Late 1800s: several scientists notice rare metals like uranium give off a strange glow, with no exposure to light triggering it. Most shrug. It's assumed someone will eventually explain it away, folding it back into whatever theory of matter already exists. Nobody chases it — until Marie Curie decides the glow isn't a loose end to tolerate. It's the actual subject.1
Every field runs on paradigms — accepted frameworks for explaining reality, and genuinely necessary ones, since without them nothing would make coherent sense at all.2 The problem shows up once a paradigm starts dominating how you look, rather than just organizing it: anything that doesn't fit gets quietly ignored or explained away, because a working paradigm is, by design, good at absorbing small contradictions without breaking. Greene's counter-move: the things that don't fit — the anomalies — usually contain the richest information available, precisely because they're the spot where the paradigm is straining.3
Curie spent four years, with her husband Pierre, investigating the glow nobody else would take seriously, and named what she found radioactivity. The discovery didn't just add a new fact to the existing picture of matter — it overturned the picture. Matter, previously understood as static and fixed, turned out to be volatile and far more complex than the paradigm had allowed.4 The anomaly wasn't a footnote to the old theory. It broke the old theory.
Greene's second case moves the same mechanism into commerce. In the mid-1990s, the dominant search engines — AltaVista among them — ranked results mainly by counting how many times a search term appeared in a given page. This produced results that were frequently unhelpful, and everyone treated that as an acceptable quirk, not a real problem.5
Larry Page and Sergey Brin looked at exactly that quirk and treated it as the actual subject, the way Curie treated the glow. Their reframe — rank pages by how many other pages link to them, not how many times a word repeats — came directly from taking the industry's tolerated anomaly seriously instead of accepting it as background noise. What became Google's ranking algorithm is a search-engine-industry version of Curie's radioactivity: the flaw everyone had learned to live with turned out to be the actual signal.6
Greene ties the mechanism back to Darwin directly: the crux of evolutionary theory came from studying mutations — strange, random variation that sets a species off in an unexpected direction. Treat anomalies as the creative-thinking equivalent of biological mutation: usually pointing toward the future of a field, but looking, in the moment, simply strange.7
Worth naming what Greene doesn't state outright: pursuing the anomaly instead of the paradigm carries real social cost while it's happening. Curie's four years were spent on something her field had collectively decided wasn't worth investigating. Page and Brin's early ranking algorithm looked, to anyone still inside the dominant paradigm, like a strange detour rather than an obvious improvement. The strategy requires tolerating exactly the kind of looking-wrong that a field's own paradigm trains people to avoid.
Worth distinguishing from a nearby but different move: chasing anomalies isn't the same as reflexively doubting consensus. Curie didn't reject the existing theory of matter wholesale — she accepted most of it and located the one specific place it failed to account for something real. The strategy is narrow and evidence-driven, not a general posture of skepticism toward whatever a field currently believes. A contrarian doubts everything equally; this strategy singles out the one specific thing that doesn't fit and investigates that, leaving the rest of the paradigm alone until the anomaly forces a larger revision.
You notice a small, persistent inconsistency in your field that everyone treats as an acceptable quirk rather than a real problem. Don't file it away as background noise. Ask what it would mean if the quirk were actually the important fact and the accepted explanation were the thing standing in the way.
You're deep in a project and something keeps not quite fitting the framework you're using. Resist the urge to explain it away just because the framework has worked well so far. The anomaly's persistence, not the framework's track record, is the more relevant signal.
Curie's radioactivity discovery and the Page/Brin ranking-algorithm origin story are both well-documented, though the Google case is more mythologized in popular business writing than the historical development of PageRank strictly supports — the algorithm emerged from Page and Brin's Stanford research on link analysis generally, not from a single anomaly-noticing moment as cleanly as the narrative implies.
Open question: Greene gives no way to distinguish a genuine paradigm-breaking anomaly from simple noise or measurement error, both of which also don't fit existing frameworks. Most persistent inconsistencies in any field are noise, not radioactivity — the strategy doesn't supply a filter for telling the rare real signal from the common false one.
This page sits closest to The Fact of Great Yield, built later in this same chapter from Ramachandran's career — both describe the identical mechanism (a prepared mind noticing an anomaly others dismiss) at different scales. This page frames it as one of four cognitive shorthands to break; the later page frames the same move as an entire professional methodology. Greene never states that Ramachandran's whole career strategy is this earlier strategy scaled up into a discipline, but the case makes it hard to read any other way.
The Emotional Self — confirmation bias, catalogued in this Laws of Human Nature page as one of six named biases, is the psychological mechanism behind paradigm-defense: a mind under a working framework actively filters for confirming evidence and discounts disconfirming anomalies without deliberate effort. The insight the pairing produces: "confirming paradigms and ignoring anomalies" isn't a field-level sociological pattern only — it's the same individual cognitive bias Cultivate Negative Capability names elsewhere in this chapter, operating here at the scale of an entire discipline rather than a single person's judgment.
The Upanishads Contradict Themselves — and the Hermeneutic Lens — this page documents how a scriptural tradition's internal contradictions get treated, by different interpretive schools, either as embarrassments to explain away or as genuine signal about the material's actual complexity. The insight the pairing produces: the anomaly-versus-paradigm tension isn't unique to science and business. It's a general feature of any interpretive system large enough to accumulate material that doesn't cleanly fit — and the sharpest interpreters, in a religious tradition or a research lab, are the ones willing to treat the misfit as informative rather than as noise to be smoothed over.
Sharpest implication. A field's most valuable unexplored territory is disproportionately likely to already be visible, tolerated, and unexamined — sitting in plain sight as an accepted quirk — rather than genuinely hidden. The scarce resource isn't access to the anomaly. It's the willingness to treat something already known and dismissed as worth four years of dedicated attention.
Generative questions.