The interpreter
Gazzaniga's name for the part of the left hemisphere that makes up a running story to explain what we do — even when the real cause sits in the other hemisphere, which the talking half cannot see. Shown a reason it has no access to, it does not say "I don't know"; it invents a plausible one on the spot and believes it. The unsettling claim is that this narrator is not a quirk of split brains but standard equipment in everyone.
The left-hemisphere system, named by Michael Gazzaniga, that constructs explanatory narratives for one's own behaviour and experience. In his 2000 Brain review he describes it as "a device that allows us to construct theories about the relationship between perceived events, actions and feelings" — the process that "keeps our story unified and creates our sense of being a coherent, rational agent." It was identified in split-brain patients, where the speaking hemisphere is repeatedly caught explaining actions driven by the mute right hemisphere it cannot access.
In the demonstration Gazzaniga ran with Joseph LeDoux on patient P.S. in the late 1970s, the right hemisphere is shown a snow scene and the left hand points to a shovel; asked why, the speaking left hemisphere — which saw only a chicken claw — answers that you need a shovel to clean out the chicken shed. It does not report confusion. It manufactures a reason, instantly and with conviction.
The interpreter is the neuroscience of confabulation, the phenomenon the black-box topic turns on — the same gap that Nisbett and Wilson found between why people act and the reasons they report, and the same post-hoc machinery Haidt found in moral judgment. Read forward, it is the human precedent for a language model that writes a fluent, confident account of an answer it reached for reasons its account never mentions.