The Atlas of Reality
Page 50
The idea of color as a power or disposition of physical objects only makes sense if there are, in addition to the physical colors, a set of phenomenal colors that characterize the sensations that colored objects are supposed to cause. Let's distinguish ‘phy-red’, the physical power to cause certain sensations, from ‘phe-red’, a certain intrinsic type of sensation. We could define ‘phy-red’ as the power to cause phe-red sensations in normal observers in normal circumstances without circularity. However, it seems clear that we do not directly perceive objects to be phy-red. Phy-red is not how objects appear to us to be in sense experience. If anything, objects appear to be phe-red.
This sort of response is also available to Indirect Realists. We could perceive mental objects to be phenomenally-red directly, thereby perceiving indirectly some physical objects to be phy-red. Indirect Realists can attribute such powers to physical objects, and we could call these a kind of “color,” but they are not what are directly perceived, since they are not what is given in sense experience.
In addition, this power theory of color runs into some difficulties in accounting for the intra-species variability and the extreme context-sensitivity of color perception (see observations #4, #5, and #6). Physical colors will have to be somewhat vague and very context-sensitive.
In summary, we have seen good reasons for thinking that colors are natural (grounding similarities), causally powerful, categorical, and intrinsic. These facts are inconsistent with the physicality of color.
THE UPSHOT FROM THE ARGUMENT FROM COLOR If colors are not physical properties, what are the options? There are four possibilities.
Colors are not properties of physical things at all but rather properties of the perceiver's mind or some part of the mind. So, we perceive directly only mental things. This would entail either Indirect Realism or the Veil of Perception. This could be extended to a Double Color Theory. On this view, there are colors of mental objects (the things we directly perceive) and colors of physical objects (that we indirectly perceive or infer), with the physical colors being either unnatural disjunctions or relational dispositions or powers.
Color Irrealism. Color Irrealism, defended by Adam Pautz (2006, 2010), is the view that colors are uninstantiated physical properties. Color Irrealism is most naturally combined with a Meinongian version of the Veil of Perception, according to which we only perceive non-existent things. Or, we could adopt an error theory of color perception, according to which we do perceive physical objects, but we systematically misperceive them as colored. This option pays a heavy price in terms of violating Sensory Error Minimization. In addition, it is difficult for Color Irrealists to give an account of how our sensations have the intentional content they do have. They cannot provide a theory of intentional content in terms of causal connection or reliable covariation, since our color sensations are not causally connected to real cases of color, nor do they co-vary with any such cases. There simply aren't any real cases of color. It remains mysterious how our senses connect us to properties that are not, and probably could not, be instantiated in the sort of world we actually inhabit.
Emergent Properties. The Joints Hypothesis is false, and colors are core physical properties, after all. They just aren't at all like the properties posited by modern theoretical physics. Suppose that there are, at the macroscopic (or mesoscopic, medium-sized) scale of colored objects and surfaces, emergent core physical properties, fully natural properties that are caused by but not reducible to or wholly grounded in the properties of their microscopic parts. The phenomenon of quantum entanglement (which we discussed in connection with Jonathan Schaffer's argument for Monism 11.2A in Section 11.2.4) could provide some support for the possibility of such irreducibly macroscopic properties. We will examine emergent properties in more detail in Section 22.6.
On this view, color perception would give us direct access to some real sensible qualities of physical objects, an access that goes beyond the resources of theoretical science.
Emergent Color Theory is a version of Powerism (4.4A.3) with sicceities (Def 6.3, see Section 6.2).
Emergent Color Theory has the advantage that colors and other sensible qualities are real, intrinsic, categorical, natural properties of physical things. In addition, they would be causally relevant, since, they confer active and passive causal powers on their bearers.
One difficulty with Emergent Color Theory is the threat of redundant causation, with human color perception's being causally overdetermined by colored surfaces and by microphysical phenomena, such as photons and electron shells of atoms. Emergentists can ameliorate this by claiming that the emergent properties of the macroscopic surfaces absorb and co-opt the causal powers of the constituent micro-particles, with the result that the powers of the micro-particles are wholly grounded in the macroscopic qualities.
Another difficulty concerns the relativity of color perception, especially across species (fact #5). They could suppose that the emergent qualities correspond to the constituent colorishnesses (like yellowishness and greenishness), whose causal powers are exercised jointly. This would mean that human color perception is incomplete, insensitive to the colorish qualities perceived by other species. Nonetheless, what we do perceive would be really there (as natural, categorical, causally efficacious properties). Our two-channel perceptual system, for example, might have evolved precisely in order to detect the distinct colorishnesses that have emerged in the environment.
Color Neutralism. According to Color Neutralism, the Joints Hypothesis is true but PPPT is false: physical objects have non-physical properties, properties that are themselves core physical properties nor are definable in terms of the core properties of theoretical physics. However, these properties are not mental, either. On this view, colors are neither mental nor physical. That is, colors are not physical in the narrow sense of being definable in terms of the fundamental properties of modern physical theory (in any of the ways 1–4 listed above). They could still be broadly physical in the sense of being natural and non-mental (although not necessarily efficacious). We turn presently to a more detailed examination of Color Neutralism.
COLOR NEUTRALISM Color Neutralism comes in two varieties: Brutely Supervenient Qualities and Panqualityism.
a. Brutely Supervenient Qualities
In Chapter 3, we introduced the idea of supervenience (Def 3.1). We offered this definition:
Def D2.6 Weak Supervenience. A set of properties A weakly supervenes on a set of properties B if and only if it is impossible for any two possible worlds to agree on which things have which B-properties but to disagree about which things have which A-properties. That is, two situations which are indiscernible in respect of the B-properties must also be indiscernible in respect of the A-properties.
Let's say that one set of properties weakly supervenes on another with metaphysical necessity when the kind of impossibility mentioned in Def 3.1 is metaphysical impossibility (see Chapter 14). Let's say that a case of supervenience is brute when it holds with metaphysical necessity but it is a posteriori (not a priori or based on conceptual connections) and not derivable from facts about identity or distinctness.
Def D13.10 Conceptually Brute Supervenience. A set of properties A brutely supervenes on a set of properties B if and only if A weakly supervenes on B with metaphysical necessity, and no predication of an A-property is a priori deducible from any conjunction of predications of B-properties, together with any set of true identities and non-identities.
According to Brutely Supervenient Qualities, colors brutely supervene on the physical properties of things. The connection between a thing's color and its physical constitution is metaphysically necessary, but we couldn't deduce a thing's color from its physical constitution, no matter how familiar one was with the intrinsic nature of that physical constitution.
There are two objections to such Brutely Supervenient Qualities. First, it's not clear that such supervenient qualities can be causally efficacious or even relevant. Jaegwon Kim (1992, 1
998, 2001) has argued that the causal relevance of brutely supervenient properties is excluded by the causal completeness of the physical realm. Since physical properties are sufficient to ground all causal interactions, there is simply no room for the causal relevance of brutely supervenient properties. This is clearly true if we assume Powerism or Nomism (4.4A.2) and if we assume that Brutely Supervenient Qualities are less fundamental than their physical bases, since it is the fundamental properties that confer causal powers on things (according to Powerism) or are situated within causal laws (according to Nomism). Kim's exclusion argument is not compelling, however, if we assume Neo-Humeism (4.4T) or Hypotheticalism (4.4A.1), since on those views neither causal powers nor causal laws are fundamental, and so it is less clear that there is anything objectionable about attributing causal power to both physical properties and brutely supervenient ones.
Second, there is a strong modal argument against the supervenience thesis entailed by Brutely Supervenient Qualities. We seem to be able to conceive of a world where color and its physical bases come apart. We can imagine something's having the physical and chemical composition of a typical ripe tomato and yet being blue rather than red. In light of Imagination as Guide to Possibility (PEpist 1), we have good reason to think that colors do not supervene on the set of physical properties.
We can avoid this modal objection by supposing that colors did not supervene with metaphysical necessity on their physical bases. We might suppose instead that there are some contingent causal laws that link colors with physical properties, in such a way that the physical properties cause their bearers to have certain colors. Although this helps with the modal argument, it makes the problem of causal exclusion much worse. It would mean that colors were causally inert by-products of underlying physical processes. This would make it a mystery how we are able to get the actual colors of things right. It's hard to see why the colors we experience would correlate with the real colors of things if those real colors are epiphenomenal, that is, if they make no difference to the causal powers of things.
Thus, the defender of Brutely Supervenient Qualities faces a dilemma: the Scylla of causal exclusion versus the Charybdis of imaginability.
b. Panqualityism
Panqualityism is a recent proposal by Brian Cutter (2013). Cutter looks back to Bertrand Russell's (1927) structuralist understanding of physical science. According to this type of structuralism, physical science tells us only about the causal powers of physical properties. It doesn't tell us anything about the intrinsic character or quality that corresponds to those properties. Yet we have good reason to suppose that there must be such intrinsic qualities, unknown to us. The physical sciences provide us with an abstract structure of properties, expressed in the form of causal laws. They do not tell us what qualities actually fill the roles specified by this abstract structure.
Structuralism corresponds to Powerism with sicceities. We don't have firsthand experience of or acquaintance with any of the sicceities of the fundamental physical properties. We know them only via a description of their causal and nomological profiles. Cutter, therefore, proposes that every core physical property has a qualitative sicceity, a micro-quality, of which we have no acquaintance. However, the colors we perceive are wholly grounded in these micro-qualities, in such a way that any subject that was acquainted with both the micro-qualities and the colors would be able to deduce (a priori) which combinations of micro-qualities ground which colors. Cutter's view is Panqualityism.
The possibility of such a strong, conceptual connection between the two rebuts the modal argument against supervenience that seemed to be decisive against Brutely Supervenient Qualities. Since we don't know what the micro-qualities are like, we can't in fact conceive of a case in which the micro-qualities are present and the color is lacking. According to Panqualityism, if we were able to conceive of the micro-qualities, we would be able to see a priori that it is impossible for those micro-qualities to be present and the color to be lacking.
In addition, such a close conceptual connection solves Kim's causal exclusion problem. Colors are intimately connected to grounding micro-qualities, and micro-qualities are causally efficacious, since they are strictly identical to core physical properties. We might plausibly say that an object's being colored just is its having the appropriate micro-quality. Since micro-qualities are causally effective, so are colors.
On Panqualityism, colors are broadly physical, since they are a priori deducible from core physical properties. However, we cannot so deduce them, since we are not acquainted with the core physical properties in their qualitative aspect.
Panqualityism has a number of advantages. Colors are intrinsic, categorical, natural, and causally relevant, just as they seem to be in our experience. But Panqualityism does face one daunting challenge: it isn't clearly plausible that colors could be a priori deducible from other qualities, even qualities with which we are completely unacquainted. Colors seem to be simple properties, neither conjunctive nor disjunctive, so the a priori deducibility of colors from micro-qualities cannot be grounded in logical relations, the way disjunctions are deducible from a disjunct, or conjuncts from conjunctions. Colors are not identical to micro-qualities, so the deduction cannot proceed by way of the substitution of co-referring terms. We are left with something of a mystery in trying to understand how the a priori deducibility could be explained.
13.4 Arguments against Phenomenalism
13.4.1 What follows from the Veil of Perception?
As we have seen, the main argument for Idealism depends on the Veil of Perception, which also provides a case for Solipsism. The two arguments for the Veil, from hallucination and from color, are inconclusive. Intentionalism, Meinongian Direct Realism, and Perceptual Dualism, in particular, have not been ruled out by the argument from hallucination, and several versions of Direct Realism, including Color Irrealism, Hidden Conjunct Theory, and Color Neutralism, are able to stand up against the pressure of the argument from color.
Nonetheless, the two arguments we have considered are not without force. Let's consider what would follow from accepting the Veil of Perception. As we mentioned above, there are two versions of the Veil of Perception: the Solipsistic Veil (SVP) and the non-Solipsistic Veil (NSVP). SVP is logically stronger than NSVP, since if each subject cannot perceive anything but parts of that subject's own mind, then no subject can perceive anything that isn't a part of some mind.
Both versions gain some support from the argument from hallucination, while the argument from color supports only the weaker NSVP.
Assuming NSVP, we have three options with regard to the non-mental and the external, namely, Inferred Anti-Idealism about the non-mental (of which more below), Phenomenalism (non-mental objects are wholly grounded in mental objects), and Eliminative Idealism (there are no physical objects at all). Phenomenalism is more moderate than Idealism, since it permits its adherents to admit that physical and external things do exist. Historically, there have been three versions of Phenomenalism, namely, Berkeleyan (Theistic) Phenomenalism, Possibilistic Phenomenalism, and Holistic (Constructive) Phenomenalism.
Berkeleyan Phenomenalism supposes that there is a God who shares our sensory experiences of the world. According to Berkeleyan Phenomenalism, God experiences every sensible object in every sensory mode and from every perspective. In fact, physical objects are really nothing but comprehensive bundles of ideas in God's mind. Our sensory perception is a kind of sampling by participation in God's comprehensive sensing of the world. The natures we attribute to physical objects in common sense and in science are merely useful fictions, constituting a kind of instrument for predicting and controlling our future sensations. Berkeleyan Phenomenalism has been summarized in a limerick by Ronald Knox (Reed 1924):
There was a young man who said, “God
Must think it exceedingly odd
If he finds that this tree
Continues to be
When there's no one about in the Quad.”
&n
bsp; REPLY
Dear Sir:
Your astonishment's odd:
I am always about in the Quad.
And that's why the tree
Will continue to be,
Since observed by
Yours faithfully,
GOD.
Possibilistic Phenomenalism was proposed by John Stuart Mill (1865/1963) as an attempt to reproduce the advantages of Berkeleyan Phenomenalism without positing God as the universal perceiver. Mill proposed that physical objects are what he called ‘permanent possibilities of perception’. We could still think of each physical object as a bundle of sensations, but now not actual sensations in God's mind but merely possible sensations in anyone's mind. Possibilistic Phenomenalism requires Hypotheticalism. There must be metaphysically fundamental conditionals of the form, if perceiver P were in condition C, P would perceive sensation S. Since there are no non-mental entities and no actual perceiver of these sensations, such conditionals lack categorical truthmakers.
Holistic (or Constructive) Phenomenalism is best exemplified by Rudolf Carnap's The Logical Structure of the World (Carnap 1928/1967) or Aufbau (in German). In Carnap's system, individual physical objects are not replaced by bundles of possible sensations. Instead, the whole physical world is replaced in one fell swoop by a single system of sensory conditionals. Carnap supposes that we posit the physical world as the simplest system of objects and states that can, hypothetically, explain, predict, and unify the sensory experiences of minds. On this view, no physical fact is fundamental.