The Atlas of Reality
Page 21
Hypotheticalism has much to be said for it, given the possibility of explaining laws in terms of conditionals. However, the problems of finks, masks and mimics suggest that we cannot reduce powers or dispositions to conditional truths. As we shall see in Chapter 27, there are also difficulties in reducing causation to conditional facts. Finally, Hypotheticalists seem to be committed to the reality of merely possible things, or to the existence of haecceities, either of which would have to be considered an ontological cost to the theory.
Nomism has a good chance of grounding the truth of most counterfactual conditionals, so long as we can add to it an account of temporal or causal priority. It does, however, face some serious epistemological challenges, since it counts the laws of nature as contingent and so must (along with Neo-Humeism) acknowledge the existence of an abundance of chaotic possibilities that mimic any apparently orderly set of observations. In addition, unlike Neo-Humeists, Nomists have no explanation of our preference for simple laws. Metaphysically, Nomism is burdened with a large number of brute necessities, linking laws with actual patterns of particular fact. In the case of functional laws (like Newton's laws of motion), Nomists are forced to posit an infinite number of laws, one for each n-tuple of values, many of which are never instantiated in the actual world.
Consequently, Powerism, the position that would seem to be closest to our common-sense view of the world, is in a relatively strong position, despite the burden of positing both natural necessities between separate things and natural intentionality. Of the three versions, Causal Structuralism, Dual-Aspect Theory, and Mixed Two-Category Theory, the possibility of symmetrical law-books provides strong evidence against the former.
Notes
1. In recent work, Nora Berenstain (Berenstain forthcoming) has argued that a structuralist should take into account more than just the causal role of properties. It is also important to look at the broader nomological role that properties play, including non-causal laws of nature (such as laws about the curvature of spacetime in general relativity, or the Pauli exclusion principle in quantum mechanics).
2. See also Schaffer (2005), who shows how each of the available anti-skeptical strategies can be adopted by the defender of sicceities (what he calls ‘quiddities’).
3. In addition, the impossibility of such divergences would deprive us of the plausible account of the semantics of counterfactual conditionals offered by David Lewis (1979b).
Part III
Universals and Particulars
7
Universals
7.1 Introduction
As has become clear in the foregoing, the world seems to be made up of both things and the ways those things are. There are things, and these things have a certain character. There are particular things, and there are the properties those things have and relations those things bear to one another. On the one hand are trees, clouds, people, fundamental particles and fields, and possibly regions or points of space. On the other hand are shapes, sizes, colors, charge, mass, and various mental states, both occurrent (being in pain) and habitual (understanding geometry). We can distinguish between particulars and properties (or attributes, features, ways), with particulars instantiating or exemplifying properties or we can speak of properties as predicated of or characterizing particulars.
However, there is substantial controversy about the nature of both particulars and properties. (The controversy about the nature of properties is sometimes called “the problem of universals” or “the problem of the one over many”.) Some philosophers, for example, think that the categories of particular and property are fundamental, that at least some of the things in both are in no way derived from or dependent on things in another category. These philosophers are Realists about both particulars and properties. Nominalists (and conceptualists), on the other hand, think of particulars as fundamental and of properties as non-fundamental, with the latter being derived from the former. This derivation has been cashed in a number of ways, and we will consider some of those ways below. Bundle Realists think the dependence runs exactly the other way: they agree with Realists against Nominalists that (at least some) properties are fundamental, but disagree with both Realists and Nominalists in thinking that particulars are derived from properties. Particulars, on this view, are simply “bundles” of properties. In this chapter and the next, we will explore this landscape in more detail. We will start this exploration by thinking about why someone might go in for Realism about properties.
But before doing that, we want to note two bits of near-universal agreement. First, very many properties, maybe even all of them, are shareable: more than one thing can exemplify one and the same property. For example, many particulars exemplify the property of being red, many exemplify the property of being a dog, and so on. Though there is controversy about the exact nature of this sharing, that there is sharing is so obvious it almost needn't be mentioned!
Second, there is some connection between something's satisfying a linguistic predicate and its exemplifying a property (or some properties). For example, one of us has a dog, a dachshund in fact, named ‘Elsie’. We can say a number of true things about Elsie: Elsie is a dog; Elsie is a dachshund; Elsie is sweet; and so on. Thus, Elsie satisfies the predicates ‘is a dog’, ‘is a dachshund’, and ‘is sweet’, among a great many others. More generally, something x satisfies a predicate P if and only if when one combines a name or specifying description for x with P in a grammatically appropriate way, one produces a true sentence.1 For example, ‘Elsie’ is a name for Elsie, and when one combines ‘Elsie’ with ‘is a dog’ in a grammatically appropriate way, one produces the sentence, ‘Elsie is a dog’, which is true. Similarly, if one combines the specifying description ‘the only pet of THP’ with ‘is a dachshund’ in a grammatically appropriate way, one produces the true sentence, ‘The only pet of THP is a dachshund.’ A simple-minded account of the truth of these sentences is just this: the property of being a dachshund is correlated somehow with the predicate ‘is a dachshund’, sort of like how Elsie is correlated somehow with the name ‘Elsie’, and thus the sentence ‘Elsie is a dachshund’ is true if and only if Elsie exemplifies the property of being a dachshund. More generally, the sentence formed by combining a name for x with the predicate P in a grammatically appropriate way is true if and only if x exemplifies the property correlated with P. (We'll consider a complication that arises from this general strategy in Section 7.2.3.)
7.1.1 What properties must explain
We have already hinted at the fundamental driving force behind the development of a theory of properties. The primary datum a theory of properties must explain is the simple, undeniable fact that many things are similar to one another in certain respects. There are many such respects of similarity, and things can be similar in some respects but not in others. Consider three things, the first of which is red and square, the second of which is red and circular, and the third of which is blue and circular. Let us call these ‘RedSquare’, ‘RedCircle’, and ‘BlueCircle’, respectively. RedSquare and RedCircle are exactly similar in color, which is to say they are both red. They are, however, dissimilar in shape, since RedSquare is square and RedCircle is circular. RedCircle and BlueCircle, on the other hand, are exactly similar in shape, since they are both circular. They are dissimilar in color, since RedCircle is red and BlueCircle is blue. RedSquare and BlueCircle, on the other hand, are dissimilar both in color and shape, which is to say that they do not share a color or a shape property.
What we see here is, first, similarity or resemblance in certain respects (e.g., color, shape, etc.) and, second, sharing of properties (e.g., redness, circularity, etc.). We might also say that, in typical cases, when there is a respect of similarity or resemblance between or among things, there will be a sharing of properties, and when there is dissimilarity or a lack of similarity or resemblance there will be a difference of properties. For it is plausible to think that RedSquare and RedCircle are similar in color because they are share a p
roperty, namely the property of being red (or redness). Likewise, RedCircle and BlueCircle are similar in shape because they share the property of being circular (or circularity). RedSquare and BlueCircle, on the other hand, are wholly dissimilar because they share no properties. These facts—facts of similarity or resemblance—are the primary bits of data a theory of properties must explain. We will focus our attention on how the theories of properties we examine explain this data.
We do not, however, want to give the impression that facts of similarity are the only bits of data a theory of properties must explain. For example, a theory of properties must also account for the relations among properties, so-called categorial relations (categorial in that they have to do with categories). For example, the property of being red and the property of being colored are related in that being red is a way of being colored. The property of being colored, we will say, is a determinable of the property of being red, and the latter is a determinate of the former. Many properties stand in these determinable-determinate relations: the property of being an animal is a determinable of the property of being a dog, which is in turn a determinable of the property of being a dachshund, which is in turn a determinable of the property of being a wire-haired dachshund. (We can turn the order around and talk instead in terms of determinates.) A different type of categorial relation obtains between other properties. Take the property of being red and the property of being blue; these are related in that nothing can have both at once (that is, nothing can be red and blue all over). We might say that being red precludes being blue. It's not difficult to identify a number of these relations of preclusion. The property of being a dog precludes that of being a cat, the property of being a circle precludes that of being a square, and so on. An interesting feature of these categorial relations is that they seem to be necessary. Given just the existence of the properties themselves, they simply must stand in these categorial relations. It just couldn't be that something could be red and blue all over at once. It couldn't be that there was something that is both a cat and a dog. And so on. It is important to consider how it is that various theories of properties fare explaining the necessity of these categorial relations (whatever one's view of the nature of the relations happens to be). Unfortunately, we don't here have space to consider these issues in more detail, but we suggest that readers take them up on their own.
Closely related to these categorial relations are predications of other properties and relations to properties. For instance, it seems clear that the property of being red is more similar to that of being orange than it is to that of being blue. Similarly, the property of having 1 kg of mass is closer to having 2 kg of mass than it is to having 10 kg of mass. These relations of relative similarity and difference among properties suggest that there must be properties to serve as the relata of these relations.
Finally, there are two other phenomena that properties have been employed to explain: the possibility of intentionality in both language and thought, and the distribution of causal powers and dispositions. Intentionality concerns the aboutness of language and thought: how it is, when we predicate some property of some entity, that our thought succeeds in being about that property. If I predicate wisdom of Socrates, what is the connection between my act of speech or my thought and Socrates' wisdom or possible wisdom (as opposed, say, to his snub-nosedness or his musical ability)? One popular theory, with roots in both Plato and Aristotle, would explain this fact by making a property (like the property of wisdom) an actual component of a thought. We'll take this idea up in Section 7.3.
We have already discussed the nature of causal powers in Chapters 4–6. We'll look at the implications of these chapters on our theory of properties in Section 7.4.
7.2 Realism
The first view, or cluster of views, about the nature of properties we will consider is Realism. Realism has two characteristic features. First, Realist views maintain that (at least some) properties ground the character of ordinary objects like rocks and tables and dogs and people. Second, Realist views maintain that properties are universals. Any view with both of these characteristic features is Realist.
7.1T Realism. Universals exist and ground the character of ordinary objects.
What, exactly, is it to think that properties ground the character of ordinary objects? Essentially, it is that ordinary objects have the character they do in virtue of the properties they exemplify; it is because ordinary objects exemplify properties that they have the features, the character, that they do. This is what we called extra-conceptual grounding in Chapter 3.4: a grounding relation between entities or facts in the world.
And what, exactly, is it to think that properties are universals? Essentially, it is that properties (of the shareable variety discussed above) are just as fundamental as the things that have them, in the sense that they are not derived from or dependent on those things. This second characteristic feature of Realism could use some clarification, for it is common to define universals as properties that can be exemplified by more than one thing. We find this characterization inadequate to fully account for the differences between believing in universals and not. To (briefly!) see that this is so, suppose that one identifies properties with classes. Two versions of Nominalism, Class and Resemblance Nominalism (see below and Chapter 8), make just this identification between properties and classes.
7.1A Nominalism. Universals do not ground the character of ordinary objects.
If one insists that universals are just properties that could be exemplified by more than one thing, Class and Resemblance Nominalism turn out to be committed to universals. Properties are classes, exemplification is class membership, and some property-classes will have more than one object in them. Thus, some properties are capable of being exemplified by more than one object, and that's just to say that there are universals. This would be fine except that Nominalism is often taken to be the view that there are no universals! Even if one rejects that particular definition of Nominalism (as we do), everyone agrees that universals can't be classes of objects. Defining universals as multiply exemplifiable properties leaves open the possibility that universals are just classes. Therefore, that definition must be rejected or supplemented. Second, starting with the thought that universals must be multiply exemplifiable forecloses a certain picture of haecceities (see Chapter 9), properties that only one possible object could exemplify. In particular, it precludes taking haecceities to be universals. We believe (though will not here defend) that one ought not foreclose this possibility at the outset. To repair these problems, we believe one ought to think of universals as fundamental character-grounders.
We thus arrive at the task of explaining this notion of fundamentality. The easiest way to appreciate what we mean by the talk of the fundamentality of properties is to contrast Realism with Class Nominalism. Class Nominalism, again, is the view that all properties—the things that are (sometimes) shared between and among things—are just classes of the things that have the property. The property of being a dog, on this view, is just the class of all dogs. The property of being red is just the class of all red things. The property of being identical to Elsie is just the class that includes only THP's dachshund, Elsie. And the property of being a color is just the class of all color properties (which would be a class of classes!). Class Nominalism, then, maintains that properties are less fundamental than the things that exemplify them, since sets are less fundamental than (at least partially grounded in) their members. There is a dependence of classes on their members that is not reciprocated. Realists deny that properties are less fundamental than other objects in this, or any other, way.
It is fairly clear that Realists accept the need to explain both character and similarity, and that they offer an explanation something like the one outlined above. The idea is that universals, in paradigm cases, ground the character of multiple objects, and in that way account for similarities among those objects. For example, the universal BLACK is exemplified by a number of things, inclu
ding THP's phone, his car key fob, and the keys on his computer's keyboard. These things, according to Realism, are similar because they exemplify one and the same universal. The universal BLACK that is exemplified by THP's phone is identical to the universal BLACK that is exemplified by the “Y” key on THP's keyboard. The similarity between these two things is to be explained by the fact that each exemplifies the very same property. Further, the reason why the exemplifying of the very same property makes for similarity, according to Realism, is that shareable properties, like the universal BLACK, ground character. THP's phone is black because it exemplifies the universal BLACK; likewise, the “Y” key on THP's keyboard is black because it exemplifies the universal BLACK. According to Realism, universals ground character, and thereby ground similarity. The picture the Realist paints appears promising, at least when it comes to the central phenomena properties are meant to explain.
7.2.1 Troubles for Realism
Regardless what variety of Realism one opts for, one must face the potential trouble spots for Realism in general. We are in the game of constructing a theory to explain a certain data set, and the process of choosing a theory is more than simply considering whether a view can explain the data. Many views can explain any set of data! We must consider a theory's strengths and its weaknesses in order to evaluate whether it best explains the relevant data. So we turn now to arguments that have been raised against Realism.