A man walks into a crowded room with a loaded machine gun. He presses the trigger and sweeps the gun back and forth, thereby killing a dozen people and injuring more. Nearly everyone would admit that what the man did was wrong. Why was it wrong to pull that trigger and move the gun around like that in that room? I'll tell you why. It was because of the consequences of his action. It is not that pushing an index finger against a piece of metal is intrinsically wrong. That action is wrong here because of the consequences or results of the action. That seems to be obvious. Any ethical theory which holds that this is the sole criterion of the rightness or wrongness of action, that whether an action is right or wrong depends ultimately on consequences of actions, is a consequentialist ethical theory.

        Now some people may even reserve judgment about my machine gun case and say that there may be rare, special circumstances that render that action permissible. But these people too are likely to give consequentialist arguments. For example they might say the people killed were a team of terrorists planning to detonate nuclear bombs in the world's most populous cities. So the argument would be that to kill those dozen people may be doing much more good than harm and thus (and here is the consequentialist reasoning) it may be the right thing to do. Nearly everyone agrees that the specific circumstances matter to the whether some action is morally right or morally wrong. Consequentialism has a way of accounting for why circumstances matter, why the details can be important, namely that what the consequences of some action are might vary significantly with the circumstances in which the action is performed. Whether it is right to serve someone a drink may well depend on whether the drink is likely to poison or refresh the person.

        Of course just saying that the moral rightness of an action is a function of the consequences of actions is not to say anything definite about what actions are right (or wrong). To provide an actual ethical theory with a determination of what is right and wrong, one would first have to say what features of the consequences of actions matter, which, that is to say, are morally relevant. Second, one would have to specify just what the function is that determines for each set of consequences with those features, whether actions having those consequences are morally right, morally wrong, or neither. Among consequentialists there are divisions on both these matters. Accordingly, there are many different possible consequentialist ethical theories. That is something of an understatement. It is not hard to specify hundreds of thousands of different consequentialist ethical theories. These theories provide different specific answers to the question "what actions are right," but they agree in where to look to find the answers, namely in the causal results of actions. And these are just the pure consequentialist theories, those that say that the rightness or wrongness of an action is determined entirely by what the consequences of the action are or would be. There are also mixed theories which say that an appeal to the consequences is part of what grounds an ethical theory but not the whole of it.

Utilitarian and Non-utilitarian Consequentialist Theories

        The prime example of a purely consequentialist ethical theory is utilitarianism. "Utilitarianism" does not denote simply one specific theory, but a large family of theories. In the words of one of its principal proponents, John Stuart Mill, utilitarianism is the view that "actions are right in proportion as they tend to promote happiness, wrong as they tend to produce the reverse of happiness."¹ This is consequentialist, since the rightness of actions is said to depend on what those actions "promote," i.e. what results, or is likely to result, from the actions. And what features of the consequences are relevant? All that is relevant, on Mill's version of utilitarianism, is what happiness and unhappiness is brought about. What on Mill's theory is the function that relates this feature of the consequences of action to rightness? It seems to be this: the more happiness that is brought about by the action, the more right it is. (Notice that Mill talks about "proportion," which suggests that he thinks rightness and wrongness are a matter of degree, a point not all utilitarians agree with.)

        But utilitarian ethical theories do not exhaust the set of consequentialist ethical theories. While it is difficult to specify exactly what must be true of a consequentialist ethical theory to count as a "utilitarian" theory (mainly because the term is used in different ways by different philosophers), clearly there are some consequentialist theories that are not utilitarian. For example, a theory that said that an action is right if and only if it leads to an increase in the amount of diversity in the world would not count as a utilitarian theory, but it would be a consequentialist one.

        Interlude on Types of Value

        Before talking more about types of consequentialism, something must be said about goodness. This is because one of the main ways in which consequentialist views differ from one another is in their accounts of what is ultimately good. We have to talk about what is ultimately good, or, as some philosophers would say, of what has "intrinsic value," because lots of things have value instrumentally (which we call "extrinsic value") because they lead to, or otherwise support, things which have intrinsic value. So, for instance, the value of medicine lies entirely in what it can bring about. Medicine can restore health and eliminate or reduce pain. It is good because these (the restoration of health and the lessening of pain) are good.

        Why is this distinction between intrinsic and extrinsic goodness important? Well, in addition to aiding the axiologist (the philosopher of value) in her search for what is ultimately valuable, it is important to the understanding of consequentialism because it helps to make sense of considerations of how much goodness (or value) would result from some actions. The point is that in measuring (if I may use that term loosely) how much goodness might come about through some course of action, we need only count intrinsic goodness. Suppose the sole value of putting a stamp on an envelope is that it helps the envelope (and its contents) get to a specified destination.² Still, putting the stamp on an envelope (at least an envelope that ought to get to its destination) is, on the whole, a good thing. If I consider how good that act is, I cannot both count its (extrinsic) goodness and the goodness of its consequences (namely that the letter gets to its destination), since that would be double counting. One way to avoid the double counting would be by just counting the value of each act (intrinsic or extrinsic) without also counting in the value of its consequences. But that would in the end yield the same result as counting the intrinsic value of all of its consequences. Therefore if we are to avoid double counting, we might as well count only intrinsic value.

       Sentience and Non-sentience Theories

        Let me refer to that subset of consequentialist theories that treat only effects on conscious beings as morally relevant as sentience theories.³ Most well known consequentialist theories are "sentience" theories in this sense. These are theories that hold that the only things that are intrinsically valuable involve how sentient beings, or some subclass of sentient beings, fare. (Sentient beings, as I am using the term, are those beings who are conscious or aware.⁴ You and I are sentient beings and so are many other animals.⁵) We will look at varieties of sentience consequentialism. But some ethicists have a broader view of what is good and think that beauty, for example, might be good quite apart from whether any one ever experiences the beauty. Some naturalists write as though a great diversity of species is good in itself (and not because of its evolutionary advantage to sentient beings), quite apart from whether anyone is ever aware of that diversity of species (or any of its consequences). If such people also held that the right actions are the ones that increase the amount of goodness in the universe, they would be what I've termed "non-sentience consequentialists." Non-sentience Consequentialism is any consequentialist ethical theory that says that what is morally right or wrong does not depend solely on the consequences that actions have for sentient beings.

        I will not say much about non-sentience consequentialism here except to mention the most famous example. G. E. Moore, a British philosopher who wrote in the early twentieth century, thought that what is morally right is to increase the amount of goodness. His is clearly a consequentialist theory. Yet Moore allows that what counts as goodness might not specifically relate to persons (or other animals for that matter). So, for example, he thinks that a beautiful world is (other things being equal) better than an ugly world even if there were no conscious beings around to appreciate it. He is thus what I call a "non-sentience consequentialist."

       Egoism and Other Single Beneficiary Theories

        Among sentience consequentialist theories the most familiar competitor to utilitarian theories is ethical egoism. Like "utilitarianism," "egoism" is not the name for a single specific theory, but a family of theories. Ethical egoists typically are relativists and claim that an action is right for a specific person if as a result that person is better off. Of course to be precise the theory would have to specify what makes a person better off, just as utilitarianism would.

        There are also non-relative versions of ethical egoism according to which what is right to do is whatever is in the best interest of the person espousing the theory. So egoist Egon may claim that what is right (for any Joe, Larry, or Shirley) are those actions that benefit Egon. (Convincing anyone else of the truth of that theory may be difficult for Egon, of course, although some charismatic leaders (David Koresh of Waco, for instance) pull this trick off.)

        The possibility of such non-relative egoistic theories, suggests a related possibility: Some consequentialists may claim that all that morally matters in action is how it affects some single specified person. It's the right thing to do, they may say, if and only if it makes the Dalai Lama happy. This is similar to egoism, in that the happiness of only one person counts, but this is not an egoistic theory, since the person whose happiness counts is not necessarily the person whose action is being considered or the person espousing the theory.

        Here is a similar theological version of a consequentialist ethical theory (again nonegoistic): an action is right to the extent that it results in God's being glorified. This is consequentialist because it looks only to the results of action. Such a theory could support the rightness of burning heretics, for instance, if the results of such a practice would be that there would be more glorification of God (since there would be fewer heretics around to get people off track).

       Utilitarianism

        Utilitarianism is distinguished among consequentialist ethical theories in that it insists that the well being (somehow specified) of everyone (somehow specified) counts positively, and counts equally. It is a very broad sentience theory. You will note that this account of utilitarianism leaves out G.E. Moore's consequentialist theory from the class of utilitarian theories. Moore's theory is definitely consequentialist, but I don't count it as utilitarian because it is consistent with Moore's theory that what happens to some people, or even all people, not count at all morally. Still, some authors call Moore's theory "ideal utilitarianism."⁶ The difference between us on this point is merely terminological.

        I say "somehow specified" (twice) in the above account of utilitarianism because there is a wide range of variation within utilitarianism. For example, there is dispute about how far the term "everyone" extends. Does that mean every person? Some may think it means every human being. This possibly differs from "every person," since there may be some human beings who are not persons (perhaps fetuses, neonates, or irreversibly comatose vegetative homo sapiens).⁷ Further, there may be persons who are not human beings (perhaps God, angels, extraterrestrial rational creatures, and maybe some other higher mammals). Other utilitarian theories include a broader account of "everyone" than either of these. John Stuart Mill, for example, a leading utilitarian, thought it was the happiness of all "sentient creation" that counted. That is, the happiness of every being who had awareness. That would include dogs, pigs, and turtles, but presumably not plants or germs. (His use of the term "creation" or "creatures" eliminates God from counting even if God were a sentient being.)

       Non-utilitarian Consequentialist Theories

        Besides forms of egoism and theories that make exactly one person's well-being count, there are other non-utilitarian consequentialist theories. Some theories may allow that only the happiness of some people counts, typically "we" count and "they" do not. You sometimes hear some people justifying use of atomic weapons against Japan near the end of World War II by speaking of how many "American" lives were saved. Similarly some racists think that only the well-being of white folks is morally relevant.

               Elitism

        Another variation on non-utilitarian consequentialism allows that everyone counts but thinks that not everyone counts equally. I call these elitist theories. Some may think that the well-being of blacks, say, counts, but not as much as the well-being of whites. Or they may think that the well-being of intelligent people counts more than the well-being of unintelligent people. Or perhaps they think that the happiness of morally virtuous people counts more than that of evil people. (So even if they think it a little bad for a evil person to suffer, they don't think it is anywhere as near bad as for a good person to be afflicted with the same amount of suffering.)

               Egocentric Consequentialism

        If elitist theories and egoist theories were to be married, they would have theories I call "egocentric consequentialism" as offspring. Like elitist theories egocentric theories say either that not everyone counts or at least that not everyone counts equally. Unlike elitism egocentric theories specify who counts by some relation to the person acting or by some relation to the person espousing the theory (in relativist and non-relativist versions, respectively). Many people think that in making moral decisions they should count the happiness of their parents, or children more than they count the happiness of others. If I win $100 and can use it to make my daughter happy or could use it to make some stranger just as happy, these theories propose that it would be "more right" to make my daughter happy. Or the relevant relation may not be "blood relation," but friendship. It may be felt by some that they should consider how some action affects their friends more than how it affects their enemies. I call these theories "egocentric" because they say that how much moral weight the well-being of anyone has depends on the relation of that person (or being) to the person performing the action. Many people implicitly hold theories of this sort. Such theories are consequentialist, since they look to the results of actions to determine moral rightness, but they do not count as utilitarian theories, because not everyone's well-being counts equally.⁸

Further Distinctions Cutting Across These Theories

        I have been talking as if these sentience theories all agree on some account of what consequences affecting individuals are morally relevant. I have been speaking of the "well being" of beings, but there are differences among theories regarding what constitutes well being. Some theories, called hedonistic theories, say the only result that ultimately counts is how much pleasure (or, in a slight modification, how much pleasure and absence of pain) is produced by the action. Jeremy Bentham's theory was of this sort and so, officially at least, is John Stuart Mill's.⁹ Theories expressed instead in the broader terms of the happiness of people (where it is assumed that some things other than pleasure may contribute to our happiness) are called eudaimonistic theories (from the Greek word for happiness or "flourishing," as it is sometimes translated). Other theories may count only whether the actions lead to satisfaction of needs. Call these "need satisfaction" theories. Not everything that makes us happy is something we need. Still other theories try to be more precise and avoid hedonism by specifying what counts as whether the actions lead to satisfaction of preferences. Let us call these "preference satisfaction" theories. According to these theories what morally matters is whether an action has the results that people prefer. Of course this doesn't mean that people can do whatever they prefer, because if the theories are other than egoist theories, the preferences of other people matter too. Further, while we might know what we want, we do not always know whether a particular action will satisfy our preferences. So, Tyler might prefer to be rid of his headache, but not know whether taking a particular pill will have that result.

        This leads to another distinction that applies to all consequentialist theories. According to consequentialism, what matters morally is what the results of actions are. But there is ambiguity here. Where we are talking of actions that are actually performed we can talk of the actual consequences of the action. But to do utilitarian calculations we must consider various options. One of these options may be taken, but surely some will not be. Those options that are not taken will not, therefore, have any actual consequences. So if we are to compare the consequences of various options, we must, at least sometimes, be thinking of something other than actual consequences. While we can talk about the actual consequences of the option chosen, to speak of the "consequences" of the other options requires some other reference. We could talk counterfactually of what the consequences these other options would be. This has the virtue of parity, since the actual consequences of the option chosen are precisely (by definition) those that "would" obtain.

        Alternatively we could speak of "probable consequences" or "expected (or foreseen) consequences" or the consequences that ought to have been expected by an informed rational being. John Stuart Mill states his theory in terms of the consequences an action "tends" to have.

        One could even talk of what consequences are simply intended to follow from each option, but such a theory takes a long stride away from the spirit of consequentialism. Consequentialism is the view that only the results of actions count for moral evaluation. If by "results" we were to mean "what was intended," then the theory amounts to saying that what matters is the intention rather than the results of the action.¹⁰

Direct and Indirect Consequentialism

        The most discussed distinction within utilitarian theories is that between what are commonly called act and rule utilitarianism. The distinction, however, cuts across all consequentialist theories, not just utilitarianism. Furthermore, there are notable variations on the "rule" side of this distinction.

        I prefer to put this distinction in terms of direct and indirect consequentialism.¹¹ So far I have been speaking as if the only kind of consequentialism were direct consequentialism. Some of my remarks would have to be modified to account for indirect consequentialism.

        Moral theories are supposed to tell us which actions are required, which permissible, and which prohibited.¹² Direct consequentialist theories do this by evaluating the options directly by appeal to the consequentialist principle. That is, a direct consequentialist theory would say, for example: choose that action among the options (possible actions) which results in the most happiness.

        Indirect consequentialism instead judges something else (other than options), such as a rule, by appealing to what the consequences of that would be, and then judging actions by appeal to this intermediate thing. Thus rule consequentialism judges what possible moral rules are legitimate (let us call them "morally valid") by appeal to the consequences of those rules. (That will take some explaining, which I will do shortly.) It then judges actions to be morally wrong if they are violating those rules.

        Naturally all this can be complicated. There are several readings of this and the differences among the readings can lead to different ethical theories. After all, what are the "consequences" of a rule? A rule is an abstract thing, not an event. So rules, as such, don't have consequences any more than numbers or letters do. But there can be consequences of people universally (i.e. without exception) following a certain rule, of people generally (even if not universally) following that rule, and simply of the rule being generally recognized in a society. One could even look at the consequences of a rule being generally taught (as a moral rule) in a society. These consequences can be different. For instance, the consequences of everyone following the rule "don't use guns" can be drastically different from the consequences of most people following this rule. Different versions of rule utilitarianism appeal to these different understandings of "consequences of a rule."

        Also, it is often not realistic to assess the consequences of people following some single rule. What perhaps really needs to be considered is an entire system of rules, a moral code. A "code" utilitarian would say an action is right if and only if it is in accord with the morally valid moral code, where the code that is morally valid is the one that if generally followed (to pick one of the above variations) would bring about the most happiness for people.

Why Indirect Consequentialism?

        Why would someone embrace a form of indirect consequentialism? Well, basically it is because sometimes what direct utilitarianism says ought to be done is hard to swallow. For example, consider whether it is wrong to kill an innocent person. Direct consequentialism would say that it is wrong if more harm than good comes from it. Yet if it turned out that the person in question lacked friends who grieved her death and for some reason her death made lots of people significantly happy, then it seems act utilitarianism would say that it is right to kill her. Many people feel that such killing of an innocent person would surely be wrong, even if it caused much more happiness than unhappiness. Killing, after all, is not on par with canceling some unpopular television show. What indirect consequentialism does is to hold on to consequentialist foundations and the insight that what morally matters is the results of actions, but to temper that with a dose of rule-governed moral constraints. So, according to rule consequentialism, everything that is wrong is wrong because it violates some moral rule or code. But not all rules matter. The rules themselves can be independently evaluated and justified by appeal to their "consequences."

        Yet how different is indirect consequentialism from direct consequentialism? It really depends on the specific form. Consider, for example, the form that declares that rule to be morally valid which if universally followed would lead to the greatest happiness of the greatest number of people. Well, presumably there is one such rule, namely the rule of utility itself! The rule that would have the best consequences is the rule that says: do the action that would have the best consequences! If that is right then rule utilitarianism collapses into act utilitarianism in the sense that it gives the same answer as the corresponding form of act utilitarianism would concerning the moral correctness of an action.

Objections to Consequentialist Theories

        Is any version of consequentialism the true moral theory, the theory that correctly specifies which actions are right and which wrong? I will not answer that question definitively here, but I will mention some objections people have raised to direct utilitarianism, perhaps the most common variety of consequentialism. Some of these objections can be modified to apply to other consequentialist theories. And, as we've seen, other consequentialist theories are beset by plenty of additional objections.

        One objection simply appeals to the gut feeling that sometimes it is wrong to do something even if the consequences on the whole would be good. This objection is usually strengthened by an example. Consider a case in which a whole city of people is nervous about going outside because there have been sporadic serial murders. Suppose the police chief knows that there is no longer any threat because he knows that the killer has died, but would be unable to prove this to anyone's satisfaction. Furthermore he finds a suspect that he can convincingly frame for the crimes. He figures that by framing this innocent person many people will be much happier. Of course the individual framed will not be as happy, but he is just one person and we can easily imagine him having few, if any, friends who care about him. The gut feeling of many people is that framing such a person would be wrong (where framing would involve eventually punishing the person by imprisonment or death) even if this led to the greatest happiness of the greatest number. If so, act utilitarianism, at least is wrong.

        Such an example could be tailored to serve as a refutation to most versions of direct consequentialism. Indirect consequentialism, of course, is designed to avoid objections of this sort. Indirect consequentialism can say that not framing innocent people is a rule that if generally or universally followed leads to better results than general or universal violation of the rule. So the rule is a morally valid rule and ought to be adhered to.

        Another objection, put forward by the contemporary Harvard philosopher, John Rawls, argues that utilitarianism does not take seriously the value of individual people.¹³ Utilitarian considerations always lump together the happiness of everyone. True, everyone's happiness counts, but it only counts as part of the total of all happiness, according to utilitarianism. For an ethic that takes seriously the value of each individual person, Rawls argues, we must look elsewhere.

        Finally, some people argue that consequentialism, at least in its direct varieties, is impractical and too demanding. It insists that one must always do the "optimific" act, the one that leads to the best consequences. But think of our ordinary lives. It seems that none of us, on this account, is really doing the right thing. Surely there is always something else we could be doing which would cause more happiness in the end than what we are doing now. Instead of reading this, you could be writing a check to some needy and honest charity.