20th WCP: Trading Lives: Consequentialism, Deontology, and Inevitable Trade-offs

It has long been one of the standard criticisms of consequentialist approaches to ethics that they too easily justify "trade-offs" that are morally unacceptable. The criticism which holds "the end justifies the means" philosophy inherent in consequentialism to be a source of great immorality is expressed, for example, in the famous scene from Dostoyevsky's The Brothers Karamazov. Remember how Alyosha reacts to the prospect offered by Ivan of a harmonious world order, a system that would bring about peace and rest and happiness for all men. A lovely idea, but the structure comes at the price of torturing one tiny child to death. And Alyosha will not consent to that "exchange."

A consequentialist response to Alyosha's refusal to consent to trade the suffering and death of one innocent in exchange for universal harmony is that, in the present inharmonious order, many innocent children will die horribly, not just one. Alyosha's tender conscience will cost thousands of innocent children their lives. And so the debate continues.

Recently, however, a proponent of consequentialism, Alastair Norcross, has sharpened the debate. (1) According to Dr. Norcross, "consequentialism tells us that if even a minor headache has some disvalue, a large enough number of such headaches must equal the disvalue of a death." So, Norcross admits that consequentialism is committed to what he calls the unpleasant conclusion, namely that "there is some finite number of headaches such that it is permissible to kill an innocent person to avoid them." Strikingly, not only does Norcross not believe the unpleasant conclusion to be really so unpleasant, but he also believes that "most of us, consequentialists and nonconsequentialists alike, accept at least some other claims that do not differ significantly from the unpleasant conclusion." (2)

His challenge to those critics of consequentialism who find the theory flawed because it may lead to the unpleasant conclusion is for the critics to find a principled way to distinguish that conclusion from what he calls the pleasant conclusion, the conclusion that we are not morally obligated to impose a national speed limit of 50 mph (or less)--even though by doing so we may save many lives.

A note about the topicality of Prof. Norcross's speed limit example: In my home state of Texas the Houston Chronicle for September 30, 1996 carried a story which begins "Traffic deaths in Texas have increased 17 percent--an average of more than 40 per month--since the state began raising highway speed limits." Perhaps we should beware of reasoning post hoc, ergo propter hoc, but it is in this case a very tempting inference to treat such data as indicating a cause and effect relationship.

Now to the main task. The way I read Prof. Norcross's argument makes it out to be a quite canny tu quoque. I take it that his main point is to defend a rather promiscuous consequentialism, that is, one which allows that smaller goods of any sort, such as the relief of headache pain, may in some circumstances be aggregated in such large numbers that their weight in toto can offset the production of a very large bad--such as killing an innocent person.

A stern opponent of consequentialism will take that trade-off to indicate a fatal deficiency in unrestrained consequentialism. Prof. Norcross's strategy, which he executes very ably in my opinion, is to seize upon an example of a particular trade-off that many of us are comfortable with, specifically that an increase in convenience due to speedier travel may be traded for a predictable increase in highway fatalities. He examines attempt after attempt to justify the trade-off on nonconsequentialist grounds--that we are not killing anyone, we are just allowing them to die, that they freely choose the risks, that there is no determinate trade-off, etc. etc.--and all of the attempts fail lamely. For example, if one argues that speed limits are set by a democratically elected government and are, thus, the end result of a process in which the victims participate, Norcross notes that many of the victims did, in fact, not participate, nor was it likely that speed limits were a campaign issue, and, if they were, some of the victims may have voted for the person favoring the safer, lower limits, etc. (3)

Norcross dismisses attempts to justify the trade-off in the speed limits case on the grounds that the legislature has it within its authority to make such decisions is by citing a lack of actual consent to the terms of the trade-off by the nonvoter, the voter against, and the minor child, none of whom explicitly endorse the trade-off, and any of whom may die violently because of it. I believe this dismissal is too hasty, but on the other other hand, we do well not to grant a blank check to any legislative body no matter how legitimate its authority. We must be able to subject legislative decisions to critical scrutiny, and that means in the matter at hand we must be able to say with some reason that the speed limit is either too high, or too low, or just about right--acknowledging all the while that changes in the speed limit arguably affect the highway fatality rate. So we need some sort of stance regarding the justifiability or lack of it of the trade-off that is independent of the issue of the legitimate authority of the legislature.

It should be plain that I found Norcross's presentation to be very stimulating. I found myself rehearsing objections only to have them anticipated and fairly countered in the next page or paragraph. However, the argument seems to be a tu quoque. Even if it is true that all of the nonconsequentialist arguments he surveys are unable to account for the speed limit trade-off, that does not automatically put a promiscuous, unconstrained consequentialism in the right. All of the views canvassed, unconstrained consequentialism included, may be, in different ways, inadequate.

Norcross's powerful point, however, is that trade-offs of lives for other values are ubiquitous, and they stand in need of some moral justification. Norcross says, "it is not obvious that we are wrong to fail to do all we can to reduce the number of deaths," (4) but the burden of his argument is that it is, at the same time, not obvious that we are right in our failure to do some of the things we could do to reduce the number of deaths, things such as reducing the speed limit.

By now I've probably tipped my hand that, in my opinion, some sort of qualification on consequentialism is worth pursuing. First, I will suggest a qualification, second I'll propose a possible line of justification for that qualification, and third I'll give an example of how the qualification might be applied.

The suggested qualification is, in part, also a response to the stance of Alyosha who would forbid any trading of a life for other lives. The suggestion is that we may be able, on nonconsequentialist grounds, to justify trading a life for two or more lives; whereas we may not trade a life for the relief of even 10,000 headaches. To motivate reflection, let us try using a moral principle that a noted consequentialist, Peter Singer, endorses. The principle of Minimally Decent Samaritanhood (my name for it, not Singer's) is that, if I am able to help someone avoid a great harm at a very small cost to myself, then I am obliged help that person. This principle would motivate someone who would feel obliged to wade into a shallow pond at the cost of getting his clothes wet, in order to save a drowning child. Singer argues this vigorously, of course, in taking his stand on a duty to aid poverty-stricken starvation victims around the world. (5)

In applying this principle, let us make the small cost extremely small, just the tiniest of twinges, hardly noticeable when it occurs and quickly forgotten. Still there is some positive value in not suffering it, and so Norcross's original argument would appear to hold that, if enough of these tiny twinges were to be relieved, say for a billion people, the positive value generated would outweigh (on some scale) the positive value of a single person's life. Let's call that person Alyosha. So the unpleasant conclusion in Alyosha's case is that it would be right to kill him if all those painful twinges would indeed be relieved.

But with Peter Singer's moral principle in place, a principle which requires only minimally decent Samaritanhood, what is the result if we compare Alysoha's situation with Ivan, one of those who will suffer the twinge if Alyosha is not killed? The result is that Ivan is obliged to suffer the twinge rather than have Alyosha killed. But a similar obligation lies upon Dimitri, Fyodor, and so forth for all the rest of the billion. It would seem then that all are obliged to suffer rather than have Alyosha killed for the sake of relieving their twinges. (Note that this is not a fallacy of composition, just a generalization based on a potentially complete enumeration.) This may not create an outright contradiction, but it does seem a bit paradoxical that the very people we are supposedly right to help by killing Alyosha, have a duty to suffer rather than have him killed.

The root of the problem, it seems to me, is the promiscuous consequentialism that allows trading off of any value against any other. There is an impersonality in this approach that tends to conflict with many of our deeply held moral intuitions. Instead of impersonality, let us see what the investigations of Frances M. Kamm into impartiality might suggest about how trading lives for lives might be justified.

Here is an "Aggregation Argument" that Kamm deploys to justify trading a life for more lives: If (1) it is worse if B and C die than if B dies alone, and (2) it is equally bad if A alone dies or B alone dies, then (3) by substitution it should also be worse if B and C die than if A alone dies. (6) Kamm is herself sensitive to a number of particular criticisms that might be made of this argument, but she does ultimately endorse it explicitly on deontological grounds. She sees the situation as presenting a conflict of interest and asserts "that as outsiders to a conflict of interest, we always move back to an impartial perspective outside that of any of the contestants." (7) When we do this, "we still do abandon someone, yet from the point of view outside that of each of the personal points of view, we see that as much is gained in helping one as in helping another and in this sense they are equivalents." The result is "the relation between candidates for help is established in matching or balancing, which involves some sort of substitution of equivalents (who are not identical)." (8)

The foregoing is of course only suggestive, and I recommend Kamm's dense, but intriguing, account of "irrelevant utilities" and the variety of blends of subjectivity and objectivity she calls "Sobjectivism 1," "Sobjectivism 2," and so on. (9)

If we are permitted on deontological grounds to make decisions to save the greater number of lives, this opens the door for a number of real world applications. Here is one I have taken from the column,"The Economic Scene," by Peter Passell in the September 5, 1996 New York Times. It concerns the increase in airport security in the wake of the TWA 800 disaster. Passell cites the work of economist Robert Hahn (10) who calculated the cost of the extra time demanded by increased airport security. Hahn took the average value of a traveler's time as $20.00 per hour and the average extra wait time at one-half hour. At $10.00 apiece for 500 million or so passengers per year, the total cost is $5 billion dollars per year. Next Hahn finds the average number of lives lost to air terrorism in the United States per year by dividing the fatality total since 1982 by the requisite number of years. The fatality total turns out to be 548, if we make the debatable assumption terrorism was in fact the cause of the TWA 800 disaster, and the result is an average of 37 lives lost per year. If the new security measures were to be totally successful, thereby cutting the rate of loss of life to zero, the cost per life saved would be about $135 million per year. Is that a good use of our resources? Perhaps not. Passell cites W. Kip Viscusi, an economist at Harvard Law, who estimates that saving a life at $50 million is self-defeating because it drains resources from other life-saving purposes. In other words, the contention is that when we spend over $50 million to save a life, there is a net loss of life.

In a similar vein, the noted political scientist, the late Aaron Wildavsky, was concerned that environmental regulation too often had the effect of causing the expenditure of large amounts of money to lessen the likelihood of only a few "statistical deaths." But the cost is not just money. In his posthumous book, But Is It True?, Wildavsky cites estimates by decision analyst Ralph Keeney that a cost of $7.5 million per hypothetical life saved will cause a death from lowered living standards. Wildavsky observes "The important thing is not the exact numbers but the view that expenditures of national wealth on regulating minute quantities of chemicals actually undermines any health improvement achieved. From any gains in health improvement due to regulation must be subtracted accidents and illnesses contracted during efforts to remove the offending substances and losses in health due to lowering the standard of living." (11)

To return to the speed limit example, one can see that to describe it as a trade-off between convenience and lives may be misleading. The trade-off is between more lives lost and more time millions of people spend not driving, that is, more time spent in a variety of potentially productive activities. When my trip from Houston to Dallas takes four hours instead of four and a half, that half hour is worth something--though whether it is $10.00 or not I don't profess to know. If lost productivity does translate into lost lives, as Wildavsky and others believe, then perhaps we have at least one route we could take to try to justify a particular speed limit on a highway system. I personally have my doubts that enough travel time is saved to make it worth as many as 40 extra deaths per month in Texas, but if time saved does translate into increased productivity which in turn saves lives, then we can understand why it is unlikely that the appropriate speed limit on the Interstate Highway System is, say, 35 miles per hour. Naturally, I am not certain that any of this is on the right track, but it does represent, however briefly and crudely, a way of thinking that will become increasingly influential in modern societies.

If these sketchy efforts to indicate how some trade-offs may be justified do not lead anywhere, perhaps we are doomed, as Calabresi and Bobbitt noted some time ago, to the making of tragic choices, choices in which our deeply held values are unavoidably in conflict. And as they say "Evasion, disguise, temporizing, deception are all ways by which artfully chosen allocation methods can avoid the appearance of failing to reconcile values in conflict. . . . Honesty is the most influential brace in the tragic equilibrium. Though subterfuge may bring us peace, for a while, it is honesty which causes tragic choices to reappear." (12) Professor Norcross has done us all an honest service by confronting us with the failure of our easy rationalizations of decisions which involve trade-offs, trade-offs that many of us are not comfortable in making, but which we are unwilling to forswear. And for that he deserves our thanks.