The Mind Projection Fallacy
The mistake of treating your representation of the world as a property of the world itself — projecting the map onto the territory.
“The only real limitations on making ‘machines which think’ are our own limitations in not knowing exactly what ‘thinking’ consists of.”
— E.T. Jaynes
The mind projection fallacy is the mistake of treating your representation of the world as a property of the world itself. You project the way your mind describes something onto the thing being described, and conclude that the description reflects how it actually is.
The coin flip is the clearest illustration.
Your friend flips a coin, catches it, and presses it face-down on their wrist, covered by their other hand. They ask: what is the probability it’s heads?
The intuitive answer — fifty percent — sounds right, but it smuggles in the fallacy. A flipped coin, in the abstract, does land heads fifty percent of the time. But this is not an abstract coin. This coin has already landed. It is lying in a definite position under your friend’s hand. The physical result is fixed. It is not in superposition; it is not wavering between heads and tails. The outcome is determined.
The fifty percent is not a property of the coin. It is a property of your knowledge — or rather, your ignorance — of the coin’s current state. You have a fifty percent chance of being correct if you guess heads. The coin has a hundred percent chance of being whatever it is.
This distinction — between uncertainty about the world and uncertainty in the world — matters more than it seems. Probability, understood correctly, is a measure of epistemic state, not of physical indeterminacy. When we say “there’s a thirty percent chance of rain,” we are not saying the atmosphere is genuinely undecided. We are saying that our model of the atmosphere, given current information, assigns that probability to rain occurring. The weather, like the coin, will do what it does.
The fallacy appears wherever we treat our descriptions of the world as though they belonged to the world. A color is not “out there” in the wavelength of light — it is a construction of the visual system. A sound is not in the compression wave — it is in the auditory cortex. Beauty, disgust, familiarity, danger: these are cognitive events, not features of the landscape. The world has properties; our minds have responses to those properties. Conflating the two produces systematic errors.
Catching the fallacy does not mean that our mental representations are useless or that the world is inaccessible. It means keeping the map and the territory distinct — which is a prerequisite for updating the map when the territory contradicts it.