What this covers
This is an ask-me-anything episode recorded in September 2021, where Sean Carroll fields questions from Patreon supporters across a wide range of topics: science, philosophy, psychology, education, medicine, politics, media, and climate. The episode amounts to an extended defense and elaboration of Carroll's core view—that the universe is intelligible through compact laws organized around a single load-bearing fact: the low-entropy Big Bang. From that foundation, he works outward through entropy and Emergence, explaining how structure and causality arise in a fundamentally reversible cosmos.
The questions let Carroll cover considerable ground. In physics, he addresses the interior structure of matter—clarifying how quarks behave in nuclei and protons, what entanglement is and how it propagates, and why black holes are asymmetric. On cosmology and foundations, he discusses the matter-antimatter asymmetry, the role of gravity in rendering the early universe's smoothness genuinely low-entropy, and the shape of Everettian Quantum Mechanics—defending its parsimony while acknowledging two hard open problems: recovering the Born rule probabilities and deriving structured worlds from the bare superposition. He also argues at length that Bayesian Reasoning is what scientists actually do, against the mythology of the formalized "scientific method," and that Falsificationism is a crude approximation of Bayesian Reasoning. Several answers tackle philosophy of science more directly—Structural Realism, the Anthropic Principle, and what emerges versus what is fundamental about causality and the Arrow of Time. A few questions venture into speculation about alien technology and how classical physics limits what is possible. Throughout, Carroll is precise about what follows from equations and what remains genuinely open.
Carroll argues that the universe is intelligible through compact laws and patterns, with entropy and emergence—anchored by the low-entropy Big Bang—doing the explanatory work across levels of reality, while Everettian quantum mechanics is defended as the most parsimonious reading of the Schrödinger equation.
- The arrow of time and irreversibility derive from the low-entropy initial state of the universe, not from time itself flowing
- Emergence (weak, not strong) lets higher-level descriptions capture real patterns while remaining compatible with microscopic physics
- The other branches of the wave function are predicted by the Schrödinger equation, so the burden is on alternatives to remove them
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
The Big Bang does not explain why ice cubes melt rather than un-melt (that follows simply from the un-melted state being lower entropy); rather, it explains why there were ever un-melted ice cubes at all—why low-entropy configurations existed in the past—because the early universe was in a highly atypical, very low entropy state, and the gradual increase of entropy since then allows ordered structures to exist and decay.
“what is being explained by the Big Bang is that there ever were un-melted ice cubes. That's the mystery”
Causality and determinism are different: determinism (knowing the state at one time lets you calculate any other time) holds at the reversible micro-level where there is no arrow of time, so one should never use causal language there; causality—where causes precede effects—is a strictly macroscopic, time-directed phenomenon that emerges through coarse-graining and entropy, and explaining why it appears at the macro level is an open problem involving physics, philosophy, and computer science.
“Causality at least in the classical sense that Aristotle would have understood... is a strictly macroscopic phenomenon, and it is a time-directed phenomenon.”
The interior of a proton is not a roiling soup of quarks appearing and disappearing; in quantum field theory the state of a proton is static and unchanging—a fixed quantum field configuration—and the language of virtual particles popping in and out comes only from interpreting that configuration via Feynman diagrams, whose sum is a static field.
“the state of something like a proton is static, there's nothing changing, there is nothing roiling, there's nothing appearing and disappearing inside a proton”
Under all conditions, an atomic nucleus like oxygen's 48-quark nucleus should be thought of as a single sea of quarks rather than separate protons and neutrons with distinct identities; the cartoon images showing discrete nucleons are misleading—the nucleus is a spherical blob with all the quarks arranged in it (subject to the Pauli exclusion principle), not quarks confined separately into nucleons that then combine.
“you shouldn't think of protons and neutrons in a nucleus as having separate identities anymore, that's all just a sea of quarks.”
Entanglement can travel at different speeds through an environment depending on the materials and interactions (Hamiltonian), with a mathematical Lieb-Robinson bound limiting the rate at which entanglement propagates through a quantum system; this differing rate can be exploited to shield systems from becoming entangled, as needed when building a quantum computer.
“there is a mathematical result called the Lieb-Robinson bound... They made a bound on the rate at which entanglement travels through a quantum system”
The matter-antimatter asymmetry is not a mere statistical anomaly we observe due to selection, because the anthropic principle, absent another mechanism, predicts the fluctuation enabling life will be as small as necessary—yet our universe's hundred billion all-matter galaxies represent a vastly larger asymmetry than life requires, so the asymmetry is much more likely produced by dynamics (requiring time-reversal violation and a departure from thermal equilibrium per Sakharov) than baked into initial conditions.
“the anthropic principle would predict, absent any other mechanism, that the fluctuation will be as small as it needs to be to get life to exist.”
Our knowledge of the universe does not scale with the universe's size or age because we collect rules and patterns, not individual facts; on a Humean view, the laws of physics are very compact restatements of everything that happens, and empirically a small number of laws and fundamental ingredients account for everything we see.
“we don't just learn a collection of facts... We're collecting rules, we're collecting patterns”
There are different kinds of entropy: the coarse-grained Boltzmann entropy (counting microstates in a macrostate) that was low at the Big Bang and grows, and the von Neumann or entanglement entropy in quantum mechanics, where two entangled subsystems have no individual wave function and are each automatically in a mixed state with an associated entropy even when the composite system's exact state is known.
“in quantum mechanics, entanglement can lead to an entropy of sub-systems, even if you know the exact state of the composite system as a whole.”
The mass of a scalar particle is the square root of the second derivative (curvature) of the potential near its minimum; for a quintessence dark-energy field that is still slowly rolling and has not reached its minimum, quoting 'the mass' at the minimum is technically cheating—one should evaluate the second derivative at the field's current location—though for a generic untuned potential the order of magnitude is comparable, so the mass serves as a quick-and-dirty proxy for how strongly the potential changes.
“the mass of the scalar field is related to the curvature of the potential near its minimum”
A classical black hole is itself not time-reversal invariant—its formation increases entropy like a melting ice cube—so it is unsurprising that test particles can cross into the event horizon but not out; the time-reverse solution, a white hole, exists and obeys the fundamental equations, with particles able to cross out but not in, making the full set of possibilities completely symmetric.
“the classical black hole is itself not time reversal invariant. A classical black hole is much like a melting ice cube in a glass of water”
The mass of the proton does not come from the masses of its three quarks (which sum to far less) nor from quark motion, but almost entirely from the energy stored in the gluon (quantum chromodynamic) field; a lone quark would have infinite field energy, but three quarks of different colors make the fields cancel at infinity, leaving a finite remnant energy that, via E=mc², constitutes most of the proton's mass.
“the mass of the proton doesn't come from the mass of the quarks; it comes from the energy contained in the gluon field”
The smooth, hot, dense state of the early universe looks high-entropy only if gravity is ignored; because gravity matters at cosmic scales, the stuff could have been arranged in far higher-entropy configurations—Penrose showed a single black hole at the galaxy's center has more entropy than the entire early universe—so the smooth early state was actually extremely low entropy.
“that single black hole has more entropy than the entire universe had at early times”
In the physicist's sense (weak emergence), higher-level descriptions speak different vocabularies but remain compatible with lower levels because they exploit patterns implicit in the microscopic dynamics; the center-of-mass example shows that throwing away almost all microscopic information still yields good predictions only because the retained information is extraordinarily special—keeping random information would predict nothing—making emergence a wonderful, non-trivial feature of nature rather than the strong emergence where macro behavior cannot be derived from micro at all.
“the higher level emergent description is taking advantage of the existence of patterns that simplify things that are implicit in the lower level dynamics”
Our own existence counts for essentially nothing when weighing theories of the universe, because it is what Bayesians call old evidence that should not be counted again—no theory worth considering predicts we are not here, so we have already taken our existence into account; thus the Fermi paradox is explained simply if the probability of intelligent life forming is sufficiently small, without needing bubble universes or a multiverse.
“I think you should just conditionalize on us being here. We already know we're here, this is what Bayesians call old evidence. You shouldn't count it again.”
Rather than Bayesianism being dressed-up induction, Popperian falsificationism is dressed-down Bayesianism: Bayes' rule is a mathematical theorem, and Bayesian reasoning assigns prior credences to propositions and updates them with data so credences never reach zero—solving Hume's problem of induction without claiming metaphysical certainty—whereas falsificationism is just a crude approximation that pretends to assign no credences when in reality we always implicitly do, favoring simpler theories with higher priors.
“I wanna say that Popperianism, falsificationism is dressed-down Bayesianism.”
Schools should teach not the formalized Baconian 'scientific method' (hypothesis, test, data, conclusion) but the real way science is done: a highly fallibilistic, Bayesian-flavored process of imagining many possibilities and updating credences—where any of your credences, likelihoods, or even your data can be wrong—and an un-orderly back-and-forth dialogue between observation, experiment, and ideas rather than a fixed sequence.
“Science is basically a highly formalized version of trial and error.”
Time does not move at the speed of light; the speed of light is how fast light traverses distance per unit time, but time does not traverse distance—it traverses time, so the only rate at which time can flow is one second per second, always, and there is no other possible rate.
“Time does not traverse distance, much less traversing distance in any particular amount of time. Time traverses time”
Spacetime is not like the fabric of a trampoline; the trampoline analogy is two-dimensional and embedded in a higher three-dimensional space, but four-dimensional spacetime is not embedded in anything, so there is no meaningful 'orientation of spacetime'—the question dissolves because there is nothing with respect to which spacetime could be oriented.
“if you take the trampoline analogy too seriously, you begin to think that space-time is embedded in something, but it's not.”
There are two good physics worries about the Everett interpretation: first, the probability question—since everything that can happen does happen in some branch, how do you recover the observed Born-rule frequencies (answered via decision theory by Deutsch and Wallace, or self-locating uncertainty by Sebens and Carroll); and second, the less-studied structure question—deriving the messy specific world from just a bare wave function and the Schrödinger equation, which is not obviously achievable and is what Carroll mostly works on now.
“the journey from the pristine purity and austerity of Everettian quantum mechanics to the messy specificity of the world is very long and potentially fraught with peril”
The relationship between decoherence and measurement is fuzzy only because the word measurement is fuzzy—decoherence is well-defined; in the Everett interpretation human measurement is just a subset of the larger class of decoherence events, so what physically happens during a measurement depends on one's interpretation, and interaction (e.g. moving a spin without entangling) is distinct from measurement because no entanglement occurs.
“The reason why the relationship is fuzzy is because the word measurement is fuzzy. The word decoherence is not fuzzy.”
Even when scientific revolutions change the ontology profoundly (e.g., Newtonian physics to relativity, classical to quantum), structural realists argue that the structure—the relationships between observations, like predictions for eclipses—is preserved, so new vocabularies do not mean we are starting over; Carroll positions himself closer to an epistemic structural realist (a statement about what we can know) than an ontic one (a statement about what exists).
“There is something that is preserved when we throw away the old picture and get the new one in. And it is the structure somehow”
You cannot assign equal weight or 'oomph' to every branch of the Everettian multiverse—both because the theory does not say that and because doing so gets the probability rule empirically wrong; branches with more amplitude count more, and for big macroscopic classical events almost all amplitude follows the classical trajectory, so worlds where the Nazis won World War II overwhelmingly have tiny amplitudes and should be given correspondingly small intellectual weight.
“You can't just assign sort of an equal amount of oomph or weight to every branch of the Everettian multiverse.”
At major research universities, what counts is research, and although administrators say teaching, outreach, and service count a little, this is false—they secretly think any time spent writing a physics book could have been spent on research, so it counts against you, whereas a hobby like playing guitar does not count against you because it is a separate category, revealing a system biased against public communication.
“if you write books about physics, they care about that, that, it counts against you.”
Speculated alien super-technologies have very different physical statuses: traveling near light speed is just a solvable technology problem; connecting with other branches of the multiverse is strictly impossible; traveling forward in time happens constantly (and faster via high speed) while backward time travel is probably impossible; and using extra dimensions, while not strictly forbidden, requires such enormous energies that no practical use seems feasible.
“These have very different statuses, these questions, or these possibilities?”
It is not that space selectively expands between galaxies; rather, in the early uniform universe everything expanded, but regions with slightly more matter and gravity pulled together, and once local gravity overcomes the expansion, the expansion of the universe has no further effect inside that region—so space expands everywhere except where there is a galaxy or other concentration of matter, and this has nothing to do with dark matter.
“it's not that space expands between the galaxies, it's that inside galaxies, space isn't expanding anymore”
It is a mistake to think of energy as a thing or substance; energy is a property of physical systems—the property that is conserved precisely because the laws of physics are invariant under time translation, per Noether's theorem—and asking 'what is energy in its essence' is misguided because for any carefully defined system you can compute its energy from the relevant formula.
“I think it's a mistake to start thinking about energy as a thing.”
Decoherence is the well-understood phenomenon in which a quantum system of interest becomes entangled with its environment; unlike classical physics where one can shield off the environment, in quantum mechanics the environment becomes entangled very quickly, and whether the system stays unentangled (coherent) or becomes entangled (decoherent) defines coherence versus decoherence.
“that process of becoming entangled with the environment is what decoherence is”
You should not say anything about what happens at T=0 because it is a singularity where physical quantities become infinite and your theoretical description breaks down; wrapping your brain around T=0 (rather than a tiny moment after) produces paradoxes because quantities take the form zero times infinity, so the size of the universe, spacing between particles, and similar concepts are simply undefined at the Big Bang.
“you shouldn't say anything about what happens at T = 0... is a singularity, by singularity we mean physical quantities become infinitely big, and therefore your theoretical description of them is wrong”
In a theory with gravity, the highest entropy state is not a black hole but empty space; Carroll and Aidan Chatwin-Davies proved that a system in curved spacetime with maximum possible entropy evolves to empty space expanding exponentially, connecting Wald's cosmic no-hair theorem to thermodynamic equilibration just as a melting ice cube reaches a uniform glass of water regardless of starting configuration.
“The highest entropy state you can get in a theory with gravity is just empty space.”
Because a black hole's Bekenstein-Hawking entropy is finite and represents the maximum entropy for a fixed volume, whereas naive quantum field theory gives infinite entanglement entropy for a region, quantum field theory is likely wrong and a finite-dimensional quantum system describes a given volume—but this finitude is about the number of degrees of freedom (dimensions of Hilbert space), not a discreteness of the wave function's amplitudes, which remain continuous.
“That seems to indicate the quantum field theory is wrong, and that there is some finite dimensional quantum mechanical system that describes what goes on inside the black hole.”
The problem with China's rise is not that it might become the top economic superpower—there is nothing intrinsically bad about another country leading, and not every country need aspire to that—but that it is an authoritarian dictatorship that may spread a model of governance worse for human rights; therefore the best US response is not adversarial policy but to be a preferable role model by being a well-functioning democracy that other countries want to emulate.
“The problem with China is not that it's potentially the top superpower, it's that it's an authoritarian dictatorship.”
Western military interventions cannot change the fabric of a society that includes many supporters of the bad actors; historical evidence suggests we cannot change countries that don't want to be changed, so the realistic lever is to be a good example—a prosperous, rights-respecting democracy others want to emulate—and where that fails, to intervene quickly and briefly rather than linger, while providing aid regardless.
“there's enough historical evidence that we're not able to change countries that don't want to be changed.”
According to our best physics, time is not something that flows—it is just a label on where you are in spacetime, and there is no difference between the physics of moving toward future or past at the fundamental level; the felt flow of time arises entirely because entropy is increasing, which produces memory of the past, causal influence toward the future, and the brain's constant comparison of remembered past, present, and predicted future.
“time isn't something that flows, that's not a fundamental feature of time. Time is just a label on where you are in the universe”
Given a fixed number of people and amount of economic activity, concentrating them in cities is better for the planet than spreading them across the land, because cities leave the rest of the land for the environment and require fewer natural resources per person—shorter driving distances, shared heating and cooling between adjacent units—so urbanization is a conservation strategy.
“you concentrate those people and that activity in a small number of locations, and you leave the rest of the land for the planet and the environment and the ecology.”
The existence of other branches of the wave function is unambiguously predicted by the Schrödinger equation, which works; the real question is why and how anyone would get rid of them, since every alternative interpretation either subtracts branches (spontaneous collapse) or adds pointers designating a real branch (hidden variables), whereas Everett simply accepts what the Schrödinger equation says at all times.
“the existence of other branches is predicted by the Schrödinger equation, and the Schrödinger equation works.”
Because university departments have nearly full responsibility to hire their own faculty and each discipline has different informal standards of what counts as good work, the academic system self-perpetuates—physicists hire physicists, biologists hire biologists—making people who work at disciplinary boundaries harder sells and causing opportunities at the interfaces to be missed.
“the departments have 98% of the responsibility to hire the new faculty members and to choose what areas they want to hire in”
Reframing introversion not as shyness or anti-social behavior but as a matter of energy—introverts expend energy interacting with others and need solitude to recharge, while extroverts gain energy from interaction—is the key insight; by this definition Carroll is an introvert who can perform socially (talks, podcasts) but finds it draining and needs alone time afterward.
“you can think of introversion as people who it costs energy for them to interact with others, whereas extroverts actually gain energy”
There are three plausible long-term COVID futures: the median scenario is it becomes like the flu but worse, with annual deaths and partial measures in a quasi-equilibrium; the hopeful scenario is global vaccination stamps it out as with smallpox; and the apocalyptic scenario is a much more deadly variant—and the longer we go without eliminating it, the greater the chance it evolves into a more virulent form.
“the coronavirus... is going to be like the flu but worse.”
There is essentially no disagreement among credentialed experts that the COVID vaccines are very good—the overwhelming majority of studies find them even better than expected—and the data show that an overwhelmingly non-representative fraction of those hospitalized with serious symptoms are unvaccinated; future historians will struggle to understand why people, given a vaccine that could wipe out a deadly pandemic, chose not to take it.
“there's not really any disagreement amongst credentialed experts about the vaccine. It's really, really good.”
As long as faculty in departments hire new faculty, the single most effective way to broaden university departments is to broaden the horizons of individual faculty members—and crucially, to do good interdisciplinary work that produces results, which is what actually changes minds rather than mere advocacy.
“the single best thing you can do to broaden the horizons of university departments is to broaden the horizons of individual faculty members”
Contrary to the questioner's framing, recent LHC and Fermilab muon experiments do not retire supersymmetry—the muon anomalies could be signatures of supersymmetry; and even if supersymmetry is never found at accessible energies, superstring theory could still be correct because its required supersymmetry might be at the Planck scale or experimentally inaccessible, so non-detection should lower one's credence in string theory but perhaps not by much.
“The recent experiments seem to show anomalies in the behaviour of muons... which could be signatures of supersymmetry”
Mainstream news media mostly do a good job reporting facts (not perfectly), so conspiracy theories of deliberate misleading are wrong, but there are real critiques concerning what the media chooses to report or not report and the spin it applies; therefore it is good to supplement mainstream sources with a wide spectrum of trusted social-media voices to surface things the mainstream misses, while remaining more skeptical of them.
“most straightforward mainstream news media mostly do a good job at reporting facts”
The notion that 'most real numbers are not computable' does not straightforwardly translate to physical phenomena; the universe obeys rules (so in that sense is computable), but the thing computing it is the world itself, which can in some sense do uncomputable things—if space is a continuum as in Newtonian mechanics, the inability to write an algorithm for every real number poses no obstacle to such a universe existing.
“the thing that is computing it is the world. So the world itself can be thought of in some sense as a computer... but it does un-computable things in some sense”
Unlike music or sports, theoretical physics cannot be done purely in a subconscious flow state—equations require conscious System 2 cognition and do not pour out automatically—but at its best it is a give-and-take between a subconscious, quasi-visualizing flow state and the deliberate writing of equations, since doing an integral demands following the steps and offers no being-in-the-zone.
“The equations are the quintessential paradigmatic example of needing your system two, needing your conscious cognition to kick in”
The theory of emergence is young and underdeveloped because the best way to understand any single level (particle physics, biology, psychology) is to study that level directly, not the connections between levels; emergence should not be thought of as deriving higher levels from lower ones (which may be effectively impossible even if consistency holds) but rather as recognizing both theories exist and finding the connections between them as a research program.
“Don't think about a direct route from the microscopic theory to the macroscopic theory. Think about both theories existing. And as a research program, find the connections between them.”
Carroll holds no convictions about the universe not based in scientific evidence—only prior probabilities that can be updated—except for self-undermining or cognitively unstable beliefs (an evil demon deceiving all sense data, or being a Boltzmann brain) which, if true, would make science itself impossible, so he assigns them essentially (but not exactly) zero prior so as not to take them seriously unless forced.
“there are some propositions which if they were true, science itself would not be possible.”
Although people do change over decades, the major changes that would substantially affect a close relationship are rare, so the better philosophy of relationships is to have relationships with good people from the start rather than partnering with not-good people and trying to fix them; rather than fighting for a relationship or changing the other person, it is more fruitful to look inward and change yourself, which is hard but easier than changing others.
“people don't change that much that quickly... the more my philosophy of relationships is just have relationships with good people from the start”
People often pursue a checklist of traits when searching for a partner, but what they actually need in a relationship is something very different, so one must be open to the possibility that initial sparks are not worth rekindling and that time is better spent finding the right person than fitting the wrong person into a situation that doesn't work for either party.
“what they look for in a person is not what they really need in a person”
Condescension and self-importance are flaws not closely related to any real virtue (unlike workaholism which has positive and negative sides), so people characterized by them rarely change; the best response is to minimize dealings with them where possible and, where not, to minimize one's reactions and let it roll off—with Twitter serving as good training ground since ignoring people is easier online than in real life.
“being condescending and self-important is just bad.”
Equations give deeper understanding than simulations because, in principle, an equation contains all possible simulations at once—an inverse-square law of gravity instantly tells you how planets orbit and how apples fall, whereas a bunch of individual simulations of planets in ellipses tells you nothing about falling apples; the rule spreads automatically across all situations.
“an equation captures every possible simulation. Simulations are done one by one.”
Science is already democratic and needs no formal extension of democracy because it is run by scientists as a collective with no Pope or emperor—individual scientists vote with their feet by deciding what to research, what papers to write, and who to hire—so ideas bubble up from the bottom and you must convince others you are right, just as in a democracy.
“Science is run by scientists as a collective, and it's not very formal.”
Gravity and electromagnetism are analogous but fundamentally different, the biggest practical difference being that electromagnetism has both positive and negative charges while gravity only ever attracts—there are no negative-gravity particles—and this single fact is crucially important to the difference between the two forces.
“gravity only has positive charges, gravity only ever attracts, one particle only ever attracts another particle, never pushes it away”
In advising the Avengers, Carroll argued that pruning alternate timelines (as in the Marvel multiverse) would be morally equivalent to genocide—ending the lives of trillions—so a better fictional approach would be to re-merge the branching timelines by aligning what happens across them, preserving a single timeline without killing anyone, though he acknowledges this is impossible in the real world.
“that would be terrible because that's literally committing genocide, you're literally ending the lives of trillions and trillions of people living in this timeline”
We will likely understand consciousness because the fact that we haven't yet is the least surprising thing in the world—it is an emergent phenomenon from one of the most complex systems we know, the human brain and body—and the road toward understanding is some combination of neuroscience, philosophy, and psychology, with the caveat that science cannot specify in advance exactly which steps will lead to the answer.
“Consciousness is one of the hardest things to try to understand because it's an emerging phenomenon from one of the most complex systems that we know about”
Within complex systems, Carroll is much less interested in any particular complex system than in the general structural, quasi-philosophical rules that find commonalities across very different kinds of complex systems, with his specific current focus being how to relate the most fundamental level of physics to higher emergent levels—what criteria make something complex, and whether the economy is similar to the nervous system or the internet.
“I'm much less interested in this or that complex systems and in rules by which we can somehow find commonalities between very, very different kinds of complex systems”
When set designers for The Big Bang Theory toured real UCLA physics labs, they were appalled at how un-glamorous the equipment was (lasers blocked by cardboard on a clothespin), which is why the show's lab scenes are more realistic than other science shows—because real science budgets are not spent on gleaming futuristic equipment.
“the set designers were appalled... "Oh my God, this place is a dump. You're blocking your lasers with a piece of cardboard attached to a clothes pin."”