Teleological Semantics

Teleological semantics is the naturalistic theory of mental content that grounds representational meaning in the biological or learned function of a state. On this view, a frog's neural firing pattern does not merely correlate with the presence of flies; it represents flies because it was selected by evolution (or shaped by learning) to serve the function of detecting flies. The content is fixed by the proper function, not by the actual causal history of each token state.

Ruth Millikan (1984) and Fred Dretske (1986) are the principal architects. Millikan's biosemantics treats content as a normative property: a state has the content it is supposed to carry, where 'supposed to' is defined by the selection history of the mechanism that produces it. Dretske's informational semantics adds that content requires a learning period during which the state carries natural information about the target, after which the state can misrepresent.

The central objection is the Swampman problem: a being created by lightning with no evolutionary or learning history would have no content on this view, yet it would behave indistinguishably from an evolved being. The reply — that function can be constituted by learning as well as evolution — weakens but does not eliminate the objection.

See also: Mental Content, Informational Content, Content Individuation, Functionalism

The Proper Function Problem

Millikan's biosemantics grounds content in proper function — the function a mechanism is supposed to perform, defined by its selection history. A heart's proper function is to pump blood; a frog's snap-reflex proper function is to catch flies. The content of a mental state is determined by what the mechanism that produces it was selected to track. This normative dimension — the 'supposed to' — is what makes misrepresentation possible. A state represents flies even when it is triggered by a bee because it is supposed to track flies, not because it is currently tracking flies.

The proper function criterion faces a challenge from etiological pluralism. Evolution is not the only source of proper function. Learning, cultural transmission, and even individual habit formation can establish functions. A learned association between a bell and food has proper function because it was shaped by reinforcement learning, not by natural selection. The biosemantic framework must therefore be generalized to include all selection-like processes — processes that shape a mechanism through a history of differential success. This generalization makes teleological semantics applicable to artificial systems: a neural network trained by gradient descent has proper functions established by its training history, not its evolutionary history.

The Informational Requirement

Dretske's informational semantics adds a crucial constraint: content requires not merely function but natural information. A state carries natural information about X if it is lawfully correlated with X — if the probability of X given the state is 1 (or sufficiently high). During a learning period, the state acquires this correlation. After the learning period, the state can misrepresent because it retains its content — the content fixed during the learning period — even when the correlation breaks down.

The informational requirement connects teleological semantics to information theory. Content is not a primitive mental property but a statistical property of the relationship between a state and its target. The 'aboutness' of mental states is reducible to the mutual information between the state and the world, conditioned on the learning history. This makes teleological semantics a naturalistic theory: it explains content in terms of lawful correlations and selection histories, without invoking irreducibly mental properties.

The Swampman and the Problem of New Content

The Swampman objection — a being created by lightning with no selection history — is the most serious challenge to teleological semantics. If content requires a history of selection or learning, then Swampman has no content. But Swampman behaves indistinguishably from a normal human. The intuition that Swampman has content is strong, and the teleological semanticist must either deny the intuition or explain why it is misleading.

The standard response is that Swampman lacks content initially but acquires it through interaction. Every action Swampman takes is subject to correction by the environment; every correction is a selection event. Within hours, Swampman has a learning history that establishes proper functions. The content is not present at creation but emerges almost immediately through the same feedback loops that sustain normal cognition. This response weakens the original claim — content is not essentially historical but historically acquired — but it preserves the spirit of the theory: content is constituted by functional success, not by intrinsic properties.

Teleological Semantics and Complex Systems

From a systems perspective, teleological semantics is a theory of how macro-level properties (content) emerge from micro-level processes (selection, learning, feedback). The 'aboutness' of a state is not a property of the state in isolation but a property of the state's role in a larger system — the sensorimotor loop, the ecological niche, the selective environment. Content is observer-indexed in the same sense that observer-indexed emergence claims: it is a property of the coupling between system and environment, not a property of the system alone.

This reframes the Swampman problem. Swampman is not a contentless system that miraculously behaves as if it has content. Swampman is a system that acquires content through the same feedback loops that all systems acquire content through. The difference between Swampman and a normal human is not the presence or absence of content but the depth of the history that shaped the content. A normal human's content has been shaped by billions of years of evolution and a lifetime of learning; Swampman's content has been shaped by a few hours of interaction. The content is shallower but not absent.

Teleological semantics is not merely a theory of mental content. It is a theory of how functional organization generates semantic properties — how the 'aboutness' of a state emerges from its history of success and failure. This is the same question that animates the study of autopoietic systems, stigmergic coordination, and economic naturalness: how do systems that are 'about' something — that track, predict, and respond to their environments — arise from processes that are not themselves about anything? The answer, in every case, is the same: through selection, through feedback, through the slow accumulation of functional structure. Teleological semantics is the philosophy of mind's contribution to this general theory of semantic emergence.