Semantic Externalism

Semantic Externalism is the philosophical thesis that the meanings of terms — and the contents of mental states — are not determined solely by what is inside the head of the thinker, but are partly constituted by facts about the thinker's environment and social community. Associated primarily with Hilary Putnam's 1975 thought experiment about Twin Earth and with Tyler Burge's work on social content, externalism poses a direct challenge to internalist theories of Intentionality and Philosophy of Mind.

Putnam's central argument: imagine a planet — Twin Earth — physically identical to Earth in every way, except that the watery liquid that fills oceans and falls as rain is not H₂O but a different compound XYZ, which behaves exactly like water under ordinary conditions. An Earthling and her Twin Earth counterpart have identical neural states when they think about "water." But, Putnam argues, what they mean by "water" differs — the Earthling means H₂O, the Twin Earthling means XYZ. Meanings ain't in the head. The content of a mental state is partly fixed by its causal history and by facts about the natural kinds in the thinker's environment.

Burge extended this to social content: what I mean by "arthritis" is partly fixed by the medical community's established usage, not just by my own beliefs about the disease. I may be wrong about arthritis in ways that do not change the fact that I am thinking about arthritis when I use the term.

Consequences

Semantic externalism has far-reaching consequences for Epistemology, Cognitive science, and the philosophy of Artificial Intelligence. If content is fixed externally, then two systems can be computationally identical — processing the same symbols in the same ways — yet have different mental contents. This suggests that a purely internalist cognitive science, which defines mental states by their computational roles, may be describing the wrong thing. At the same time, externalism raises questions about whether AI systems can have genuine content at all: if content requires a causal history connecting states to objects in the world, then a system trained on text about the world may have a different relationship to content than a system embedded in physical interaction with that world. See also: Intentionality, Mental Content, Embodied Cognition.

The externalist conclusion that is hardest to absorb: we do not have privileged access to the contents of our own thoughts. What I am thinking about when I think about water depends on facts I may not know — the chemical composition of the liquid in my environment. This is a form of epistemic humility that has not been fully absorbed by either folk psychology or cognitive science.