Drowning in the Shallows

It must be hard to cover philosophy as a journalist. Your subject hardly moves. Not only does it fail to move the world, it barely budges from its study. Thus the movement called Effective Altruism has been a godsend for philosophical journalists like David Edmonds, best known for co-writing the best-selling Wittgenstein’s Poker in 2001. Billionaires; TED Talks; young people at elite universities eager to commit themselves to doing good by doing well, financially; hundreds of millions of dollars raised. A bonanza of tempting themes.

Edmonds’s new book Death in a Shallow Pond chronicles the rise of Effective Altruism while rooting that movement in a single thought experiment by the influential philosopher Peter Singer: the Shallow Pond. Edmonds seems impressed and sometimes convinced by the core principles of Effective Altruism, which include the moral imperative for charitable giving to others and the need for that giving to be calibrated for maximum effect. But because Death in a Shallow Pond appears at a moment when the Effective Altruism movement has been attacked and, in the eyes of some, fully discredited, the book reads as a defense of the altruists against their critics. What could be wrong with giving? I will suggest that the problem is not altruism, ultimately, but the machinery of effectiveness.

Edmonds begins with Singer’s original thought experiment: You are walking to work in the morning, and you pass a shallow pond in which a child struggles not to drown. You can save the child at the cost of some light personal distress — damaging your fine shoes or clothing, say — and you obviously should save them.

As Edmonds acknowledges, this is not much of an experiment. Singer simply meant that our (“our” implying decently well-off people living in the developed world) obligation to people suffering far away, in great poverty or facing rampant illness or war, is akin to our obligation to the drowning child. Only the myopias of comfortable living could hide this from us. So the Shallow Pond is less a thought experiment than an extended “gotcha” exercise — “caught you out, didn’t I, you bourgeois oaf,” it seems to say. Philosophers and social theorists, including David Hume and Adam Smith, have long reflected on how geographic distance may limit our sympathy to far-off suffering. The Shallow Pond is an effort to expose our moral psychology and take us to task over our limitations. Morality, Singer holds, is about the consequences of our actions and not about the mindset in which we take those actions. Nor does it matter if we ever witness the good we do, up close and personal, or if we are ever thanked.

The Shallow Pond experiment asks us to accept the moral equivalency of two types of good-doing: saving a child’s life with our hands and donating money so that a stranger’s life (or, potentially, a lot of strangers’ lives) is saved from disease or malnutrition elsewhere in the world. Effective altruists spend considerable time trying to calculate just how much good they can do with the resources available to them, but it is worth considering a basic feature of Singer’s Shallow Pond. Singer is committed to the equivalency of saving with our hands and saving with our wallets. This, in turn, reflects his utilitarianism, one feature of which is consequentialism. Utilitarians view consequences (such as lives saved, happiness produced, suffering prevented) as the only morally significant things; actions have no intrinsic moral value for them. Utilitarianism has other features too: universalism, or preferring no one’s interests over anyone else’s; welfarism, meaning considering whether or not people’s needs are satisfied. Lastly, Utilitarianism is aggregative, meaning that it measures the well-being of all persons, seeking to minimize suffering and maximize happiness for as many as possible. Not all effective altruists are utilitarians, but as a movement Effective Altruism has been strikingly influenced by Utilitarianism, and particularly by its interest in outcomes and measurement. It is always better to save more children, or more adults or — depending on your attunement to nonhuman animals — more chickens.

The Shallow Pond shares a basic feature with Singer’s well-known arguments on behalf of nonhuman animals: the effort to catch people in their moral slumber. His 1975 book Animal Liberation argues that the only thing that allows us to treat nonhuman animals immorally (by treating them cruelly as food animals and by experimenting on them for medical and cosmetic research) is “speciesism,” a term coined by the experimental psychologist and animal-protection activist Richard D. Ryder. Singer wants people to abandon their speciesism and bring nonhuman animals within their circle of moral concern.

The Shallow Pond and Animal Liberation both seek to help the reader locate a prejudice within themselves and overcome it. Those who contemplate the Shallow Pond can act more altruistically through charitable giving. Readers of Animal Liberation might cease to eat animals and campaign to end factory farming, as well as animal experimentation. Philosophers have long worked this way, trying to find moral failure within themselves and in others, and helping people to improve themselves. Historically speaking, utilitarians have opposed slavery and promoted women’s liberation; their philosophy has been very useful for cutting through beliefs and prejudices that impede social and moral progress. I know readers of Animal Liberation who are sure that, in decades to come, eating animals will look as morally wrong as owning other humans under slavery.

Singer published the Shallow Pond thought experiment quite early in his career, as a very small portion of a 1972 article called “Famine, Affluence, and Morality” in the journal Philosophy & Public Affairs. Much of Edmonds’s book deals with the decades-later development of Effective Altruism through the work of two latter-day Oxford students, Toby Ord (an Australian, like Singer) and the Scottish William MacAskill. The movement Ord and MacAskill have built since the mid-2000s (the term Effective Altruism stuck in 2011; the previous organizations were called Giving What We Can and 80,000 Hours) has grown and grown, attracting membership around the world. Elite colleges and universities often have local chapters.

The movement’s presence in Silicon Valley is palpable; I first encountered effective altruists (or EAs) when I was researching efforts to grow meat in laboratories, which, if successful, could radically reduce animal suffering by eliminating industrial-scale animal agriculture. More conventional effective-altruist giving might be directed toward treating diseases in underresourced parts of the world or eradicating hunger, which are obviously noble goals. Many EAs take a “giving pledge,” such as promising to contribute 10 percent of their earnings to worthy (and efficacious) causes; the giving pledge has received some public notice because some very wealthy people, including Bill Gates, have made their own versions of it. The altruism of Effective Altruism is very much like what Singer promoted in 1972; the emphasis on effectiveness, on the other hand, seems to me a different kind of inflection, precisely because it leads to dreams of scale. One conspicuous feature of Effective Altruism has been helping young elites choose careers that empower them for moral action. The slogan is “Earning to Give,” after all.

As Edmonds explains, some EAs go far beyond the intentions of Singer’s Shallow Pond by embracing what is called “Longtermism.” Longtermists are inspired by the work of another Oxford philosopher, Derek Parfit, who argued in his 1984 Reasons and Persons that not-yet-existing human beings should have moral standing in our choice-making. The Shallow Pond is about caring for people far away: Longtermists replace space with time. This opens up a Pandora’s box of futurisms, which emerge and grow to a scale typically treated by science-fiction writers. Longtermists think about high-stakes issues like the extinction of the human species, or our expansion throughout our galaxy, or the development of self-aware and powerful artificial intelligence that might help us or kill us. Existential risk provides all the gravity longtermists need, because it means that the actions we take to reduce risk have considerable moral standing — potentially more than helping our fellow beings suffering in the here and now.

Parfit believed that we live during an historical phase of great long-term importance because of the changes technology brings. He called this “the hinge of history.” In the hands of some less intellectually serious futurists, this view produces dreams of a “singularity” arriving at some point in the mid-21st century, either welcomed or feared depending on whether you think a godlike AI will smile upon us or smite us. Longtermists want to act correctly during the “hinge” to ensure the ongoing flourishing of the species, which may settle other planets and produce more billions or trillions of people capable of happiness. Their happiness must be protected from pandemics, from all manner of natural disasters, and yes, from potentially harmful AIs. (Edmonds notes, helpfully, that Singer has kept his distance from the longtermist version of EA, preferring “Mali to Mars,” the developing world over worlds to come.)

So far, so weird. But the strangeness of Longtermism has not inflicted the same level of reputational harm on Effective Altruism as Sam Bankman-Fried, whose disastrous career in cryptocurrency followed from coaching he received from MacAskill while Bankman-Fried was an undergraduate at the Massachusetts Institute of Technology. Bankman-Fried, the child of Stanford Law professors themselves sympathetic to Utilitarianism, wholeheartedly embraced Effective Altruism and its goal of “earning to give.” One reading of his financial success coupled with his well-known perpetration of fraud is that Bankman-Fried simply did not believe that actions themselves were wrong if they could lead to productive acts of charity. This is a vivid version of a problem that arises from both Utilitarianism and Effective Altruism. Who gets to act as the judge and administrator of moral good? Who gets to tally up all that happiness and suffering and decide what to do? How can we trust their judgment?

In recent years, and beginning before the Bankman-Fried case, many criticisms of the Shallow Pond and Effective Altruism have appeared. Edmonds devotes the second half of Death in a Shallow Pond to those criticisms, each getting a separate short chapter. Many of these are technical in character: Can we really calculate the good we do effectively? Indeed, Edmonds takes more seriously the challenges to Effective Altruism from developmental economists such as Angus Deaton than he does the philosophical challenges. I was reminded of the philosopher Bernard Williams’s well-known “A Critique of Utilitarianism,” (which Edmonds does cite but does not discuss at length) in which Williams argued that utilitarians turn morality into a matter of technical complexity precisely because they are afraid of moral ambiguity. This might apply to the overlapping set of effective altruists too; some of them may be people who want there to be morally correct answers, the more precisely calculable, the better.

One of the most interesting criticisms targets not Effective Altruism as a movement but thought experiments in general. They are, Edmonds explains, designed to expose our moral intuitions, but they can instead focus our attention on “edge” cases that generate more heat than light. Thus Edmonds recounts that in one debate, MacAskill insisted that a person in a burning building should save a Picasso painting rather than a child, on the grounds that the money from the sale of the painting could be used to save more children later. MacAskill seems to be saying that our moral intuition (save the kid!) tricks us into doing the wrong thing. But does it? A similar logic would excuse the longtermist entrepreneur who passes a child drowning in a shallow pond and who decides that impressing a venture capitalist with their clean shoes matters more than saving a child’s life. Think of all the lives venture-capitalist money will ultimately help them to save! Personally, I think that the moral intuition that one should help another human being in the present moment is correct. The suspension of moral intuition through thought experiments is what ultimately lets us down. Do you want to trust the person whose head is full of math instead of feeling? Myself, I would prefer a less-abstract pattern of moral responsiveness — call it analog morality, if you will, the kind that happens in our real lives outside of thought experiments.

The critique of Effective Altruism that I find most pressing comes from the point of view of democracy. In his 1996 book Living High and Letting Die, which anticipated some aspects of Effective Altruism, Peter Unger advocates for charitable giving at a morally and financially athletic level, beyond what Singer suggested in his 1972 article. In a harsh review of Unger, Martha Nussbaum pointed out that if everyone in the world suddenly gave a sizeable percentage of their incomes to global NGOs, those organizations would acquire power beyond the scope of governments. Nussbaum’s criticism of Unger applies to Effective Altruism, too. Effective Altruism has attracted some spectacularly wealthy people, and while their charitable giving (think of the Gates Foundation) may seem positive on its face, there is the question of how very large-scale giving intersects with government and democratic processes. It is the administrative question again. Who, after all, gets to decide what counts as an ultimate good, and thus a good use of funds? Do the people receiving aid have some say in the way aid reshapes their society? The potter’s hands are on the clay, but the clay is other people’s lives.

Edmonds calls Singer, Ord, and MacAskill “activists.” He credits Singer, who has written critically about Karl Marx, with following Marx’s famous Eleventh Thesis on Feuerbach: Philosophers must change the world. Perhaps crediting the effective altruists with activism is an easy way to circumvent the more complex story of the relationship between intellectuals, money, and power. Edmonds’s skillfully written book shows that Effective Altruism did not originate in the clean room of a thought experiment but in the matrix of compromise we call a social world. Effective Altruism emerges from conversations among elites and has necessarily been molded by the pressures on them and the opportunities afforded them.

After reading Death in a Shallow Pond, I was left thinking that Effective Altruism is in fact the perfect philosophy for our time, given that our time is wallpapered by instrumentality, afflicted with a craze for the STEM-ish value of measurability, and obsessed with something — I am still murky about what it is — called “impact.” Can anything stand as an end, in and of itself, anymore, or are we content with treating everything as a means only?

I am skeptical about the moral failings that seem to me to follow from the relentless pursuit of effectiveness. And yet the philosophy of altruism itself retains a certain power. Edmonds has convinced me that we should not ignore Singer’s original provocation, the Shallow Pond, even if it fails to critique the economic and political configurations that make altruism necessary. Our goodness, after all, is a matter of personal comportment and self-reflection, and not simply a function of our life under the systems that shape us. We retain agency, even if we do not make choices under conditions we choose.

The thought experiments of moral philosophy are useful for helping us to stay clear about what we value and what we mistrust. I mistrust a mind that abstracts itself from immediate moral issues — like a drowning child — because of the temptation to imagine that saving their clothes, or saving a few minutes, will allow them to help more people later on. I mistrust where the moral psychology of a person who thinks this way may take them, and I do not think that morality “scales” well. After all, the pond before us is shallow. Would I trust them with the ocean?