The Atlas of Reality
Page 93
As Sider (2001: 191–192) puts it, the difference between Worm Theory and the Stage Theory lies in the realm of semantics rather than metaphysics. Both agree exactly about what sorts of things exist. The difference between the two lies simply in their varying accounts of the correct truth-conditions to give to ordinary language assertions about the existence and diachronic identity of persisting things. For Stage Theorists, names pick out temporal parts; for Worm Theorists, they pick out spacetime worms.
Worm Theory has an obvious solution to all of these problems. Distinct spacetime worms that wholly (or nearly wholly) share a common segment throughout some temporal interval. However, Worm Theory cannot explain how Statue and Lump could differ if they in fact coincide throughout their entire lifetimes. If, in other words, the clay and the statue were created and destroyed at exactly the same times, they would still be distinct worms because they would still differ in their modal properties: Lump could have survived crushing, but Statue could not.
Here, Worm Theorists (and Stage Theorists as well) must appeal to Counterpart Theory (16.1A.1). Statue and Lump have different counterparts in other possible worlds. Specifically, in a world in which Statue/Lump are crushed, the counterpart of Lump in that world persists, while the counterpart of Statue does not. Worm Theorists could take the relata of the counterpart relation to be spacetime worms in the various worlds if they then insist that different counterpart relations be used for different sortals (like statue and piece of clay). Thus, Counterpart Theory offers Worm Theorists a simple response to the grounding objection: it is our concepts and interests that make different counterpart relations relevant for statues and for lumps of clay. There need be no deep, metaphysical facts about which ground the difference.
A great advantage to Stage Theory is that it can make use of counterpart relations both for de re modality and for temporal persistence. A given time-slice can, under a given sortal, have both temporal counterparts at other times in the same world and modal counterparts in other worlds.
Peter Unger's Problem of the Many (Unger 1980) illustrates some common ground between Worm Theory and Stage Theory. Suppose that we have a single, vaguely composed and vaguely located object. On both theories, there are multiple, equally good candidates for the job of being the referent of the relevant singular term (that is, all of the spacetime worms that include candidate precisifications of the object's location and composition or all of the contemporary stages of those worms). In giving a semantic account of assertions involving the term, Stage Theorists have two choices: (i) a supervaluationist semantics, in which assertions come out as super-true or super-false just in case they are true (or false) under each of the candidate referents or (ii) a semantics that makes use of a notion of approximate truth or near truth. On the second account, it is literally true that there are trillions of persons in the room, but it is almost true that there are 14, since there are 14 clusters of nearly-identical, highly-coincident spacetime worms.
The difference in the two accounts comes out in the case of persistence through time. Unlike Worm Theorists, Stage Theorists deny that people, organisms, artifacts, and other ordinary objects really persist through time at all. Instead, each of necessity exists for only an instant. For this reason, Stage Theorists must produce a semantic account of our ordinary assertions of persistence that enables most of them to come out as true. This is where a temporalized version of Counterpart Theory comes into play. To say that entity x will have some property P at future time t2 is to say that there is a counterpart of x that exists at t2 and that has property P. In other words, ordinary objects “persist” by having temporal counterparts at other times. In addition, Stage Theorists can make use of a family of counterpart relations, each corresponding to a different sortal.
Sider argues (Sider 2001: 194) that Stage Theorists can be neutral on the question of whether there are brute facts involving such counterpart relations or whether all such counterpart relations must be reducible to (or at least supervenient on) other, purely intrinsic and qualitative or spatiotemporal relations. (This is an issue that divides Classical Perdurantists 24.1T.1T.1A.1T from Classical Genidentity Theorists 24.1T.1T.1A.1A.) That is true, but a primitive counterpart relation, or perhaps a family of primitive counterpart relations, each bearing some further primitive relation to some sortal, would seem to be a highly mysterious and obscure sort of thing. In fact, it's clear that Sider believes that counterpart relations are reducible (via some analytic meaning postulates) to some logical complex of other relations.
A difficult problem for Stage Theory is that of cross-time counting. For example: fewer than a trillion people have lived in North America. Strictly speaking, this is false if Stage Theory is true, since there have been infinitely many instantaneous person-stages living in North America (or, more precisely, in past continent-counterparts of North America). At this point, Sider recommends a hybrid theory, according to which the word ‘person’ sometimes picks out person-stages and sometimes person-worms (Sider 2001: 197).
How do Stage Theorists deal with the various paradoxes? In the case of Statue and Lump, they simply make use of two different counterpart relations—one for statues (or artifacts in general) and another for lumps of clay (masses of stuff). While they coincide, Statue and Lump are fully identical. However, use of the two names ‘Statue’ and ‘Lump’ will trigger (as a pragmatic matter) the use of different counterpart relations, making it possible to say that Lump will, and Statue will not, survive the smashing. The paradox ofTib and Tibbles) can be dissolved in a similar manner. There is one counterpart relation for cats (organisms) and another for sums of atoms or other cat-parts.
In the case of the fission and fusion paradoxes, Sider argues (2001: 202–203) that Stage Theory is superior to Worm Theory. Worm Theory fails to satisfy Fission-Product Interest below, given the assumption that Interest Requires Genidentity (a thesis suggested by the work of Derek Parfit 1985):
Fission-Product Interest. For any person x each of x's future fission products matters to x.
Interest Requires Genidentity. Person x matters to person y if and only if x and y are genidentical.
Suppose Alpha will undergo fission, and that two fission-products, Beta and Gamma, will result. On Worm Theory, there are, before fission, two distinct, overlapping persons corresponding prospectively to Beta and Gamma. Beta matters only to the first, Gamma only to the second. On Stage Theory, on the other hand, at each time pre-fission, there is only one person counterpart-related to both Beta and Gamma.
In the case of fusion, we have a similar difference. For Worm Theory, there are two distinct, coinciding persons post-fusion, whereas on Stage Theory there is only person, counterpart-related to the two prior branches. Sider sees this as another advantage of Stage Theory, remarking that it would be just to punish the post-fusion person for crimes committed by either pre-fusion branch. Not everyone will share this intuition, reflecting the lack of grip that the fusion story has on our considered judgments.
Stage Theory has no difficulty with the longevity or Methuselah paradox (see Section 24.2.2.2), since nothing prevents our using a non-transitive counterpart relation. In fact, such non-transitive cases are tailor-made for the use of temporal counterparts.
The paradoxes of vague and conventional identity similarly pose no problem for Stage Theory, since there are plenty of eligible candidates for the interpretation of names and singular terms, corresponding either to non-functional counterpart relations or to alternative eligible counterpart relations. Temporal Counterpart Theory offers a unified account of temporal and modal phenomena, since it is readily combinable with cross-world, modal counterpart relations. The modal version of the Statue/Lump case, in which Statue and Lump coincide throughout their actual life-spans, is thus no problem.
In fact, the paradoxes and puzzles dissolve so readily under Stage Theory that one has to suspect that Stage Theory is simply too good to be true, as Dean Zimmerman has suggested (reported in Sider 2001: 206–207). The natur
e and number of relevant counterpart relations is so unconstrained that it is hard to imagine any combination of intuitions about individual cases that Stage Theory couldn't easily accommodate. A very similar phenomenon is evident in the case of modal Counterpart Theory as developed in Lewis (1986a). We would put Zimmerman's complaint this way: Counterpart Theory is so flexible that, not only does it permit us to treat our ontological and modal intuitions as reliable, it actually renders them infallible in principle. This kind of epistemological overkill is exactly analogous to the mistake of the logical positivists and Phenomenalists (13.2T), who so tailored their metaphysical theories to our experience as to render all assertions of protocol (observation) sentences true by definition. (In Chapter 5, we voiced similar objections to the Ramsey/Lewis Theory of laws of nature, which renders our non-empirical criteria for theory choice infallible.) A complaint like Zimmerman's relies on the following principle of meta-metaphysics:
Reasonable Reliability. Any metaphysical theory that entails that we are either more or less reliable than we have a right to suppose is to be rejected.
Thus, metaphysical theories that entail unreasonable forms of skepticism or of infalliblism are equally objectionable. Let's illustrate some of the unpalatable consequences of the flexibility of Counterpart Theory. Imagine some metaphysically sophisticated defenders of slavery in the antebellum South, who regularly insist that slavery is just because all slaves are natural slaves. They defend the latter thesis by claiming that all slaves are essentially slaves. Moreover, they claim that emancipation is tantamount to murder, since no slave could survive such a process. It is child's play to show that what the slavery apologists say is true, given Counterpart Theory. All we have to do is to make use of a Counterpart Theory that provides each actual slave with no free counterparts, and each present slave with no future, free counterparts. Similarly, it would be easy to defend a position that rejects any anti-poverty measures by simply asserting that poor people are essentially poor and that no poor person could survive a significant increase in income. Sider could certainly reply that the use of such counterpart relations is wicked and reprehensible, but he could not take seriously even the possibility that their users were guilty of any intellectual error.
Finally, we should bear in mind that both Worm Theory and Stage Theory inherit all of the problems with Replacementism discussed in Chapter 24, especially the problem of making sense of the intrinsicality of motion.
25.3 Conclusion
In this chapter, we considered two questions about the relation between composition and persistence, namely those of constancy and coincidence. We canvassed important puzzles and paradoxes for those who believe that mereological inconstancy and mereological coincidence are possible. Three views emerged as front-runners, though none are absolutely unscathed. First, there is van Inwagen's Near-Nihilism, which countenances only material simples and organisms. Second is an Aristotelian view nearby Near-Nihilism that countenances homogeneous blobs as well as simples and organisms. Third, there are Worm and Stage Theory, which are Temporally Plenitudinous Replacementist views most plausibly coupled to Concretism and Counterpart Theory. In the final three chapters, beginning with the next, we will turn to a final set of issues, those concerning the relations of causation. There we may find evidence that will tilt our views in one direction or another.
Note
1. One might object that human beings are not organically symmetrical. The left hemisphere of the brain is significantly different from the right hemisphere, for example. However, let's suppose that Alpha is a member of some humanoid species on another planet that is perfectly symmetrical or that Alpha is an abnormal, perfectly symmetrical human being. Either of these seems to be metaphysically possible.
Part VIII
Causation
26
The Existence and Scope of Causation
The nature of causation has been one of the central questions of metaphysics since ancient times. For example, Plato (in The Laws, Book X) uses an argument about the cause of motion to prove the existence of God and the soul. Even earlier, Pre-Socratic philosophers like Thales, Anaximander, and Empedocles speculated about the cause of the origin of the universe and of natural phenomena, like the eclipse. We start in Section 26.1 by asking why we should believe that there are any causes at all. What does causation do for us? How does it help us make sense of the world? Then, in Section 26.2, we consider the issue of the scope of causation: how much of the world is caused? Is there a principled reason why some things are caused and others are not? In the following chapter, Chapter 27, we will be ready to take on the nature of causation. Is causation fundamentally a logical relation between truths or propositions or is it a real relation between things? If it is a relation between things, what sort of things—facts or events? Can absences and other negative things be causes? Finally, in Chapter 28 we will look at the relationship between causation and time, distinguishing between discrete causation (as an instantaneous relation between two things) and continuous causation (involving an infinite number of events in a temporally extended process).
26.1 Are there Causes?
We begin with our intuitive, commonsensical ideas of causation and ask: why think that anything causes anything? Should we opt for Causal Realism or Anti-Realism:
26.1T Causal Realism. Some things are caused.
26.1A Causal Anti-Realism. Nothing is caused.
26.1.1 Arguments against Causal Realism
Let's start by looking at the arguments for Causal Anti-Realism. Historically, there have been two main arguments for this view.
1 CAUSATION NOT NEEDED IN FUNDAMENTAL PHYSICS. CAUSATION IS SCIENTIFICALLY OBSOLETE. In 1913, Bertrand Russell published an article (Russell 1913) in which he argued that causation was an obsolete concept. Russell pointed out that modern physics, including Newtonian mechanics and Maxwellian electromagnetic theory as well as Einstein's special and general theories of relativity, describe the natural world by means of mathematical functions and equations. The words ‘cause’ and ‘effect’ do not appear in any of the fundamental laws of nature, nor do any of the other verbs that express causal relations, like ‘push’, ‘pull’, ‘repel’, ‘attract’, ‘fracture’, ‘merge’, and so on. The laws simply describe how certain quantities, like mass, energy, position, and velocity, change and do not change over time. To make this an argument for Causal Anti-Realism, one must assume that the best metaphysical theories should only appeal to the fundamental types of things in our best current physical theories. This is a non-trivial assumption.
2 CAUSATION REQUIRES NECESSARY CONNECTIONS BETWEEN SEPARATE EXISTENCES. The second principal objection to Causal Realism builds on the work of David Hume (Hume 1748). Hume observed that our idea of causation seems to involve some notion of a necessary or non-accidental connection between cause and effect. We think that if the cause is fully present and operational, and there are no obstacles or interferences, then in some sense the effect must follow. In addition, we think that if the cause had not occurred, the absence of the cause would have made the occurrence of the effect impossible in some sense.
At the same time, Hume argued that our ordinary conception of causation involves the separateness of the cause and effect. We don't think of things' causing themselves, either in whole or in part. Instead, one thing is thought to cause some other, completely separate thing. We think of the occurrence of a spark as causing a subsequent explosion, because the existence of the spark is wholly separate from the existence of the explosion. In contrast, we don't talk of John's loving Mary as the cause of Mary's being loved by John, since the two facts are inseparable, perhaps even identical.
These two facts about causation pose a problem. How can the existence of one thing necessitate the existence of another, separate thing? And how can the non-existence of something necessitate the non-existence of the other? Hume argued that there cannot be necessary connections of this sort. When two separate things are in view, we can always imagine o
ne of the things existing without the other, and vice versa. By Imagination as Guide to Possibility (PEpist 1), we should conclude that there are no necessary connections between separate existences.
As we have noted, though, causation seems to be a relation between separate existences. For example, Hume argued that we could always imagine the cause's occurring without the effect (the spark without the explosion) or the effect's occurring without the cause (the explosion without the spark). If causation is a relation between separate existences, then causation cannot be a relation of necessitation. The two components of our intuitive notion of causation—necessity and separateness—cannot both be maintained. Our notion of causation, then, is incoherent.
However, we have seen that imagination is a fallible guide to possibility. We can in certain cases imagine even logical impossibilities, like the non-identity of Mark Twain and Samuel Clemens. Thus, it isn't entirely clear how much stock we should put in our ability to imagine a cause's existing without its effect, and vice versa, and more generally in Hume's claim that there can be no necessary connections between distinct existences. The argument is interesting, no doubt, but may not be ultimately successful, especially given the centrality of causation to our thinking.
Hume had a further, closely related argument against the reality of causation. He argued that our experience provides us with no direct acquaintance with necessary connections like causal relations. When we reflect on our experience of apparent causation, Hume claimed, we find that all we experience is the succession of one event by another. We never see (hear, feel, etc.) the connection between the events. If we have no acquaintance with causation or the associated necessary connections between events, then we must lack even a concept of causation. Our idea of causation is merely a confusion of several distinct concepts, namely, the concepts of regular succession and contiguity among events, and the concept of a felt propensity on our own part to infer one thing from another. Hume argued, in effect, that our concept of causation cannot be defined in terms of other, more fundamental concepts, and we can have no empirical knowledge of instances of causation. By Conceptual Acquaintance (PEpist 5), we ought to deny that we can know the truth of any proposition involving that concept.