A prize for disagreement: the 2026 Lebowitz goes to two rival theorists of memory

Academic prizes usually reward a person. The Lebowitz Prize rewards a disagreement.

The rule is plain. The winners must be two philosophers holding opposing views on a question of current interest in the discipline. The two present their positions and debate one another at the annual Lebowitz symposium, held at a divisional meeting of the APA. Each receives an honorarium of $25,000.

The prize is given jointly by the American Philosophical Association and the Phi Beta Kappa Society.

The 2026 winners were announced on 26 August.

Felipe De Brigard — professor of philosophy and of psychology and neuroscience at Duke University, a member of the Center for Cognitive Neuroscience and director of the Imagination and Modal Cognition Laboratory (the IMC Lab) at the Duke Institute for Brain Sciences.

Sarah Robins — professor of philosophy at Purdue University and founding co-director of Purdue's Center for Cognition, Agency and Intelligence (CAIC).

What is the disagreement about?#

The first comment posted under the announcement asked exactly the right question: what precisely do these two disagree about?

The answer lies on the axis that has organised the philosophy of memory for twenty years: causation, or simulation?

Causalism#

The traditional view — the line Robins defends — holds that for a mental state to count as remembering there must be an appropriate causal connection between the event remembered and the content present at the moment of recall.

What makes that connection possible is a memory trace which preserves the content. The event you live through leaves a trace in the brain; when you remember, you gain access to the trace.

The strength of the view is that it can separate remembering from imagining. To “remember” something true by accident is not really to remember it, because the causal link is missing.

Robins's 2016 paper “Representing the past: memory traces and the causal theory of memory” is the contemporary defence of that position.

Simulationism#

The opposing view rests on a finding from neuroscience: remembering and imagining draw largely on the same neural system. The networks that fire when we recall the past fire again when we envisage the future.

The thesis drawn from this is radical. Remembering is not the reading of a record of the past but the reconstruction of it. In which case neither the condition of an appropriate causal connection nor memory traces are needed.

The view explains the fallibility of memory well. Misremembering is not a malfunction but a consequence of the system working normally — because the system was never built to keep records but to construct.

Where De Brigard stands#

What makes De Brigard interesting is that he is not a pure simulationist.

He argues that simulationism is probably wrong to posit a single cognitive system for mental time travel. And further: simulationism cannot rid itself of memory traces altogether, since traces remain explanatorily indispensable.

He proposes instead that traces be rethought so as to be compatible with a constructive account of memory. In recent work he argues that the debate has largely been set up wrongly, and offers a model of reconstruction capable of reconciling the causal and the simulationist views.

The “opposition” the prize rewards is therefore not a simple affair of two camps. Robins defends the necessity of the condition; De Brigard questions the way the condition has been framed.

Why the argument matters#

This may look like a technical matter confined to the laboratory. It is not.

Law: the reliability of eyewitness testimony depends on whether memory is a record or a construction. If misremembering really is the system working normally, the hierarchy of evidence in criminal trials needs rethinking.

Identity: since Locke there has been a tradition of grounding personal identity in memory. If memory is continually rebuilt, what is the ground of being “the same person”?

Artificial intelligence: this is exactly the side question raised by the debate about consciousness we take up today. What is it for a language model to “remember”? Is there an appropriate causal connection between the training data and the output — or is there only reproduction?

The name of De Brigard's laboratory — Imagination and Modal Cognition — already announces this breadth: research on memory is treated as part of the capacity to think about what might have been.

On the design of the prize#

The most interesting thing about the Lebowitz Prize is less the money than the form it establishes.

Academic prizes usually consolidate a position: the winner was right. Lebowitz refuses to determine who is right, and makes a well-constructed disagreement visible instead.

It institutionalises a virtue peculiar to philosophy. The health of a field is measured not by how many people agree but by how well its disagreements are put together.

The Philosopher's Annual selection we discussed a few days ago was a list of verdicts. Lebowitz does the opposite.

As a model for philosophy prizes in Turkey it is worth watching: a prize can be built to make an argument visible rather than to raise up a name.

---

The Lebowitz symposium is held at an APA divisional meeting in the year the prize is awarded. Past winners are listed at apaonline.org.