Dad says enough of this biblical exegesis and hermeneutics nonsense. He wants more science and history of science for iconoclasts and Kuhnians. I said that if prophetic exegesis was good enough for Isaac Newton – who spent most of his writing life on it – it’s good enough for me. But to keep the family together around the spectroscope, here’s another look at what’s gone terribly wrong with institutional science.

It’s been thirty years since John Horgan wrote The End of Science, arguing that fundamental discovery was nearing its end. He may have overstated the case, but his diagnosis of scientific fatigue struck a nerve. Horgan claimed that major insights – quantum mechanics, relativity, the big bang, evolution, the double helix – had already given us a comprehensive map of reality unlikely to change much. Science, he said, had become a victim of its own success, entering a phase of permanent normality, to borrow Thomas Kuhn’s term. Future research, in his view, would merely refine existing paradigms, pose unanswerable questions, or spin speculative theories with no empirical anchor.

Horgan still stands by that thesis. He notes the absence of paradigm-shifting revolutions and a decline in disruptive research. A 2023 Nature study analyzed forty-five million papers and nearly four million patents, finding a sharp drop in genuinely groundbreaking work since the mid-twentieth century. Research increasingly consolidates what’s known rather than breaking new ground. Horgan also raises the philosophical point that some puzzles may simply exceed our cognitive reach – a concern with deep historical roots. Consider consciousness, quantum interpretation, or other problems that might mark the brain’s limits. Perhaps AI will push those limits outward.

Students of History of Science will think of Auguste Comte’s famous claim that we’d never know the composition of the stars. He wasn’t stupid, just cautious. Epistemic humility. He knew collecting samples was impossible. What he couldn’t foresee was spectrometry, where the wavelengths of light a star emits reveal the quantum behavior of its electrons. Comte and his peers could never have imagined that; it was data that forced quantum mechanics upon us.

The same confidence of finality carried into the next generation of physics. In 1874, Philipp von Jolly reportedly advised young Max Planck not to pursue physics, since it was “virtually a finished subject,” with only small refinements left in measurement. That position was understandable: Maxwell’s equations unified electromagnetism, thermodynamics was triumphant, and the Newtonian worldview seemed complete. Only a few inconvenient anomalies remained.

Albert Michelson, in 1894, echoed the sentiment. “Most of the grand underlying principles have been firmly established,” he said. Physics had unified light, electricity, magnetism, and heat; the periodic table was filled in; the atom looked tidy. The remaining puzzles – Mercury’s orbit, blackbody radiation – seemed minor, the way dark matter does to some of us now. He was right in one sense: he had interpreted his world as coherently as possible with the evidence he had. Or had he?

Michelson’s remark came after his own 1887 experiment with Morley – the one that failed to detect Earth’s motion through the ether and, in hindsight, cracked the door to relativity. The irony is enormous. He had already performed the experiment that revealed something was deeply wrong, yet he didn’t see it that way. The null result struck him as a puzzle within the old paradigm, not a death blow to it. The idea that the speed of light might be constant for all observers, or that time and space themselves might bend, was too far outside the late-Victorian imagination. Lorentz, FitzGerald, and others kept right on patching the luminiferous ether.

Logicians will recognize the case for pessimistic meta-induction here: past prognosticators have always been wrong about the future, and inductive reasoning says they will be wrong again. Horgan may think his case is different, but I can’t see it. He was partially right, but overconfident about completeness – treating current theories as final, just as Comte, von Jolly, and Michelson once did.

Where Horgan was most right – territory he barely touched – is in seeing that institutions now ensure his prediction. Science stagnates not for lack of mystery but because its structures reward safety over risk. Peer review, grant culture, and the fetish for incrementalism make Kuhnian normal science permanent. Scientific American canned Horgan soon after The End of Science appeared. By the mid-90s, the magazine had already crossed the event horizon of integrity.

While researching his book, Horgan interviewed Edward Witten, already the central figure in the string-theory marketing machine. Witten rejected Kuhn’s model of revolutions, preferring a vision of seamless theoretical progress. No surprise. Horgan seemed wary of Witten’s confidence. He sensed that Witten’s serene belief in an ever-tightening net of theory was itself a symptom of closure.

From a Feyerabendian perspective, the irony is perfect. Paul Feyerabend would say that when a scientific culture begins to prize formal coherence, elegance, and mathematical completeness over empirical confrontation, it stops being revolutionary. In that sense, the Witten attitude itself initiates the decline of discovery.

String theory is the perfect case study: an extraordinary mathematical construct that’s absorbed immense intellectual capital without yielding a falsifiable prediction. To a cynic (or realist), it looks like a priesthood refining its liturgy. The Feyerabendian critique would be that modern science has been rationalized to death, more concerned with internal consistency and social prestige than with the rude encounter between theory and world. Witten’s world has continually expanded a body of coherent claims – they hold together, internally consistent. But science does not run on a coherence model of truth. It demands correspondence. (Coherence vs. correspondence models of truth was a big topic in analytic philosophy in the last century.) By correspondence theory of truth, we mean that theories must survive the test against nature. The creation of coherent ideas means nothing without it. Experience trumps theory, always – the scientific revolution in a nutshell.

Horgan didn’t say – though he should have – that Witten’s aesthetic of mathematical beauty has institutionalized epistemic stasis. The problem isn’t that science has run out of mysteries, as Horgan proposed, but that its practitioners have become too self-conscious, too invested in their architectures to risk tearing them down. Galileo rolls over.

Horgan sensed the paradox but never made it central. His End of Science was sociological and cognitive; a Feyerabendian would call it ideological. Science has become the very orthodoxy it once subverted.