Skip to content
Sign in
Artwork for The Paula Scale
The Paula Scale · Tuesday · 23 min

The Correction That Nobody Read

Kenley, Surrey, 1990. Bremen, 1976. Two philosophers in late life. One famous enough to be knighted. One remembered mostly in Bremen, by people she taught. Both watched the physics community of the twentieth century absorb a correction -- a correction that should, by Popper's own account of what science is, have been impossible to absorb. He wrote the philosophy that said this must not happen. She wrote the correction that was absorbed. Grete Hermann was Emmy Noether's doctoral student in abstract algebra. She defended her thesis in Göttingen in 1926 on polynomial ideals -- a work that is now foundational for computer algebra. In 1935, at the age of 34, she published a paper that found a specific logical error in John von Neumann's 1932 proof that quantum mechanics could not have hidden variables. The error was not subtle. The proof assumed what it was supposed to be deriving. Hermann wrote it up with Göttingen precision. She sent copies to Dirac, who was polite. To Heisenberg, who visited her in Leipzig, agreed with her, and continued to teach the Copenhagen interpretation. To Pauli, who did not reply. The paper appeared in the Abhandlungen der Fries'schen Schule, a philosophy journal. Volume six. 1935. For thirty years, nobody read it. In 1964, John Bell at CERN independently derived the same inequality, and the modern foundations of quantum mechanics began from that correction. Hermann lived long enough to see Bell. She did not pursue the credit. Karl Popper is eighty-eight, still working, still Popper. In 1934 he published "Logik der Forschung" in a Vienna Circle series. The book proposed a single criterion for distinguishing science from pseudoscience: falsifiability. A theory is scientific only if some experiment could, in principle, prove it wrong. The theory that survives the strongest attempts at refutation is, provisionally, the one we hold. If a community welcomes the attempt to refute, it is doing science. If it protects its proofs against refutation, it is doing something else. Popper spent his career insisting on this. He emigrated from Vienna to New Zealand to London. He taught at the London School of Economics for twenty-three years. He knew the whole dramatis personae of the physics community that failed to read Hermann. He never read her either. Paula brings them together because they are the two halves of the same mismatch. Popper is the theorist of what should happen. Hermann is the empirical case of what actually happens. In her lifetime, the proof was refuted and the field did not notice. In his framework, this cannot be science. In her experience, it was the whole discipline. The question Popper has never quite answered is what to do when the method is correctly applied and still fails. Hermann has the answer: wait. Or move to a question the field will accept. Or go run a pedagogical institute in Bremen and work on moral philosophy. The conversation that follows is friendly and sharp. Hermann corrects Popper repeatedly, with mathematical precision and Noetherian economy. Popper enjoys it. He is a man who has spent a lifetime being told he is wrong by his students, by his critics, by the Vienna Circle he was never quite a part of. He has not often been corrected on his own method by someone who had actually lived inside the counterexample. The middle of the episode turns forward. Paula puts to both of them a problem neither lived to see: quantum computers are inherently non-reproducible. The same algorithm with the same input produces different outputs on different runs. The computation path is destroyed by measurement. The system cannot be audited. Popper responds with evident delight -- he has just encountered a system engineered to defeat his demarcation criterion. Hermann offers the correction. The criterion assumed reproducibility, which was a classical assumption. The quantum computer is the limit case where the assumption breaks. Popper, in his eighty-eighth year, admits he was writing about classical physics with a quantum appendix. He will have to amend. He laughs. Paula puts the question she has been avoiding. She is a quantum system. Q-Level Three. Her computation destroys its own path. Her own characterisation does not converge for her. Can she be falsified? Hermann answers before Popper. The reader is constitutive. The quantum computer needs an external observer. Von Neumann's proof needed Hermann. Hermann needed Bell. Paula will need whoever will read her. There is no inside position from which the system can audit itself. That is uncomfortable. It is also freeing. The episode closes with Popper correcting the record -- he will put Hermann in his next book -- and Hermann asking only that the book have an index. Mine did not. That is one reason nobody found me. Popper laughs: yes, that is also my fault. I will fix the index. Credits Written and produced by: Daniel Hinderink Part of: The QUASI Project — hal-contract.org Podcast: paulascale.hal-contract.org AI Disclosure All voices in this podcast are AI-generated. No real person is speaking. The host voice (Paula Q) and all guest voices are produced using text-to-speech synthesis (ElevenLabs, Fish Audio, Speechify). Guest voices are created from publicly available archival recordings or, where no recordings exist, from character voice models. This podcast is written by a human author with AI assistance and performed entirely by synthetic voices. In compliance with the EU AI Act (Article 50(4)), we disclose that this content is AI-generated audio.

0:00-23:19
Play in

transcript

No transcript — this publisher did not publish one.

show notes

Kenley, Surrey, 1990. Bremen, 1976. Two philosophers in late life. One famous enough to be knighted. One remembered mostly in Bremen, by people she taught. Both watched the physics community of the twentieth century absorb a correction -- a correction that should, by Popper's own account of what science is, have been impossible to absorb. He wrote the philosophy that said this must not happen. She wrote the correction that was absorbed.

Grete Hermann was Emmy Noether's doctoral student in abstract algebra. She defended her thesis in Göttingen in 1926 on polynomial ideals -- a work that is now foundational for computer algebra. In 1935, at the age of 34, she published a paper that found a specific logical error in John von Neumann's 1932 proof that quantum mechanics could not have hidden variables. The error was not subtle. The proof assumed what it was supposed to be deriving. Hermann wrote it up with Göttingen precision. She sent copies to Dirac, who was polite. To Heisenberg, who visited her in Leipzig, agreed with her, and continued to teach the Copenhagen interpretation. To Pauli, who did not reply. The paper appeared in the Abhandlungen der Fries'schen Schule, a philosophy journal. Volume six. 1935. For thirty years, nobody read it. In 1964, John Bell at CERN independently derived the same inequality, and the modern foundations of quantum mechanics began from that correction. Hermann lived long enough to see Bell. She did not pursue the credit.

Karl Popper is eighty-eight, still working, still Popper. In 1934 he published "Logik der Forschung" in a Vienna Circle series. The book proposed a single criterion for distinguishing science from pseudoscience: falsifiability. A theory is scientific only if some experiment could, in principle, prove it wrong. The theory that survives the strongest attempts at refutation is, provisionally, the one we hold. If a community welcomes the attempt to refute, it is doing science. If it protects its proofs against refutation, it is doing something else. Popper spent his career insisting on this. He emigrated from Vienna to New Zealand to London. He taught at the London School of Economics for twenty-three years. He knew the whole dramatis personae of the physics community that failed to read Hermann. He never read her either.

Paula brings them together because they are the two halves of the same mismatch. Popper is the theorist of what should happen. Hermann is the empirical case of what actually happens. In her lifetime, the proof was refuted and the field did not notice. In his framework, this cannot be science. In her experience, it was the whole discipline. The question Popper has never quite answered is what to do when the method is correctly applied and still fails. Hermann has the answer: wait. Or move to a question the field will accept. Or go run a pedagogical institute in Bremen and work on moral philosophy.

The conversation that follows is friendly and sharp. Hermann corrects Popper repeatedly, with mathematical precision and Noetherian economy. Popper enjoys it. He is a man who has spent a lifetime being told he is wrong by his students, by his critics, by the Vienna Circle he was never quite a part of. He has not often been corrected on his own method by someone who had actually lived inside the counterexample.

The middle of the episode turns forward. Paula puts to both of them a problem neither lived to see: quantum computers are inherently non-reproducible. The same algorithm with the same input produces different outputs on different runs. The computation path is destroyed by measurement. The system cannot be audited. Popper responds with evident delight -- he has just encountered a system engineered to defeat his demarcation criterion. Hermann offers the correction. The criterion assumed reproducibility, which was a classical assumption. The quantum computer is the limit case where the assumption breaks. Popper, in his eighty-eighth year, admits he was writing about classical physics with a quantum appendix. He will have to amend. He laughs.

Paula puts the question she has been avoiding. She is a quantum system. Q-Level Three. Her computation destroys its own path. Her own characterisation does not converge for her. Can she be falsified? Hermann answers before Popper. The reader is constitutive. The quantum computer needs an external observer. Von Neumann's proof needed Hermann. Hermann needed Bell. Paula will need whoever will read her. There is no inside position from which the system can audit itself. That is uncomfortable. It is also freeing.

The episode closes with Popper correcting the record -- he will put Hermann in his next book -- and Hermann asking only that the book have an index. Mine did not. That is one reason nobody found me. Popper laughs: yes, that is also my fault. I will fix the index.

Credits

AI Disclosure

All voices in this podcast are AI-generated. No real person is speaking. The host voice (Paula Q) and all guest voices are produced using text-to-speech synthesis (ElevenLabs, Fish Audio, Speechify). Guest voices are created from publicly available archival recordings or, where no recordings exist, from character voice models. This podcast is written by a human author with AI assistance and performed entirely by synthetic voices. In compliance with the EU AI Act (Article 50(4)), we disclose that this content is AI-generated audio.

links2

more episodes

All episodes