/// What follows is a peice which presents a series of examinations of utilitarianism from Harwood’s seminal essay, “Eleven Objections to Utilitarianismâ€, along with my demonstrations of how these objections are unsound. Previous installments can be found here and here. I apologize for the length. ///
5. Average and total utilitarianisms produce absurdities. Now we come to what are, I think, the single most powerful arguments against utilitarianism. They arise from the desire to answer the question, “how exactly is a society’s state of happiness calculated?” We have at least two answers, called “total utilitarianism” and “average utilitarianism”. But both seem to produce absurdities.
5A. Total utilitarianism tells us that we calculate a world’s happiness by adding up the happiness of each individual member. Take a world of four people: a, b, c, and d. Each one has a certain amount of happiness (hedons) on a scale of one to ten, where “one” means “utterly miserable”, “five” means “so-so”, and “ten” means “bliss”. Reputedly, this is the view which classical utilitarians like Bentham had.
But this view has interesting philosophical implications. Let’s also say that these people in this world, as they are examined in a particular scratch of time (say 20 years), live in a post-apocalyptic landscape. Each person is miserable: for the sake of argument, let’s say that each has only one hedon, which means the world has four out of a possible forty hedons. The theory would suggest that reproduction would increase happiness, and thus, be mandatory, even if the child conceived would be doomed to a hellish world. The drawbacks are obvious. The duty to copulate seems wrongheaded on the face of it, since it is seemingly increasing misery, not pleasure. And it would ultimately recommend overpopulation, since each new baby would increase happiness just that much more; but in fact, it would just be a multiplication of misery.
Harwood offers some interesting replies to the overpopulation objection. “In real life, we cannot jam the planet full of people and expect to retain enough control of the situation to maximize total satisfaction. First, the more people there are to satisfy, the harder it is likely to be to satisfy them. Second, a world where everyone ekes out a life barely worth living is likely to be unstable, presenting a great danger of disease a chain reaction of catastrophes…”
These are fairly good replies, I think. But their power is restricted to the ‘overpopulation’ objection. We need more than that. The core of the objection, it seems to me, is that there’s something wrong with condemning one’s child to live in a horrible world. In this vein, I can supplement Harwood’s account with three more replies. One I think is good, and the other two are bad.
First, the sustaining of life opens the door to future improvement (a long-term consequence). But this is problematic, because the scenario suggests that something really has been done right in the time frame being analyzed, yet the treatment just mentioned suggests that the goodness of the situation is pushed off to the future. Surely these are incompatible.
Second, the objection might be interpreted more strongly. It may want us to imagine a new world where not only is the actual state of affairs poor, but moreover, it would be impossible to improve the states of affairs. This would be a foul situation indeed. However, again, it presupposes an evil scenario in order to make a point about the good, confusing our understanding of both. The absurdity of such techniques lies, in this case, in the notion that human beings cannot, and do not, by their nature, engage in creative striving to improve their lives. If the situation were such that no creative striving were possible (i.e., if all shelter were measly, all beneficial technology impossible to produce, etc), then once again, we should not be surprised that our intuitions are scandalized, since the very world under examination provides inescapable moral horror. Still, if we were to grant that the scenario is uncooperative, and not a candidate for defeat of utilitarianism, it still runs into the same problem as the first objection. Our concern is about the effects of birth in hell upon the world’s level of hedons, not with the future conditions. And the account just seems to get something wrong if it says nothing about that.
Third, perhaps it’s the method with which we’ve described the situations which is in error; maybe it would be better to register happiness and misery on a positive-negative scale. In which case, the happiness of the hellish world of four would be -40 out of a possible 40 hedons. This new characterization would suggest that the introduction of a baby into the world would decrease utility (-50 out of 50), and thus, be wrong; and it seems to capture the language of the situation in a much more felicitous way. There seems to be no problem with this new interpretation. The objection seems to have been defused.
But, getting back to the original intent of the objection, what if we revised the scenario under the new description? In this case, the world would have 4 hedons on a scale of -40 to 40. A so-so world. Are people then compelled to produce offspring to make it immediately better? It would seem so. I have no particular intuition or care about this result. It seems acceptable enough. But other questions do arise from having made this elaborate philosophical journey.
First, all other things equal, we have assumed that, for the sake of demonstration, the people of a world (a, b, c, and d) had only one value of happiness assigned to each for the duration of the time period under examination. Thus, (a) has one hedon, (b) has one hedon, and so on. But this is, of course, foolish. Peoples’ moods change all the time. Second, we have also forgotten all the contingencies related to childrearing; the joys of parenting and of adolescence. The very postulation of some significant act, also demands that we factor in the hedonic consequences of the act. But we haven’t, in these cases; and that makes them ineffective. Third, in order for the former hellbound scenarios to make any sense, we would essentially have to pretend that the world was inhabited by miserly mannequins whose happiness is not contingent upon personal expectations of the world. But in many cases, a world populated by persons who had gotten used to certain conditions, would adapt to those conditions. This is both an evolutionary blessing (in the sense of it may protect the population’s eyes from gazing across the true and terrible vistas of their world by comparing it to other, better possible worlds), and a damnable curse (since, as any activist will tell you, the banality of evil and the amorality of apathy are often the products of the merest lack of imagination for an easily attainable and better life).
With all of the above considerations in mind, the reason why I ultimately find total utilitarianism dissatisfying is that it fails to take into account a very significant fact: that, for any state of affairs, there is a pre-arranged amount of possible hedons that the world could have, given the number of persons. What I mean is, I have misleadingly presented the data in terms of 4/40 hedons, or somesuch, but in fact, total utilitarianism only cares for the actual amount that occurs — 4 hedons — and doesn’t seem to care in its numerical analysis about the number of possible hedons involved. But it should make a difference whether or not the 4 hedons are there out of a possible 40, or 60, or 80, because that tells us, in at least abstract and restricted terms, how far away from achieving utility we are at; and that constant striving is a characteristic of utilitarianism. This revised view doesn’t just tell us the state of affairs at some particular time; it tells us how far off we are from the mark, and so, seems more acceptable in the abstract.
The outcome of a stronger analysis, I think, is that while total utilitarianism cannot be soundly accused of absurdity, it may be lacking in other ways.
B. Average utilitarianism is our second alternative. It makes up for the deficiency noted just lately by taking into account not just the sum of hedons from the aggregate of persons, but also by factoring in the population size. Thus, in a world where there are four people (a,b,c, and d), where (a) has 3 hedons, (b) and (c) have six apeice, and (d) has one hedon, our analysis would say that the world has, on average, four hedons (16/4).
The objection to average utilitarianism is that it seems to suggest that the world would be happier if (a, b, and c) were to gang up on (d) and kill her, thus increasing the average to five hedons. This, Harwood suggests, is a decisive objection against average utility.
This is deeply unconvincing for the same reasons that have been addressed in previous posts: it relies on falsehood. The reason why we follow rules is the same reason why we keep promises and the same reason why we don’t harvest organs from the living: because human life is oriented significantly around trust. Indeed, I would go so far as to say that trust is an incorrigible value, along with agency (as we have seen). To commit cold-blooded murder of another human being is to destablize relations of trust in a community. Murder decreases happiness for the survivors.
An analogy might help to illustrate. A woman has an affair with a married man. Soon afterward, he leaves his wife for his partner. The partner dumps him. He asks her, “Why?”. She tells him: he cannot be trusted. He’ll just cheat on his partner, too. It’s not worth her while. If they had all made better decisions, they could have been happier; and most of wisdom comes in understanding these game-theoretic lapses of insight.
It does not matter who the person is that dies, it does not matter if they are a pariah or outcast. The integrity of a social system can be understood, without exception, by examining how its members consciously treat its outsiders. This is not just a quaint philosophical musing, but a fact about human group relations. Reciprocity creeps inward, eventually: a certain comfort and feeling of entitlement that is associated with the enactment of certain punishments infects the norms in one’s own life and community. One cannot express surprise when a violent cop turns out to be abusive at home, because when a person’s personality is geared towards violence, its outlet is of little concern to him.
Granted, the human mind is magnificient in its ability to compartmentalize various duties and identities. Thus, what I am prepared to do to a stranger, I might not do to a friend. Moreover, it seems that some polities which are externally aggressive – i.e., contemporary Israel – are constituted by a societies which have relatively low rates of violent personal crime. These facts seem to give the lie to the bold statements I made in the previous paragraph. But these are facts which traffic in human frailty involving a state of affairs that is several times removed from everyday experience; people don’t usually identify with the social systems they belong to, they only identify with those systems which give them identities. This is Hannah Arendt’s ‘banality of evil’; it is what I have called a derelection of agentic duties; and it is not virtue. Still, in this situation, reciprocity might not creep inwards, so long as relatively innocent persons remain ignorant of violations of their trust. All I can say, in these instances, is that only so much of an imbalance can be tolerated or legitimated by any society without reinforcing either anomic or xenophobic conditions, and empowering institutions (the church/media and the nation, respectively) that feed upon those conditions. Such things are not conducive to utility.
None of this is to say that life is of incorrigible value. (For example, passive euthanasia is still morally permissible. It is an act of mercy to relieve a voluntary person of incurable and intense pain.) It is just to say that only through the protection from misery, short-term and long-term, that anything of that sort can be entertained.
6. Rule utilitarianism is incoherant or redundant.
Rule utilitarianism is the idea that at least some actions are justified by the rules which they are driven by, and those rules are in turn justified by the principle of utility. Harwood points out four objections, each of which either conclude that rule utilitarianism is either incoherant or redundant. It is not necessary to go into the details of all four objections, because they at core reduce down to two mutually exclusive possibilities: incoherance or redundance. (They are mutually exclusive because any particular aspect of a theory cannot both be criticized of both being redundant and being incoherant at the same time. Redundancy is mere repetition, and repetition is logically sound (though trivial), while incoherance is illogical.)
1. Rule utilitarianism is, indeed, redundant: it is usefully redundant. Take the utterance, “I see an animal”. If I were to utter it, and then right afterwards say, “I see a bear”, I would be committing redundancy. However, one utterance would be more specific than the other; it would carry a bit more information. The former comment is akin to rule utilitarianism, the latter is akin to act utilitarianism. However, in the end, it’s the latter which is most correct.
Harwood brings an argument to bear on the matter which formulates the most plausible form of rule-utilitarianism as if it were an absurdity. His is the objection based on extensional equivalence. Harwood explains that two moral systems are extensionally equivalent if “they always agree about what we should do in any case”. If rule-utilitarianism were extensionally equivalent to act-utilitarianism, Harwood suggests (following R.M. Hare), it would not seem to reconcile our ordinary moral beliefs with the propositions introduced by utilitarianism.
But what rule-utilitarianism really provides is an emphasis upon a certain bit of information about human social reality which might be overlooked otherwise: namely, that humans are meaning-makers, and to a large extent, are rule-following animals. Act-utilitarianism can recommend rule-following when we put it in practice, but the point needs to be driven home to would-be detractors, so we invent this thing called “rule-utilitarianism” to clear up any lingering confusion.
Harwood’s most interesting comment on this score comes at the end of the section, when he writes: “…if a rule really were so useful that its adoption would maximize satisfaction, then act-utilitarianism would require us to do the act of adopting that rule and taking that rule to heart”. This is plausible, and it says some things that are true. But it suggests that absolute rule-following is mandated by utilitarianism, and this is wrong.
2. One of the most excellent features of utilitarianism is that it specifies, not just our rules, but also the exception to those rules. Thus, when confronted with the question: “what are the acts which ought to guide my life?” We answer: those which are in line with rules which are compatible with utility. And when confronted with another: “What are the exceptions to the rule?” We answer: when the violation of the rule would maximize utility.
3. A third question would be, “What about rights? Isn’t the idea of a right that it admits of no justifiable exception?”
A right is a very special kind of rule which functions much in the same way as a stop sign. Whenever stop signs are effective, it is because they are respected without question. Even a car who sits idling in front of a stop sign at midnight, with no police around to observe him cheating, obeys the law in order to maintain his good habits, and to avoid attention from his peers. When stop signs are violated, the person has a chance of being punished. Sure, the threat at the level of the individual seems measely. But the necessity of the law is such that, with each flouting of it, the law itself becomes less and less powerful. In the same way that the laws of traffic become moot when a traffic accident makes all travel impossible, the flouting of a right threatens the very potency of rights in the first place.
What rights have in common with stop signs is that there is no other way in which either could function except through near-absolute regard, and through true absolute renown. Essentially, what’s especially significant, and which separates a right from a mere law, is that rights operate in the way they do, because there would be no other way in which they could operate. The difference between them is that there is direct moral significance to the goal of a right which is not present in the laws of traffic. Any violations must either be in terms of established exceptions, or barring that, as a last resort in the name of utility proper.
[Edit: It is worth noting, and emphasizing, that the act of violating a rule is not in and of itself significant to utilitarianism. Rather, among other things, it is the habit which offends the principle of utility, since it guarantees more than one future act, which will (presumably) raise even more consequences besides. It would not be much of an exaggeration to say that, when it comes to understanding rules, the act is just a means of observing the habit at work.]
I say all of this, purposefully using a hedge-word like “near-absolute”, though I full well know that the language of rights has an air of absolutism about it. This is, at first glance, because rights must be treated as absolute, for the same reason that Sherwood Forest must be treated as haunted. The consequences of hedging and the use of hypothetical phrasing may be unhappy. Absolutism, like a haunted forest, has a quasi-mystical quality to it, and inspires more awe than tales of drunken rascals hiding in the bushes.
The former comments ought to cause anyone who pledges allegiance to the publicity condition to twitch. I have been sardonic enough, I think, to make twitching justified in every quarter. But I must also grant that my comments do not seem to be kind to publicity. I will have to have more to say about this as one of my last topics of the series.
[Edit: I think I need to say a lot more about rights.
Rights can be analysed across at least two dimensions: at the system level, and at the level of the individual. The system-level is characterized by a bifurcation between nominal and defacto rights: if you violate a person's rights in fact, you do not necessarily alienate them from the right in name. By contrast, an individual-level of analysis is characterized by a collapse of this distinction. To violate a person's right, is to alienate them from it.
We understand the notion of a rule as if it were non-defeasible: that is, as if it always applied to the letter except when certain noted exceptions amend the law (call them RULES). By contrast, we understand mid-level principles to be susceptible to weighting (call them PRINCIPLES). Act-utilitarianism's treatment of rule-following behavior may make room for RULES, but its list of worthy RULES will be slender and subject to qualification. For it, regular everyday rules act as if they were things to be weighted against one another, as PRINCIPLES (thus, for instance, the duty to not engage in civil disobedience may be weighed against the duty to hold solidarity with justified activists, and come out in favor of the latter). Only rights could be given the nominally infallible status that we're looking for.
However, it's not entirely clear what the de facto status of rights are: it will depend on our chosen level of analysis. In other words, it is not immediately clear to me what the utilitarian must do in an emergency no-win scenario, on pain of incoherance. Take an emergency scenario, where a person must choose to either kill an innocent person and save five others, or refrain from killing the one and letting the five die. At the level of the system, if the person chooses to kill the one to save the five, they chip away at the de facto right of the victim without necessarily affecting their nominal right. It is worth emphasizing that, at the level of the social system, the utilitarian may claim that causing the death of the victim is of equally negative consequence as letting five others die, because at that level, the de facto/nominal distinction holds up to scrutiny; and this distinction gives us two models of the best possible outcome. The drawback is that they would also have to be committed to the notion that violation of rule-following behavior would produce equal or more long-term misery than the death of the five, which is only feasible if one has a romantic view of things. Moreover, conflicting advice between defacto and nominal senses of 'right' seem to make the theory incoherant in an irresolvable way.
Meanwhile, at the level of the individual, the nominal/defacto distinction seems paltry. Invocation of a right without fear of consequence seems hollow. And I must confess that, though I once thought that the system-level of analysis could be defended in a robust sense -- that the breaking of a right was, in some sense, "world-breaking", either as an affront to integrity or as a catalyst for social systems failure -- I now see that as a romantic vision of the world which must be deflated by a more realistic perspective. I find the system-level sort of analysis, so long as it is taken for granted, to be less plausible than the analysis at the level of the individual. For it seems ridiculous to say that a serf in the middle ages, who suffers from every pox and plague and horror, and who is alienated from the kind of modernity which would make it even possible for rights to be protected, stands unalienated from his rights. So much the worse for rights, it seems.
A solution at the level of the individual might be found by investigating the doctrines of positive and negative responsibility under it. As we have seen, because of the incorrigible value of agency, we can assign a slight difference between the two: causing harm is slightly worse than letting harm happen. Allowing harm to happen is a violation of one of one's duties, which may have grisly results. A possibility of argument which I would like to leave open is that, while rights do sometimes have something like the character of prime-facie duties -- that is, it is susceptible to weighting -- they only have this quality in situations where the agent will have to assume negative responsibility for some future behavior. By contrast, rights have the status of near-absolute rules if and whenever the agent shall assume positive responsibility for some potential action: an actor does not seem to be able to ever plausibly knowingly will the violation of a right and be moral. In this sense, "doing" is more plausibly constrained by rules (to the extent that we value agency), and "allowing" may be constrained by either rules or acts, depending on utility. The difference is that rights will tend to have more weight in the former than the latter.
This runs directly out of the real insight behind the incorrigible value of agency: namely, that rights begin with me and my actions, and emerge from there. There is no collective agency, there is only individual agency, and I must act according to what I, as an individual, may know. This position is consistent with utilitarianism, if we understand that the greatest happiness for the greatest number does not compel one to prefer one option over the other when their relative ranking as perveyors of lesser and greater misery are hidden from the actor's eyes. And the consequence of flouting a rule, because of our individual conscious minds, is largely opaque to us at the level of the system.
There is, however, a limit to the extent to which utilitarianism can tolerate this flexibility. If a great calamity were to compel the violation of a right, and though the consequences of violation of the latter is unknown while the consequences of violation of the former are known, then all virtue rests on moral luck. All that may be held to justify an action is the conviction that obeying the rule will tend to increase happiness, and not otherwise, though no knowledge enters the fray. But in this case it seems that the subjective utilitarian can justify themselves one way or the other, and blame it on the limits of their agency. I think this is mostly intuitive.
I have lingering worries that my commitments to liberalism have clouded my judgment, and made my resistence toward appealing to the system-level of analysis premature. I want to emphasize two points in this regard. One: I remain committed to the view that things can be analyzed at the level of the social system, so long as our ambitions are restricted to terms of trust (and power). Two: so long as we strive towards realism and not romanticism, if we find that the system-level of analysis yeilds wildly different results from the individual-level of analysis, then we have reason to redouble our efforts to find a lucid picture of the social world which draws upon the relative contributions of both. This is not necessarily "incoherance", any more than the visual data provided by my left eye is "incoherant" with the data provided by my right eye; they do not cohere only insofar as we fail to understand the two well enough that a synthesis can be developed from both. In the meanwhile, though, we must settle for individual-level, and seemingly, to act-utilitarianism, because we are naturally endowed with knowledge at the individual level which we lack at the system level. This does not excuse us from our obligations to learn more, of course; one day, when social science is more mature, we may have a richer complement of system-level insights that may guide our actions.]
