Mary's Room

In brief

Frank Jackson published the thought experiment in "Epiphenomenal Qualia" (1982) and sharpened it in "What Mary Didn't Know" (1986). Imagine Mary, a brilliant scientist who has spent her whole life in a black-and-white room, learning about the world through a black-and-white monitor. She specializes in the neuroscience of color vision. She knows every physical fact there is: the wavelengths of light, the behavior of the retina, the firing of neurons, the words people use when they see red. She knows all of it. Then one day she leaves the room and sees a ripe tomato for the first time. Does she learn something new?

Jackson's intuition, and the intuition the case pumps in most people, is that she does: she learns what it is like to see red. But if she already knew every physical fact and still learns something on leaving the room, then what she learns is not a physical fact. There are facts about conscious experience that all the physics in the world leaves out. And that, Jackson argued, means physicalism (the view that everything is physical) is false. The strange twist is that Jackson, roughly twenty years later, decided the argument was wrong and went back to being a physicalist.

The full treatment

The problem it answers

The dominant view in modern philosophy of mind is physicalism: the mind is the brain, mental states are physical states, and a complete physics would leave nothing out. The knowledge argument is an attempt to refute this by way of a single vivid case. It targets the felt, qualitative character of experience, what philosophers call qualia: the redness of red, the sting of pain, the taste of coffee. Jackson's claim is that no amount of physical information captures these, and Mary is the proof.

The argument belongs to the same family as Thomas Nagel's "What Is It Like to Be a Bat?" (1974) and anticipates the framing David Chalmers would later give it as the hard problem of consciousness. All three press the same worry from different angles: objective, third-person science seems structurally unable to reach the subjective, first-person fact of what an experience is like from the inside. Nagel asks whether we could ever know what it is like to be a creature radically unlike us. Chalmers asks why there is any felt experience at all. Jackson's contribution is the most surgical of the three: a case designed to isolate exactly the residue that physics allegedly cannot touch.

How the argument works

Set out plainly, the reasoning has three steps.

1. Before leaving the room, Mary knows all the physical facts about color and color vision.
2. On leaving the room and seeing red, Mary learns something new (what it is like to see red).
3. Therefore there are facts about color experience that are not physical facts, and physicalism is false.

The force is in the first two premises. Premise one is stipulated to be as strong as it can be: Mary is not merely well informed, she is complete on the physics. Premise two rests on the reader's own reaction: nearly everyone agrees that Mary would react to her first sight of red with genuine discovery, an "ah, so that's what it's like." Jackson's move is to say that discovery is the acquisition of a fact she did not have before. If she gains a fact after already having all the physical facts, the new fact is non-physical.

The key thinker and the recantation

Frank Jackson (born 1943), an Australian philosopher, meant the argument to support a specific and unfashionable position: epiphenomenalism about qualia. On this view qualia are real, non-physical byproducts of brain processes that themselves have no effect on the physical world. The title "Epiphenomenal Qualia" is the point of the paper, not a footnote to it. Jackson accepted the awkward consequence that these felt qualities do no causal work, which is part of why the position drew fire.

Then Jackson changed his mind. In papers through the late 1990s and most fully in "Mind and Illusion" (2003), he concluded that his own argument had to be unsound, because its conclusion (that qualia are causally idle non-physical extras) is ultimately unbelievable. If qualia have no physical effects, he reasoned, then they cannot be what causes us to talk about them, and Mary's very exclamation on seeing red would have some purely physical cause that has nothing to do with the quale itself. That is too strange to accept. Jackson decided the intuition behind premise two, however powerful, must be a kind of illusion about the difference between representing something and having a new fact about it. It is rare and instructive to watch the author of a famous argument dismantle it in print.

Distinctions that matter

Everything turns on what "learns something new" means. There is a difference between learning a new fact (propositional knowledge, knowing that) and gaining a new ability or a new way of representing an old fact. Mary clearly gains something. The whole debate is whether that something is a new fact about the world, which would sink physicalism, or a new skill and a new mode of access to facts she already possessed, which would not. Sorting this out is what the replies below are about.

Lineage

The knowledge argument descends most directly from Thomas Nagel's "What Is It Like to Be a Bat?" (1974), which first pressed the gap between objective description and subjective experience. Its ancestry runs deeper. A structurally similar case appears in C. D. Broad's work in the 1920s, and the general worry that the qualitative feel of experience escapes physical description is a modern descendant of the mind-body dualism of Rene Descartes (1596 to 1650), for whom mind and matter were distinct substances. It also inherits an empiricist thread from David Hume (1711 to 1776): the thought that some ideas can only be had by first having the corresponding impression, that a person born blind cannot form the idea of scarlet from description alone. Jackson turned that scattered intuition into a tight modern argument, and in doing so set the agenda for two decades of debate collected in the 2004 volume "There's Something About Mary."

The strongest case for it

The argument's power is that it needs almost nothing from the reader except an honest reaction to a simple story. It does not depend on any contested theory of mind; it depends on the near-universal sense that Mary's first sight of red is a genuine revelation. That reaction is robust: change the color, change the sense (imagine a congenitally deaf expert on the acoustics and neurology of music who then hears for the first time), and the intuition holds. This suggests the gap it exposes is real and general, not an artifact of one example.

It also draws a clean line that physicalism has struggled to erase. Physical facts are, by their nature, the kind of thing that can be learned from a book, a diagram, or a monitor: they are structural and relational. What red looks like does not seem to be that kind of thing. You can be told everything about it and still not know it until you see it. That asymmetry is exactly what the case makes vivid, and no amount of insisting that "it must all be physical" makes the asymmetry go away. Even Jackson's recantation concedes the intuition is strong enough that a physicalist owes a serious story about why we are so tempted by it.

The strongest case against it

The replies are among the sharpest in analytic philosophy, and several are formidable.

The most influential is the Ability Hypothesis, developed by David Lewis (1941 to 2001) in "What Experience Teaches" (1988) and by Laurence Nemirow. They grant that Mary gains something but deny it is a fact. What she gains is know-how: the ability to recognize, imagine, and remember the experience of red. Knowing what an experience is like is a bundle of abilities, not a piece of propositional information, and abilities are not facts that physicalism must include in its inventory. Mary learns no new truth about the world; she acquires new skills.

A second line, the phenomenal concepts strategy associated with Brian Loar (1943 to 2006) and later John Perry, grants that Mary acquires new knowledge but denies she learns a new fact. On this view she comes to know an old physical fact under a new mode of presentation, a "phenomenal concept" available only from the inside. Just as one can learn that Hesperus is Phosphorus (two names for the one planet Venus) and gain knowledge without adding a new object to the universe, Mary gains a new way of thinking about a brain state she already knew about under its physical description. Same fact, new concept; physicalism is safe.

Daniel Dennett (1942 to 2024) pressed a blunter objection in "Consciousness Explained" (1991) and elsewhere: we cannot actually imagine someone who truly knows all the physical facts, and we sneak in the assumption that Mary knows only book learning. Dennett argued that a Mary who genuinely knew everything physical about color could, on stepping out, correctly identify red on her own and would not be surprised. Our confidence that she learns something, he claimed, comes from failing to take the "all the physical facts" premise seriously.

Finally, the epiphenomenalism the original argument was built to defend is widely rejected on independent grounds, and this is the point that moved Jackson himself: if qualia have no physical effects, they cannot cause Mary to say anything about them, which makes our own reports of our experiences mysteriously disconnected from the experiences they report.

Where it stands now

The knowledge argument remains one of the most discussed thought experiments in philosophy of mind, taught in every course on consciousness and still generating fresh literature. It has not settled the question. Physicalists overwhelmingly reject its conclusion, but they disagree about where it goes wrong, and the fact that so many distinct replies are needed is itself taken by anti-physicalists as evidence the case is onto something real. The most striking feature of its afterlife is its creator's defection: Jackson now thinks Mary learns nothing non-physical and that the intuition of discovery, though powerful, misleads us. That an argument can outlive its author's belief in it, and keep working on new readers who feel its pull the first time they meet Mary, is a fair measure of how deep the puzzle runs.

Test yourself

You were Mary for a moment: you were told she knows every physical fact about color, then asked whether she learns something on seeing red. Notice which way your gut jumped. Now press on it. If you think she learns a new fact, can you say what that fact is, and why a complete physics could not already contain it? If you think she learns only a new skill or a new way of thinking about something she already knew, can you explain why the sense of discovery feels so much like finding out something true? The whole debate lives in that one honest hesitation.