though, which I hold for reasons I can’t fully explain here, is that we should be skeptical of grand philosophical theorizing, and just try to get judgments on down-to-earth questions–like whether we should legalize torture of terror suspects, given the actual circumstances we find ourselves in, as opposed to hypothetical ones–get those questions right. We may not be in a position to get the grand theorizing right.