Questions: Natural Kind Terms and Semantic Externalism

Question 1

On Twin Earth, a liquid exists that looks, smells, and behaves exactly like water but has chemical structure XYZ. A student argues: 'Since Earth and Twin Earth speakers have identical mental concepts of water, their terms must refer to the same thing.' What does Putnam's argument show is wrong with this?

• A. Nothing — identical mental concepts do imply identical reference, so both terms refer to the same kind
• B. The extension of a natural-kind term is determined by the actual molecular structure of the world, not by internal psychology; the terms have different extensions despite identical mental states
• C. The terms would refer to the same thing only if both speakers knew chemistry
• D. Reference is determined by functional role, so XYZ and H2O both qualify as water

Answer: (B)

This is the core of Putnam's semantic externalism. Psychological duplicates — speakers with identical internal states — can use the same word with different extensions if their environments differ at the molecular level. 'Water' on Earth refers to H2O; 'water' on Twin Earth refers to XYZ. The mental concept alone doesn't fix the extension. Meaning 'ain't in the head': what matters is the actual nature of the kind in the world.

Question 2

Putnam argues that 'Water is H2O' is a necessary truth discovered empirically, not a definitional stipulation. Why?

• A. Scientists chose to define water as H2O for practical purposes, making it true by convention
• B. 'Water' rigidly refers to whatever natural kind has the same molecular structure as our original paradigm samples; H2O was discovered to be that structure through chemistry, not stipulated to be it
• C. Necessary truths are always knowable a priori, so once known empirically 'Water is H2O' must become a priori as well
• D. H2O is simply a more precise way of restating the observable definition of water

Answer: (B)

If 'water' simply meant 'clear, drinkable, odorless liquid,' then 'water is H2O' would be a definitional truth and chemistry would have discovered nothing. But we did discover something: the molecular structure underlying all the observable properties. 'Water' refers to the natural kind, not to its description. Because the kind is H2O in every possible world, the identity is necessary — but it required empirical investigation to find, making it a posteriori necessary.

Question 3

Natural-kind terms like 'water' refer to whichever substance satisfies the central descriptions and stereotypical properties that speakers associate with them.

• T. True
• F. False

Answer: False

This is the internalist view Putnam's Twin Earth argument refutes. If 'water' referred to the stereotypical description (clear, tasteless, drinkable), then XYZ on Twin Earth would qualify as water, and fool's gold would qualify as gold. But these are not the same kinds — their microstructure differs. Natural-kind terms refer rigidly to underlying essence; the observable properties help us identify the kind but don't constitute what the kind is.

Question 4

According to the division of linguistic labor, an ordinary English speaker who cannot identify gold's atomic number still uses 'gold' with the same extension as a chemist.

• T. True
• F. False

Answer: True

The division of linguistic labor means extension is maintained across a linguistic community through a chain of use: ordinary speakers use 'gold' deferentially to experts who can identify the kind, and the whole community's use is anchored by that expert knowledge plus the causal-historical chain. You don't need to know atomic number 79 to refer to gold — reference is fixed by the community's collective practice and the actual nature of the substance, not by each individual's mental concept.

Question 5

Why does the Twin Earth thought experiment show that semantic content is not 'in the head'? What feature of the scenario makes it an effective argument against internalist theories of meaning?

Model answer: The scenario constructs a case where two speakers are psychologically identical — same beliefs, concepts, and internal states — yet their terms have different extensions because their environments differ at the molecular level. An internalist theory predicts that identical psychology implies identical meaning. The Twin Earth case shows this prediction fails: 'water' on Earth refers to H2O, 'water' on Twin Earth refers to XYZ, despite identical psychology. Therefore something external to the mind — the actual structure of the world — must partially determine semantic content.

The thought experiment's force comes from isolating the variable: holding psychology constant while varying the environment. Any theory explaining reference purely through mental states must say both uses of 'water' refer to the same thing — but this is intuitively wrong (they are different substances with different microstructures). The argument works because our intuition about natural kinds is stronger than our commitment to internalism.