A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Multiple realizability, introduced by Hilary Putnam, is the observation that a single mental state type (such as pain) can be physically realized by very different material substrates across different species, and potentially in artificial systems. A human, an octopus, and a hypothetical silicon robot might all be in pain, yet have entirely different physical implementations. This undermines type identity theory, which requires a one-to-one correspondence between mental types and neural types. Multiple realizability became a cornerstone argument for functionalism and for the autonomy of the special sciences (psychology, economics) from lower-level physical sciences.
Consider Putnam's original argument and Fodor's defense of the special sciences. Then engage with responses: Kim's challenge that multiply realizable properties cannot be causally efficacious, and the suggestion that mental types should simply be construed more locally (human pain, octopus pain).
To see why multiple realizability matters, you need to hold two things in mind simultaneously: the functionalist picture you've already studied, and the identity theory it was partly designed to replace. Type identity theory made a bold, clean claim: pain just *is* a particular brain state — say, C-fiber firing. Mental state types are neural state types. This is appealingly simple and clearly physicalist. The question is whether it's true.
Hilary Putnam's multiple realizability argument starts with a thought experiment. Consider pain across species. Humans feel pain; so (we have good reason to believe) do octopuses, mice, and potentially many other creatures. But human nervous systems and octopus nervous systems are radically different. The neural state that realizes pain in a human is almost certainly a different physical state from whatever realizes pain in an octopus. If pain were identical to a specific neural type, it would follow that human pain and octopus pain are simply *different mental states* — which contradicts our reason for calling them both pain in the first place. Extend this to hypothetical silicon-based intelligences, and the point becomes sharper: if anything that functions as pain counts as pain, then mental types cannot be identified with any particular physical type.
This is not a claim that mental states are non-physical. Each particular instance of pain — your pain right now — is physically realized in your neural tissue. Multiple realizability is a point about types, not tokens. It says there is no single physical type that corresponds to the mental type "pain." Every token pain is physical; no physical type exhausts what pain is. The distinction matters: you can be a thoroughgoing physicalist and accept multiple realizability, as long as you give up type reductionism.
The philosophical payoff is enormous. If mental state types are defined by their functional roles — what causes them, what they cause, how they interact with other states — rather than by physical composition, then psychology becomes an autonomous science. You can have genuine psychological laws ("pain causes avoidance behavior") that hold across species with different physical implementations. This is Jerry Fodor's point about the special sciences: psychology, economics, and ecology study patterns that cannot be reduced to physics without loss, because the same pattern can be physically realized in many ways.
The main challenge, pressed by Kim among others, is that multiply realizable properties may not be causally efficacious at the type level — if "pain" picks out different physical states in different creatures, what single causal power does it name? Critics also question whether human pain and octopus pain really do share a common mental type, or whether we are just using the same word for functionally similar but ultimately distinct states. These challenges don't defeat multiple realizability, but they show that the functionalism it supports still faces serious pressure — pressure you'll examine further when you reach the Chinese Room and related debates.
Topics in reflective domains aren't scored by quiz answers. Read, reflect, and mark when you've thought it through.