Four prizes: De Brigard's MacArthur, Danks's AI2050, Norton's Lakatos, a lycéens' vote

Prizes in philosophy tend to arrive in clusters, and this autumn's cluster has a shape. Of the four confirmed in the past month, two went to philosophers whose daily work is done with scanners, datasets and engineers rather than with texts; a third went to a historian of physics for an argument that logicians have been refusing for twenty years; and the fourth was decided by a vote of French seventeen-year-olds. We wrote four days ago about Nussbaum's Kluge Prize, Vetter's Leibniz and Dixsaut's grand prix. This is the next instalment, with the usual note at the end about what we could not stand up.

De Brigard's MacArthur: memory as a kind of imagining#

The John D. and Catherine T. MacArthur Foundation announced its 2026 class of fellows on Tuesday 29 September: twenty people, each receiving 800,000 dollars paid over five years with no conditions attached. One of them is a philosopher. Felipe De Brigard is professor of philosophy and associate professor of psychology and neuroscience at Duke, where he runs the Imagination and Modal Cognition Lab, and the foundation's one-line citation describes him as "a Duke University philosopher using cognitive science to investigate the relationship between memory and forgiveness".

The citation undersells the programme. De Brigard's work over nearly twenty years has been an attack on the picture of memory as an archive, a store from which the past is retrieved more or less intact. His behavioural and imaging studies, with collaborators at Duke and Harvard, found that the brain regions active when people remember an episode overlap heavily with those active when they imagine how the same episode might have gone otherwise. Remembering and counterfactual thinking look, on this evidence, like two uses of one system whose job is to simulate possibilities rather than to replay the past. Errors, distortions and confabulations are then features of a flexible machine rather than failures of a faithful one. He told a North Carolina television station this week that the comparison he prefers is with vision: perception can mislead without making all perception unreliable, and memory is the same.

The foundation telephoned him, he said, during a working lunch on 27 August; he still does not know who nominated him. Most of the money will go to the Memory and Forgiveness project, which for five years has run field studies in his native Colombia with victims of political violence, asking whether changing the emotional charge attached to a memory of harm, without changing its content, is what forgiveness consists in. Readers of this site met the same philosopher in a different role earlier this year, when he and Sarah Robins shared the Lebowitz Prize for defending opposite answers to the question whether remembering is a kind of imagining. Two major awards in a year for the philosophy of memory is a signal that the field has moved from the margin of philosophy of mind to somewhere near its centre.

Danks's AI2050: what if the values move?#

Schmidt Sciences, the research funder established by Eric and Wendy Schmidt, announced the fifth cohort of its AI2050 programme on 1 October: twenty-eight fellows, twenty-two early-career and six senior, with up to 16 million dollars committed to the cohort over three years and 80 million to the programme so far. The six senior fellows are Francis Bach of Inria, Mikhail Belkin of UC San Diego, Sendhil Mullainathan of MIT, Alondra Nelson of the Institute for Advanced Study, Bernhard Schölkopf of the ELLIS Institute Tübingen and the Max Planck Institute for Intelligent Systems, and David Danks, professor of philosophy, artificial intelligence and data science at the University of Virginia. Danks is the only philosopher among the twenty-eight. Daily Nous, reporting the award on 6 October, put his senior fellowship at one million dollars over three years; Schmidt Sciences' announcement gives only the cohort total, and we were not able to find a per-fellow figure on its pages, so the million is on Daily Nous's authority.

The project is titled "human-AI value coevolution", and the title carries a complaint about the alignment debate. That debate assumes the values are fixed and the task is to fit the machine to them. Danks's project starts from the observation that AI systems and the values they are built to serve change one another over time, as earlier technologies changed the societies that adopted them. The question then becomes not how to align systems to human values but how to understand, and to some extent steer, a process in which the two change each other. Schmidt Sciences' summary says he intends to turn the findings into practical guidance for AI design, development and governance. James Manyika, the programme's co-chair and a Google senior vice-president, offered the philosophers' own line in the announcement: "the most important questions to get right about AI cannot be answered by computer science alone".

Danks is a plausible person to carry that brief. He spent years at Carnegie Mellon working on causal learning and cognitive science before moving to UC San Diego and then, in 2025, to a foundation-endowed chair at Virginia split between the arts and sciences college and the school of data science; he has sat on American federal advisory bodies on AI. The sum is unusual. Philosophical research grants are usually counted in tens of thousands; a million-dollar award to a philosopher-led project, if the figure stands, is one of the larger single grants the discipline has seen, and it sits alongside the hiring of philosophers by AI companies as evidence that the funders of the field have concluded value theory is not optional.

Norton's Lakatos: induction without rules#

The Lakatos Award is the philosophy of science's book prize. It is administered from the London School of Economics by an international management committee, funded by an endowment from the Latsis Foundation in memory of Imre Lakatos, and worth 10,000 pounds; the committee decides on the strength of anonymous selectors' reports and may make no award at all. On Monday 14 September the department announced that the 2026 award goes to John D. Norton, Distinguished Professor at the University of Pittsburgh, for two books from the University of Calgary Press, The Material Theory of Induction (2021) and The Large-Scale Structure of Inductive Inference (2024). Norton will give a public lecture at the LSE on a date still to be fixed.

The selectors called the two volumes "a very strong case for the material account of inference", "a work of exceptional scholarly rigour" and "unusually clear and readable, indeed almost uniquely so". The argument they are praising is one of the few genuinely new answers to Hume's problem in a century. Philosophers have spent that century looking for a formal rule that licenses inductive inference whatever its subject matter: Mill's methods, Bayesian conditionalisation, inference to the best explanation. Norton's claim is that there is no such rule and cannot be. Every real inductive inference in science is warranted by facts about its domain. What lets a chemist infer the melting point of bismuth from a single sample is not a logical schema but the material fact that elements are chemically uniform. Remove the fact and the inference collapses; supply a different fact and a different inference is licensed. Justification, on this view, happens inside science, by science, and the second book takes on the obvious objection, that the whole structure then hangs in the air, by arguing that the local warrants support one another in a web that is coherent without being founded.

Norton is best known outside the field for his work on Einstein, and the material theory grew from that work: a historian who had watched physicists actually reason found that none of the philosophers' schemas described what they did. The prize is a departure for the award, whose recent winners, Chirimuuta, Hoefer, Massimi, Dutilh Novaes, Plutynski and Shea, were mostly honoured for books about particular sciences. It is also a pointed choice for an award named after Lakatos, whose research programmes were themselves a reply to the collapse of a single formal criterion, as we noted in our piece on Popper's falsifiability. Norton's answer is more radical than Lakatos's: no criterion, local or global, only facts.

The lycéens choose argument over contempt#

France makes philosophy a compulsory subject in the final year of the lycée, and the public-school philosophy teachers' association, the APPEP, runs a book prize to go with it. The Prix lycéen du livre de philosophie works like the Goncourt des lycéens: a short selection of recent books is read in class over the winter, the pupils vote in May, and the author of the winner receives the prize from the pupil jurors in the autumn. This year's selection was Estelle Ferrarese's Une philosophie des sanglots (Rivages), Jean-Marc Mouillie's Cinq apparitions du ciel étoilé: Kant et le sens de l'existence (Les Belles Lettres) and Antoine Vuille's Contre la culture du clash: débat d'idées et démocratie (éliott). The pupils chose Vuille. The prize will be presented on Tuesday 3 November at 14:00 at the Bibliothèque nationale de France, François-Mitterrand site.

Vuille is not a name the profession knows. He holds a doctorate from the University of Neuchâtel, where he was an assistant and then a lecturer, and now teaches philosophy and French in a Swiss gymnase; he has also published a collection of short stories. His book's starting point is blunt: "we too often content ourselves with despising the political ideas we do not share". Its proposal is to put argument back at the centre of public debate, on the assumption that if a disagreement persists, each side probably has some reasons on its side. A reader of MacIntyre will hear the echo of After Virtue's interminable arguments; a reader of Rorty will hear the hope that a public can keep talking without first agreeing on the foundations. That seventeen-year-olds, offered a book on grief and a book on Kant's starry sky, chose the one about how to disagree is the kind of result that tells you something about the year.

What we could not confirm#

The Turkish-language reporting that reached us gave vote totals for the lycéens' prize, 820 pupils in 44 schools and 387 votes for Vuille. We could not open the school pages those figures came from and the APPEP's own site does not publish them, so they are omitted here. On the AI2050 award, as noted, the one-million-dollar figure is Daily Nous's and not Schmidt Sciences'. The Royal Institute of Philosophy's Nayef Al-Rodhan Prize, whose longlist we covered, has not yet announced a winner; a report that it will do so on 4 November is plausible but not on the Institute's pages, and we have left it out. A separate report that Japan's Nakamura Hajime Eastern Studies Prize has gone to the Indologist Kiyotaka Yoshimizu is unconfirmed on the awarding body's site and is not included.