Every theory is an approximation: why good physicists admit ignorance

The Short Version

Five physicists and astronomers, asked separately, describe the same working rule: a scientific claim is trusted to the extent it has been measured, and everything beyond that has to be labeled as unknown. Andrew Strominger puts the principle most bluntly, saying "Everything is an approximation" [1]. Michelle Thaller says science is deliberately limited to what can be reproduced, Carlo Rovelli says certainty is what blocks understanding, Brian Cox says a scientist must be "delighted when you're wrong," and Lisa Randall says you must believe your ideas "while questioning it all the time" [2][3][4][5]. That shared stance gives a reader a practical test for any claim: ask how it was measured, whether independent checks agree, and whether the person making it can say where their knowledge stops [2][4][5]. The five still disagree about particular questions, including what an electron is between measurements and what happened before the Big Bang, and they say so openly [3][5].

Table of contents

What They Said

Andrew Strominger: every theory has an edge, and the edge is the good part

Strominger, a theoretical physicist at Harvard, states his view as a general law about laws: "there's no exact equation. Everything is an approximation" [1]. He runs through the history to show it. Newton's gravity needed tiny corrections, and the first sign was a measurement of the drift in Mercury's orbit made in 1859 by Le Verrier, who, Strominger says, made one of the greatest discoveries in the history of science without knowing what good it would be [1]. Maxwell's electromagnetism needs quantum corrections, and they are not small: without quantum mechanics the electron would spiral into the proton and the atom would collapse [1]. His summary: "So every theory gets corrected as we learn more" [1].

General relativity is his special case, because of how its limit shows up. A perfectly normal system can evolve into a singularity where Einstein's equations become infinite and no longer apply [1]. Strominger does not treat that as a defect. He says it is a good thing that the theory "predicts its own demise," because the singularity shows exactly where improvement is needed, and that the improvement must involve quantum mechanics [1]. Asked whether the failures of a theory are wonderful to him, he answers, "Edges are wonderful because that keeps us in business" [1].

He applies the same standard to success. He calls the standard model "the most accurate theory in the history of human thought," with theory and experiment agreeing to 16 decimal places, and in the same conversation calls it a laundry list of hundreds of free parameters with no organizing principle [1]. Of dark energy he says he trusts the astronomers' measurements but "we have no friggin idea why it's there" [1].

He is just as plain about his own field. String theory is the one fully consistent framework that reconciles quantum mechanics and gravity, yet that only means it cannot be ruled out, and he does not expect a yes-or-no experiment in the foreseeable future [1]. Asked whether the theory is dead, he calls the question a logical error: "What it is is a stepping stone" [1]. He even grants it is logically possible that it will turn out to be a dead end [1].

Strominger extends the tolerance for error to people. Einstein wrote his field equations in 1915 and about 25 years later wrote a paper arguing that black holes did not exist; in the 1930s he wrote another arguing gravitational waves did not exist [1]. Strominger's verdict on how to score a scientist: "It's the number that you got right" [1]. Of himself he says, "people are always sort of disagreeing with me. And they're usually right, but I'm right enough" [1].

Michelle Thaller: science is limited on purpose

Thaller, an astrophysicist and retired NASA executive, makes the limit a matter of principle. The summary of her conversation with Joe Rogan quotes her: "You have to be honest about what you don't know" [2]. She calls this one of the most beautiful and one of the most frustrating, even scary, things about being a scientist [2].

Her rule is that science confines itself to "what is reproducible, consistently reproducible" [2]. In her words, "Science is limited deliberately" [2]. She explains what that looked like at NASA, where calls from people who said they had traveled in time or been abducted by aliens were often forwarded to her [2]. She listened, and told them they had had a profound experience, but also that "there's not much as a scientist that I can do with an individual experience," because she cannot run an experiment on it or have colleagues around the world repeat it [2]. She is careful about what that does and does not imply: "it's not that these aren't real" [2].

She lists the edges of her own field without hedging. On neutron stars, pretty much every model so far fails to match what is measured: "We cannot describe them yet. We need better physics" [2]. On the first instant of the universe: "We don't have the physics to describe high density, high gravity conditions," and "I don't think we understand how the Big Bang went off yet" [2]. She takes theories about the first fractions of a second with a big grain of salt and treats inflation as one possibility [2]. On time itself, the summary quotes her: "I don't think we have an answer to what time is" [2].

Thaller also shows what the discipline looks like when a scientist badly wants something to be true. Everyone she worked with at NASA believes there must be life elsewhere, and she says, "I always keep a bottle of champagne chilling" for clear evidence [2]. But fighter-jet UFO videos do not qualify, because unexplained is not the same as alien [2]. She describes living on "this sort of skeptical tightrope with hope" [2].

Carlo Rovelli: certainty is the obstacle

Rovelli, a theoretical physicist who works on quantum gravity, tells a family story to make the point. His father grew up believing his name was Franco until his birth certificate showed a different first name, and Rovelli's conclusion is "You can be wrong about everything, even about your own name" [3]. Certainty, he argues, is what blocks further understanding, and curiosity is what drives discovery [3].

He holds physics itself to that. He calls the field overrated in one sense: knowing the fundamental equations is like knowing the rules of chess, which is very different from knowing how to play, and chemistry, psychology, literature and emotion are all needed to understand the world [3]. He also describes what the best-tested theory leaves unsaid. Quantum mechanics "is solid," used everywhere, including in the microphone he is speaking into, yet it tells you what you will see if you do this or that and "it doesn't give a story of what happened" in between [3].

About his life's work, loop quantum gravity, he asks and answers his own question: "Are we sure? No, not yet" [3]. The reason is a rule: "Science is not sufficient to have ideas," they have to be tested [3]. Asked what happened before the Big Bang, he begins, "So first, we don't know. I think it has to be said clearly," and only then offers the bet he would make if forced, a collapsing universe that bounced [3]. Of the idea that black holes lead to other universes, which he credits to his friend Lee Smolin and calls beautiful, he says, "I am not convinced" [3].

Brian Cox: be delighted when you are wrong

Cox, a particle physicist at the University of Manchester, worries that science can look arrogant in public, as if scientists were announcing how the world is [4]. He says the reverse is the case: "I think it's a humble pursuit, ultimately science," and the reason for its success is that "you have to be delighted when you're wrong" [4]. If dark matter and dark energy turned out to be nonsense, he says, good scientists would be absolutely delighted, because ruling out the picture would be tremendously exciting [4].

His example is his friend Brian Schmidt, who measured supernova explosions in the 1990s and found the universe's expansion speeding up when no one expected it [4]. Schmidt thought the result was wrong but could find nothing wrong with his data [4]. Cox gives his reasoning: "I don't think this is right, but I can't see anything wrong with it. I'll publish it. Someone else will tell me where my mistake was" [4]. There was no mistake, and Schmidt won the Nobel Prize [4].

Cox is exact about what is and is not known. Ordinary matter is less than 5 percent of the universe, with about 25 percent dark matter and 70 percent dark energy [4]. Dark matter is inferred from several observations, but the particle has never been detected, and "until you really find that particle, then you don't know" [4]. On time he says, "And we don't actually know what it is" [4]. He admits that he and his colleagues do not understand Roger Penrose's cyclic cosmology [4]. He defines science as the best picture of nature we can manage at any moment, and says the key to it is being comfortable with not knowing, not being clever [4].

Lisa Randall: believe and question at once

Randall, a theoretical physicist at Harvard, opens her conversation by saying that physics teaches both our limitations and our abilities [5]. Asked for the limits of science, she answers, "I'm smart enough to know. I have no idea" [5]. A century ago, she notes, no one knew the universe was expanding or knew about quantum mechanics or special relativity, so "it would be premature to say we know limitations" [5].

She labels her own speculation before anyone else can. Of her idea that a disk of dark matter may have raised the odds of the impact that killed the dinosaurs, she says, "This was a speculative theory" [5]. What she likes about it is that it can be tested by mapping the positions and velocities of stars [5]. When critics said it made dark matter too weird, she replied, "We don't know what it is. So there's no normalization here" [5].

Her cautionary tale is the Large Hadron Collider. Finding the Higgs boson, proposed 50 years earlier, was a major victory [5]. But many physicists assumed in advance that the collider would also find what lies beneath the Standard Model, often supersymmetry, and that confidence was one factor in it becoming "the only machine in town" after the U.S. Superconducting Super Collider was cancelled [5]. Her lesson: "you don't want to assume things before they've been actually found" [5]. She says string theorists were early on "a little bit overly arrogant" about the problems they could solve, while adding that the theory has given useful tools and insights [5].

Her summary of the stance is advice to young people: "you have to believe really strongly what you do while questioning it all the time" [5].

Where They Agree, and Where They Don't

The agreement is broad. All five say that current theories are incomplete and name specific gaps [1][2][3][4][5]. All five treat that as normal. Strominger says there is no reason to suppose theories should ever stop being corrected, Randall is not sure physics will reach a bottom level, and Cox calls science the best picture available at a given moment [1][4][5].

They agree that being wrong is not shameful. Strominger scores scientists by what they got right, Cox wants them delighted to be wrong, and Randall says a scientist must work "at the same time knowing they could all be wrong" [1][4][5].

They agree on several of the same unknowns. Thaller and Cox both say we do not know what time is, and Strominger goes further, saying of space and time, "They're both illusions" [1][2][4]. Thaller, Rovelli and Cox all say the beginning of the universe is not understood [2][3][4]. Strominger and Randall both doubt a final theory, and Strominger recalls asking Steven Weinberg, "Why should there ever be a final theory?" [1][5].

They disagree on specifics, and two of them disagree with each other directly. Rovelli argues that asking where a particle is when it is not interacting with anything is a badly posed question, and that there is no single objective reality beneath our perspectives: "The dream of an absolute objectivity is a mistake" [3]. Randall pushes back on his claim that electrons exist only when they interact [5]. She accepts that an electron has no definite position until it is measured, but says something was there, which physicists describe with a wave function, and that an electron in an atom is really there [5]. Each has a reason. Rovelli starts from the fact that the theory predicts outcomes of interactions and tells no story in between [3]. Randall starts from the view that the world does not come into being because she measures it [5]. She adds that this is a dispute about interpreting quantum mechanics, not about the Standard Model, so both accept the same measured results [5].

They differ on the earliest universe. Thaller takes theories of the first fractions of a second with a big grain of salt [2]. Cox describes inflation as part of the standard picture and notes that it predicted the pattern of tiny ripples in the microwave background before that pattern was observed, while also saying it is unclear whether the universe had a beginning [4]. Rovelli would bet on a bounce [3]. Strominger says string theory has said almost nothing compelling about the Big Bang [1].

They work on rival approaches to quantum gravity, and each downgrades their own. Strominger works on string theory and says he does not believe it himself without an experiment [1]. Rovelli works on loop quantum gravity and says it is not yet confirmed [3]. Randall says judging string theory by success or failure misses the point [5].

They also use the word "limits" in two senses. Thaller means a boundary science draws for itself, around what can be reproduced [2]. Randall means how far science may eventually reach, which she declines to guess [5]. These do not conflict, but they should not be confused.

Finally, they differ in how the unknown feels. Cox says the most human feeling is to "stand on the edge of the known," and that knowing everything would make existence pointless [4]. Thaller finds the same edge beautiful and also frustrating and scary [2]. Randall says, "When things don't make sense, it really bugs me" [5].

Why It Works

Admitting ignorance is what makes the rest believable

Thaller and Rogan arrive at this together. Rogan says the admission matters because "otherwise we're not going to believe you with stuff that you do know," and Thaller agrees that this is why science has to be limited [2]. A field that claimed to explain everything would give no way to tell its tested results from its guesses [2].

Reproducibility filters out the single observer

Thaller's test is whether colleagues all over the world can do the same experiment [2]. Her LIGO example shows it working: two detectors recorded the same signal from two merging black holes, offset by exactly the light-travel time between the sites [2]. Her example of the first instinct is the discovery of the cosmic microwave background. Penzias and Wilson found the same faint noise in every direction and assumed, as she says a good scientist would, "that's probably a problem with your telescope" [2]. Only after they had cleaned out the pigeon droppings and the signal remained did it count [2].

Independent lines of evidence have to agree

Cox says the strength of the dark matter model is that many independent observations fit together, in the way the ages of the Earth, the sun and the universe all agree, and that alternatives which fix one observation tend to break others [4]. Randall says there are eight or nine independent ways to infer how much dark matter there is, and they all agree [5]. That agreement is why both accept dark matter while saying they do not know what it is [4][5].

Contradictions point to the next theory

Strominger says general relativity grew out of a contradiction between Maxwell's theory, in which nothing outruns light, and Newton's gravity, which acts instantly [1]. He adds a caution: not every contradiction is deep, and many are just mistakes [1]. The clash between quantum mechanics and general relativity is one that has resisted every attempt to explain it away [1].

Overconfidence has a cost

Randall's collider story is the evidence here: assuming what would be found contributed to having one machine instead of a more robust test [5]. Strominger gives the reverse case, where honesty about a flawed theory paid off. Einstein's 1914 version of general relativity predicted the bending of starlight wrong by a factor of two, and an expedition to measure it was captured during the war before it could, which gave him time to fix the theory [1].

Where It Applies, and Where It Doesn't

Humility about theory is not doubt about measurement. Thaller says time dilation is everyday engineering: without correcting GPS satellites for both speed and gravity, positions would be off by roughly six miles within a day [2]. Strominger says black holes definitely exist, and Rovelli recalls that as recently as 2000 a respected colleague told him they were not real [1][3]. An approximate theory can still be right to 16 decimal places inside its range [1].

It is not a license to believe anything. Thaller's point about UFO videos is that an unexplained observation stays unexplained [2]. When a Webb result hinted at organic molecules on a planet around another star, she reports that the rest of the scientists said "the data is not good enough yet" [2]. Randall says the right approach to a far-out idea is to exhaust the possibilities within existing frameworks first [5]. Cox notes that conspiracy theories about space assume governments are far more competent than they are [4].

It does not rule out what science cannot measure. Thaller says experiences outside the realm of measurement are not thereby unreal, and that she hopes we live in a larger reality than she can measure [2]. That is a statement about the reach of a method, not a finding.

It does not apply evenly to open research. Strominger admits nobody knows exactly what the holographic principle is, and says whether the soft hair result is the key error in Hawking's argument or a technical point remains to be seen [1]. Cox says physicists now think, to most people's satisfaction mathematically, that information does get out of black holes, though the physical picture is not understood [4]. Claims like these are moving, and several results mentioned in these conversations were described as disputed or unconfirmed when recorded, so check the current position before relying on them [2][4].

What To Do With It

1. Ask whether a claim has been reproduced by people other than the one making it. Thaller's standard is "consistently reproducible" [2].
2. Count the independent lines of evidence. Dark matter earns trust from eight or nine methods that agree, even with no particle found [4][5].
3. Separate what is measured from what is inferred from what is conjectured. Thaller does this for black holes, where she says our physics simply gives up on what is inside [2].
4. Look for the label. Randall calls her own idea speculative before describing it, and Rovelli says "we don't know" before offering a bet [3][5].
5. Ask where a theory stops working. Strominger's view is that every theory has such a place, and a source that cannot name it has not earned much trust [1].
6. Treat an unexplained result as unexplained. First suspect the instrument, as Penzias and Wilson did, and do not jump to the most exciting cause [2].
7. Publish or share the result you cannot fault, even if you doubt it, and let others look for the mistake, as Brian Schmidt did [4].
8. Do not assume what an experiment will find before it has found it. Randall's collider story shows the price [5].
9. Notice when you are defending an idea because it is yours. Rovelli's rule is that you can be wrong about anything, and Cox's is to be delighted when it happens [3][4].
10. Keep working on questions whose end you cannot see. Strominger says "the fact that we can't see to the end isn't a reason not to do it" [1].

Sources

1. Andrew Strominger on Black Holes, String Theory, the Holographic Principle, and Soft Hair · Lex Fridman Podcast
2. Michelle Thaller on Black Holes, Time, Quantum Weirdness and the Limits of Science · The Joe Rogan Experience
3. Carlo Rovelli on Relational Reality, Quantum Weirdness, AI Consciousness and the Case Against War · The Joe Rogan Experience
4. Brian Cox on Black Holes, the Great Silence and the Meaning of Life in the Universe · The Joe Rogan Experience
5. Lisa Randall on Dark Matter, the Dinosaur Extinction, and the Edges of Physics · Lex Fridman Podcast