Putnam's Functionalism and Semantic Externalism

Michael BrenndoerferUpdated October 13, 202613 min read

Hilary Putnam’s functional account of pain and his Twin Earth argument for how a speaker’s environment helps determine what a word refers to.

PhilosophyAll chapters
Reading tools Notes & comments
Notes & comments
Part 7 of 12Mind, Self, and Experience
Chapter 82Mental roles and the world beyond the mind › Two Ways to Look Beyond a Brain State
Reading the full chapter · 9 sections · 13 min9 sections · 13 min

All 9 sections · 13 min read

Mental roles and the world beyond the mind › Two Ways to Look Beyond a Brain StateFull chapter · Sections in this chapter1 / 9 ↓

Two Ways to Look Beyond a Brain State

Suppose two creatures pull away from an injuring stimulus, remember where it happened, and later avoid the spot. Their nervous systems may do this work in different ways. Must the state we call pain be the same physical kind of state in both? To answer that question we would need to know what unifies the cases. One possibility is that the same condition of their tissues occurs in both creatures. Another is that different physical processes play the same role in each creature's life.

Now compare two people who use the same word, “water.” They have matching expectations and reactions, but the samples they learned to call water come from different environments. Must the word pick out the same substance in both cases? This time we have reason to look beyond the person's own organization. What they encounter may help fix what the word applies to.

The two examples pose different questions. Hilary Putnam develops the proposal about pain in The Nature of Mental States, sections II–IV, and the case about the word “water” in The Meaning of ‘Meaning’, the Twin Earth discussion. Putnam's first question is about the state an organism is in. His second asks what gives a word its extension, which is the range of things to which it applies in a given use. Reading the two essays together is a way to see how the answer can depend on organization in one case and environment in another.

What a Functional Explanation Adds

Imagine that a researcher proposes to identify pain with one precise physical and chemical type of brain state. The proposal is stronger than the claim that our pains have a bodily basis. It says that organisms capable of the same kind of pain will qualify in virtue of the relevant physical and chemical state. If two cases really share the psychological state and differ in the details the theory takes to be essential, that identification fails. This is why a common bodily indicator in one species would not finish the argument.

Putnam makes the point by comparing creatures including mammals and octopuses in The Nature of Mental States, section III, pp. 76–77 in the linked reprint. The possibility of a shared physical condition is allowed. His objection is that it is demanding to require the same physical and chemical condition to occur in all creatures to which the psychological kind applies. Even if a common condition were found for pain, an identity theory proposed for mental states in general would face the same test for other states it claims to identify.

Multiple Realizability

A state individuated by its role may be realized by different physical arrangements. This is the possibility that motivates the functional hypothesis. Showing that two physical arrangements differ does not yet establish that both realize a mental state. We also need grounds for classifying their states together.

What would the role amount to? Putnam uses a model of an organism as a probabilistic automaton. A description of such a system specifies states and inputs and gives the probabilities of resulting states and outputs. A given stimulus need not produce a single inevitable response. The response depends both on inputs and on the organism's other states. An account of its organization can describe how those factors fit together across cases. Putnam treats this as a hypothesis about the sort of system an organism is, with details to be made precise. The Nature of Mental States, section II, pp. 75–76

For pain, the proposal looks to a class of sensory inputs that play a characteristic role in the whole organism, including that the organism evaluates those conditions negatively. This need not make every case of pain look the same from the outside. A person might choose to endure discomfort rather than give up a goal they value more. A description that mentions both the aversive input and the competing goal can explain why behavior changes without treating the discomfort as having disappeared. Putnam expressly distinguishes assigning high disvalue to an input from always avoiding it. The Nature of Mental States, section IV, p. 78

Why Behavior and Structure Are Different Clues

Suppose one person grimaces and another keeps still after the same sort of injury. Merely observing two different reactions leaves open whether they have a similar experience. There might be a relevant difference in their sensory capacities. There might instead be a difference in what each is trying to do. On a functional proposal those other factors belong in the explanation. If the state has a characteristic role across the system, we can ask how that state together with changes in the other factors could yield divergent behavior.

This also separates the evidence for a state from the proposed explanation of that evidence. Putnam observes that we often judge whether an organism is in pain from its actions without knowing what state its brain is in. That ordinary route to a judgment does not decide what pain is. A theory of pain should be able to explain why the indications we use are informative, insofar as they are, even when the observer knows nothing about the underlying theory. This is part of Putnam's case against treating pain itself as a disposition to produce certain peripheral behavior. The Nature of Mental States, sections IV–V

The distinction is important when comparing creatures with very different bodies. A similar outward reaction suggests a question to investigate. We can ask whether similar inputs have similar consequences within two organisms and how those consequences change when goals or circumstances differ. A reflex that occurs under one condition is too little by itself to establish the similarity Putnam needs. It provides a possible starting point for trying to describe the organization of both systems. In section V he calls the identification tentative and evaluates it partly by whether it leads to fruitful predictions and questions.

There is a challenge on the other side as well. If a whole-system description can always be tailored after the fact to include whichever cases we wanted to count together, it may explain little. The hypothesis needs an informative role that holds across systems and a reason to prefer it over the available physical and chemical descriptions. Putnam acknowledges that his proposal leaves the appropriate organization underspecified. The open work is to describe it without quietly replacing the question of which organisms can feel pain by an assumption that we already know the answer.

When an Implementation Matters

Suppose two machines open a valve when their sensors register the same input. One passes an electrical signal to a motor, and the other uses a mechanical linkage. This example does not tell us whether either machine has a mind. It lets us isolate a simpler point. We can ask whether the two machines have the same causal role at one level of description even though a physical description distinguishes their components. In a larger system the similarity might break down. If the electrical machine also retains a record that affects how it responds on the next occasion, the one shared output conceals a difference in organization.

A functional account therefore has to identify which relations belong in the description. The pain proposal cannot be vindicated just by finding some resemblance between any two physical systems. It has to say what the inputs and other states do in an organism and why those roles explain the pattern we classify as pain. The differences between implementations matter when they change those roles. Putnam's project is to find a description that captures the important regularity across varied cases. The choice of a useful description is part of the explanatory work, not something settled by declaring in advance that the organization must be the same.

Two Communities Saying “Water”

In The Meaning of ‘Meaning’, pp. 4–6 of the MIT excerpt, Putnam imagines Twin Earth. In the story there is a liquid that resembles the water we use in ordinary conditions. Twin Earth speakers call it “water,” and it plays the same familiar role in their surroundings. Its chemical composition, abbreviated XYZ, is different from H₂O. The resemblance is part of the setup. We are not meant to conclude that such a liquid has been discovered.

Putnam then sets the comparison around 1750. The speakers in his imagined case do not know the composition of either liquid. Oscar on Earth and his counterpart can match one another in their inward experience, thoughts and expectations while learning the word from different samples. Setting the example before the relevant chemical knowledge matters. If the speakers knew the chemistry and explicitly meant different compounds, it would be easy to find a difference in what each knew. Putnam is asking us to keep that knowledge fixed and vary what the word has been used to identify in each environment.

What should we say about the extension of the word? Putnam thinks the Earth speaker's use applies to water and the Twin Earth speaker's use applies to the different liquid that occupies the corresponding place in their environment. The difference turns on the actual substances, although the speakers do not yet know how to tell those substances apart. The argument relies on holding fixed a suitably narrow description of each speaker's psychological state. If we build which external substance a belief is about into our description of that state, we have already allowed the sort of external difference the example is designed to bring out.

Extension

The extension of a term in a particular use consists of the things it applies to. Putnam's Twin Earth argument concerns how that extension gets fixed. It leaves room to distinguish a term's extension from what a speaker knows or ordinarily associates with the term.

Following a Sample

Picture a person learning the word by pointing to a glass. Their words concern the liquid, rather than its container or its temperature. They also take the sample to stand for the same kind of liquid as other samples people around them call water. This is an assumption that can be wrong. Putnam's illustration imagines that the sample unexpectedly turns out to be gin. We have reason to revise our judgment about the sample instead of using it to redefine every instance of water. The Meaning of ‘Meaning’, pp. 5–6 of the MIT excerpt

The example shows why visual resemblance is insufficient in the proposed account. What matters for the term's application is whether further instances are of the relevant kind, and that may take investigation. Early users of the word might sort everyday samples very successfully while having a mistaken view of a test that would distinguish the kind. In Putnam's scenario a later discovery about the liquid can correct a judgment about a sample without constituting an arbitrary change in what earlier speakers were talking about.

Expertise is a related but distinct issue. Putnam also imagines an exchange with a community whose use of the term “aluminum” concerns what ours calls molybdenum. He supposes, for the sake of the example, that a nonexpert cannot ordinarily distinguish items made of the two materials, while a metallurgist can. A speaker can use a word that applies to a kind without personally being able to perform the distinguishing test. The 1750 case adds a different point. It asks how extensions could differ even before the relevant ability to discriminate is available within the communities. Explaining the water example solely by appeal to experts who already know the formula would miss that point. The Meaning of ‘Meaning’, pp. 6–8 of the MIT excerpt

What the Argument Does and Does Not Settle

A critic can grant that the two terms have different extensions and still ask how much follows about what a speaker understands. Oscar and Twin Oscar may both associate the word with an ordinary clear liquid. Perhaps those shared expectations explain why they talk and reason in parallel. That leaves an external component in any account of which substance their words apply to. It raises the further question of whether that component is part of what we should call meaning or something a theory of meaning must connect to.

The worry can also be pressed against the example's verdict itself. Why say that the speakers' terms concern different kinds instead of saying that both use a term broad enough to cover whichever liquid behaves in the familiar way? Putnam's answer relies on a feature of the kind term as he understands it. Ordinary recognition supplies evidence for applying it, while belonging to the same kind as the relevant samples takes priority when evidence of a difference comes in. A critic can contest that priority. This identifies a point of disagreement more exactly than treating the imagined planet as though it were a laboratory experiment that confirmed a general theory of language.

Nothing requires us to apply that answer to every expression. Consider “clear liquid.” Two chemically different substances can both fall under that description if they are clear and liquid in the circumstances. The water example makes its point by distinguishing such a description from putatively naming a kind. The example's conclusion would be much weaker if every description of an object were already a name for whatever chemical substance happens to be present.

Reading the Two Arguments Together

Place the examples side by side. In the pain case an identity theory insists on the same physical and chemical realization where Putnam thinks a common functional role might be enough. In Twin Earth a theory of extension insists that a narrow psychological description is enough where Putnam thinks the historical use of different substances makes a difference. He changes what information counts as relevant because the questions ask about different properties.

An account of how two speakers respond to words, remember their earlier encounters, and revise their expectations can tell us a great deal about why their behavior matches. It can still leave out which liquid they have been identifying. It would take another argument to show that no functional description can include relations with a surrounding world. The water case challenges the narrower assumption that what happens inside the individual already fixes the extension of the term.

This leaves us with a practical way to compare explanations. When two cases share a label, ask what must be held fixed for the label to continue to apply. For a proposed account of pain, the answer might be an organization of states and inputs that different physical systems share. For a word used to pick out a natural kind, one thing to check is the sample history or surrounding practice that gives the use its object. Putnam's arguments give reasons for placing those boundaries differently, and they make it possible to say where a critic would have to disagree.

Further Reading

  • Hilary Putnam, The Nature of Mental States. Read section II for the functional hypothesis and sections III–IV for the comparison with physical and behavioral alternatives.
  • Hilary Putnam, The Meaning of ‘Meaning’ (1975), in excerpt. Read the Twin Earth comparison before the example involving metals. The cases test different ways a speaker can use a word without knowing what distinguishes every instance.
Join the discussionAsk a question or share what you learned.

Comments

Add to the conversation

A question, a correction, or an example from your own work.

No comments yet. Be the first to share your thoughts!

A different angle
All writing
Philosophy

Aristotle's De Anima: Soul, Perception, and Intellect

Aristotle treats soul as the form of a living body and asks how living things perceive, move, and think.

14 min read
Philosophy

Husserl's Cartesian Meditations: Transcendental Phenomenology

Edmund Husserl's five meditations on transcendental phenomenology, from the epoché to the experience of other minds and a shared world.

14 min read
Philosophy

Husserl's Ideas I: The Epoché and Transcendental Phenomenology

Edmund Husserl's 1913 systematic presentation of transcendental phenomenology, introducing the epoché (bracketing) and eidetic reduction.

13 min read

About the author

Reference

Citation details

Cite or share this article.

BIBTEXAcademic
@misc{brenndoerfer2026putnams, author = {Michael Brenndoerfer}, title = {Putnam's Functionalism and Semantic Externalism}, year = {2026}, url = {https://mbrenndoerfer.com/writing/putnam-functionalism-meaning-semantic-externalism-guide}, organization = {mbrenndoerfer.com}, note = {Accessed: 2026-11-19} }
APAAcademic
Michael Brenndoerfer (2026). Putnam's Functionalism and Semantic Externalism. Retrieved from https://mbrenndoerfer.com/writing/putnam-functionalism-meaning-semantic-externalism-guide
MLAAcademic
Michael Brenndoerfer. "Putnam's Functionalism and Semantic Externalism." 2026. Web. November 19, 2026. <https://mbrenndoerfer.com/writing/putnam-functionalism-meaning-semantic-externalism-guide>.
CHICAGOAcademic
Michael Brenndoerfer. "Putnam's Functionalism and Semantic Externalism." Accessed November 19, 2026. https://mbrenndoerfer.com/writing/putnam-functionalism-meaning-semantic-externalism-guide.
HARVARDAcademic
Michael Brenndoerfer (2026) 'Putnam's Functionalism and Semantic Externalism'. Available at: https://mbrenndoerfer.com/writing/putnam-functionalism-meaning-semantic-externalism-guide (Accessed: November 19, 2026).
SimpleBasic
Michael Brenndoerfer (2026). Putnam's Functionalism and Semantic Externalism. https://mbrenndoerfer.com/writing/putnam-functionalism-meaning-semantic-externalism-guide
Newsletter

Stay up to date

Get new articles, book updates, and notes on AI and technology.

Unsubscribe at any time. Privacy policy