CONCEPT
The Principle of Charity (Davidson)
Donald Davidson's name for the unavoidable interpretive assumption that any speaker we understand must be assumed largely rational and largely right about the world—and, by extension, the precise mechanism by which fluent AI outputs exploit our interpretive reflexes to project minds where none may exist.
The principle of charity is
Donald Davidson's answer to the question of how interpretation ever gets started. To assign meanings to a speaker's words, an interpreter must already know the speaker's beliefs; to assign beliefs, the interpreter must already know the meanings. There is no neutral ground to stand on while working out both simultaneously. Davidson's escape is to hold one variable steady by stipulation: the interpreter assumes, before any evidence, that the speaker's beliefs are largely true and largely coherent—that the speaker reasons in recognizable ways and gets the world mostly right. Only against this background assumption can behavior be used to triangulate what the words mean. The principle has two strands: logical coherence (the speaker's beliefs hang together by the rules of inference) and empirical correspondence (the speaker believes what a well-placed observer would believe about the shared environment). Both are preconditions of interpretation, not