Suffering in the MultiverseABSTRACT
The Abolitionist Project outlines how humans and transhumans may use biotechnology to abolish suffering in all sentient life. The last experience below hedonic zero in our forward light-cone may be a precisely dateable event a few centuries hence. Sadly, this utopian-sounding outcome may not be nearly as wonderful as it sounds. For if a "block-universe" conception of spacetime is true, then suffering occurring in what we naively call "the past" is as real and unalterable as what we call "the present". Moreover, post-Everett quantum mechanics suggests that Darwinian life abounds elsewhere in the Multiverse. In the vast majority of quasi-classical macroscopic branches in which sentient life arises, no hominin-like creatures will evolve capable of rewriting their own source code and abolishing pain, misery and malaise. So "future" suffering persists indefinitely too. Worse, if Linde's chaotic eternal inflation scenario should turn out to be true, then the amount of suffering in Reality is increasing exponentially. Its extirpation in any one pocket universe like our own would be a purely local phenomenon. The only crumb of comfort to be drawn from this analysis is that the scenarios sketched are all extremely speculative.
The Abolitionist Project lays out the case for abolishing suffering via biotechnology; and cautiously predicts that our posthuman descendants will live, in effect, happily ever after. A heart-warming tale? Yes, in a sense. However, any rosy conception of the world which this scenario inspires is potentially misleading. Here are three depressing reasons why:
First, on a "block universe" interpretation of the world - mandated by the Theory of General Relativity - the Darwinian Era perpetually occupies the space-time coordinates it does. The pain and suffering of primordial life can't be erased. At best, we are poised merely to determine its boundaries. A full scientific understanding of time remains elusive. Yet barring some unimaginable revolution in our entire logico-conceptual scheme, rational agents can't extinguish the frightful events happening elsewhere in space-time. The past is fixed and unalterable. Admittedly, the feasibility of backward-causation suggested by quantum-mechanical "delayed-choice" experiments is a tantalising complication to this generalisation. Some advocates of time-symmetric causality like Lev Vaidman stress how well the two-state vector formalism dovetails with Everett's multiverse. Quantum cosmology suggests that future and past quasi-classical histories alike are non-unique. But we can safely state that even the most godlike of our post-human successors can't eradicate their terrible origins.
In practice, even professed utilitarians are much more relaxed about tragedies occurring in what we call the past, especially the distant past, than about what unfolds in the future. This statement of human psychology is reflected in our asymmetric attitudes to our own past and future pain. Compare one's joyful relief on leaving the dentist with one's dread at an imminent dental appointment. By the same token, mature posthumans - for whom the Darwinian Era belongs to distant antiquity - may find the ghastliness of their ancestors' suffering seems less important, "less real" - than everyday Heaven-on-Earth, assuming (problematically) that future life chooses to contemplate its dreadful birth-pangs at all. But whether such primitive nastiness is commemorated or forgotten, the horrors of Darwinian life are a fixture of Reality; and these horrors are not diminished by spatio-temporal distance. Sub specie aeternitatis, all here-and-nows are equally real.
Secondly, our best fundamental theory of the world is quantum mechanics; and our best understanding of the quantum formalism suggests that we live in a Multiverse rather than a classical universe. Post-Everett quantum mechanics (i.e. the universal Schrödinger equation or its relativistic analogue without any ill-motivated "collapse of the wave function") discloses the existence of a multitude of macroscopic branches rather than a single unique history. The fact that most of these classically inequivalent branches interfere only minimally with each other explains the popular soubriquet "Many Worlds", though the term can mislead the unwary. In the vast majority of these macroscopic world branches, no complex structures can arise, let alone sentient life; the coupling constants of the forces of Nature and other "fundamental" parameters in such branches are wrong. Their sterility still leaves googols of branches where information-bearing self-replicators evolve via natural selection. Critically, in only a small minority of these populated branches of the Multiverse can intelligent agents arise who are able to eradicate the biological substrates of their own suffering. In branches where, for example, a meteorite didn't wipe out the non-avian dinosaurs, Darwinian life "red in tooth and claw" presumably continues indefinitely. This is because only language-using tool-users can ever master the rudiments of science; and then go on to devise the biotechnology needed to rewrite their own genetic code and redesign their global ecosystem. To the best of our knowledge, no reptile could ever do this. Yes, we should beware of naïve, anthropocentric definitions of intelligence; but this cognitive constraint rules out self-emancipation in the overwhelming bulk of branches of the Multiverse that support life.
To advance such a conjecture isn't dogmatically to claim that only members of the genus Homo could ever initiate a post-Darwinian transition. Passage through this choke-point may be possible via species in other biological taxa thanks to the phenomenon of convergent evolution. We simply don't know. Thus if ape-like marsupials were ever to evolve in Australia, for instance, then it is possible that one species would also have stumbled on the suite of adaptations needed to liberate their own phenotypes and then the rest of the living world. But either way, most life-supporting branches of the Multiverse are inaccessible to techno-scientific agency. And TV sci-fi dramas aside, we can't do anything about life in this (comparative) abundance of god-forsaken worlds. Interstellar rescue missions are in theory feasible if sentient life exists elsewhere in our galaxy, and maybe even our local galactic supercluster. [Unless our understanding of physics is fundamentally wrong, the accelerating expansion of the universe precludes full-blown cosmic-engineering] But we can't tamper with other branches of the universal wavefunction. The evolution of the universal wavefunction is continuous, linear, unitary and deterministic. One may hope modern physics is mistaken; but if it isn't, we're stuck.
What might be the practical implications of post-Everett quantum mechanics for intelligent moral agents? One practical implication of the reality of other macroscopic branches might be to compel a systematic reassessment of our notions of "acceptable" risk. Recognition of the freakish unlikelihood of various desired outcomes does not stop most of us playing the National Lottery; but the converse doesn't hold. Thus we are accustomed to thinking that various nasty scenarios are of negligible likelihood, and even vanishingly small possibility; and then disregarding them altogether in the way we behave. Yet if a realistic interpretation of quantum mechanics is correct, then all these physically possible events actually happen, albeit only in low-density branches of the universal wavefunction. So one should always act "unnaturally" responsibly, driving one's car not just slowly and cautiously, for instance, but ultracautiously. This is because one should aim to minimise the number of branches in which one injures anyone, even if leaving a trail of mayhem is, strictly speaking, unavoidable. If a motorist doesn't leave a (low-density) trail of mayhem, then quantum mechanics is false. This systematic re-evaluation of ethically acceptable risk needs to be adopted world-wide. Post-Everett decision theory should be placed on a sound institutional, research and socio-economic footing, not just pursued by responsible Everettistas on an individual basis. The ramifications of the Everett interpretation of quantum mechanics are ethically too momentous for purely private initiative. Our moral intuitions fail because natural selection equipped us to deal with a classical world rather than a Multiverse. Human beings tend to discount "remote" risks by treating the probability of such events as zero. Ultimately, perhaps ethical decision-making should be performed by quantum supercomputers doing felicific calculus across world-branches; quantum ethics may be computationally too difficult even for enhanced post-human brains. For it's worth stressing that Everett's relative state interpretation of quantum mechanics doesn't propose "anything goes". The branching structure of the Multiverse precisely replicates the probabilities predicted by the Born rule. There are no branches supporting civilisations in the middle of the Sun. Nor are there any branches where, for instance, one of the world's religions is true (as distinct from believed to be true): Everett is not a theory of magic. But the universal wavefunction does encode hell-worlds beyond our worst nightmares, albeit at very low density.
Perhaps it's worth noting, too, that many physicists still don't accept Everett, or at least suspend judgement. Yet this is typically more out of incredulity at what the equations [and experimental evidence] are telling us, not because of any evidence that the unitary Schrödinger dynamics breaks down at large scales.
Thirdly, some speculative work in contemporary theoretical physics suggests that even the multiverse of Everettian quantum mechanics doesn't remotely exhaust the totality of suffering. For there may be googols of other multiverses. Suffering may exist in other post-inflationary domains far beyond our light cone; and in countless other "pocket universes" on variants of Linde's eternal chaotic inflation scenario; and in myriad parent and child universes on Smolin's cosmological natural selection hypothesis; and among a few googols of the other 10500+ different vacua of string theory; and even in innumerable hypothetical "Boltzmann brains", vacuum fluctuations in the (very) distant future of "our" Multiverse. These possibilities are not mutually exclusive. Nor are they exhaustive. Thus some theorists believe we live in a cyclic universe, for instance; and that the Big Bang is really the Big Bounce.
Of course, the theories alluded to above are highly speculative. They are far removed from our everyday experience and experimental test. Even if one or more of these theories is correct, it is tempting implicitly to suppose that the suffering of sentient beings occupying such realms is (somehow) less real than our own: metaphysical theories imply, in some sense, only metaphysical suffering. This comfortable assumption would be wrong-headed, not to say complacent. If any of the above hypotheses are substantively true, then the suffering of victims embedded therein is no less real than our own. Moreover in the case of other branches of "our" multiverse, it's debatable whether the branches are even "metaphysical". Not merely is their existence implied by empirically well-attested theory. Strictly speaking, interference effects from other quasi-classical branches never disappear; they merely become vanishingly small. Interference effects between different "worlds" can in principle be quantified by decoherence functionals. Their inferred real existence isn't just airy philosophising.
* * *
Faced with this fathomless immensity of suffering, a compassionate mind may become morally shell-shocked, numbed by the sheer enormity of it all. Googolplexes of Holocausts are too mind-wrenching to contemplate. We might conclude that the amount of suffering in Reality must be infinite - and hence any bid to minimise such infinite suffering would still leave an infinite amount behind. A sense of moral urgency risks succumbing to a hopeless fatalism.
Thankfully such moral defeatism is premature. For it is not at all clear that physically realised infinity is a cognitively meaningful notion. Infinities that crop up in the equations of theoretical physics have always hitherto turned out to be vicious; and yield meaningless results. Doubts about physically embodied infinity arise, not because one presumes to "tell God or the Devil how to construct the world" ("Einstein, stop telling God what do do" - Niels Bohr) but because of doubts that the claim of physically realised infinity is well-defined or even intelligible. Granted, some kind of "as if" platonism - and "Cantor's paradise" [or demonic bestiary] - may be mathematically fruitful. Yet it's doubtful whether Reality supports any abstract objects, let alone an ontology of physically realised infinities, notionally large or small. If Reality is indeed finite rather than infinite, then the suffering in the world is presumably infinitesimal compared to truly infinite suffering. So we should be grateful for small mercies. But the sheer scale of such suffering as indubitably does exist still swamps human comprehension. Mercifully, we can't grasp the potential real-world implications of our own notation.
The tenor of this narrative would be contested by many non-utilitarians. Why focus exclusively on suffering? Lighten up! There is so much more to life. Why not think about life's joys? Temperamental optimists will tend to have mood-congruent thoughts about the plenitude of unsuspected wonders that modern physics' expanded vision of Reality entails rather than focus on the nasty side of existence. But if you are an ethical utilitarian, then the relative importance of anything isn't a mere subjective value-judgement but a matter of objective fact, written into the fabric of the world. Extreme emotional intensity of experience morally matters most. Since the extremes of suffering dwarf the mundane pleasures of Darwinian life, they should presumably dominate any narrative outline of its features. And Darwinian life is statistically far more common in the Multiverse than post-Darwinian life.
A classical utilitarian might respond that it is more appropriate to focus on the unimaginable glories of our superhappy descendants rather than dwelling on the nastiness of Darwinian life. Yes, branches supporting such sublime superhappiness may be unrepresentative of sentience in the Multiverse as a whole - though the numbers become complicated if superintelligence hypothetically converts the accessible universe into blissful computronium. But assuming that the intensity of well-being of posthuman superbliss will surpass the comparatively dim awareness of ancestral minds, possibly by several orders of magnitude, then such superbliss matters far more than dim Darwinian consciousness too. As a consequence, posthuman superbliss should dominate our narrative. Analogously, any history of contemporary life on Earth should focus, not on its inordinate number of beetles, but on human beings. The negative utilitarian, for whom minimising suffering is the absolute moral priority, will of course find such a response unsatisfactory. There is nothing dim about Darwinian consciousness if, say, you are a grieving parent who has just lost a child. Or, more prosaically, if you have a toothache.
This discussion contains a controversial assumption which if confounded will make the story sketched here even darker. The controversial assumption is that when intelligent agents have attained the technical means to abolish the biological substrates of suffering, they will almost invariably do so. Thus by implication, suffering will be abolished in the great preponderance of branches where humans [or their functional counterparts] decipher their own genetic source code and develop biotechnology. A subsequent cross-branch reproductive revolution of designer babies is effectively inevitable. This generalisation might seem an extraordinarily reckless prediction. Forecasting is perilous enough even if one is a classical one-worlder. So predicting that a highly speculative scenario (i.e. the abolition of suffering) will eventually play out in the vast bulk of branches of macroscopic worlds with inhabitants attaining our level of technological development - and conversely, predicting that only a vanishingly small density of such branches will retain suffering indefinitely - might seem foolhardy in the extreme. Perhaps so. Recall how opiophobia still retards the medical treatment of even "physical" pain. Yet let's suppose instead that the analogy with anaesthetics holds up. After the discovery of general anaesthesia, its surgical use was contentious for a decade or two. But pain-free surgery soon became universally accepted. In our current state of ignorance, there is no way we can rigorously calculate the probability density of branches of the Multiverse where anaesthesia was discovered and rejected. But at worst, it's fair to say the proportion of branches is extremely small. Branches where governments outlaw pain-free surgery aren't sociologically credible. Of course the abolition of psychological distress is a less clear-cut case than anaesthesia. Technologies to abolish mental pain are in their infancy. But let's assume that in future they can be made as technically clean and successful as surgical anaesthesia. In what proportion of such branches will some or all people reject mental superhealth indefinitely? Again, a case can be made (though it won't be attempted here) that the proportion will be vanishingly small. Unfortunately, the proportion of life-supporting branches of the Multiverse whose dominant species reaches this stage of technical development is extremely small too. So the anticipated local success of the abolitionist project touted here is not as wonderful news as it sounds.
What practical lessons, if any, should be drawn from this bleak analysis of Reality? Assume, provisionally at any rate, a utilitarian ethic , or simply the weaker assumption that, other things being equal, suffering is best avoided. The abolitionist project follows naturally, in "our" parochial corner of Hilbert space at least. On its completion, if not before, we should aim to develop superintelligence to maximise the well-being of the fragment of the cosmos accessible to beneficent intervention. And when we are sure - absolutely sure - that we have done literally everything we can do to eradicate suffering elsewhere, perhaps we should forget about its very existence.
(2008, last updated 2015)
Leiden im Multiversum (German)
El sufrimiento en el Multiverso (Spanish)
Sofrimento no Multiverso (Br. Portuguese)
Social Media (2022)
The End Of Suffering
The Pinprick Argument
The Abolitionist Project
The Reproductive Revolution
Physicalistic Idealism (2015)
Infinite Ethics (Nick Bostrom)
On Infinite Ethics (Joe Carlsmith)
Quotations On Pain and Suffering
Abolitionism: a definition of terms
The Abolitionist Project (podcast; 15.3 MB)
Population Ethics and the Repugnant Conclusion
Better Never To Have Been: The Harm of Coming into Existence
¿Una ética cuántica? El sufrimiento en el Multiverso (Spanish translation)
The Everett Interpretation of Quantum Mechanics FAQ (Michael Clive Price)
Ética Quântica? Sofrimento no Multiverso (Brazilian Portuguese translation)