The Atlas of Reality

by Robert C. Koons,Timothy Pickavance

  In addition, if epiphenomenalism were true, we would have to take seriously the possibility that everyone else is a zombie—a body with no attached, experiencing soul—since zombies would behave no differently in the absence of a soul than would en-souled bodies. Finally, we all have direct experience of acting upon the world by moving our bodies in intentional ways. All of this would have to be an illusion for epiphenomenalism to be true.

  The evident failure of epiphenomenalism illustrates the difficulties with positing entities solely on the basis of those entities' supposed passive powers. Consequently, we can plausibly strengthen Redundancy, restricting the permissible entities to those with non-redundant active and immanent powers. In addition, if an entity is wholly redundant with respect to its active and immanent powers, then its tendencies to use those powers would also seem to be redundant, consisting entirely in the tendencies of the constituent parts to use their powers.

  Finally, suppose that an entity's active powers are redundant. Could it have non-redundant immanent powers? It is hard to see how it could, since it would have no active powers of its own to affect. So, its immanent powers would have to be powers only over other immanent powers, keeping the effectiveness of the entity forever trapped in a self-referential circle. Let's restrict the Redundancy principle, then, to active powers alone:

  Active Power Redundancy. Reject any theory that posits fundamental entities whose fundamental active causal powers are redundant, given the other fundamental entities posited by that theory.

  When do the active powers of the parts of a composite thing make the active powers of the whole redundant? What would it be for a whole to have non-redundant active powers? Some philosophers, going back to the British Emergentists of the early twentieth century, have called such powers ‘emergent’. A whole has non-redundant active powers when it has emergent powers.

  A whole has emergent active powers when it is able to do things that cannot be explained in terms of the powers of its parts. By saying that this cannot be explained, we do not mean merely that it cannot be explained by us right now, due to the complexity or other practical obstacles involved in formulating such an explanation. We mean something like that the emergent power cannot be explained in principle, not even by God, in terms of the powers of the parts. We are interested in ontological emergence, not merely some sort of human-centered epistemological emergence.

  When a whole has emergent powers, actions take place that are not merely the sum of the individual actions of the component parts. The whole is greater in active power than the sum of the active powers of the parts.

  Could we ever be in a position to know, or at least to believe with good reason, that some whole has emergent powers? One worry is that whenever we see a whole seeming to do something over and above what could be done by its parts, we could always describe what's going on as the exercise of hidden powers of the parts, powers that those parts never exercise except when they are put together in such a way as to constitute a whole of this kind. Since that always seems to be a live option, we would never be able to tell whether it is really the whole exercising some emergent power or just the parts jointly exercising some hidden powers.

  We could escape from this impasse if we could find certain powers that are essentially unitary, powers that could only be exercised by some one, unified entity, and that could never consist in the joint possession of individual powers by the members of a plurality. Here are several possible candidates for essentially unitary powers:

  The power of self-reproduction, as exercised by living organisms.

  The power of growth, self-development, and self-repair through the assimilation of new material.

  The power of self-determination as exercised in free, conscious choices.

  Each of these powers makes reference to the self that exercises it. This self-reference is what seems to make each essentially unitary. The particles making up an organism's body cannot exercise the power of self-reproduction, since it is the organism and not any of the particles that is reproduced. Similarly, the particles do not grow, either individually or collectively. If the organism contained a billion billion particles before the episode of growth, then those billion billion particles are no larger or more massive afterward than they were before. It is the organism that grows, not the particles. Finally, the particles making up a person's body do not make up their collective mind to do one thing rather than another. The particles don't have minds to be made up—only the person does. Therefore, we will propose that a theory may reasonably posit the existence of a fundamental composite entity only if it credits that entity with powers that are essentially unitary. It seems that these essentially unitary powers will be either immanent or passive powers, not active ones. Anything that can be done to some other thing could, it seems, by done either by a single agent or by a plurality of agents acting jointly.

  A passive power, in contrast, might be essentially unitary. The power of sensation, for example, seems essentially unitary. Being affected with a particular sense-quality is something that can happen only to a single thing because sensory consciousness is essentially unified. Immanent powers can also be essentially unitary, like the power of self-reproduction or organic growth. Thus, in order for a theory's postulation of composite entities to be justified, the composite entities must be assigned both emergent active powers and essentially unitary passive or immanent powers.

  Def D22.1 Emergent Power. A power of a composite entity x is emergent if and only if the power of x is not wholly grounded in the sum of the causal powers of x's parts, together with the intrinsic qualities and mutual relations of those parts.

  Def D22.2 Essentially Unitary Power. A power is essentially unitary if and only if it is a fundamental power that could by its very nature be possessed only by a single entity, not collectively possessed by a plurality of entities.

  PMeth 4 Redundancy for Composite Entities. Reject any theory that posits any kind of fundamental composite entity without both emergent active powers and essentially unitary passive or immanent powers.

  Redundancy for Composite Entities presupposes that we can distinguish between fundamental and derived powers. This distinction is implied by Strong Powerism (4.4A.3), the view that some causal powers are fundamental properties.

  A similar distinction could be made on the basis of either Strong Nomism (4.4A.2) or Strong Hypotheticalism (4.4A.1). Suppose, first, that Nomism were true. This would entail that there are fundamental truths concerning the existence of certain laws of nature, where each law of nature involves some necessary connection among two or more properties. Each law of nature could be thought of as grounding a set of relatively fundamental causal powers. Suppose, for example, we had a law requiring that F's change themselves into G's. This would correspond to a fundamental immanent power possessed by all F's, namely, the power to become a G. Other powers, involving properties not linked together directly by laws, would count as derivative powers. For instance, if F's have the power to become G's and also the fundamental power to become H's, then they would have the derived power to become G and H.

  Similarly, Strong Hypotheticalism posits certain counterfactual conditionals as fundamentally true. Each fundamentally true counterfactual would correspond to a set of relatively fundamental powers. Counterfactual conditionals that are only derivatively true would correspond to derivative causal powers.

  Suppose, however, that we were to embrace Neo-Humeism (4.4T). According to Neo-Humeism, no causal powers are metaphysically fundamental, and neither are any causal laws or counterfactual conditionals. The only fundamental truths involve the distribution of fully categorical qualities (non-powers) in space and time. Powers, laws, and conditionals are determined by whatever set of global regularities is simplest and most powerful, that is, by the fact that some scientific theory is the best fit to the world's basic facts. On such a theory, there would be no basis for a distinction between fundamental and derived causal powers. All powers are equally derived. There
is no one-to-one correspondence between fundamental facts and relatively fundamental powers either, unlike with Nomism and Hypotheticalism.

  Without a distinction between fundamental and derivative causal powers, the very idea of redundant powers makes no sense. Both simple entities and composite entities derive their powers from the overall pattern of facts throughout the world. On Neo-Humeism, there is no real redundancy in assigning powers with equal status both to a whole and to each of its parts, even if we could explain the same behavioral regularities in terms of the powers at either level alone.

  Neo-Humeists should, we think, embrace the thesis of Compositional Equivalence. They should deny that parts are prior to their wholes, and that wholes are prior to their parts. Instead, the existence of both parts and wholes is wholly grounded in facts about the Humean mosaic, the distribution of qualities in spacetime. They must think of this mosaic either nihilistically (as a thing-free world) or atomistically (as consisting of simple, point-like entities that cannot be combined into composites). As we saw in Chapter 11, Nihilism runs up against some serious problems with Ockham's Razor, with the attainment of qualitative economy. Thus, the Neo-Humeists' best bet would seem to be one of treating the pixels in the mosaic as simple and uncombinable. The entities that can be either parts or wholes should be treated as derived entities, wholly grounded in the facts about the Mosaic.

  There is one further wrinkle to mention. It might be possible for Powerists and Nomists to reject Redundancy for Composite Entities and to embrace instead the first form of the Ontological Free Lunch principle, that is, Compositional Equivalence (22.1A.1T.1). They can do this if they can claim, with plausibility, that composite entities have causal powers, but that those powers are identical to the powers of their parts. This would require, in turn, both that the composite entities instantiate power-conferring universals, and that the instantiation of such universals by a composite entity is nothing over and above the instantiation by its proper parts of that or other power-conferring universals. For this to make sense, we would have to make sense of one case of instantiation's being identical to a number of other cases of instantiation.

  In order to investigate whether we can make sense of such reducible or non-fundamental cases of instantiation, we will have to break the investigation up into two parts, one relying on Relational Ontology (9.1A) and the other on Constituent Ontology (9.1T). Consider first the case of Relational Ontology, specifically Classical Relational Realism (9.1A.1T). Since Relational Realism takes instantiation itself to be a fundamental relation, it would seem to be impossible for Relational Realists to count some cases of instantiation as identical or equivalent to other cases of instantiation. There may be an exception, however, in the case of those Relational Realists who believe in nexuses, as entities that connect particulars to universals. If instantiation consists in the existence of a nexus, then we could suppose that some nexuses are composite, that is, that they have proper parts. If so, it could be that the existence of a nexus N connecting a composite substance C to a universal U just is the existence of the proper parts of N, each of which connects some proper part of C to U.

  This can work only in cases in which a composite entity instantiates some quality that is perfectly homogeneous, in the sense that every part of the entity necessarily instantiates the same quality whenever the whole does. Having a certain uniform density or uniform temperature would be examples of such homogeneous qualities. However, this strategy of composite nexuses will not work in the case of non-homogeneous properties, like size, shape, or total mass. The proper parts of a cubical whole do not have to be cubical themselves, and the proper parts of a body with mass of one gram must have a mass that is less than one gram. In order to deal with such non-homogeneous properties, the Relational Realists will have to suppose that the corresponding universals are themselves composite. So, when a composite body B is connected by nexus N to some quantity Q (like BEING ONE GRAM), then if B is composed of two discrete parts, B1 and B2, then N must also be composed of two parts N1 and N2, and the universal Q must consist of two parts Q1 and Q2, such that N1 connects B1 to Q1, and N2 connects B2 to Q2.

  However, this raises an obvious problem. How can a universal being BEING ONE GRAM be composed of two parts, each of which is identical to the BEING ONE-HALF GRAM universal? What would it mean for one universal to be composed of two “copies” of another universal? We encountered exactly this problem in Section 10.3, when we investigated structural universals. Our solution was a resort to amphibian theory. So, we could suppose that the universal BEING ONE GRAM consisting of two amphibians of the universal BEING ONE-HALF GRAM.

  Will amphibian theory enable Realists to embrace Compositional Equivalence? The answer is unclear. There is a difference between instantiating a universal and instantiating an amphibian of that universal. Therefore, it does not seem that a nexus connecting a whole to some structural universal could be merely the sum of the nexuses connecting the parts to the amphibians of that structure. It seems that instantiating a universal would be a more fundamental fact than merely instantiating an amphibian, in which case the whole that instantiates a structural universal would be metaphysically prior to its parts (which merely instantiate amphibians of parts of that structure). This would seem to exclude Compositional Equivalence.

  Now, let's turn to Constituent Ontologists who are also Realists (7.1T), whether Classical Bundle Theorists (9.1T.1T.1A) or Classical Substrate Theorists (9.1T.1A.1A). Let's consider homogeneous qualities, like color or density first. It seems that a Constituent Realist cannot embrace Compositional Equivalence, since the instantiation of a universal by a particular just is (according to Constituent Realism) that universal's being an immediate proper part of the particular: that is, its being a proper part, and not by virtue of its being a part of some other proper part. Consequently, whenever we find a universal as an immediate proper part of some particular, that particular must be metaphysically fundamental, and the fact that its proper parts also instantiate that universal must be wholly grounded in the instantiation of that universal by the whole.

  22.4 Fundamental Heaps

  We conjecture that the only sort of composite things that could have essentially unitary powers are living things, artifacts, and their functional parts (like organs and so on). Any other sort of thing is a heap.

  Def D22.3 Heap. x is a heap if and only if x has actual parts and x is not a living thing, an artifact or a functional part of a living thing or an artifact.

  One faces a choice with respect to the fundamental existence of heaps, and a further choice if one denies that there are fundamental heaps:

  22.2T Fundamental Heapism. Some heaps exist fundamentally.

  22.2A Anti-Heapism. No heaps exist fundamentally.

  Examples of heaps abound. All positive natural formations are heaps. These include rocks, clouds, grains of sand, mountains, rivers, lakes, oceans, planets, stars, galaxies, crystals, and literal heaps. Some very exotic things, if they exist at all, would also be heaps. If, for example, there existed something consisting of THP's left big toe and the Eiffel Tower, that thing would be a heap. So would the thing consisting of all of the earth's water if there is such a thing. Arbitrary parts of things, if they exist, are also heaps. RCK's left side is a heap of organic matter, as it is neither itself a living thing nor is it a functional part of my body, like my heart or liver.

  22.4.1 Arguments against fundamental heaps

  Do heaps really exist fundamentally? There are three objections to Fundamental Heapism: the appeal to Redundancy (PMeth 4), the problem of identifying the fundamental occupiers of space, the appeal to Ockham's Razor (PMeth 1). In addition, we will consider in Section 2.7 an objection to the existence of heaps of any kind, whether fundamental or not—namely, the difficulty of finding a principle of composition that matches our pretheoretical convictions.

  1. Appeal to redundancy. Redundancy for Composite Entities (PMeth 4) makes trouble for heaps. Heaps seem to be causally redundant. T
here is no reason to think that they have either emergent active powers or essentially unitary passive or immanent powers. Anything that a heap does or causes is really done or caused by its constituent parts acting in concert. For example, to use Merricks's (2003) example, when a rock breaks a window, it is really the congeries of particles arranged in a baseball-pattern that shatter the window or, to be more precise, that scatter the particles arranged window-wise. The only agency rocks have is that of physical matter: energy, momentum, angular momentum, temperature, and so on. All of these powers are simply the aggregation of the powers of the rock's constituent particles. The rock itself is causally superfluous.

  Similarly, rocks and other heaps have no essentially unitary passive powers either. Everything that can be done to a heap can always be re-described, apparently without loss of information, as something that is done to its constituent parts. Heaps can be changed in shape, scattered or re-assembled, but each of these processes consist in nothing more than the movement of the particles. Heaps can be painted or made radioactive, but these changes also involve nothing over and above certain changes to the simplest parts. Heaps can grow or diminish in size, but only as a result of particles' being spread out or pressed together. If more particles are added to a heap, we could always say that, strictly speaking, we have a new heap, composed of a new set of particles.

  But don't mountains, rivers, and crystals grow, just as organisms do? When a mountain grows, its constituent particles don't grow (either individually or collectively). It grows by adding particles, just as organisms do. However, there is a crucial difference. When a mountain “grows”, this is merely the accidental by-product of more fundamental processes, like continental uplift. In contrast, it is at least possible that organic growth is a fundamental process in its own right, with the growth and development of the organism as its proper effect. Organic growth is much more than the mere accidental accumulation of additional matter.