Is Longtermism Such a Big Deal?

What is the attitude that most people might have about our responsibilities to distant generations, humans living far in the future? Many might say that they deserve some consideration but are not as important as people alive now. Occasionally, different attitudes to ethics and time are revealed. I was once in a seminar with a philosopher colleague with whom I tended to find a lot of agreement, and the question arose of whether present-day people should be willing to eventually get out of the way—leaving room and resources for new humans—if sudden medical advances meant that our current lifespans could be greatly extended. To me, the answer seemed an obvious, “Yes, we should.” To my colleague, the answer was an equally clear, “No.” What responsibilities could we have to merely possible people that might outweigh the interests of living, breathing humans?

My colleague’s response might have been on the extreme side, and it’s certainly common to give weight to people in the near-term future—think of the way future generations are invoked in discussions of climate change and economic policy. But the very long term, with a span of hundreds of thousands of years, might seem to fade into an unknowable fog. William MacAskill’s What We Owe the Future, along with other works written under the longtermist banner, aims to clear this fog and motivate us to think hard about the long term, as well as the present and the next few generations.

Longtermism has quickly become a movement. MacAskill’s book was publicly endorsed by Elon Musk (“a close match for my philosophy”), and it arouses a good deal of distaste and alarm in other quarters. In Salon, for instance, Émile P. Torres wrote of it as a “quasi-religious worldview” that asserts that “there could be so many digital people living in vast computer simulations millions or billions of years in the future that one of our most important moral obligations today is to take actions that ensure as many of these digital people come into existence as possible.” This, Torres argued, was “bizarre and dangerous.”

But Torres’s is not at all an accurate description of the movement. His is a destination that a long chain of arguments within the longtermist project could possibly lead us to if you add plenty of other assumptions—about the lives that “digital people” might enjoy, for example. But this is nowhere near where MacAskill’s book lands.

What the longtermists want us to do is think carefully about the distant future—which contains possibilities of human extinction, stagnation, vast populations, and so on—take those possibilities seriously in moral reflection and choice, and do so in a rigorous, quantitative way wherever we can. The project does conjure up a picture that will set off alarms for some people. We imagine philosopher-technocrats looking dispassionately over the tops of immediately pressing concerns, weighing enormous but abstract possibilities, and perhaps finding that their calculations give little weight to what previously seemed most important: the tangled and messy political problems of the now. But even if the alarms ring a little, isn’t this project worth a look?

What We Owe the Future, William MacAskill, Basic Books, 352 pp., $32, August 2022

MacAskill motivates his project by considering two kinds of distance between us and the people our choices might affect. Though it can take some argument, we do now tend to take seriously the interests of humans who are far away in space—people in distant countries we will never meet. We should do the same for those who are far away in time. They are, MacAskill writes, as real and as deserving of consideration as those across the globe: “Distance in time is like distance in space. People matter even if they live thousands of miles away. Likewise, they matter even if they live thousands of years hence.”

The idea that future people are distant in time in a way that’s comparable to distance in space is one that quite a few physicists and philosophers would say is factually correct. We are often told that time is another physical dimension, akin to the spatial dimensions, and people in the future are real parts of the universe. But it’s also true that whoever is around in the future will be there as a consequence of the choices we make now and tomorrow. Some of these choices would result in fewer, more, or no people at all being around a thousand years on.

How can future people be real, merely distant, and yet have an existence dependent upon our current choices? There’s no paradox here, because our choices, along with their consequences, can also be seen as laid out in the grand mosaic of events. But we can’t approach our own decisions that way, as if making our own choices is a process of discovering which decisions will exist in the near future. Making choices, deciding what to do, requires that we see ourselves as more active than that. When we choose, we set some processes in motion and rule out, or make less probable, many others. From this point of view, future people affected by our decisions are not just out there, as individuals, waiting to be counted. Their existence is contingent on the choices we make, and we are trying to work out which choices to make.

We learned not to ignore people across the world, and MacAskill’s aim is that we learn not to ignore another kind of distance. But being in the future is not just another way of being far away. Moral concern about future people, especially many years down the line, is, or can be, different from other kinds of moral concern. We saw that in my exchange with the colleague at the start of this review. My colleague said we don’t have a responsibility to people who don’t yet exist. Maybe we don’t, but we can still decide that it would be a good thing, better than other options, to act in a way that sets in motion some futures rather than others.

Suppose we agree that long-term consequences should be factored into our decisions. A next question is: What difference does this make? Suppose one person says that the future is so unknowable (imagine how 2022 would look to someone even a century or so ago) that there’s not much to gain from lifting our attention from the near and medium terms (the scale of decades, say). And someone else insists that our present actions will have long-term effects that are so enormous that we simply must try to factor them in. I want to know what decisions would be made differently by those two people. What kind of trade-off might there be between our familiar concerns and these new ones? Here we reach a surprise—for me, at least—in MacAskill’s book. The differences don’t loom very large at all.

Quite a lot of MacAskill’s book is not about topics of this kind—which things we might do differently. A long early discussion looks at moral change and progress. This includes a detailed look at some stages in the abolition of slavery, especially at how reflection and moral commitment made a difference. Slavery, though, was always the most immediate of moral issues. Its harms, or many of them, were not at all deferred.

MacAskill also discusses climate change, and says at one point that decarbonization of our energy supply is a “proof of concept for longtermism”: If someone doubts that we can give good arguments about policy based on long-term effects, this should convince. Perhaps, but by MacAskill’s own lights, this is another case where good choices look similar using both short- and long-term reasoning, because decarbonization is a “win-win,” as he puts it. It will reduce air pollution and stimulate innovation in the short term, as well as help in the long term. It’s not a case that shows how longtermists might act differently.

Here and elsewhere, the longtermist perspective generates new reasons to do some things, even if we’d already want to do them. The new reasons often have to do with averting various kinds of total collapse. We should leave some accessible coal in the Earth’s crust, for example, because we might need to burn it later if human technology nose-dives and must recover. But it would not be a good idea to burn the coal now even if we knew the collapse would never happen. MacAskill also thinks we should reject the idea that having children is an unethical choice. But again, he thinks that having children looks good from both short-term and long-term perspectives. The longer-term benefit comes from the need to reduce the chances of a future population collapse.

Setting aside the question of trade-offs, one policy area where time does matter in interesting ways is around what MacAskill calls “lock-in” of values. Lock-in is the closing of debate and questioning around moral issues, making decisions about values that leave no room for revision. MacAskill thinks—more than I’d have expected—that the lock-in of values is a fairly likely future event, occurring perhaps through advances in technology and the centralization of power. What we should do is delay it, if we can’t prevent it, or at least try to prevent “bad” values from being locked in.

The idea of preventing a lock-in of “bad” values, as a goal, shows why this issue is not straightforward. We’ll always want to prevent what we now think are bad ideas from taking hold. But if values do eventually change in a way that is reasoned rather than compelled, future people will surely look back and be glad of the changes. We look back at earlier periods today and are very glad that slavery, to use MacAskill’s example, is no longer seen as defensible. If values change again, people in the future are likely to feel the same way as they look back on our present.

The big question, then, is not whether we should try to avoid the lock-in of “bad” values, but whether we should regard lock-in as undesirable in general. On some issues, it seems we might fervently hope for lock-in: democracy, free expression, rejection of racism. Here, we might want to say, any future move away would be a move in the wrong direction. (As an aside, I wish that recognition of the value of individual liberties and free expression had been a bit more locked in, or at least had more inertia, over the last couple of years.) MacAskill, though, thinks that it’s reasonable to only want lock-in of some very broad values that bear on how decisions are reached—free speech and free migration, for example. He would work against lock-in more generally, even if we think we now have some things right. Earlier lock-ins would, in general, have been mistakes, and that weighs heavily with him. I tend to agree, though a genuinely locked-in future also seems less likely to me than to MacAskill.

What We Owe the Future is a good book in a lot of ways. It includes a particularly good introduction to “population ethics,” an important and rather intricate part of recent philosophy. I have expressed my reaction to the absence of striking practical differences between short- and long-term perspectives as one of surprise, and in one way as an intellectual disappointment—it was a surprise how little in the way of notable new choices came out of putting the long-term front and center. But in many ways, disappointment is the opposite of the reaction we should have. We should be glad if many important goals make sense in both the short and long terms: eliminating factory farming, protecting the environment, reducing economic inequality, protecting democracy and individual liberties. It is good news, even if a less intellectually exotic outcome, when the differences between short- and long-term choices are not too acute.