Def 1: Type-identity theory: brain states and mental states are the same thing. The mental state "I love Metallica" is nothing more than the neural configuration of a Metallica fan.

Def 2: Multiple Realizability: that the same mental state can be realized in more than one type of system, a human brain, an animal brain or an AI.

I was listening to a lecture on the mind-body problem, and the speaker stated that the concept of Multiple Realizability definitively refuted the type-identity theory of the mind. That once the Multiple Realizability argument was made, people started looking for other possible physicalist answers to the mind body problem (e.g. functionalism, the computational theory of mind, etc...) and gave up on type-identity.

What I can't wrap my head around is, I don't see why Multiple Realizability refutes type-identity.

1 - There is no reason why Multiple Realizability should hold: for all we know, human brains are the only systems capable of realizing human mental states.

2 - The fact that animal brains, AI or alien minds made of some other form of matter can have thoughts and sensations analogous to human thoughts and sensations doesn't mean that human mental states form an ontologicaly distinct category from human brain states. For example, the fact that there are blonde wigs made of nylon and lions with yellow fur, doesn't mean that "blondness" is a different category from "having hair with a certain type of pigmentation".

So my questions are: What arguments are given for the necessity of Mutiple Realizability? And, even if it does hold, why does Multiple Realizability imply that mental states are different from brain states?

4 Answers 4


As you already wrote, MR doesn't specifically contest Identity Theory. The two concepts are contrary, however. Type-Identity theory is just another way of saying that there must be, for each and every unique mental state, an equally unique brain state, and vice-versa. Without an exact, physical copy of some brain state, the mental state that obtains will be different. MR claims the opposite, that different mental states can be instantiated by the same brain-state. If one is false, then the other is true.

Hence, the argument against type-identity theories supports MR. The original arguments are recorded in the SEP on Multiple Realizability.

In a series of papers published throughout the 1960s, Hilary Putnam introduced multiple realizability into the philosophy of mind. Against the “brain state theorists,” who held that every mental kind is identical to some as-yet-undiscovered neural kind, Putnam (1967) notes the wide variety of terrestrial creatures seemingly capable of experiencing pain. Humans, other primates, other mammals, birds, reptiles, amphibians, and even mollusks (e.g., octopi) seem reasonable candidates. But then for the “brain state theory” to be true, there must be some physical-chemical kind common to this wide variety of pain-bearing species, and correlated exactly with each occurrence of the mental kind.

  • 1
    it seems to me as if the quoted SEP argument against brain state theorists is not an answer to @AlexanderSKing, but on the contrary, reinforces his question; to put it in simple words - what do we care if monkeys experience something that looks like pain from the outside but is infact a little different; we call each other humans although we know we are all different from each other, sometimes drastically, in various capacities, and properties; if that is true for people, why can't it be true for mental states?
    – nir
    Commented Aug 26, 2015 at 8:00
  • @nir I agree, and if you follow that line of thought you may end up concluding that mental states are personal and transient: my pain is not necessarily the same state as your pain any more than your pain is the same state as a monkey's, and my next pain is not necessarily the same state as my previous experience of it. Furthermore, this arguably extends to very specific states: my knowledge that 1 + 1 = 2 changed as I learned something of number theory. Once you realize that mental states are only approximately comparable between individuals and over time, type-identity seems just irrelevant.
    – sdenham
    Commented Jul 31, 2019 at 15:35

For this question, it might help to parse what type-identity theory means and what multiple realizability means.

type identity

The idea is that a type, i.e. a mental state, maps onto a physical brain state in a numerically identical fashion.

Translated into plain English, this is the view that any mental state is exactly and in every respect one brain state.

multiple realizability

multiple realizability is the claim that the same mental state can be realized in multiple physical ways.

Regardless of how, this would imply that the identity is broken, because the same thing can occur in ways that differ.

It might help to think of a pair of six-sided dice.

Let's first consider the number 2. The only way to get the number 2 on two dice is 1 and 1. Thus, there's an identity between rolling a 2 and having the physical state of two dice each showing 1. These two things are thus identical.

But then let's consider the number 6. We can get 6 by having a 1 and 5, 2 and 4, or 3 and 3. Thus, rolling a six is multiply realizable. This means its non-identical with any of these particular physical arrangements of dice.

Thus if multiple realizability is true about mental states, then type identity is false. And if type identity is true, then multiple realizability is false.

  • 2
    suppose according to type-identity, 1M neurons are generating what we call PAIN, and suppose one of these neurons stops functioning; then, according to type-identity, can we no longer have PAIN?
    – nir
    Commented Aug 26, 2015 at 8:11
  • that would seem to follow though I don't know if type-identity theorists mean to assert down to that level of specificity.
    – virmaior
    Commented Aug 26, 2015 at 8:39
  • I think it bears on the question and on the answer. @AlexanderSKing, do you know what would theorists of type-identity, say in such a case?
    – nir
    Commented Aug 26, 2015 at 9:53
  • 1
    @virmaior In your example, a unique object the number "6" maps to multiple dice configurations. Implicit in MR is that a unique entity, the mental state pain, maps to multiple physical configurations. What I am challenging is that that mental states, "pain", "love", "belief that the sky is blue", etc..are not necessarily unique entities. The term pain, instead of corresponding to a unique entity, can correspond to a class of many possible configurations. If this is the case then MR and type identity are no longer mutually exclusive. Commented Oct 21, 2015 at 22:23
  • 1
    I think everyone would reject that fear means containing a certain particular hydrogen atom on a single molecule in the brain. But as we scale up eventually some will say fear is having a particular type of brain state and thus take a type identity view while others will maintain multiple states at that level of precision are fear and be respective "multiple realizability" theorists.
    – virmaior
    Commented Oct 21, 2015 at 23:06

Why they are incompatable

If you have any math background, there is an easy analogy to explain the difference. Given A is the set of all physical states, and B the set of all mental ones, Type Theory prescribes a bijection, a 1-to-1 relationship, while MR only holds that there is a projection from A to B. The difference comes in the backwards implication. Both say that a specific physical state of the brain implies a mental state. But Type Theory says that the mental state also implies the physical state, while MR doesn't require this to be true.

In your example of blondness, the way you phrased it, what you've actually done is defined a yellow wig and a lions fur as the same physical state, "having hair with a certain type of pigmentation." Then you only have one physical state, the pigment, and one descriptive state, being blonde, to map to. Both fields agree having hair of yellow pigment implies that you are blonde. The way to extend this example would be to say that for Type Identity, being blonde also implies that your hair has yellow pigment, while MR would say that that's not true, if you have black pigment but died your hair yellow you are also blonde, so being blonde doesn't imply your physical state.

Reasons to believe in MR

Defining physical state is hard

How specifically do you define a physical state? Say it's by the neuron. An exact arrangement of neurons in a specific form is a physical state. But consider when you were a child. Your brain was smaller and significantly different. By Type Identity, this would mean that the feelings you felt as a child are not the same ones you feel now. If pain is a certain arrangement of neurons in your brain right now, then you never felt "pain" as a child. Similarly, no two people ever experienced the same mental state. Even if this is true, it's not a very useful way to examine the mind, as it ends up being fatalistic.

But if you get too broad, you also run into issues. Say physical state isn't by neuron, it's more just the outline of brain structure. Then people could share mental states. But potentially so could animals, if we're only worried about broad structure. And could there even be enough broad structures to encompass every possible mental state? Even more important, if you change mental states, i.e you stop liking Metallica, this would mean a large physical change has occurred in your brain, but this is something that's never been detected.

There is no good resolution to define physical state as that doesn't bring up some complications and hard questions.

  • "Even more important, if you change mental states, i.e you stop liking Metallica, this would mean a large physical change has occurred in your brain, but this is something that's never been detected." Brain activity is a well-measured phenomenon, though. Brains do go through physical changes based on what they are thinking. An example would be something like ncbi.nlm.nih.gov/pubmed/29573366, though I guess that's not exactly what you meant as you're talking about a higher-level concept. Put another way, brain state isn't just neuron arrangement but also neuron activity.
    – JAB
    Commented Apr 6, 2018 at 15:35
  • Of course you still felt pain as a child, pain is something very broad and vague. It’s like saying you never experienced fruit juice as a child because you moved to another country where the oranges are a different species. Commented Jun 1, 2019 at 13:55
  • @MichaelAngelo Exactly. You felt pain as a child, and you feel pain now, but clearly your brain state can't be the same as it was when you were a child on too granular a level. This then implies that two different brain states both caused "pain", aka pain was Multiply Realized.
    – Cain
    Commented Jun 3, 2019 at 13:16
  • @Cain yes, and I’m saying that this is not a problem because there no sensible reason to assume that the pain you experience now is the same as what you experienced as a child. Commented Jun 3, 2019 at 13:30

Type-Identity and Multiple-Realiazibility offer two mutually exclusive possibilities; the truth of which we cannot definitively assert.

A: To paraphrase the first, we can say animals have brains, and brains have minds; for this we have empirical evidence.

B: To paraphrase the second, we can say possibly minds can occur in artificial brains; we have no empirical evidence for this supposition; only certain circumstantial evidence: Turing machines, simulations on computers, a physical theory of matter and force.

If we had definite evidence we can definitively rule out one or the other; thus they would be definitively mutually exclusive possibilities; but we don't have this now, we may have this in the future (unless it is an undecidable problem); thus all we can say now, is that both possibilities are possible; or both A or B are possibly true.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .