Peter Singer’s essay “Famine, Affluence, and Morality” originally was published in 1972. It has been anthologized frequently,1 and most recently was republished by Oxford University Press as the first chapter of a small, stand-alone book (Singer 2016) with the same title as the essay. The page numbers for these notes are keyed to the book version. These notes do not try to go through the essay page by page, but rather try to reconstruct Singer’s central argument as rigorously as possible and to consolidate possible objections and replies that are scattered throughout the book. I should be clear at the outset that, to keep things organized, I sometimes take points of clarification or exposition by Singer (e.g. that our obligation to others does not decrease with distance) and recast them as responses to possible objections (e.g. the objection that one ought to help those nearby before helping those far away).
To put into current context the problems Singer tries to tackle in his essay, I encourage you to consult the 2015 UNICEF – WHO – World Bank Group joint child malnutrition estimates and the 2015 Millennium Development Goals Report.
3. The central argument
Let us call a person destitute if that person is at risk of dying from hunger or malnutrition, from exposure to the elements, or from poor access to medical care. A person who chronically is at such risk is destitute, but so is a person who suddenly is thrown into such risk by, for instance, natural disaster or war.
Here is how I would reconstruct the central argument of Singer’s essay:
- There are destitute people.
- It is bad for anyone to suffer or die from destitution. (p. 5)
- (The principle of preventing bad occurrences:) If a government or person can prevent something bad from happening, without either (a) causing something just as bad to happen (i.e. creating an offsetting bad), or (b) giving up something good enough to completely offset the bad (i.e. forgoing a compensating good), then that person or government ought to do so. (pp. 5-6)
- Therefore, governments and people ought to help as many of the destitute as possible over the long run, as efficiently as possible and to the fullest extent possible without creating an offsetting bad or forgoing a compensating good.
- If the governments and typical citizens of developed nations were to forgo most of the things on which they spend their resources, and instead direct those resources to helping as many of the destitute as possible over the long run, as efficiently as possible, that diversion would not create an offsetting bad or forgo a compensating good.
- Therefore, the governments and typical citizens of developed nations ought to forgo most of the things on which they spend their resources, and instead direct those resources to helping as many of the destitute as possible over the long run, as efficiently as possible.
- The most efficient way for the governments and typical citizens of developed nations to help as many of the destitute as possible over the long run is to substantially increase the amount they give to foreign aid and to foreign disaster relief.
- Therefore, the governments and typical citizens of developed nations ought to substantially increase the amount they give to foreign aid and to foreign disaster relief.
Most of the above should be self-explanatory, but I will make a few comments before turning to objections and replies.
4.1. When Singer talks about what one ought to do, he is talking about what is morally required.
Singer is not merely saying that it would be commendable to give more to help the destitute. He is saying that to fail to give more would be morally wrong. As he puts it, it is indefensible to regard the kind of aid he advocates as charity, with its connotation of generosity beyond what is required; on the contrary, he takes the kind of aid he advocates to be a duty. (p. 15)
4.2. What does it mean to “substantially” increase the amount one gives?
The maximum that the principle of preventing bad occurrences can require one to give is an amount that reduces one to the level of marginal utility—to the level at which one would oneself become destitute if one gave any more. Since, in Singer’s moral framework, being destitute oneself is just as bad as someone else’s being destitute, giving more than this would violate the principle of preventing bad occurrences.
Is one always required to give enough to reduce one to the level of marginal utility? In theory, no. Precisely how much one ought to give depends in part on how much others are giving. If enough people and governments were to coordinate and each give an appropriate share, then the burden placed on each individual would be modest.2 In the here and now, however, the resources actually earmarked for the destitute fall so far short of what is required that the burden on each individual to pick up the slack becomes enormous, requiring them to reduce themselves to the level of marginal utility.
4.3. In the move from 3 to 4, where do “as many as possible,” “over the long run,” and “efficiently as possible” come into the picture?
Suppose one helps fewer than the maximum number of destitute people one can without creating an offsetting bad or forgoing a compensating good. This means that one can, without creating an offsetting bad or forgoing a compensating good, help additional destitute people. Since the suffering and death of these people is bad, then one has trivially violated the principle of preventing bad occurrences. Similar reasoning applies if one saves fewer of the destitute than possible over the long run, or saves them less efficiently than one can, since the resources wasted could otherwise have been used to prevent other bad things from happening or to create other goods.
4.4. The “moderate” version of the principle of preventing bad occurrences.
Singer thinks that the conclusion of his argument follows even if he replaces the original formulation of the principle of preventing bad occurrences (the strong version) with a somewhat weaker principle (the moderate version). According to the moderate version, individuals and governments are obligated to prevent bad things from happening when they can do so “without thereby sacrificing anything morally significant” (p. 6). Since Singer believes that the strong version is in fact correct (p. 28)—he even says that he “can see no good reason for holding the moderate version of the principle rather than the strong version” (p. 29)—I will stick to the strong version in these notes, and (in the time-worn tradition of mathematicians) leave it as an exercise for the reader to reflect at each point whether the moderate version changes anything important.
4.5. The principle of preventing bad occurrences applies universally.
Although Singer’s essay is directed to the governments and typical citizens of developed countries, the principle of preventing bad occurrences has a far more general scope. It is likely, for instance, that the principle obligates poor nations and individuals to assist even poorer nations and individuals. It is even possible that there are circumstances under which the principle would obligate the destitute to divert to others whatever scarce resources they might have. Singer does not pursue these points, so I will not either, but it is worth mentioning to make sure you understand the principle.
5. Objections and Singer’s replies
5.1. Shouldn’t we help those who are nearby before we help those who are far away? (pp. 7-9)
Singer points out that the principle of preventing bad occurrences does not distinguish between bad occurrences that take place near you and bad occurrences that take place far away. He seems to find this consequence obvious even when considered in itself, separately from the principle. He grants that we feel less inclined to help people with whom we lack personal contact, but takes it as obvious that this psychological fact has no bearing on what we ought to do.
5.2. But aren’t we in a better position to help those who are nearby, both because it is easier to determine what they need and easier to get them the resources they need? (p. 8)
Singer believes that modern communication has made this objection irrelevant: modern communication allows us to judge the needs of those far away as easily as we can the needs of those who are nearby. Though he does not say so in this essay, Singer must also think that money and time directed to helping developing nations has a greater impact per dollar or unit time on the destitute than does money and time directed at helping even the most disadvantaged citizens of a developed nation.
5.3. Does it matter if no one else cares? (pp. 7, 9-11)
Singer points out that on the principle of preventing bad occurrences, any obligation you have to help those in need cannot be lessened by the refusal of other people to help. As with the irrelevance of distance, Singer seems to find this consequence obvious even when considered in itself, separately from the principle. Likewise, he grants that seeing others contribute less than they ought to probably will make you feel less inclined to contribute, but thinks it is obvious that this has no bearing on what you ought to do.
5.4. Does the requirement create a paradox? (pp. 11-13)
This worry is related to the point made above in section 4.2. If you look at how many destitute people there are right now, and you assume that you are the only one who is interested in helping them, then the principle of preventing bad occurrences would seem to tell you that you must give until you have reduced yourself to the level of marginal utility. The problem is that everyone who is not yet at the level of marginal utility will come to this same conclusion. But if everyone who came to this conclusion were to give until they reduced themselves to the level of marginal utility, then the total amount given to the destitute would vastly exceed what was needed, while everyone who formerly had some personal excess would now be on the brink of becoming destitute. This situation arguably is worse than the original situation.
Singer concedes that this could happen if everyone had to decide how much to give without knowing what anyone else was giving. However, in the first place, he denies that such a situation would be paradoxical, because the principle of preventing bad occurrences does itself take into account how much others will give; you might, from lack of information about what everyone else is doing, incorrectly conclude that the principle requires you to give more than it actually requires, but there is no self-contradiction in such a state of affairs (p. 13). More importantly, Singer does advocate that one check to see how much others are giving, to avoid such situations in practice.
5.5. Don’t we have special obligations to members of our own society? Isn’t that the entire point of moral requirements? (pp. 17-19)
In what probably is the most technical objection Singer considers, J. O. Urmson argues that the entire purpose of moral requirements is to help people to function as a society. If this is the case, then helping people outside of one’s society may be supererogatory (above and beyond the call of duty), but cannot be obligatory. Singer asserts that Urmson is mistaken, and that morality does indeed require us to look beyond our society; Singer, however, does not provide any immediate counterargument, so one must infer it.
I suspect Singer’s counterargument would take the form of a standard argument for utilitarianism: even Urmson surely admits that the suffering and death of people outside of his society is just as bad as the suffering and death of people in his society. Thus, how good or how bad the world is depends on the total amount of suffering and death in the entire world. With this acknowledge, the question becomes whether people ought to make the world as good as they can? To deny the principle of preventing bad occurrences, Urmson would have to give a negative answer to the former question. I suspect that Singer would take such an answer to be absurd on the face of it.
5.6. Are such extreme requirements too much for people to handle? (pp. 19-20)
There are two worries here:
First, most people seem to be dispirited by requirements that impose great hardship upon them. Since, in the here and now, the principle of preventing bad occurrences probably requires people to reduce themselves to the level of marginal utility, there is the worry that the requirement, if advertised, will cause people who might otherwise give something to the destitute to lose their motivation to give anything at all (p. 19). Singer grants that it is important in practice to find out what kinds of moral requirements people would be able to live with so that we can avoid these kinds of effects, but he does not think this has any bearing on what actually is morally required of us; if we are psychologically unable to act according to the full demands of the principle of preventing bad occurrences, then this is an indictment of us, not an indictment of the principle.
With this said, Singer also believes that our feelings about what kinds of demands we can live with are in part determined by the general attitude of our society, and that this general attitude can shift as more individuals decide to give an amount closer to what is morally required.3
The second worry is that considering how many people are destitute, the principle of preventing bad occurrences may seem to require that one devotes every moment of one’s time to helping them. Since some of the destitute will die while one sleeps, one might think the principle demands that one should never even sleep, but instead work oneself without pause into a very early grave. Is this actually required?
Here is where the “over the long run” and “efficiently as possible” clauses (see section 4.2 above) come into play. Since working oneself into an early grave almost certainly would mean that one would save fewer of the destitute over the long run, it is very unlikely that the principle would demand such a schedule. The actual schedule the principle demands (in the absence of effective and widespread coordination) is likely to be very harsh, but Singer sees no reason why this should this count as an objection: there is no reason why the demands of morality should not be very harsh.
Again, Singer advocates coordination to reduce the burden on each person, but as long as such coordination continues to be inadequate the burden on each individual should be expected to be severe.
5.8. Is the requirement just too counterintuitive to be correct? (pp. 16-17, 22-23)
Sometimes an argument has such an absurd conclusion that one feels confident that the argument goes wrong somewhere, even if one can’t see exactly where. One of the things that makes Singer interesting in general as a philosopher is that most of his arguments start from very intuitive premises, but end up with conclusions that contradict moral conventions that still pass for truisms in most circles. The key argument in “Famine, affluence and morality” is no different: one common reaction to it is that Singer’s position is so “wildly out of line with what everyone else thinks and has always thought that there must be something wrong with the argument somewhere” (p. 618).
Singer’s response to this in his essay is at first glance surprisingly limited: all he does is quote a passage from Aquinas which draws a conclusion similar to his own. However, it is difficult to see what more Singer could possibly say against such an objection, since he has developed his argument as carefully as possible and the objection does not even try to specify a point where the argument goes wrong. If one presents a good argument with a counterintuitive conclusion, and no one is able or willing to show where it makes a misstep, presumably all one can do is leave it at that and hope that the argument gradually helps to change people’s intuitions.4
5.9. Shouldn’t governments, rather than individuals, be responsible for foreign aid? (pp. 24-26)
If the contention here is that increasing private aid lessens government aid, Singer says this has not been demonstrated, and that probably the reverse is true. If the contention is that one can best help the destitute by using one’s time and money to lobby one’s government to increase foreign aid, Singer grants that this may be true; however, if one determines that it is true, then one must actually spend that time and money lobbying one’s government rather than using the analysis as an excuse to do nothing.
5.10. Won’t helping the destitute just increase their population and make things worse? (pp. 26-27)
Singer admits that this may be true for some kinds of aid, but points out that it is not true for all kinds. If population control ultimately is the most efficient way to minimize the suffering and death of the destitute over the long run, then one should give one’s time and money to organizations that promote population control until one is reduced to the level of marginal utility—there still would be no justification in keeping anything beyond the level of marginal utility for oneself. Furthermore, as people become more economically secure, they typically have fewer children, so forms of aid that promote self-sufficiency (aiding agricultural development, for instance) generally will alleviate destitution in the short term and reduce population growth at the same time.
5.11. Won’t giving the amount required damage the giving nations and individuals too much financially? (pp. 27-30)
This objection sometimes is offered in the same spirit as the objections in sections 5.6, 5.7, or 5.8 above, and Singer’s response here is no different than before, except that he notes that it may actually be a good thing in itself for economic growth to slow down and for consumer society to disappear.5
The objection sometimes also is made on the ground that economic damage to oneself or to one’s country may reduce one’s ability or one’s country’s ability to give more over the long run. Singer admits that this needs to be taken into account when one tries to figure out exactly how much to give, but the matter is purely academic at this point for the governments and citizens of developed nations, because they clearly waste most of their money.
1 In LaFollette (2014), Cahn and Markie (2015), and Shafer-Landau (2012), to name just a few widely-used ethics anthologies.
2 Much of Singer’s non-academic work tries to build such coordination.
3 In the preface to the 2016 book, Singer offers some anecdotal evidence that his essay has been helping to bring people’s intuitions around.
4 See the previous note.
5 He only footnotes others here, though, and does not explain even in broad strokes why this would be good.
Cahn SM and Markie P. (2015). Ethics: History, Theory and Contemporary Issues. Oxford: Oxford University Press.
LaFollette H. (2014). Ethics in Practice: An Anthology: Fourth Edition. Malden, MA: Wiley Blackwell.
Shafer-Landau R. (2012). Ethical Theory: An Anthology: Second Edition. Malden, MA: Wiley Blackwell.
Singer P. (1972) Famine, affluence, and morality. Philosophy and Public Affairs 1(3): 229-43.
Singer P. (2016) Famine, Affluence, and Morality. Oxford: Oxford University Press.