Naturalizing Intentionality

Core Idea

The project of naturalizing intentionality seeks to explain how physical systems can have mental states that are about things, using only the resources of the natural sciences — no irreducible mental notions allowed. Informational semantics (Dretske) holds that a mental state represents what it reliably carries information about: a neural state means 'there is a fly' if it is reliably caused by the presence of flies. The disjunction problem arises immediately: since the frog's detector also fires for BBs and shadows, it seems to represent the disjunction 'fly or BB or shadow.' Teleosemantics (Millikan) addresses this by appealing to biological function: the detector's content is 'fly' because that is what it was selected for by evolution. Fodor's asymmetric dependency theory offers another route: the fly-to-detector connection is basic, and the BB-to-detector connection depends on and is parasitic upon it.

How It's Best Learned

Start with the basic puzzle: a thermometer 'represents' temperature, but we do not think it has genuine intentionality. What would need to be added? Then work through each naturalization strategy (information, teleology, asymmetric dependence) and its specific vulnerabilities. Dretske's Knowledge and the Flow of Information and Millikan's Language, Thought, and Other Biological Categories are the foundational texts.

Common Misconceptions

• Naturalizing intentionality is not the same as denying that intentionality exists; it is the attempt to show that aboutness is a natural phenomenon explicable in physical and biological terms.
• Teleosemantics does not claim that all meaning is biological; it uses evolutionary function as a model for the kind of normativity that distinguishes correct from incorrect representation.

Explainer

You know from your study of intentionality that mental states have the remarkable property of being about things — a belief is about the election, a desire is about coffee, a perception is about the red apple in front of you. You also know from physicalism that the brain is a physical system describable in terms of neurons, electrochemical signals, and causal mechanisms. The project of naturalizing intentionality is the attempt to bridge these two: to explain, in purely physical and biological terms, how it is possible for a physical state to be about anything at all.

The challenge is sharpest when you consider a simple case. A thermometer's reading tracks temperature — the mercury rises when temperature rises. In that sense, the thermometer's position carries information about temperature. Fred Dretske proposes building mental content on this foundation: a neural state means 'there is a fly' if it reliably carries information about flies — if it is caused by flies across a range of circumstances. This is informational semantics, and it captures the intuitive idea that representation is a kind of natural tracking. The thermostat tracks temperature; the frog's retinal state tracks flies.

But the disjunction problem immediately threatens this account. The frog's fly-detector also fires for small dark blobs moving on a light background — BBs, pellets, even optical illusions. So what does the detector state mean? 'Fly'? 'Small dark blob'? 'Fly or BB or shadow'? The information-theoretic account cannot cleanly select one content over the others, because the state is caused by all of them. Ruth Millikan's teleosemantics addresses this by invoking biological function. The question is not what the state is caused by, but what it was *selected for* over evolutionary history. The frog's detector was shaped by selection because it reliably triggered fly-catching behavior that caught flies — not BBs. So the content is 'fly,' because that is what the mechanism was designed by evolution to detect. This introduces a crucial distinction: a detector can be *triggered* by a non-fly while still *representing* a fly — because representing, on this view, is about function, not just causal history.

Jerry Fodor's asymmetric dependence theory takes a different route, staying closer to causal information while addressing the disjunction problem. The key insight is that the 'BB causes detector to fire' connection is not independent — it depends on and is asymmetrically parasitic on the 'fly causes detector to fire' connection. If there were no flies, BBs would not cause firings (because the detector would never have developed); but even if there were no BBs, flies would still cause firings. Genuine content is fixed by the basic, non-derivative causal connection. Each of these strategies tries to locate, in purely natural terms, the normativity that distinguishes correct from incorrect representation — what the state is supposed to represent, even when it misfires. Whether any of them fully succeeds remains one of the deepest open questions in philosophy of mind.