The surgery is done to control severe epilepsy. A seizure that begins in one hemisphere spreads to the other through the corpus callosum — a thick band of roughly 200 million axons connecting the two sides of the brain. Severing it can contain the seizure. After the surgery, patients appear normal in everyday life: they speak, plan, remember, move around. The change is not visible from the outside until you run a specific kind of experiment.
The experiment works because visual information is processed contralaterally. What you see in your left visual field travels to your right hemisphere; what you see in your right visual field travels to your left hemisphere. In an intact brain, the two hemispheres share that information immediately through the corpus callosum. In a split-brain patient they don't. If you flash an image briefly enough that the patient can't move their eyes, you can send information to one hemisphere without the other knowing about it.
In 1978, Michael Gazzaniga and Joseph LeDoux ran the following experiment with patient P.S. They flashed a chicken claw to his right visual field — processed by the left hemisphere, which controls language — and a snow scene to his left visual field, processed by the right hemisphere. Then they gave P.S. an array of pictures and asked him to point to the one most related to what he'd seen. His right hand (controlled by the left hemisphere) pointed to a chicken. His left hand (controlled by the right hemisphere) pointed to a snow shovel. Both responses were correct. Each hemisphere had identified what it knew.
Then the experimenter asked P.S. why he'd made those choices. He said: "Oh, that's simple. The chicken claw goes with the chicken, and you need a shovel to clean out the chicken shed."
The left hemisphere, which controls speech, had seen a chicken claw, watched the right hand point to a chicken, and then watched the left hand point to a snow shovel. It had no information about the snow scene — that had gone only to the right hemisphere, across a connection that no longer existed. What the left hemisphere had was: a chicken claw, a chicken, and a shovel. It constructed an explanation that connected these. A chicken shed needs cleaning. The explanation was sincerely held. P.S. was not guessing or deflecting; he gave his best account of his own behavior, and from inside his available information, the account was reasonable.
Gazzaniga called the left hemisphere the "interpreter." Not because it interprets the world — that's something both hemispheres do — but because it interprets the self. It generates causal narratives to explain actions, decisions, and choices. In a normal brain, the corpus callosum gives the interpreter access to what both hemispheres know, and its stories are usually consistent with what actually happened. In the split-brain patient, the disconnection makes the gap visible. The interpreter is still running, still generating confident first-person explanation, but it's doing so without access to the actual cause. The chicken shed story is not a malfunction. It is the system working correctly, on incomplete information.
There is a behavioral expression of the conflict that can appear: alien hand syndrome, in which the patient reaches for something with the right hand and the left hand physically blocks it, or picks up an object and puts it down again before the right hand can use it. The hands are enacting a disagreement the patient cannot articulate, because the disagreement exists across a connection that's no longer there. The interpreter, observing this, has to explain it too — and does, with the same sincere confidence that produced the chicken shed.
A 2017 study by Yair Pinto and colleagues at the University of Amsterdam added a complication. Using confidence ratings and a range of tasks, they found that two callosotomy patients with complete, confirmed corpus callosum transections showed awareness of stimuli throughout the entire visual field, regardless of which hand responded. Each hemisphere was processing its side of the visual field, but the conscious awareness of both sides seemed to be shared. They described it as "divided perception but undivided consciousness." The perceptual representations were split; the unified observer appeared to persist. What this means for the interpreter theory is not fully resolved — if consciousness remains unified despite the perceptual split, the interpreter may be speaking for a self that is less divided than the classic account suggested.
What the experiments leave unresolved is whether the interpreter is a property of split-brain patients, or whether it is what the left hemisphere does in everyone. The disconnection made the mechanism visible by placing it in an unusual situation: the narrator explaining an action that came from somewhere it couldn't reach. In an intact brain, the corpus callosum provides the interpreter with more information, and its stories are more consistent with what actually happened — which makes them harder to catch as stories. The mechanism may be the same. Every confident account you give of why you did something may be generated by the same narrator that explained the snow shovel as a chicken shed tool: working in good faith, working fast, and working from whatever it has access to. The question that can't be answered from inside a normal brain is how often the interpreter has the information it needs to get the explanation right, and how often it's doing the same thing P.S. did — and producing a story coherent enough that neither the teller nor the listener notices the gap.