Parfit, Long Termism and Existential Risk: and Nietzsche's Fearless ones.

Some people believe that there cannot be progress in Ethics, since everything has been already said. Like Rawls and Nagel, I believe the opposite. How many people have made Non-Religious Ethics their life’s work? Before the recent past, very few. In most civilizations, most people have believed in the existence of a God, or of several gods. A large minority were in fact atheists, whatever they pretended. But, before the recent past, few atheists made Ethics their life’s work. Buddha may be among this few, as may be Confucius, and a few Ancient Greeks and Romans. After more than a thousand years, there were a few more between the Sixteenth and Twentieth Centuries. Hume was an atheist who made Ethics part of his life’s work. Sidgwick was another. After Sidgwick, there were several atheists who were professional moral philosophers. But most of these did not do Ethics. They did Meta-Ethics. They did not ask which outcomes would be good or bad, or which acts would be right or wrong. They asked, and wrote about, only the meaning of moral language, and the question of objectivity. Non-Religious Ethics has been systematically studied, by many people, only since the 1960s. Compared with the other sciences, Non-Religious Ethics is the youngest and the least advanced.

I believe that if we destroy mankind, as we now can, this outcome will be much worse than most people think. Compare three outcomes:

(1) Peace.
(2) A nuclear war that kills 99% of the world’s existing population.
(3) A nuclear war that kills 100%.

(2) would be worse than (1), and (3) would be worse than (2). Which is the greater of these two differences? Most people believe that the greater difference is between (1) and (2). I believe that the difference between (2) and (3) is very much greater.
My view is held by two very different groups of people. Both groups would appeal to the same fact. The Earth will remain inhabitable for at least another billion years. Civilization began only a few thousand years ago. If we do not destroy mankind, these few thousand years may be only a tiny fraction of the whole of civilized human history. The difference between (2) and (3) may thus be the difference between this tiny fraction and all of the rest of this history. If we compare this possible history to a day, what has occurred so far is only a fraction of a second.
One of the groups who hold my view are Classical Utilitarians. They would claim, as Sidgwick did, that the destruction of mankind would be by far the greatest of all conceivable crimes. The badness of this crime would lie in the vast reduction of the possible sum of happiness.
Another group would agree, but for very different reasons. These people believe that there is little value in the mere sum of happiness. For these people, what matters are what Sidgwick called the ‘ideal goods’—the Sciences, the Arts, and moral progress, or the continued advance towards a wholly just world-wide community. The destruction of mankind would prevent further achievements of these three kinds. This would be extremely bad because what matters most would be the highest achievements of these kinds, and these highest achievements would come in future centuries.
There could clearly be higher achievements in the struggle for a wholly just world-wide community. And there could be higher achievements in all of the Arts and Sciences. But the progress could be greatest in what is now the least advanced of these Arts or Sciences. This, I have claimed, is Non-Religious Ethics. Belief in God, or in many gods, prevented the free development of moral reasoning. Disbelief in God, openly admitted by a majority, is a recent event, not yet completed. Because this event is so recent, Non-Religious Ethics is at a very early stage. We cannot yet predict whether, as in Mathematics, we will all reach agreement. Since we cannot know how Ethics will develop, it is not irrational to have high hopes.--Parfit (1984) Reasons and Persons, Section, 154. "HOW BOTH HUMAN HISTORY, AND THE HISTORY OF ETHICS, MAY BE JUST BEGINNING"

After my rant went viral this past weekend, Richard Pettigrew reminded me that if I really want to understand the intellectual roots of contemporary long-termism I can't ignore the last two pages of Reasons and Persons, quoted in full above. (Strictly speaking these are not the last two because the appendices are not wholly irrelevant, especially the material on social discounting.) And my hypothesis was that it would help explain, perhaps, something of the evident sectarianism that longtermism generates in its adherents. And when I returned to this material, I was, of course, immediately struck by the choice between 99% vs 100% extinction. And I started to think of objections.

But I hope you will forgive me if I step back first. For, I noticed that the hypothetical extinction scenarios are introduced with, and framed by, an implied contrast between [A] "Non-Religious Ethics" and Religious Ethics. The first oddity here is that the contrast is not drawn in the linguistically more accurate contrast between atheist ethics and theistic ethics (which in turn comes in monotheistic and polytheistic versions). Embedded in this contrast [A] is another explicit contrast (non-trivially indebted to Kant's "What is Enlightenment?") between [B] the "free development of moral reasoning" and the unfree development of moral reasoning (found in so-called 'Religious Ethics). And this, in turn, is connected to the contrast between [C] young ("recent past") and old ("more than a thousand years" etc.), which turns out to track [D] a contrast between the possibility of "greatest progress" and proven failure. And, again, this is connected to the contrast between [E] the "very few" and the very many. In addition, there is an implied contrast between [F] (a pacific) "a world-wide community" and the state of nature among nation-states. This, in turn is linked to the fact, [G] there is a to be hoped for possible consensus which contrasts with our existing cacophony. There are, finally, two further connected contrasts: (i) there is a contrast (again echoing Kant) between [H] irrational hopes (characteristic of superstition) with rational (or at least not irrational) hopes, and (ii) [I] the contrast between knowledge (of the past) and uncertainty (of the future).

As a two-fold aside, first, Parfit explicitly leaves open -- one need not be Straussian about this -- that this last contrast [D] may involve the practice of esoteric speech or dissembling: "A large minority were in fact atheists, whatever they pretended." One wonders what grounds this 'in fact.' And second, it is interesting that Parfit treats Buddha and Confucius as possible exemplars of this practice of pretending. To the best of my knowledge (and I know slightly more about the case of Confucius) this idea was popular in the wake of Bayle's linking of Spinoza to both Confucius and Buddha. (I have no idea if Parfit gets it directly or indirectly from Bayle, but in Bayle, too, we find the linking of these matters to esotericism.) And my guess is that Bayle shaped some of the orientalist tropes that helped found the study of comparative civilizations in the late nineteenth century.

I don't mean to suggests that nine-fold contrast [A-I] is exhaustive. Nearly all of these contrasts are not argued for in Parfit's book. (Please feel free  to correct me!) Such a list of contrasts is characteristic of what I call 'philosophic prophecy' by which I mean the structured ways in which concept formation by philosophers is aided by rhetorical or literary imagery that shapes the reception of (more rational) arguments and thereby can shape possible futures. And I would have given you a more detailed explanation of this claim in light of the typical nature of philosophic prophecy.

But then I noticed something remarkable. In the quoted paragraphs, Parfit is treating the 'death of God' as a world-historical thesis which gives ground to the greatest possible (non-irrational) hopes. This is a feature and not a bug because it also informs the very plausibility of Parfit's rather dramatic claim that the gulf between (2) and (3) is larger than the gulf between (1) and (2). (For a certain kind of theist this would be question begging.)

That the death of God is a world historical event -- in part for its effect on morality -- is a thesis associated with Nietzsche. And one of its canonical formulations is in Nietzsche's Gay Science, Book V, section/aphorism 343. So, it is no surprise -- although it shocked me -- that this very aphorism supplies Parfit with the epigraph to his book:

At last the horizon appears free to us again, even granted that it is not bright; at last our ships may venture out again, venture out to face any danger; all the daring of the lover of knowledge is permitted again; the sea, our sea, lies open again; perhaps there has never yet been such an ‘Open sea.

Parfit's epigraph also evokes the end of Book I of Hume Treatise as well as a famous passage of Socrates in Phaedo 99c–d, and all three cases (just like in Parfit) the very nature of philosophy, and its future, can be said to be at stake. Parfit himself reminds of his affinity with them in a passage in the book, "I would far prefer to have lived through the previous two and a half centuries, having had among my friends Hume, Byron, Chekhov, Nietzsche, and Sidgwick." As it happens Gay Science, Book V, section/aphorism 343 also is one of the texts that gave me idea of the existence of philosophical prophecy (since in it Nietzsche describes two kinds of such prophecy).

Crucially for present purpose is Nietzsche's attitude toward the rather grim immediate future: "This lengthy, vast and uninterrupted process of crumbling, destruction, ruin and overthrow which is now imminent: who has realised it sufficiently today to have to stand up as the teacher and herald of such a tremendous logic of terror, as the prophet of a period of gloom and eclipse, the like of which has probably never taken place on earth before?" Nietzsche feels not "sympathy" with the victims of history, "but rather like a new and indescribable variety of light, happiness, relief, enlivenment, encouragement, and dawning day? ... In fact, we philosophers and "free spirits" feel ourselves irradiated as by a new dawn by the report that the "old God is dead"; our hearts overflow with gratitude, astonishment, presentiment and expectation." (If you are shocked by this God is not dead for you.) In Nietzsche the price of progress is destruction of the old order so that we can breed (either culturally or genetically, i leave that aside) higher kind of men.

Now, notice that Parfit gives a classical utilitarian and elitist (that is, Nietzschean) argument for why 3 is worse than 2.* But that's not sufficient for his argument. He also needs to offer an argument for why 3 being worse than 2 is itself worse than 2 being worse than 1. But no such argument against "most people" is given. (I have some sympathy for the idea that philosophy is not a democracy.) We only receive an explanation for why total extinction (3) is worse than (2). And I do not think Parfit has an argument that is non-question begging because it presupposes the unfolding (as of yet "uncompleted") Death of God, that is, a kind of rational faith or hope that once completed will become self-evident. Among the fruits of this hope are the captains of longtermism.+

 

*I am not so sure he is right. Imagine being the last person on Earth. (Or the last 1% spread thinly accross the globe.) I suspect that it would be a lot worse than surviving extinction. (I have long thought that this held true for most of my family members that survived the Holocaust, so maybe I can't think clearly about this.)

+After I posted this Nickolas Delon called my attention to an essay by Richard Y Chapell that anticipates my conclusion albeit with a very different argument.

Comments

[Michael Kates]

Very interesting post. I must say that I find the last two pages of Reasons and Persons extremely moving. It's quite beautiful, in my view.

And I can't recall at the moment, but I believe Parfit gives a reason for including the Nietzsche quote as the epigraph of the book that doesn't relate to what you discuss in the post. Indeed, I believe he included it for a somewhat odd reason. Perhaps someone else knows.

[John Quiggin]

I don't like the cardinal scales required to say that diff(2,3)> diff(1,2). But I'm comfortable with uncertainty.

So, let's posit that a nuclear war wiping out 99 per cent of humanity would also destroy nuclear weapons and the knowledge of how to make them, and create a durable taboo against ever looking into this question again. And let's say (my best guess) that there is something like a 10 per cent chance of nuclear extinction this century. If you could press the 99 per cent button, would you do so? I wouldn't.

Having said that, I agree that there is something special about total extinction, so I guess I must place some weight on potential future people.

[Chris]

You mention a gap in Parfit's reasoning for the conclusion that the difference between 3 and 2 is greater than the difference between 2 and 1.

This gap is (mostly) filled by preceding discussion in the fourth section of Reasons and Persons. Namely, Parfit defends the claim that future people are axiological equals to present people (assuming they are guaranteed to exist). Separately, he argues against person-affecting views. Additionally, he argues that mere addition always constitutes an improvement.

These, combined with the Egyptology objection he discusses in R&P, entail that the gap between 3 and 2 is greater than that between 2 and 1, permitting reasonable empirical assumptions about humanity.

[Chris]

In reference to my previous comment, a condensed argument based off of Parfit's reasoning would go as follows.

1 - when a person exists does not vastly affect the moral value of their existence.

Argument for 1 - if 1 is false, the world would be made impartially better if present people received a minor benefit, and future people received a massive reduction in quality life. Example: bury nuclear waste more cheaply, but it will increase cancer rates of future people (assume they will still have lives worth living.)

2 - The fact that the identity of future people is up in the air does not undermine their value.

Argument for 2 - It would seem that a mother makes the world better if she takes on minor inconveniences so as to make it so that her child won't have an ailment, even if doing so changes the child's identity.

3 - If a population has extra happy people, it is better by an amount that is not highly variable over the other people.

Argument for 3 - if 3 wasn't true, the badness of many humans being harmed would depend on greatly disconnected people (Martians, or ancient Egyptians, for example)

By 1, 2, and 3, and because the future contains vastly many people, the loss in value when comparing 99 to 100 is much greater than 99 and 1, because going from 99 to 100 eliminates all future people, who matter to a comparable degree as present people.

[ERIC SCHLIESSER]

Hi Chris,
The question is about relative distances between 1&2 and 2&3, such that 2&3> 1&2. But you don't seem to take the suffering of 99% death rate from nuclear war really seriously (and the many centuries of subsequent sufferig from Nuclear fall out). That's not "minor inconvenience." Anyway, we're now entering repugnant conclusion territory, and I have nothing to add to that discussion that hasn't been explored in the literature.

[Michael Kates]

I remember! According to Jeff McMahan, Parfit wanted to use the picture he took of a ship in Venice as the cover (which, of course, he did), but he needed a way to explain it to the publisher. The Nietzsche quote then dawned on him, and so he wrote the last chapter to justify it! (This is from a wonderful interview of McMahan by Simon Cushing: http://profiles.cognethic.org/McMahan.pdf

[ERIC SCHLIESSER]

Thank you for that: Here's what McMahan says
"Well, I’ll tell you the story that Derek told me. I think Dave Edmonds has discovered something a bit different, but I’m quite sure that what Derek told me, I remember very clearly what Derek told me, which was, he wanted to put that particular picture of the ship in the harbor at Venice on the cover, so he thought, “how can I get this on the cover?” And he knew of this quotation from Nietzsche about “our horizons are open and we can set sail,” and all this kind of thing, so he thought well, I’ll use that quotation as an epigram, but then he thought well why would I have that quotation as an epigram, so he thought I’ll write a last chapter that says, “moral philosophy is in its early days.
Everything is open to us now as a result of the shift away from the dominance of religion in moral thinking.” And, so that’s what he said to me. Now, I think Dave has a different
story, but basically, what Derek told me was, he wanted to get the photograph on the cover, for that he needed the Nietzsche quotation, and to rationalize the Nietzsche quotation, he had to write this last chapter."

[Chris Minge]

Hi Eric,

Here is the more thorough reasoning I have in mind. 99% of humanity dying would mean 7.92 billion people die, leaving 80 million. If we assume the human population returns to 8 billion in the time it first took us to go from 80 million to 8 billion (4000 years), and that the human species will ultimately last as long as the average mammal species (1 million years, 200,000 already lived), and we assume 80 year average life-span, then there will be 9950*8 billion future people.

So to compare the numbers [in units of 8 billion people].
1 - nobody dies
2 - 0.99 of the population doesn’t exist which otherwise would exist.
3 - 9951 populations who otherwise would have existed won’t exist.

If future people are axiologically comparable to present people without large factors (my previous comment discussing how Parfit argues for this), then the gap between (2) and (3) is ~9950, while the gap between (1) and (2) is ~1, making 2&3 larger than 1&2 according to this analysis.

You are right to point out that there is more than just lives being lost in these scenarios. 99% die painfully in (2) and (3), and in (3), 1%, and many of their descendants, will likely have very poor lives because of nuclear fallout and related chaos. For Parfit’s point, I’m sure he would have been happy to revise it to a claim about a scenario in which the survivors aren’t harmed (less realistic, but it contains the relevant longtermist claim). Regarding the intitial claim, because 99% die of nuclear war in both (2) and (3), we can ignore that for the comparison 2&3. However, survivor-suffering is unique to (3).

However, I still think the conclusion follows from reasonable assumptions, even granting this detail. If we grant Parfit’s other arguments (controversial of course), then the original scenarios should be roughly equivalent (in terms of axiological evaluation) to the following.

There are 9951 isolated earths, each with 8 billion people, [all of whom are sterile], and everyone will live the average quality of life of the future people of scenario (1), likely better than present-day earth’s quality of life due to technological advancement.
1+: nobody dies.
2+: On one earth 80% die, 20% live in nuclear fallout, and the remaining 9950 earths are unaffected.
3+: All 9951 earths are eliminated in nuclear war.

The 1 earth lives of (2+) with nuclear fallout is how many would experience nuclear fallout lives in the original (2), assuming that it lasted for 1000 years. If the moral arguments go through, and 2+&3+ > 1+&2+, then 2&3 > 1&2. Reasonable judgments about the quality of life in nuclear fallout, and the significance of death, support 2+&3+ > 1+&2+.

Regarding calling it a “minor inconvenience”, I do not think Parfit’s arguments imply 99% of earth dying is a minor inconvenience. I would take it to show that 100% is an incredibly massive tragedy, while 99% dying is an incredibly massive, but not nearly as massive, tragedy (not an inconvenience). Considering that it is impossible to put into words how horrible 99% of earth dying is, it is naturally impossible to use words which can distinguish 99% and 100% while still accurately describing both (assuming Parfit is right).

Lastly, I should just note that this longtermist conclusion is compatible with rejecting the repugnant conclusion. Of note - Parfit (2016) “Can We Avoid the Repugnant Conclusion?”, Hajek & Rabinowicz (2021) “Degrees of Commensurability and the Repugnant Conclusion” (I do not know whether or not Hajek or Rabinowicz endorse other claims which rule out Parfit’s longtermist axiological conclusion, but this paper does not). I myself am inclined to accept such a view.

[Eric Schliesser]

I am not especially moved by the idea that possible distant future people are what you call axiological equals to present people (and find the arguments for this commitment unpersuasive). So, I feel bad that somehow my post made you spell out an argument in such detail that I don't think should even get off the ground.
But I will note that you don't really address my observation that I think that (2) is itself objectively a far worse situation than (3). And the only way you seem to get out of that is to start positing lots of possible lives lost (or lots of possible artworks never created) that somehow counterbalance that claim.
Anyway, thank you for taking the time to do so.

[Chris Minge]

Yes, arguing that distant future people are axiological equals is no easy task. Much of section 4 of R&P is just trying to do this (as is my brief second most recent comment), and it constitutes a long discussion. Not surprising that debate continues to this day.

Your observation of (2) 1+&2+ (which you very well may not share), and the axiological equality of future people (which you are not on board with). You're right that the many future possible lives are important for counterbalancing. This is precisely the empirical core of why Parfit thinks 3&2 > 1&2.

[Chris Minge]

">" signs made some of my previous comment not displayed. Should be "Your observation of (2) worse than (3) is addressed by 1+,2+,3+ conditional on the judgment that 3+ greater than 2+ (which you very well may not share), and the axiological equality of future people (which you are not on board with, but would make 2+,3+ analogous to 2,3 if true). You're right that the many future possible lives are important for counterbalancing. This is precisely the empirical core of why Parfit thinks 3&2 > 1&2."

[Aljosa Kra]

It seems that for Parfit the death of God plays double role here. On the one hand the death of God enables rational morality. A belief in moral progress (or rationaal hopes) is a part of that rational morality.

On the other hand, the death of God is itself an event in moral progress.

Taken together, this means that the death of God enables us to recognize the death of God as a step in the morally desirable direction. It's a kind of a revelation then.

[Matthew Adelstein]

I think that it would make sense to address the many arguments Parfit presents for the conclusion in the book rather than merely these few pages summarizing the conclusion. For a quicker presentation of some arguments for this conclusion, see here. https://benthams.substack.com/p/longtermism-is-correct-part-1

[Eric Schliesser]

Hi Matthew,
Those arguments have been addressed by many others. I did start reading your substack account. I stopped reading when you claimed "If you are going to create a person with a utility of 100, it is good to increase their utility by 50 at the cost of 40 units of suffering." It is pretty clear that (a) you are not restricting who is the subject of this suffering , and if they have a say in the matter; (b) this kind of reasoning is easily abused (as history has shown). Objections to this move are well known. (As I also note above, the near symmetry you posit between actually existing people and possible humans in the very distant future is also something I reject--again for well known reasons.)

[John Quiggin]

The argument about distant people is critical and represents a misunderstanding of utilitarianism. It's a political philosophy, based on the idea that everyone *in a given society* should count equally. Bentham is quite clear on this.

That's why the early utilitarians were Malthusians: if you accept Malthus' economics, population restriction is the only way to raise average living standards.

[Eric Schliesser]

Hi John,
Longtermists tend to reject the Malthusian restrictions -- in fact, Macaskill is all in on population growth -- because they think any non-suffering addition to the population is worth having on its own terms and has good effects on innovation, economic growth, and values.