Internalism and Externalism about Mental Content

Content internalism holds that a subject's mental content is fixed by what is inside the subject, so intrinsic duplicates share their thoughts. Content externalism holds that environment, causal history, or social practice partly determines what a subject thinks. Intrinsically identical subjects can therefore differ in content.

The dispute concerns constitution, not merely influence. Everyone agrees that the world causes beliefs and that communities teach words. The externalist claims these relations enter the identity of the content itself. This promises a robust word-world connection but complicates self-knowledge and causal explanation: how can a subject know their own thought directly if what they think depends on facts they may not know?

Twin Earth

Putnam asks us to imagine Twin Earth, molecule-for-molecule like Earth except that the clear, drinkable liquid in rivers and taps is not $H_{2}O$ but a different compound, XYZ. Before chemistry, Oscar on Earth and Twin Oscar are internally identical and use "water" in the same ways.

Putnam argues that Oscar's term refers to $H_{2}O$ while Twin Oscar's refers to XYZ. Their meanings differ because reference depends on the substance in their environments:

$$Internal(Oscar)=Internal(TwinOscar)$$

while

$$Content(wate{r}_O)\ne Content(wate{r}_T).$$

"Meanings just ain't in the head" follows. Applied to mental content, Oscar believes that water is wet while Twin Oscar believes that twater is wet, even though neither can distinguish the liquids internally.

The case most directly establishes externalism for natural-kind terms in public language. Extending it to all thought requires the premise that the content of belief inherits the externally fixed content of the concepts deployed.

Burge and social externalism

Tyler Burge's arthritis case shifts from nature to community. A patient mistakenly believes they have arthritis in the thigh. In the actual linguistic community, "arthritis" refers to joint inflammation, so the belief is false. In a counterfactual community where the same word correctly covers rheumatoid ailments of joints and muscles, an internally identical person expresses a different concept.

The actual subject's content depends on deference to communal experts even though the subject incompletely understands the term. This social externalism explains how ordinary speakers think about elm trees, electrons, and arthritis without possessing expert criteria.

The view makes error possible through partial understanding. It also raises a boundary problem: which community and experts fix content when usage is contested, multilingual, or changing? Social dependence must be structured rather than a vague appeal to what people around the subject say.

Natural and social externalism

Two externalist mechanisms should be separated.

• Natural-kind externalism: causal relations to substances or kinds partly fix reference.
• Social externalism: division of linguistic labour and communal norms partly fix concept content.

A subject can depend on both. Their term "gold" reaches a natural kind through a social chain of deference and historical contact. Neither mechanism requires the subject to know the microstructure or history.

Externalism can also extend beyond language. A perceptual state's content may depend on what its biological or learning history was selected or trained to track. Two current duplicates with different histories can then represent different properties.

Narrow and wide content

Wide content includes environmental or social relations. Narrow content is the internally fixed component shared by duplicates. A two-factor view uses both:

$$\text{narrow role}+\text{external relation}=\text{wide content}.$$

Narrow content explains why Oscar and Twin Oscar reason and behave alike locally; wide content explains why their thoughts have different truth conditions. This division is attractive but difficult to specify. If narrow content has no truth conditions, it may not be content in the ordinary sense. If it has truth conditions over abstract or centred possibilities, the account must connect them to the ordinary objects of thought.

Some externalists reject narrow content and explain local psychology through syntax, computational role, or content types coarse enough to be shared. Internalists instead treat externally individuated reference as a semantic fact about expression while preserving an internally fixed thought content.

Self-knowledge

A subject seems able to know without empirical investigation that they are thinking that water is wet. Externalism says the content depends on whether their environment contains $H_{2}O$, a fact they may not know. This creates the compatibility problem between externalism and privileged access.

The standard response distinguishes knowing what one thinks from knowing the conditions that make the thought have that content. A speaker can know that they think about water while not knowing water's chemistry or reference-fixing history. Self- knowledge is authoritative at the content level without containing a semantic theory.

Sceptical scenarios sharpen the issue. If one cannot know from the armchair whether one is on Earth or Twin Earth, how can one know whether the thought is water or twater? A disjunctive or deferential self-ascription may preserve authority, but the content of the self-ascription is itself externally fixed. Externalism makes self-knowledge socially and environmentally embedded even when it remains non-inferential.

Mental causation

Wide contents may differ while intrinsic physical states and immediate behaviour remain the same. If causes must be local intrinsic properties, wide content appears causally irrelevant. Oscar reaches for a glass because of an internal state shared with Twin Oscar, not because one state concerns $H_{2}O$ and the other XYZ.

Externalists answer in several ways:

• causal explanation can cite relational properties, as a key opens a lock partly because it matches that lock;
• content explains action across a broader environmental timescale, including learning history and successful interaction;
• narrow role explains bodily movement while wide content explains success, reference, and rational evaluation.

The division can seem to concede that meaning is causally idle at the moment of action. The alternative is to reject a purely local account of causation and treat organism- environment systems as the relevant causal units.

Slow switching

Suppose a subject is moved unknowingly between Earth and Twin Earth while their internal state remains initially unchanged. Does content switch immediately, gradually, or only after new causal relations stabilise? Slow switching cases test how historical and current environment contribute.

An instantaneous switch makes content depend on surroundings too directly; a delayed switch preserves history but needs a threshold. Memory creates a further puzzle: after switching, does the subject remember thoughts with old content while currently deploying new concepts?

These cases show that externalism is a family of dependence claims. Different contents may be fixed by current environment, causal origin, biological function, or social deference on different timescales.

Externalism and hallucination

Perceptual content must accommodate states caused abnormally or without their ordinary objects. A hallucinating subject may share an internal neural state with a veridical perceiver. If content is wholly wide, the states can differ despite subjective indistinguishability.

This can be a feature: veridical perception genuinely relates the subject to an object, while hallucination does not. It can also make common-cause psychological explanation difficult. Narrow representational content or disjunctive theories offer different ways to preserve either shared appearance or world-involvement.

The issue connects externalism with representational theories of consciousness: if phenomenal character is fixed by wide content, internal duplicates may differ phenomenally; if character is narrow, content and phenomenality can come apart.

The scope of externalism

Externalism is strongest for natural kinds, names, and socially deferred concepts. It is less obvious for logical concepts, sensations, or novel thoughts. A subject's pain may have bodily or environmental content while its phenomenal character remains internally fixed. Mathematical content may depend on inferential role rather than causal contact.

A pluralist view can therefore assign different fixation mechanisms to different kinds of content. The cost is loss of one general theory; the benefit is avoiding the claim that every thought depends on the environment in the same way.

Relation to public meaning

The public-language versions of causal-historical reference and social meaning are treated in Semantics and the Theory of Meaning. The present question is what follows for the states of speakers. Public terms can have externally fixed meanings while an individual's associated conception remains narrow, partial, or mistaken.

This distinction matters for machines and learners. Competent participation in a linguistic practice may confer externally anchored sentence content even when speaker meaning, internal concepts, or conscious understanding remain disputed. Semantic externalism can establish aboutness without establishing phenomenality.

Assessment

Externalism establishes that thought is not simply an inner inscription interpreted in isolation. It does not yet explain how causal and social relations become normative content or how misrepresentation is possible. Those are the tasks of naturalisation.

Selected references

• Burge, Tyler. "Individualism and the Mental" (1979).
• Fodor, Jerry A. Psychosemantics (1987).
• Putnam, Hilary. "The Meaning of 'Meaning'" (1975).
• Segal, Gabriel. A Slim Book about Narrow Content (2000).