World and Head
When Hilary Putnam made the claim that meanings are not “in the head” he emphasized the indexical character of natural kind terms. His point was that terms like “water” have their reference fixed by demonstratives like “that liquid”, and demonstratives have their reference as a function of context not descriptions in the mind of the speaker. In David Kaplan’s terminology, context yields content in conjunction with character—character alone cannot determine reference. Thus two people could be internally indistinguishable and yet refer to different things with “that liquid” depending on the actual physical context (H2O or XYZ). Since the anchoring demonstrative fixes the reference of “water”, that term too will vary its reference according to context—hence Twin Earth cases. Strictly speaking, Putnam overstated his conclusion, since all that follows is that an aspect of meaning is not in the head—the aspect Kaplan calls content; character does not vary in this way, being completely context-independent. He should have said that part of meaning is not in the head, the part that is “in” the context. The meaning of an indexical is always a two-component affair: the component that results from context and the component that is “in the head”. In the case of a demonstrative like “that liquid”, the first component corresponds to the external natural kind being demonstrated, while the second component may be identified with something like the perceptual appearance presented by the demonstrated object or kind. If someone were to claim that the meaning of a demonstrative is completely outside the head, the reply would be that an aspect of its meaning is clearly inside the head—the aspect Kaplan calls character. Indexical meaning is double-aspect.
So far, so familiar: what I want to ask is whether the natural kind itself is wholly “outside the head”. Is what we mean by “water” something completely divorced from its appearance to human minds? Is the reference of “water” purely an objective matter? Or is the reference partly constituted by what is in the mind? Water has both a hidden essence, captured by “H2O”, and a superficial appearance, captured by “transparent, tasteless liquid”: is only the first of these constitutive of water? It might be tempting to suppose so—it is necessary and sufficient for something to be water that it be H2O. But that has to be wrong: in a possible world in which something is H2O but has none of the manifest properties of water we would not say that that stuff is water. If it had all the manifest and dispositional properties of honey, it would be wrong to classify it as water, no matter that it is composed of H2O. This might prompt the retort that nothing could be H2O and have the manifest properties of honey, because of the necessary connection between chemical composition and manifest properties; but (a) not all the manifest properties of water follow simply from its being H2O, and (b) this is to concede that manifest properties also count as necessary conditions of being water. In fact, being tasteless and colorless are parts of the nature of water, along with its chemical composition. Natural kinds really have a double nature—underlying real essence and apparent nominal essence (to use Locke’s terminology). Both are necessary to being water, neither being sufficient.
Now we come to the more difficult question: is nominal essence tied essentially to the mind? Is it “in the head”? There is an obvious precedent for such a claim, namely colors and other secondary qualities: objects only have colors in relation to minds—so colors are “in the head”. That means that colored objects (qua colored) are partly in the head: an object is red only because of its relation to a mind that perceives it as red. Colored objects are partly in the world and partly in the head—they have both aspects. In the case of water we can likewise claim that being colorless and tasteless are also secondary qualities: things only have these qualities because there are minds that respond with certain types of sensory experience to them. But there is a further point: the very concept of a manifest property is tacitly mind-relative. Manifest to whom? Properties are only manifest in relation to minds that can perceive them; to beings with different senses from ours chemical composition might be manifest while color and taste are inferred. Real and nominal essence could conceivably be reversed. As things stand, however, the properties we attribute to the surface of water are mind-relative, so that this aspect of the nature of water is mind-dependent. It is a function of how our senses represent the world. Water, as we commonsensically conceive it, is thus partly “in the head”. In fact, the part that is in the head coincides with the part of meaning that is in the head, viz. the sensory appearance. Thus meaning has two aspects, one in the head and one not, and likewise natural kinds have two aspects, one outside the head (chemical composition) and one not (sensory appearance). Meanings have a worldly component and a mental component, but so do the objects we refer to. In the sense that meaning is not “in the head”, so objects are not “in the world”, i.e. they are partly so. It is the fact that objects have a foot in both camps that enables meaning also to have a foot in both camps. Surface properties correlate with demonstrative modes of presentation, while hidden properties characterize the reference, as it exists independently of such modes of presentation. The duality of meaning thus maps onto a duality at the level of reference: the inner component of meaning contains the manifest properties of the reference, while the outer component corresponds to the non-manifest properties. Meaning is a hybrid of internal and external, as the character-content analysis of demonstratives suggests, while objects themselves divide into an objective aspect and an aspect that is tied to perception. The distinction between the scientific image and the manifest image is mirrored in the distinction between content and character, respectively—what is fixed by context or environment or causation and what is wholly subjective or “in the head”. To put it differently, the objects of common sense have a dual component analysis corresponding to real and nominal essence. The abstract structure of objects is thus reflected in the abstract structure of meaning: a bit in the head plus a bit not in the head. Character goes with surface and content goes with hidden. There is more to meaning than what is in the head, and there is more to objects than what appears to heads—but we must not forget that the head-involving aspects are also part of meaning and of objects. Meaning is not only what lies beyond the head, and objects are not only what is independent of appearance; meanings and objects are both hybrids. Meanings are composites of two factors, as Putnam taught us (with a little help from Kaplan), but so also are objects.
 Here we might be reminded of Eddington’s two tables: the table of science and the table of common sense. The former has nothing mind-dependent about it while the latter is a projection of mind. Eddington in effect reifies the distinction of aspects or components that I am suggesting. Frege’s sense-reference distinction also finds a counterpart in the distinction between real and nominal essence—roughly, the distinction between the known and the unknown properties of objects. Senses are necessarily known by one who grasps them, but it is possible to refer to things and know little about them. Then too, we have Kant’s bifurcation into the phenomenal and the noumenal. In philosophy later distinctions often trace back to earlier distinctions. And what seems unitary often turns out to be divided.