What this covers

Eric Schwitzgebel and Sean Carroll discuss the claim that our most fundamental theories of reality—from quantum mechanics to consciousness—are irreducibly weird, and that this weirdness should reshape how we evaluate competing explanations. Schwitzgebel argues that Universal Bizarreness and Universal Dubiety are both inescapable: every interpretation of quantum mechanics violates common sense, and every theory of consciousness does likewise. Since all available options are bizarre and none commands decisive rational grounds, the fact that a view sounds outlandish cannot by itself disqualify it. The conversation moves between abstract epistemology and concrete cases—the many-worlds interpretation, panpsychism, and the question of whether a country or a sophisticated AI system could be conscious under standard materialist premises.

The episode ranges across philosophy, physics, and how we should actually live with radical uncertainty. Schwitzgebel distinguishes Grounded vs Ungrounded Skepticism, arguing that dreams warrant nonzero credence because we dream nightly and some neuroscience suggests dream experiences can rival waking life in stability—a point with potential behavioral consequences. The pair examine why Three Grounds for Theory Choice—common sense, empirical evidence, and theoretical elegance—all fail to settle between competing bizarre theories, making tradeoffs inevitable rather than clear winners. They also probe whether bizarreness Is Not Absurdity, unpacking why the weirdness of a view need not make it false, and how we might reserve modest credence (~0.1% to 1%) for skeptical scenarios without paralysis. Carroll defends his 90–95% credence in many-worlds interpretation while engaging seriously with the counterargument that it strains common sense further than alternative frameworks.

Sharpest takeaway

Schwitzgebel argues that our best theories of reality—from quantum mechanics to consciousness—are irrevocably 'weird': they each violate common sense AND remain dubious, so since every available option is bizarre, bizarreness alone cannot be a knockout argument against any view.

  • Every interpretation of quantum mechanics and every theory of consciousness violates common sense (universal bizarreness).
  • We lack decisive grounds—common sense, direct scientific evidence, theoretical elegance are all imperfect—to choose among them (universal dubiety).
  • Grounded skeptical scenarios (dreaming, simulation) warrant a small but nonzero credence (~0.1–1%) that should subtly influence behavior.

The argument · threads6 threads · 25 claims
0.80

Common sense evolved for middle-sized, low-energy Earth conditions and fails at extreme scales and novel domains like consciousness and AI.

5 pointscentrality 5/5
  • Our common sense and intuitions are built on a very limited range of experience—relatively low-energy, middle-sized, slow-moving stuff on Earth that we evolved and were socially trained to handle—so at radically different scales (much larger, smaller, or more energetic), and for unfamiliar cases like sophisticated AI or consciousness, there is no good evolutionary, developmental, or social reason to expect our common-sense judgments to be well-tuned.

    “Our common sense is trained upon, built upon a very limited range of experiences... it's relatively low energy, middle-sized slow-moving stuff on earth.”

  • Bizarreness can dissipate not by being refuted but because our intuitions and sense of common sense change over time—just as a heliocentric Earth orbiting the Sun was bizarre when Copernicus proposed it but is no longer felt to violate common sense—so it is possible that some day people will find it commonsensical that the United States is conscious or that the universe is conscious (panpsychism); common sense is not a fixed point.

    “the idea that the earth moved around the sun was bizarre when Copernicus suggested it, but we no longer seem to find that a sharp violation of common sense. Common sense has changed over time.”

  • We should take seriously our gut reaction against bizarre-seeming views (many-worlds, panpsychism, only-humans-have-souls), but not rely on it alone, allowing that some theories we think absurd may be merely bizarre and not actually absurd—so they deserve more than reflexive dismissal.

    “the main reason I think people reject panpsychism... is that it just seems so contrary to common sense. But if I'm right, well, something contrary to common sense is probably true.”

  • A GPT-3 model fine-tuned on roughly 2 million words of Daniel Dennett's writing ('DigiDan') produced answers to philosophical questions that Dennett experts could only distinguish from Dennett's real answers about 50% of the time (versus 20% chance), and on one question Dennett himself conceded the model's answer was arguably better.

    “The experts got it about 50% right. So they did better than chance, but still half the time they chose the model's answer over Dennett's answers.”

  • Large language models will for sure change the world dramatically, though it is hard to know exactly how that change will unfold.

    “But for sure it's going to change.”

0.71

Rational agents should reserve modest credence in grounded skeptical scenarios like dreaming while adjusting behavior only when stakes are low.

6 pointscentrality 4/5
  • Conditional on living in a simulation, there is a substantial chance (perhaps 10–90%) that it is small or unstable rather than large and stable, because the simulations we run tend to be small and unstable and a full galaxy-from-Big-Bang simulation would be enormously resource-intensive when a researcher might only care about, say, a few people having a conversation—so contra Chalmers, this would constitute a genuinely radical skeptical scenario in which most of one's ordinary beliefs could be false.

    “if we are in a sim there's a decent chance that it's a small or unstable one. So if we think about the simulations we run, they tend to be small and unstable.”

  • Reserving a small credence for skeptical scenarios can rationally influence behavior in low-cost cases: if you might be dreaming, with no cost and no observers, it can be rational to try to fly because the upside (flying in a dream) is large and the downside is nil—though one should not try this on a cliff edge, and a failed attempt should reduce one's credence either that it is a dream or that flying would work.

    “I did try to fly to the library and I failed. But I think that was a rational decision because I could have been dreaming and then it would've been awesome.”

  • It is rational to assign roughly 0.1% to 1% total credence to some radically skeptical scenario being correct, leaving about 99–99.9% credence to the world being basically as we think for everyday middle-sized objects; this small reserved credence is large because, for example, if you give ~20% credence to the dream theory that we sometimes have mundane stable experiences in dreams, the resulting per-moment dream credence (~2%) cannot be plausibly knocked down by many orders of magnitude.

    “I think it's rational to assign about a 0.1 to 1% credence to some radically skeptical scenario or other being correct.”

  • Skeptical scenarios should be divided into ungrounded skepticism (hypotheticals like being a brain in a vat or an inception scenario, for which there is no positive reason beyond mere logical possibility) and grounded skepticism (starting from ordinary background assumptions and finding a positive reason for doubt)—and dream skepticism is grounded, because we dream every night and some mainstream theories of dreaming hold that we sometimes have stable, mundane experiences indistinguishable from waking life.

    “Grounded skepticism starts with our ordinary background assumptions and says, looking at these assumptions, there's some positive reason to give some credence to skeptical doubt.”

  • We should reserve some doubt about our ordinary assumptions on the basis of 'wild card skepticism'—the recognition that there may be grounds for doubt that we are not even capable of conceiving, just as ancient skeptics could not conceive of the simulation argument that future philosophers and physicists may yet devise scenarios we cannot currently imagine.

    “I should have a certain amount of doubt about my ordinary assumptions about the world, just on the basis of the fact that there's some skeptical possibility that I'm not even capable of considering.”

  • A theory positing a lower microscopic level beneath the manifest image (e.g., everything being wave functions or strings) does not count as a radically skeptical scenario, because in such cases the macroscopic world still exists and obeys its rules—the coffee mug is still there, Earth still existed for billions of years, the sun still rises—so our everyday beliefs remain mostly true.

    “I don't count that as a radically skeptical scenario of the relevant sort. 'Cause there still would be... Our everyday beliefs would still mostly be true.”

0.70

Every major theory of consciousness commits to bizarre consequences, leaving no single non-bizarre option available.

3 pointscentrality 5/5
  • A substance dualist who locates consciousness in souls faces four bizarre options about which entities have souls: only humans (implying dogs and cats are not conscious, as in Descartes), everything including protons (panpsychism), a sharp arbitrary line across the continuum of animals (e.g., one genus of toad has a soul and another does not), or graded/partial souls (a frog having an eighth of a soul, which is incoherent since souls seem discrete)—and every horn is strikingly bizarre.

    “The dualist faces two questions, where all of the answers seem to be bizarre. One concerns what kinds of entities have souls and the other concerns the causal relationship between material world and souls.”

  • Every historical attempt to make sense of how consciousness fits into the world—substance dualism, materialism, idealism, and compromise/rejection views—ends up committing to bizarre and dubious theses; even dualism and materialism, which seem unproblematic on their face, force bizarre consequences once their metaphysical details are worked out, so there is no live, non-bizarre option.

    “every single attempt to make sense of this is jaw droppingly bizarre.”

  • On a broad class of standard materialist theories of consciousness taken 'out of the box' without post-hoc adjustment, the United States—conceived as a concrete entity with people as its parts—plausibly counts as conscious, because it processes vast information, represents itself, and responds intelligently to its environment; one can react by abandoning materialism, by tweaking theories ad hoc to exclude the case, or by accepting group consciousness, but each route forces further bizarre commitments.

    “according to a broad class of materialist theories it's very plausible that the United States has a stream of conscious experience.”

0.56

Choosing among bizarre theories requires trading off imperfect common sense, limited evidence, and incomplete theoretical elegance.

3 pointscentrality 4/5
  • Because every available theory of consciousness (and of quantum mechanics) is bizarre, the fact that a view violates common sense cannot serve as a decisive defeater; we must still rely on imperfect common sense to some extent, but we cannot use bizarreness alone to reject panpsychism, many-worlds, or any other option, since something contrary to common sense is probably true.

    “the main reason I think people reject panpsychism... is that it just seems so contrary to common sense. But if I'm right, well, something contrary to common sense is probably true.”

  • We have basically three broad epistemic grounds for choosing among bizarre theories of consciousness or quantum mechanics—common sense (imperfect and bound to be violated), direct scientific evidence (currently unable to settle questions like whether electrons have souls or which QM interpretation is correct), and theoretical elegance (indecisive because panpsychism, materialism, and many-worlds each have elegant aspects)—so the choice is a tradeoff among imperfect tools, which justifies holding non-extreme credences across competing views.

    “we have basically three broad types of epistemic grounds for choosing among these theories... One is common sense... Another is scientific evidence... And then the third tool we have is something like theoretical elegance.”

  • There is no established philosophical toolkit for reasoning about scenarios where the credence of an outcome is incredibly tiny but the consequences are incredibly large (e.g., an LHC black hole eating the world at one in ten billion, or AI extinction risk), and our intuitions lose their grip on magnitudes above about a trillion; one proposed approach is 'Nicolausian discounting,' where below a low enough credence (e.g., 10^-30) one simply ignores the possibility, but this remains contested.

    “There's this idea that's sometimes called Nicolausian discounting, which is the thought that once something has a low enough chance or you give a low enough credence to it, you just ignore it completely.”

0.51

Bizarreness and dubiety are endemic to quantum mechanics; all interpretations are dubious and none compels acceptance over others.

2 pointscentrality 4/5
  • Every viable interpretation of quantum mechanics is bizarre—there is no common-sense way of thinking about quantum mechanics—and all are dubious in the sense that there are reasonable grounds for doubting any of them, so we are not epistemically compelled to accept many-worlds over the alternatives; this makes quantum interpretation a prime illustration of the universal bizarreness and universal dubiety claims.

    “Every viable interpretation of quantum mechanics is bizarre. There's no like common sense way of thinking about quantum mechanics.”

  • The 'weirdness' of the world is defined as the conjunction of two conditions: bizarreness (our common-sense understanding is sharply violated by how the world actually is) and dubiety (the situation exceeds our ability to fully comprehend, so doubt is justified)—which is stronger than mere bizarreness, since some bizarre things like special relativity can be fully understood once learned.

    “our common sense understandings of how the world is are gonna be sharply violated by how the world actually is. That's the bizarre-ness element, and then there's also an element which I call dubiety”

0.29

Ethical philosophers behave morally no better than peers despite endorsing moral truths, aiming to match rather than exceed peer standards.

3 pointscentrality 2/5
  • Across many empirical studies, professional ethics professors behave about the same morally as comparison groups of other professors—including the case of vegetarianism, where ethicists were far more likely to call eating mammal meat morally bad yet reported eating meat at their previous evening meal at essentially the same rate (37% of ethicists vs 38% overall).

    “what I found over and over again, almost without exception, is that they behave about the same as comparison groups of other professors.”

  • People generally do not aim to be morally good or bad by absolute standards; instead they aim to be about as morally good as their peers—unwilling to make the sacrifices of being much better, but also unwilling to be the worst—which explains why ethicists who discover a moral truth (e.g., that eating meat is wrong) lower their opinion of their own and peers' behavior without changing their behavior, since their peers still eat meat.

    “people don't generally aim to be good or bad by absolute standards. Instead, they aim to be about as morally good as their peers.”

  • Long-termists typically assume human extinction would be bad for future history, but we are radically ignorant about whether currently available actions have good or bad very-long-term effects; for instance, if humanity is a violent, technological species likely to destroy itself explosively (ruining Earth for others), peacefully bowing out now—as antinatalists suggest—might leave Earth in better shape for non-technological species like dolphins to flourish for a billion years, so it is very hard to know what is objectively good over a billion-year horizon.

    “The long termists will typically say or assume that human extinction is likely to have a bad effect instead of a good effect on future history. I don't think that that's necessarily true.”