What this covers
This episode is an Ask Me Anything session in which Sean Carroll fields 24 questions from Patreon supporters across physics, philosophy, politics, and related domains. Rather than a sustained argument, the session reads as a collection of mini-lectures, each clarifying a persistent misconception or resolving an apparent paradox by sharpening the terms involved. The recurring theme is poetic naturalism: the framework that different levels of description are all valid once properly specified, and that concepts like 'mass,' 'entanglement,' 'entropy,' and 'complexity' require careful unpacking of what we mean rather than pursuit of answers to muddily posed questions.
The physics ranges widely: Carroll explains how mass arises in quantum field theory through the Higgs mechanism and potential energy, why asking whether the universe's wave function is entangled is meaningless, and how decoherence—entanglement with an environment—explains why classical chaos emerges from quantum linearity. He addresses gravitational waves as energy carriers capable of collapsing into black holes, clarifies that coarse-graining is what entropy requires (not observation in a mystical sense), and distinguishes the many-worlds interpretation from stochastic quantum mechanics when thinking about branching and moral weight. Beyond physics, he argues that sociology and psychology are genuine sciences by the standard of model-building and empirical testing, that Bayesian reasoning explains how people with different prior probability distributions converge toward truth, and that precognition claims are better explained by the brain's documented quirks than by quantum exotica. Several claims invite disagreement—notably his contention that a higher-level explanation is not "more valid" than a low-level physical one, only more efficient—and his skepticism about neutrino telescopes yielding major cosmological insights.
Carroll argues that most apparent paradoxes in physics, identity, and morality dissolve once you adopt poetic naturalism: there are multiple valid levels of describing reality, terms like 'alive,' 'intelligence,' 'entropy,' and 'self' are constructed conceptual tools rather than nature-given essences, and the right move is usually to analyze what we mean rather than answer the question as posed.
- People are physical processes with no metaphysical essence, so identical copies bear identical responsibility
- Higher-level explanations can be more efficient without being more 'valid'; validity is just truth
- The arrow of time, complexity, and entropy are observer/coarse-graining dependent, not fundamental flows
Reality has a deep mathematical structure underlying quantum mechanics, spacetime, and all observable phenomena.
- There is a kind of complexity that is observer-dependent because it involves coarse-graining; even the well-defined concept of entropy is observer-dependent in this sense — Laplace's demon, seeing every microstate, would never need entropy and would see no arrow of time, just different states at different moments. So a coarse-graining observer thinks about complexity differently than a hypothetical (physically implausible) observer who sees every microstate.
“there is absolutely a kind of complexity that is observer-dependent, that involves a coarse graining”
- Entropy is not a relational phenomenon; for a given system with a given coarse-graining, entropy is simply high or low, full stop. What it requires is a coarse-graining — a chosen way of grouping microstates into indistinguishable macrostates — which essentially everyone agrees on but which we nonetheless must choose; once chosen, entropy is just a calculable number not related to anything else.
“entropy is not a relational phenomenon. There is... a given system in a given coarse-graining is just at high entropy or low entropy, full stop.”
- Mathematically, 'straight' in a curved space means a geodesic — the shortest (or extremal) distance path. Two procedures define it: pulling a string taut between two points (shortest distance) or pointing in a direction and walking forward continuously; in a curved background neither looks straight to an outside observer but both are the closest analog to Euclidean straight lines used in general relativity.
“Straight means what we call a geodesic, that is to say the shortest distance path or the extremal distance path”
- All electrons have the same charge because they are all excitations of the single electron field — not, as Wheeler joked to Feynman, because they are all the same electron moving back and forth in time (Wheeler was not correct). But this only explains the uniformity of electron charge, not why the electron's charge is exactly equal in magnitude and opposite in sign to the proton's; that remains unanswered, with grand unified theories or gauge anomaly cancellation offering partial relationships.
“the reason why all electrons have the same charge is because they're all part of the electron field. But that is not enough to tell you why the charge in the electron field is exactly equal in magnitude and opposite in sign to the charge in the proton field”
- Quantum mechanics is linear with no chaos when following the universal wave function, yet classical chaos (which is nonlinear) emerges as a limiting case because we follow individual decohered branches. The tumbling of Saturn's moon Hyperion illustrates this: in isolation its orientation wave function would spread into a fuzzy blob, but constant environmental monitoring (radiation entangling with it) decoheres it so that within our observable branch it follows a definite chaotic trajectory — no inconsistency between quantum linearity and classical chaos.
“classical mechanics can be highly non-linear, quantum mechanics is linear, but classical mechanics also emerges from quantum mechanics. It's a limit, it's a limiting case.”
- Irrational numbers are real numbers that cannot be expressed as a ratio of two integers (like pi or the square root of 2), not merely weird or special numbers; among any span of real numbers there are far more irrationals than rationals, making them a crucial part of constituting the continuum and hence central to physics and math.
“real numbers like pi or like the square root of 2, but numbers that cannot be expressed as a fraction of two integers.”
- The fine-structure constant is uniquely the most obviously dimensionless of the fundamental constants (about 1/137), independent of units. Other dimensionless constants like the weak mixing angle are constrained as angles, and the strong coupling appears dimensionless classically but requires a dimensionful QCD scale because it blows up at zero energy under the renormalization group. The hype is largely historical because we figured out quantum electromagnetism first.
“The fine-structure constant is interesting because it is dimensionless, so it's a number, which means that it doesn't depend on what units you use to measure it.”
- A black hole cannot spin so fast it tears itself apart; there is an extremal limit beyond which it would cease to be a black hole and become a naked singularity. In practice you cannot reach this extremal point because adding spin requires throwing in spinning matter, which also adds mass, so spin never overtakes mass — consistent with the cosmic censorship conjecture. Spinning black holes have oblate-spheroid geometry, not perfectly spherical, though the event horizon retains spherical topology.
“It is not possible for a black hole to spin so fast it would tear itself apart, but there is a limit that you reach where it's the fastest a black hole can spin, and this is what is called an extremal black hole.”
- Apparent precognition or seeing future/distant events is best explained by the brain's known tendencies to hallucinate, fill in gaps, misperceive (optical illusions), and retrospectively overweight surprising events, rather than by quantum mechanics or many-worlds; the brain is far more complicated and harder to understand than quantum mechanics.
“I would blame the tricks of the brain for that rather than the rules of quantum mechanics.”
- Lee Smolin's cosmological natural selection — universes born at black hole centers inheriting slightly different physics, selected for reproductive success — never caught on because the mechanism for making new universes inside black holes and varying their laws was never specified (standard physics expects a singularity, not a new universe), and because the theory would naively optimize for making black holes, yet our universe has far fewer black holes than stars or planets and one can easily imagine laws making black holes more easily, so the data seem to argue against it.
“This idea never really caught on for two reasons... the mechanism by which you made a new universe inside a black hole, and the mechanism by which that new universe has slightly different laws of physics was just never specified.”
- Even if space emerges from quantum entanglement (as in quantum circuit cosmology or de Sitter as a tensor network), it does not provide an easy or novel explanation of inflation or the initial low-entropy state, because you still have to posit that the entanglement had a particular structure at early times.
“it does not, as far as I know, provide any easy ways or novel ways of addressing the fine-tuning problems of the early universe. Even if you are thinking of space as created from quantum entanglement, you still have to just posit the fact that that entanglement had a certain structure at early times.”
- It is misleading to think of 'billions of future me's' as multiplied copies; in many-worlds each future person who counts your present self as their past self is a different person, and crucially each is 'thinner' in a quantitative sense — the worlds they live in are tinier fractions of the total. This is distinguishable from imagining many copies in a single-universe stochastic ontology.
“all of those people are different, but they're all thinner in a very quantitative sense. The worlds they live in are tinier fractions of the net total of the world out there.”
- Unitarity is the quantum-mechanical conservation of information (reversibility). The intuition that an expanding universe needs more information and thus violates unitarity is overblown — it comes from intuition not the equations, and there is no necessary reason expansion violates unitarity. However, given the mysteries of quantum gravity and initial conditions, Carroll remains open to abandoning unitarity if there is a strong reason, while warning it would be a very large departure.
“At face value, I think that that particular tension is overblown, it comes from our intuition, not from the equations. There's no necessary reason just because the universe is expanding that you need to violate unitarity.”
- Priors are not developed with no data whatsoever; our biology, psychology, education, upbringing, and evolutionary history all count as data feeding intuitive priors. Different people starting with different priors is a feature not a bug, and rationality lies in the reasoning process from those priors; with enough data good Bayesians converge on the same answers.
“These ways of thinking about things that seem naturally intuitive to us, feed into our priors, they're not completely arbitrary priors.”
- The perceived slowdown in fundamental physics is not due to lacking another Einstein (there are many brilliant people, and others would have made the same discoveries) but because the first half of the 20th century was extraordinarily and atypically productive, with multiple revolutions (relativity, quantum mechanics, quantum field theory, Big Bang, modern particle physics) in a few decades. The subsequent decades of coming to terms with those discoveries represent normal science, which is the historically ordinary rate of progress.
“it's that the first half of the 20th century was just absolutely amazing and atypical in the history of science.”
Consequentialist reasoning justifies action: climate requires stopping emissions, democracy requires stopping lies, and utilitarianism requires quantifying utilities.
- Carroll's credence on building a high-resolution neutrino telescope to see beyond the CMB is low; while detecting even the cosmic neutrino background would be awesome and would let us see past the surface of last scattering to the first fraction of a second (since neutrinos freely streamed rather than being trapped in an opaque plasma), in terms of physical distance it would only be a couple hundred thousand light years further, where conditions are essentially the same as the microwave background, so no astonishing discoveries are expected beyond what is already implicit in cosmology.
“I think that my credence is pretty low. I think it would be awesome just to detect the cosmic neutrino background, which we've not been able to do”
- There is a correlation between central supermassive black hole size and galaxy size, though the origin is not fully understood. Central black holes are extremely small compared to their galaxies, so it is not the black hole governing the galaxy's evolution; rather the galaxy affects the evolution of the black hole such that bigger galaxies host bigger black holes. Smaller galaxies may still have central black holes that are simply harder to detect.
“there is a correlation between central black hole size and the size of the galaxy. And this is a bit mysterious, people don't completely understand the origin and evolution of central black holes.”
- Antimatter is not mysterious or hard to find; it appears naturally on Earth all the time. Cosmic-ray showers produce positrons and anti-muons, and many radioactive decays give off antimatter — rarely positrons but very commonly anti-neutrinos. Such antiparticles don't last long: positrons annihilate with electrons and anti-neutrinos escape.
“many kinds of radio activity give off antimatter particles. There's the rare, rare, rare event where you can give off a positron, but it's extremely common for radioactive decays to give off anti-neutrinos.”
- The most obvious technology for climate change is not a technology at all — just stop spewing greenhouse gases into the air, which we know how to do and should prioritize. Developing ways to remove greenhouse gases from the air is also good, but it is not a replacement for stopping emissions in the first place.
“There's one very, very obvious technology, which is not even a technology, it's just stop spewing greenhouse gases into the air, and that's something we know how to do, so I would prioritize that.”
- It is a misapprehension to treat Aristotle's four causes as cause-and-effect relationships in our modern sense; Aristotle was trying to understand change very generally, in Ancient Greek, without modern physics, calculus, or differential equations, and did not distinguish biological, physical, intentional, and natural changes the way we do. His classification does not carry over to modern physics, like most ancient philosophy.
“it's a slight misapprehension to think that Aristotle's four causes are causes as in cause and effect relationships that we have. Aristotle had a much more expansive notion”
- Even if the world contained much suffering, extinction would not be the more moral choice, because the world is also rife with joy, triumph, and accomplishment, and there may be far more good in the future; moreover extinction removes any chance of future improvement, so eliminating that possibility is a strong reason against it.
“I don't agree that the world is rife with human anguish and agony... I think that the world is also rife with joy and triumph and human accomplishment, and furthermore, there might be a lot more of that in the future.”
- Academia over-emphasizes individual metrics like citations and the achievements of particular people, while undervaluing the support system — mentoring, public communication, acknowledgments, helping colleagues and students — that makes good science possible. These contributions are very hard to quantify (unlike citations, an imperfect impact measure), but evaluations should still try to capture them, which is partly why hiring and promotion use expert recommendation letters, not just publication lists.
“we far over-emphasize the specific contributions or achievements of this or that individual person and completely forget about the huge support system that goes into doing this”
- Punishment should be justified by deterrence and rehabilitation — making the world better — rather than revenge; there is no metaphysical correctness to which person 'deserves' punishment, so anyone in the same mental state as the criminal should be treated the same way under a consequentialist view of law.
“I'm very much in favor of consequentialism here, we should do it for deterrence and rehabilitation purposes, we're trying to make the world better, not just getting some sense of revenge.”
- The fundamental badness of the January 6th attack was not the storming of the Capitol per se but the promulgation and belief of the lie that the election was stolen; in a hypothetical where an election truly had been stolen by rigging voting machines, people would have a right and even obligation to protest, possibly violently. Extra-judicial violence is justified only when a society purporting to be a democracy is not really one.
“the origin of the badness isn't the attack, it's the fact that they promulgated and believed that lie about the election being stolen, that was the problem.”
- Merely ranking outcomes as better/worse/equal is not enough for utilitarianism, because the whole point is to compare situations where many small goods may outweigh one big good; that comparison requires quantifying utilities, not just ordering them. You can reject utilitarianism, but if you want it, simple ranking cannot do the work of deciding how many small goods overcome one large good.
“simply ranking good and bad things is not good enough, because the whole point of utilitarianism is to try to provide a way of comparing situations where you do have many little goods versus one big good.”
- The Goldilocks-zone framing (a planet too close or far from its star being too hot or cold) has been somewhat deprecated by exobiologists because some of the most promising venues for life in our own solar system are the moons of Jupiter and Saturn, which the old framing would place outside the habitable zone, so the search should be more open-minded about where life could exist.
“the idea of the Goldilocks has been somewhat deprecated in the thinking of exobiologists... some of the most promising venues for life appearing are places like the moons of Jupiter or Saturn, which might have been considered in the old days outside the Goldilocks zone.”
- Although honoring a homophobic figure with a telescope name is bad, realistically there is almost nothing listeners can do to get the James Webb telescope renamed because bureaucracies are hard to change; a better use of energy is to keep the issue visible and use it to talk about ongoing discrimination rather than banging one's head against a wall trying to change the name.
“realistically, there's almost nothing. I think, realistically, that ship has sailed.”
- Peace talks in the abstract are good, but no peace agreement is realistic now because Ukraine and Russia's differences are too large. What matters most is what Ukrainians want since they are the ones invaded, so the US has no right to demand a settlement they oppose. Everyone shares an interest in avoiding nuclear war, but emphasizing the catastrophic consequences of any first-strike use of nuclear weapons is a more stable long-term policy than brokering a peace the invaded country rejects, since allowing nuclear powers to invade and get their way sets a dangerous precedent.
“what has to matter the most is what people in Ukraine want, they're the ones who are being invaded. So I don't see any Ukrainians asking the US to negotiate a peace settlement right now, so I don't think that we really have the right to demand such a thing.”
- If cosmic large-scale structure arose from quantum fluctuations, then under many-worlds there are many (perhaps not infinitely many) other branches with different structure. These superpositions decohere via the environment — modes of quantum fields outside the horizon and short-scale photon modes act as the environment — rather than being magically measured, as many physicists loosely assume in inflation discussions.
“they often treat different regions with different values of quantum fluctuations as sort of as if they were magically being measured by something... But they're not of course being magically measured, but they are decohering.”
- Decoherence is not just entanglement; it is entanglement with the environment, which has many degrees of freedom, making it irreversible (like scrambling an egg) and raising entropy. This is why entanglement can be both common as dirt (uncontrolled environmental entanglement) and a precious resource (controlled entanglement between tracked subsystems of a quantum computer). Crucially, becoming entangled with the environment performs an effective measurement that breaks the entanglement between other subsystems, which can damage a quantum computation.
“decoherence is not just entanglement, decoherence is entanglement with the environment.”
- Black holes can form in general relativity with no matter, because gravitational waves carry energy (as seen in binary inspirals losing energy to waves), so a sufficient concentration of colliding gravitational waves could collapse into a black hole. By the no-hair theorem such an empty black hole is indistinguishable from a matter-formed one from outside, since black holes are not solid objects but just spacetime inside; however, because gravity is weak, such formation is very rare.
“it's absolutely possible to make black holes in general relativity without having any matter, because gravitational waves and similar kinds of distortions of spacetime can effectively have energy.”
- Long-running network TV shows like Lost struggle to deliver satisfying finales because stories are shaped by their endings, yet showrunners often don't know how many seasons they'll get and set up mysteries without predetermined payoffs; mystery genres especially demand a payoff, and shows like Lost, How I Met Your Mother, and Game of Thrones illustrate how getting locked into early setups or losing quality can ruin the conclusion.
“stories are very often shaped by how they end... But when you're in season 3 of a TV show and you don't know how many seasons you're going to get... it can be really challenging to set yourself up for a successful finale”
- Carroll is happy to say hi to listeners and patrons if they run into him, though he may not have time to chat at length.
“I'm very happy to say hi to anyone who supports Mindscape on Patreon. I'm even happy to say hi to those listeners out there who don't support on Patreon.”
- Caliban the cat adapted to the Baltimore move right away while Ariel was traumatized and hid but has nearly returned to normal.
“Caliban, the boy kitty, adapted right away. He's very, very chill. Ariel, the girl kitty, is much more high strung, it's taken her a long time.”
Understanding emerges through rigorous truth-seeking, model-testing across domains, and trusting expert consensus over non-expert opinion.
- Formal debate is good training for public speaking and logical thinking but poor at finding truth; it is more entertainment and inspiration for an audience than an intellectual exercise. Hashing out disagreements and digging into why people differ is valuable and is what academics do daily, but it should not be conflated with debating, and one should engage with good-faith interlocutors rather than those seeking to demolish opponents.
“What it's not especially good at is a way to find the truth.”
- Global warming does not mostly come from the heat we directly generate (e.g., car brakes or engines); it comes from the greenhouse effect. Earth receives visible light from the sun and radiates back infrared; greenhouse gases like CO2 are transparent to visible light but absorptive to infrared, so incoming sunlight passes through while outgoing infrared is delayed in escaping, heating the planet. It is the change in the atmosphere's infrared transmissivity, not our heat generation, that is the dominant contributor.
“global warming doesn't mostly come from heat, and it's a little counter-intuitive... we get primarily visible light from the sun, and we radiate back primarily infrared light”
- We accept LHC findings partly because they can be replicated by two independent general-purpose detectors (CMS and Atlas), which was crucial for the rapid acceptance of the Higgs discovery since both saw it in the same place; beyond that, credence depends on Bayesian priors and track record, and lay people reasonably trust expert consensus because experts, while never perfect, are more likely correct than non-experts.
“there's more than one detector at the Large Hadron Collider. There are two large general purpose detectors, the CMS detector and the Atlas detector, and this was in fact historically crucially important to why we accepted the discovery of the Higgs boson so quickly”
- If you devise a new theory (e.g., a quantum interpretation) and want experts to review it, your responsibility is first to do the work of learning all the existing interpretations and literature; the chances a non-expert who hasn't done that work solves a problem experts haven't are overwhelmingly low, and doing the work lets you critique your own idea, which is more useful than asking random people online.
“If you're going to ask other people to do the work to understand your stuff, then your responsibility is to do the work to understand their stuff first.”
- What gets new people interested in science is the idea of science and chasing down big questions, almost irrespective of the actual subject matter; many physicists in narrow subfields were first inspired by sweeping popularizations (like Brian Greene on string theory) or science fiction. Big questions like 'where did the universe come from' require little background to ask, which is part of why physics popularizes well.
“It's the idea of science and chasing down these big questions and trying to answer them that gets people interested, almost irrespective of what the actual subject matter is.”
- There is a fundamental rock-bottom makeup of the universe — there is something the universe does, whether or not it is a wave function or quantum mechanics — and it is in principle figure-out-able. The universe has given no indication of being unintelligible; on the contrary it has revealed an enormous number of its secrets, so whatever the final description is, it is out there lurking somewhere.
“I think that the universe has given us no indication whatsoever that it is somehow unintelligible. Exactly the opposite, the universe has really revealed an enormous number of its secrets to us.”
- Being tiny and short-lived compared to the universe should not be a rational source of anxiety, because this is simply our world and what it is — something to accept and, if one chooses, to cherish and appreciate; meaning is found at human scale rather than cosmic scale.
“there is no rational reason to feel anxious, because this is our world, this is just what the world is. It's something that should be accepted.”
Higher-level explanations are epistemically valid; categories and essences are useful human constructs despite having no independent metaphysical foundation.
- A higher-level explanation (e.g., '13 is prime') is not more valid than a low-level physics explanation tracing particles from the Big Bang; both are true. What differs is efficiency: the higher-level explanation requires much less input information to yield the prediction, but validity is simply whether it is true, not a measure of efficiency.
“It might be more explanatory. You might imagine that there is some sense in which we can rank the explanatoriness of different explanations by some version of how much you get out versus how much you put in.”
- Mass is the rest energy an object has just sitting there (E=mc²). In quantum field theory, all particles are excitations of fields, and the mass of a field comes from its potential energy — the energy associated with the field having a value. The photon is massless because its field value carries no potential energy (a consequence of gauge invariance), while the electron acquires mass because its interaction with the non-zero vacuum value of the Higgs field turns what was an interaction into a potential energy.
“the mass of a field comes from its potential energy”
- The word 'quantize' is misleading: quantum mechanics has nothing fundamentally to do with making things discrete. If anything quantum mechanics turns everything into waves, and discreteness arises only because waves under certain boundary conditions take on discrete sets of wavelengths (like harmonics on a string). To quantize means to associate a classical system with a quantum theory — an imperfectly-defined procedure — not to make it discrete.
“quantizing has nothing whatsoever to do with making discrete. Quantum mechanics doesn't have anything fundamentally to do with making things discrete. If anything, it's the opposite. Quantum mechanics turns everything into waves”
- Whether a quantum state is entangled is not a well-posed question for a single system; you must divide a quantum system into subsystems and then ask whether those subsystems are entangled. Therefore asking whether the wave function of the universe is entangled is meaningless — there is no state with 'no entanglement,' only a set of sub-states that are not entangled with each other.
“whether a quantum state is entangled or not is actually not a well-posed question. You have to take a quantum system and you have to divide it into sub-systems, and then you can ask whether those sub-systems are entangled or not”
- Gender is a social construct, but so are tables, chairs, the Earth and Jupiter — these are concepts humans invented to divide the world into stuff, and 'social construct' does not mean made up, illusory, or fake. Social constructs can refer to something real; postmodernism's useful insight is that while reality may or may not exist, our ways of talking about it are invented, which matters because it reveals how language can be used to manipulate reality.
“I absolutely think of gender as a social construct. Almost everything that we talk about is a social construct, tables and chairs are social constructs, the Earth and Jupiter are social constructs.”
- For confusing questions like 'must an entity be alive to be intelligent,' the right philosophical move is not to insist on an answer but to step back and analyze what we mean by 'alive' and 'intelligence,' which each have multiple aspects relevant in different circumstances; the mistake is assuming these words denote single rigorous things that must either match up or not.
“the right thing to do is not to insist on trying to answer a question... but rather taking a step back and thinking about what we mean by being alive and having intelligence.”
- There is no metaphysical essence of self that follows a person's physical atoms; people are processes and collections of matter. Therefore an atom-for-atom copy of a criminal with the same memories and psychology is, for all intents and purposes, the same person and bears the same responsibility. If any one such collection of matter deserves a punishment, all identical copies (with those memories) deserve it equally; you cannot offload your sentence onto a copy, and copies without the criminal memories should not be punished.
“I don't think there's any essence of self-identity that follows the actual physical atoms of a person, that doesn't make sense to me.”
- Consciousness can arise in many substrates, so artificial minds could be conscious, but humanity cannot simply transport itself unaltered into a virtual realm because having a body is crucial to being human: bodies embody our minds and metabolic processes and supply motivation (hunger, thirst, tiredness). Without those, a virtual mind would be a very different thing than a human; one could simulate hunger and thirst, but it's unclear one would want to.
“human beings are very, very, not controlled, but affected, determined, embedded in the fact that we have bodies, okay. Having a body is crucial to being a human being.”
- There is a transferable 'physics thinking' that Carroll tries to teach: beyond general scientific thinking, its essence is the search for the correct simplifications — identifying which features of a system you can ignore and modeling it with a much simpler toy system that still captures essential behavior. This works well in physics because it's often clear what to ignore; applied to complex domains like democracy it may yield insights under the right circumstances, but one should not be surprised if it fails and should then use other methods.
“the essence of it, besides just being science... specifically physics thinking... it comes down to the search for the correct simplifications, what are the features of the system that you can ignore”
- Sociology and psychology are genuine sciences because science is the attempt to explain the world by building models and testing them against data, regardless of subject. Physics seems 'easy' not because becoming a physicist is easier but because making progress is easier, since physics deals with very simple systems (a pendulum, a vibrating field) compared to society or the brain.
“science is at the attempt to explain the world by developing models and testing them against data, that's what science is”
Time's apparent flow is a psychological impression grounded in thermodynamic entropy; quantum many-worlds is mathematically sound but lacks practical consequences.
- There is no fundamental flowing of time in physics — time is just a label in the equations. The flow of time is a psychological impression, and every such impression ultimately reduces to the thermodynamic arrow of time, the increase of entropy. We know the past but not the future, and feel causal influence over the future not the past, because entropy was lower in the past; without increasing entropy we would have no conscious experiences at all, including the experience of time flowing.
“I would argue that the flow of time is an impression that we human beings have psychologically, and every psychological impression that we have ultimately comes down to the thermodynamic arrow of time, the fact that entropy is increasing.”
- The apparent inconsistency — energy not conserved in changing spacetime, yet conservation-of-energy violation rendering an early Einstein equation incorrect — resolves because conservation of energy means different things in different contexts in general relativity. There is a rule for how energy changes in response to spacetime curvature, called covariant conservation of energy: energy need not be globally constant but varies in a specific, predictable way tied to how spacetime itself varies.
“It's rather covariant conservation of energy, which means that the energy might not be constant, but it varies in a very, very specific predictable way in relationship to how spacetime itself is varying.”
- Energy is conserved for pulsars: they do lose energy through radiation and genuinely slow down over time, but the slowdown is extremely tiny because there is enormous spin energy and only a small fraction leaks out, so the slowdown can be accounted for and pulsars remain excellent clocks.
“It's a well-known fact that pulsars slow down over time... the slowdown is very, very tiny. There's a lot of energy available in that spin of the pulsar and only a tiny amount of it leaks out”
- Physicists who agree many-worlds is the best interpretation can still disagree about how to attach words to it; everyone agrees on the math. Carroll disagrees with Deutsch's way of describing a pre-decoherence system as already two universes, because before decoherence there is no unique way to divide a system into universes — it can branch along different axes, so it is just one thing.
“we all agree on the math, there's no distinction and no disagreement about how to use the equations, but how best to interpret them in terms of words changes from person to person.”
- The Everett many-worlds interpretation has essentially no implications for practical reason or morality, because quantum mechanics' relevant feature — that observers cannot predict the future exactly — holds regardless of interpretation. Just as energy is conserved (not doubled) when the wave function branches, utility should be divided among branches rather than multiplied, so branching the wave function is not a way to do moral good. Only an exotic absolutist ethics (where some outcome is impermissible if it ever happens) would yield a difference between many-worlds and stochastic quantum mechanics.
“I don't think that Everett has many, if any, implications for practical reason and morality.”
- Newtonian physics treats the second derivative of position (acceleration) as fundamental because the state of a system is given by position and velocity, so a formula for acceleration (F=ma) determines everything; you don't need higher derivatives. Acceleration is special because, unlike position and velocity, it can be absolutely measured — there is no absolute position or velocity, but you can tell whether you are accelerating. Any such equations can equivalently be rewritten as purely first-derivative equations (as Hamiltonian mechanics does), so why position and momentum are treated so differently is a deeper unsolved question.
“Acceleration is the first of these quantities where you can absolutely measure it. You know whether you're being accelerated or not, according to traditional classical physics.”
- The theological argument that an infinite past means you could never reach today (because it would take infinitely long) is a bad argument analogous to claiming zero can never be reached from minus infinity. The point is you don't start from minus infinity; if time is infinite there was no starting point, and 'time is infinite' and 'time started' are mutually incompatible, so the present arriving is not problematic.
“It's like saying the number zero can never be obtained, because starting from minus infinity, it would take infinitely long to count to zero. The point is, you don't start from minus infinity”
Systems and incentives shape behavior more than individual excellence; institutions matter more than power inequality for democratic erosion.
- Academic conservatism causes well-established, high-prestige subfields to be perpetuated because they have well-understood criteria for success, while new upstart fields require a sales pitch to colleagues since their value is harder to demonstrate; the Nobel-recognized quantum entanglement experiments were once a niche, low-priority pursuit, illustrating that more openness to young fields is warranted even though the current arrangement is understandable.
“there is a kind of conservatism among academics, there's certain sub-fields that have been successful and high prestige for a long time, and it's very easy to just keep doing those over and over again.”
- Students fear being wrong and want to be told what to think largely because the educational system trains them that way with high-stakes grades; the better way to encourage curiosity is not to find more inspiring adult role models but to adjust the system to reward curiosity and play without punishing being wrong as often, while acknowledging grades serve real evaluative functions that make threading this needle hard.
“The system is kind of set up to make them think that way. We give kids grades and those grades can be very impactful, they're high stakes.”
- The spread of literacy is not an unalloyed good because most goods are not unalloyed, but it is an overwhelming good; even if it pushes thinking toward reductionism, the benefits far outweigh any disadvantage.
“literacy is an overwhelming good. Maybe there are some disadvantages, maybe it has made us lead toward being reductionist, lean toward being reductionist, but man, it's worth it.”
- The biggest societal problem is that too many people feel they have no control over their environment — economically, politically, or culturally — despite humanity generating enormous wealth over the last century or two. It is not inequality per se (Carroll minds that not everyone is rich, not that some are richer); debt, changing jobs, and uncontrollable healthcare costs funnel resources from many to a few and produce powerlessness, which erodes faith in democracy.
“If I had to pick a single societal problem, it's the fact that we have arranged things in our society where too many people feel there's nothing they can do to control their environment, that they are subject to forces that are much greater than themselves”