Smolin, Lee
▶ Watch on YouTubeCore ideas
Smolin opens by declaring himself a realist: he believes in an objective world independent of human perception, and that scientific progress means coming asymptotically closer to a true description of that world, even though certainty about "arriving" there is never fully attainable (a question he says is "for people 5,000 years in the future"). He rejects the idea of a single, well-defined "scientific method" (citing philosopher Paul Feyerabend's argument, building on Karl Popper, that historically significant scientific progress has often required violating naive methodological rules), instead framing science as a community bound by shared ethical norms — honest reporting of results, rigorous cross-checking — enforced through training (a PhD representing the ability to defend a claim against expert scrutiny) rather than a fixed algorithm. He illustrates the messiness of theory-testing with Galileo's falling-ball experiment, which was consistent with both Aristotelian and Galilean physics depending on background assumptions (like inertia), showing that decisive theoretical transitions (Aristotelian to Newtonian physics) require assembling many interlocking cases over long periods, not single crucial experiments.
Smolin distinguishes realism from the anti-realist views held by many quantum mechanics founders (citing Niels Bohr's view that physics describes not nature itself but our interactions and conversations about nature, freely using useful concepts like "particle" or "wave" without claiming they describe mind-independent reality). He personally maintains realism as an aspiration while acknowledging what we perceive is heavily constructed and processed by the brain. Invoking his friend Eric Weinstein's observation that geniuses across fields (Einstein, Freud, Edison) all "carried their luggage" for centuries before anyone thought to put wheels on it, Smolin reflects that revolutionary insight requires some combination of luck, stubbornness, confidence bordering on arrogance, and above all sustained, almost obsessive intellectual "ferocity" of the kind he attributes to Einstein — while cautioning that the same traits which enable breakthroughs can just as easily lead a scientist astray.
The core of the conversation covers what Smolin calls Einstein's "unfinished revolution": the early-20th-century twin breakthroughs of relativity and quantum mechanics remain unreconciled, since general relativity (governing large-scale gravity and cosmology) and quantum theory (governing small-scale phenomena) have never been successfully unified, and quantum mechanics itself contains an unresolved "measurement problem" — its mathematics specifies two incompatible modes of time evolution, one for ordinary (unobserved) quantum systems and a different, discontinuous one for measurement/observation, with no accepted account of how or why the second applies. Smolin (sharing, he says, Einstein's own eventual view) believes quantum mechanics is consistent but incomplete, requiring additional hidden variables or structure to become a fully realist theory, and frames his own research program (developed with collaborators Marina Cortês, Roberto Mangabeira Unger, and Carlo Rovelli) around treating causality — the fundamental idea that discrete "events" cause other events — as truly fundamental, while space itself (and even the combined four-dimensional spacetime of general relativity) is merely an emergent, approximate description arising from sufficiently complex networks of causally-connected events; time, in this view, is fundamental and consists in the "continual creation of events from existing events," while the future does not yet exist and only the past (and a present) are real.
Connected to this, Smolin discusses locality, distinguishing the well-tested notion from quantum field theory (spacelike-separated field operators commute) from a stronger notion tested via Bell's theorem — the assumption that the "reality" of a distant entangled particle cannot be affected by a distant observer's choice of what to measure — which experiments (over distances up to roughly 170 kilometers) have confirmed is violated by many standard deviations, meaning nature is fundamentally non-local in this Bell-theorem sense. He argues that abandoning strict locality and treating causality (rather than fixed geometric spacetime) as fundamental are the two central moves needed to complete Einstein's revolution. Asked about Sean Carroll's favored many-worlds interpretation, Smolin explains his skepticism stems partly from a philosophical starting point he and Unger adopted (that there is only one universe, which happens once) and partly from unresolved technical difficulties in deriving the Born rule (quantum probabilities) once the measurement postulate is discarded in favor of pure Schrödinger evolution — noting that even sophisticated modern defenses of many-worlds (from Oxford philosophers like David Wallace and Simon Saunders, using decision theory to justify effective "betting" probabilities within an Everettian multiverse) remain, by those researchers' own private admission to him, not yet fully successful.
On theoretical methodology, Smolin argues (again following Einstein's example) that progress in a genuinely new domain of physics requires first discovering foundational principles, then building mathematical models that apply them — contrasting this with the current state of quantum gravity research, where more than a dozen partially-successful approaches (string theory, loop quantum gravity, causal dynamical triangulations, causal sets) each solve some problems while getting stuck on others, without yet establishing the right foundational principles. He recounts the origin of his controversial 2006 book "The Trouble with Physics": a personal conversation with string theorist Andy Strominger about the difficulty of counting or classifying the vast number of possible string theory solutions (Calabi-Yau manifolds and their generalizations) inspired Smolin, then reading extensively in evolutionary biology, to propose "cosmological natural selection" — a speculative hypothesis (published 1992, expanded in his 1997 book "The Life of the Cosmos") that universes reproduce via black holes and undergo a selection-like process, with each "offspring" universe having randomly varied fundamental constants, potentially explaining why our universe's parameters appear fine-tuned without invoking the unfalsifiable anthropic principle. Smolin says his critique of string theory's community dynamics in "The Trouble with Physics" was intended methodologically rather than personally (criticizing papers and community incentive structures, not individuals), regrets that readers didn't recognize the same critique applies equally to his own quantum-gravity community, and expresses cautious optimism that younger researchers today show markedly less tribal attachment to a single approach (loops vs. strings vs. causal sets), increasingly attending cross-community conferences — a sociological shift he sees as necessary preparation for eventually finishing Einstein's unfinished revolution.
Excerpts
On why space (unlike time) is not fundamental in his view:
“I think time is fundamental and goes all the way down, and space does not... there's a level of description in which there are events, and events create other events, but there's no space — they don't live in space, they have an order in which they cause each other, and that is part of the nature of time for us. But there is an emergent, approximate description [of space].”
On Bell's theorem and non-locality:
“You make that assumption — that's called Bell locality — and you derive a certain inequality... you can test that pretty directly in experiments which create pairs of photons... and it's wrong by many sigma. It's not true that whatever is real about particle B is unaffected by the choice that the observer makes as to what to measure in particle A, no matter how far away they've propagated from each other.”
On the origin of "cosmological natural selection" as an alternative to the anthropic principle:
“I took their concept — the word 'landscape' from theoretical biology — and made a scenario about how the universe as a whole could evolve to discover the parameters of the standard model... almost every universe would share the property that its fitness was maximized to some extent... and I could deduce predictions that could be tested from that, and I compared it to the anthropic principle, where you weren't able to make tests or falsifications.”