Michael Graziano in The Atlantic:
When Isaac Newton was 17 years old, he performed a series of experiments with prisms and light beams. Within weeks he discovered the scientific explanation for color, invented the reflecting telescope, proposed the particle theory of light, and deduced that the human eye contained three receptor types corresponding to the three primary colors. Not bad for a teen.
Newton’s insights were not easily accepted. At the time, the prevailing theory of color was metaphysical. White light was thought to be pure, heavenly, and scrubbed of all contaminants, whereas colored light was contaminated by the worldly surfaces it touched. To scholars, the exact process by which white light became dirtied was a philosophical hard problem worthy of debate.
We now know why that hard problem was so darn hard. The brain processes the world in a simplified and inaccurate manner, and those inaccuracies gave people the wrong idea about color. Deep in the visual system, the brain reconstructs information about light. In that simplified code, white corresponds to the color channels registering zero and the brightness channel cranked up high. Pure luminance without color is a physical impossibility, because white light is a mixture of all colors. The pre-Newtonian problem of color was hard because it had no possible solution.
Why would the brain evolve such an inaccurate, simplified model of the world? The reason is efficiency. The brain didn’t evolve to get all the scientific details right. That would be a waste of energy and computing time. Instead, it evolved to process information about the world just well enough, and quickly enough, to guide behavior. All the brain’s internal models are simplified caricatures of the world it models. Arguably, science is the gradual process by which the cognitive parts of our brains discover the profound inaccuracies in our deeper, evolutionarily built-in models of the world.
More here.