Quasi-hylomorphism

Quasi-hylomorphism is presented as a kind of middle ground between two well-known approaches to the metaphysics of ordinary objects. On the one hand is the approach that takes classical extensional mereology (or something very like it, such as Judith Jarvis Thomson’s (Reference Thomson1983) cross-temporal calculus of individuals) to be the theory of the only kind of composition there is and that consequently identifies complex, ordinary objects – giraffes, tables, mountains – as mereological sums. What the parts of such objects are differs according to different varieties of this approach. Perhaps the best-known version, and one widely accepted, is associated with David Lewis (Reference Lewis1986). Lewis is a four-dimensionalist and takes the parts of complex, ordinary objects to include their temporal cross-sections at each instant they exist. This yields a view of ordinary objects that is quite bizarre. Since there is just a single, transitive conception of parthood, if B is a part of C and A is a part of B, then A is a part of C in exactly the same way that B is. So the idea that an object like a watch has an ontologically (rather than just pragmatically) significant decomposition into cogs and springs but not into random bits of cogs and springs (or their sums!) cannot be maintained. Nor do the sorts or kinds to which parts or wholes belong have any ontologically interesting role to play in understanding such objects. This is because such approaches take composition to be universal. Whether two things compose (are parts of) a third has nothing at all to do with the natures of the objects, and hence with the kinds to which the objects belong. So the approach as a whole ends up according no ontological significance to kind membership, which is merely an accidental feature of various quantities of matter.

On the other hand, contemporary hylomorphists in the vein of Kit Fine (Reference Fine1999) and Kathrin Koslicki (Reference Koslicki2008) present a view of ordinary, complex objects as involving both formal and material parts. The formal parts specify idiosyncratic conditions under which some things, specified in terms of the kinds to which they belong, compose (or are the material parts of) complex objects. Kinds thus take on a crucial role in understanding how some objects compose others; there are privileged ways in which complex objects can be decomposed into parts. These views, then, do well in just the areas where the classical mereological view, as understood by someone like Lewis, yields counter-intuitive results. But, predictably, hylomorphic views come with their own problems. Sattig is especially concerned with the apparently mysterious nature of the way in which formal parts, or structures, combine with material parts. Composition, in Lewis’s classical mereology, namely summation, applies universally, without any distinctions. It is thus relatively straightforward, Sattig claims, by contrast with the hylomorphist’s kind-sensitive composition.

Sattig positions quasi-hylomorphism as a view that combines the simplicity of the classical mereologist’s conception of composition with the richness of the hylomorphist’s account of the role of structure and kinds in understanding ordinary objects. It has the virtues of each of those approaches without their liabilities (though, as we shall see, it has those virtues in rather different ways). To pull this off, he employs a familiar philosophical move: if you can’t get what you want from an operation or relation (here the operation of composition), put what you need into the operanda or relata! Ordinary objects, according to Sattig, are compounds of what he calls material objects and K-paths.Footnote ¹ K-paths are the magic operanda that will import the resources of hylomorphism into a strictly classical mereological framework. But before we look at them, I will say a few brief words about compounding and material objects. Compounding is simply classical mereological summation, but restricted to the right kinds of operanda, material objects and K-paths that bear a certain relation to each other, to be specified shortly. As long as a material object and a K-path do bear the right relation to each other, their compound, in Sattig’s account, is simply their classical mereological sum. This is why q-hylomorphism has the virtues of Lewis’s austere conception of composition and does not have to truck with the allegedly more mysterious kinds of composition invoked by hylomorphists.

The expression ‘material object’ is used by Sattig to contrast with ‘ordinary object’ (evidently, since the latter is analysed partially in terms of the former). This usage is a little confusing since it is quite common to describe ordinary objects like tables and mountains as material objects. They are not material objects, on Sattig’s usage. Material objects are characterized by him in two ways. First, they are things that have material, spatial properties non-derivatively. Ordinary objects, by contrast, have those properties derivatively, in virtue of their material object parts. Secondly, material objects, as Sattig puts it figuratively but helpfully several times, go where their parts go and their parts go where they go. They are subject to the principles of classical mereology. Neither of these features unequivocally distinguishes them from ordinary objects. The first depends on the theory of q-hylomorphism and hence, while a valid distinction, is dialectically compromised. The second, mereological way of distinguishing them is difficult to assess, for in a sense (we shall see below what this ‘in a sense’ amounts to), the same can be said of ordinary objects. They too are wholly describable by classical mereology, go where their parts go, etc. Ultimately, the difference between material and ordinary objects comes to this: that the former have nothing like K-paths as parts whereas the latter do, and that it is owing to having a K-path as part that, in a different sense from the ‘in a sense’ above, ordinary objects are not wholly describable by classical mereology, do not go where their parts go, etc. In an effort to give his approach the broadest possible appeal, Sattig allows that material objects may be, as far as perspectival hylomorphism is concerned, three-dimensional endurants, four-dimensional space-time worms, or instantaneous stages related by counterpart relations to other stages. If one takes them to be four-dimensional space-time worms, then one could say that material objects are like what Lewis and others wrongly took ordinary objects to be.

Now, what about K-paths? These are the crucial conceptual novelty of the book, allowing the formulation of both q-hylomorphism and semantic perspectivalism. The K in ‘K-path’ is a place-holder for a kind term, and an object that has a given K-path as a part is an object of kind K. K-paths are sets of K-states that bear certain relations to each other. A K-state characterizes an object of kind K at a time. Each K-state is a large conjunctive fact about a material object, to the effect that it instantiates certain properties. Sattig takes a great deal of care in characterizing exactly what those properties are and, since I see no way to convey that short of reproducing several very dense pages of his book, I shall ask the reader to rest content with the very rough and reductive claim that these properties are intimately bound up with being a K, either through being intrinsic properties of Ks or through a sort of conceptual link to the kind. A K-path contains a succession of K-states that, intuitively, would track what the folk would take an object of kind K to be like over its lifetime. So for a bunch of K-states to constitute a K-path, those states much be such that later ones exhibit law-like dependence on earlier ones, successive ones are qualitatively highly similar (objects do not change too rapidly) and no larger collection of K-states that meets the conditions includes them (arbitrary temporal parts of a K are not, at least ordinarily, themselves Ks).

Where a material object is the subject of one of the states in a K-path, that K-path and material object have a compound. Ordinary objects are such compounds. (Not all such compounds are ordinary objects, however. Some K-paths may not correspond to any familiar kinds.Footnote ²) Thus, an ordinary object will, strictly, be decomposable into parts according to the principles of classical mereology and will not have any special features owing to the kind K of its K-path. But the K-path itself will tell a different story, in which many features over time are linked in terms of what it is to be a K and many parts of the material object in question will bear structural and functional relations to other parts. This story, in which kinds are ontologically relevant and objects have privileged decompositions into parts, the story which hylomorphism holds to be unconditionally true of ordinary objects, will, according to q-hylomorphism, be thrown across ordinary objects like a shadow cast by their K-paths. This is the sense in which q-hylomorphism has the virtues of hylomorphism in its account of ordinary objects.

As noted above, Sattig formulates his view so that it can incorporate a view of material objects as three-dimensional endurants, momentary stages, or four-dimensional perdurants. Combining a material object with a K-path to get an ordinary object will yield very different looking theories depending on which of these options respecting material objects we take. (And the different-looking theories will vary in how they handle particular problems.) If we take the four-dimensional view of material objects, such objects themselves may, while being mereologically rigid, have different parts at different times. Hence the mereological variation in the associated ordinary object can all be taken care of by the material object itself. In that case, the resulting theory will look something like this. An ordinary object, say a piano (I choose this example since the book is graced with a picture of an amazing wire sculpture of a piano by Fritz Panzer), will be the mereological sum of a single material object, that has different parts at different times, and a piano-path. The latter will consist, roughly, of piano-relevant facts about that material object at different times and the ways in which its material properties ground its piano-relevant properties. Since these facts all concern the same material object, it is obvious that there will be law-like connections between the facts about it at one time and the facts about it at another. Ordinary objects will be like Lewis’s conception of ordinary objects together with a set of facts about them and their properties. If material objects are seen as three-dimensional endurants, however, the picture that emerges looks very different. Since material objects are mereologically rigid and since, on the three-dimensional view they do not have different parts at different times, to capture what we think of as an ordinary object, a single piano, we will have to invoke a large number of distinct ordinary objects in Sattig’s sense. Suppose, as we would casually say, a single piano persists over a time during which it becomes a little worn (some of its molecules are lost). There is a material object m1 that contains those molecules and a distinct material object m2 that does not. Suppose now it is a fact that m1 is F and a fact that m2 is G and that these two facts are linked in a law-like way. Then, simplifying greatly, they will form components of a piano-path, PK. There is now one compound which is the sum of m1 and PK and a distinct compound that is the sum of m2 and PK. Each of these is what Sattig calls an ordinary object, but neither is the whole original piano as recognized by the folk. They are ordinary only by definition of the theory. How do these two distinct compounds relate to the single piano we thought we had? Is there any sense in which, for Sattig, we can truly say there is a single piano which persists through the loss of some molecules? For the answers to these questions, we must turn to semantic perspectivalism.

Semantic perspectivalism

Semantic perspectivalism is the view that what we say (and to the extent that thought is linguistic in nature, what we think) can be assessed from different perspectives. Specifically, Sattig postulates two different modes of predication, what he calls formal and material predication, and develops them for all the possibilities concerning the nature of material objects (i.e. as three-dimensional, four-dimensional, or instantaneous stages). (He actually postulates three modes of predication, but I leave my introduction of the third mode till later.) When a term referring to an ordinary object (as I shall say for short, when an ordinary object) is the subject of a formal predication, the resulting claim will be true (or false) in virtue of the object’s component K-path. When such an object is the subject of a material predication, the claim will be true (or false) in virtue of the object’s component material object. To illustrate, consider the case of the piano introduced in the previous paragraph, only now, let us consider it during the process of repair, in which one of its keys is replaced. So, take the sentence, uttered before the replacement:

‘P’ refers to an ordinary object that exists at the time of the utterance of CHANGE. That object is, as we have just seen, a compound of a material object, m1 and a piano-path, PK. If the predication in the sentence is material predication, it must be judged according to how things are with m1. It will be true if and only if m1 later has a new key as part. This might be true if we take the material object in question to be four-dimensional, so (since our goal here is to bring out the difference between formal and material predication) let us assume one of the other possibilities, namely, that m1 is either an enduring three-dimensional object that does not have parts relative to times, or an instantaneous stage-like thing, that also does not have parts relative to times because it exists at only one time. In either of those cases, if CHANGE expressed a material predication, it would be false. That CHANGE is false, however, conflicts with the folk’s view that a piano can indeed come to gain or lose parts. That is because, Sattig thinks, the folk are hearing CHANGE as a formal predication. On that reading, its truth must be assessed in terms of the piano-path PK that is a part of P. If PK contains the fact that some other material object than m1 instantiates certain properties (such as having a key as part that is not part of m1 while having many other parts that are also parts of m1), then the predication is true. Our second compound, of m2 and PK, fits the bill. The fact that m2 has a key as part that m1 does not is itself a component fact of PK. Hence, the formal predication expressed in CHANGE will be true. Effectively, the compound P has a ‘double life.’ It has one life in virtue of its material part and in that life, it does not come to have a new key as a part; it has another life in virtue of its formal part and in that life, it does (here I use formal predication) come to have a new key as part. (Expressed without formal predication, this amounts to: its piano-path PK includes the fact that some other material object has a new key as part.) The two lives are lives associated with the two modes of predication, formal and material, with which things can be said about the object. These modes of predication offer differing perspectives on the things that can be their subjects.

Putting together q-hylomorphism and semantic perspectivalism: an example

Perspectival hylomorphism is the conjunction of q-hylomorphism and semantic perspectivalism. This conjunction presents a strange and revisionary view of ordinary objects but nonetheless makes true many of the things we ordinarily say and think about ordinary object. Sattig uses this powerful combination to address a number of metaphysical puzzles concerning coincidence, modality, fission, determinism, indeterminacy and space-time. Generally speaking, his treatment of these puzzles takes the following form. When the folk think about whatever the issue is, they are conflicted and subscribe to apparently contradictory judgments. The apparent conflict is between particular judgments and general principles. Traditionally, philosophical attempts to resolve these conflicts take what Sattig calls an incompatibilist approach. They either require us to reject the particular judgments (or certain features of them), or they require us to reject the general principles. Contrary to these incompatibilist responses, Sattig offers, on the basis of perspectival hylomorphism, a compatibilist response. The particular judgments are taken as involving formal predication and are, when so taken, true. The general principles involve material predication and are also, when so taken, true. Since the claims that apparently conflicted turn out to exhibit different modes of predication, the apparent conflict is merely apparent.

To illustrate this strategy, I will consider just one example, but the reader should bear in mind that Sattig deals with many different areas in philosophy and so has a lot to say on particular issues that will be abstracted from here. The case I shall describe is a very simple and familiar kind of case from the metaphysics of ordinary objects concerning coincidence. (It is case (A) in Section 3.1 of Sattig’s book.) A child takes a piece of paper and makes a paper aeroplane out of it by folding it. On the one hand, we are likely to judge (according to Sattig) that the piece of paper continues to exist and that a distinct object, an aeroplane, comes to exist and is coincident with the paper. The objects are distinct because one existed at a time at which the other did not and hence, by Leibniz’s Law, the paper and the aeroplane do not have all their properties in common and so cannot be identical. On the other hand, there is the pull of a general principle, dubbed (AC):

According to (AC) there cannot be distinct objects that coincide even for a limited part of their existence. Two distinct things can never coincide. There isn’t room for two different things to occupy exactly the same space at the same time. Since the paper and aeroplane in our case do coincide at some time (namely, after the paper is folded), they cannot be distinct. So, here is our apparent conflict, between the intuition (based on Leibniz’s Law and the view that the aeroplane did not exist before the paper was folded) that the paper and aeroplane are distinct and the general principle (AC) that prohibits coincidence.

Incompatibilist solutions either deny the intuitive description of the case or deny (AC). Among the former, Michael Burke’s (Reference Burke1994) ‘dominant-kind’ view holds that when the aeroplane comes to exist, the piece of paper no longer exists – hence we do not have any time at which allegedly distinct objects coincide. Another incompatibilist solution that denies the intuition about the case might hold, alternatively, that the aeroplane is not an object distinct from the piece of paper. If the paper existed before it was folded, so did the aeroplane – it just wasn’t aeroplane-shaped at that point of its existence. Among those who deny (AC) are the myriad constitution theorists who accept the two distinct objects, paper and aeroplane, but deny there is a problem with their coinciding, so long as one (the paper) constitutes the other (the aeroplane). (Full disclosure: this is the view I advocate in my Reference Evnine2016.)

Sattig raises problems for all these options. (Indeed, his objections to attempts to invoke constitution play a very important role in the structure of the book since that is the relation that hylomorphists like Fine and Koslicki use to understand the relation between an ordinary object and its matter or, as Sattig would put it, its material object part.) But ultimately, the main point in favour of his own solution is that allows us to keep both our intuitive description of the case and our general principle (AC). Let t be a time after the paper has been folded. According to perspectival hylomorphism, at t, there are two distinct objects, the paper and the aeroplane. The paper is the mereological sum of a given material object, m and a piece-of-paper-path that includes the fact that m is currently shaped like an aeroplane, but also includes the fact that some material object m’ (possibly m itself, possibly distinct) was not aeroplane shaped at some earlier time t’. The aeroplane is the mereological sum of m and an aeroplane-path, which also includes the fact that m is aeroplane-shaped at t, but does not include the fact that some other object m’ was not aeroplane-shaped at t’. The intuitive description of the case, as involving the coincidence of two distinct objects, is to be understood in terms derived from the divergent paths that are parts of the respective objects. (AC), on the other hand, is to be understood in terms of material predication. Essentially, it tells us that two distinct ordinary objects cannot coincide at a time where that would require the coincidence of two distinct material objects. This principle is not violated by the case as we have it since it is the same material object that is part of both paper and aeroplane at the time they coincide.

I have given a very rough and truncated sketch of the way that the details of Sattig’s ingenious views are developed through a whole panoply of cases and with respect to a wide variety of philosophical issues. As far as I can tell, it all works. The details are all there, they do what they need to do, and the promised results are indeed delivered. But is the exercise worth it? Is the combination of revisionary metaphysics and novel semantics just an extravagant Rube Goldberg machine, clever, but not what one ultimately wants to use to boil an egg as long as there is a good saucepan around? Or is this a theory we should want to adopt? Is it preferable to the alternatives and good enough in its own right? I said above that if at least one of three conditions is satisfied, that will be reason seriously to consider adopting the theory. The three conditions were (1) that the metaphysical part of the theory is superior, in its metaphysical virtues, to alternatives; (2) that the semantic part of the theory has some independent motivation beyond getting the revisionary metaphysics to cohere with ordinary judgments; (3) that ordinary judgments seem to be sufficiently confused and that no better way of saving them is at hand. In the remainder of this essay, I will speak to each of these points.

Semantic issues

I begin with the semantic issues since clearing up an important point here will smooth the treatment of metaphysical issues to follow and allay a concern that many readers may already be experiencing. As we have seen, Sattig distinguishes between two modes of predication, the formal and the material. One very attractive feature of his whole project is to tie this distinction to work in psychology, associated with Elizabeth Spelke, about object representation. According to this research, infants begin by representing objects only in terms of spatiotemporal (i.e. material) properties that are independent of the ordinary sortals objects fall under. Only later do we come to individuate and represent objects as falling under those various ordinary sortal kinds like dog and table. Sattig conjectures, very plausibly, that the infant practice of representing objects in ways that abstract from the sortals to which the objects belong does not get replaced by sortal-sensitive identification but persists alongside the later addition. If that is right, then the semantic distinction Sattig appeals to between material and formal predication would be underwritten by independently attested psychological capacities (sortal-abstract object representation versus sortal-sensitive object representation) in a very satisfying way.

There is, perhaps, some question over how exactly the principles of sortal-abstract object representation, as detailed by Spelke and others, corresponds to the details of material predication as developed by Sattig.Footnote ³ But let us set aside this question and grant the details to Sattig. There is, still, a potential worry about this justification for the semantic complexity introduced by Sattig. The metaphysician must have the resources to say things like ‘ordinary objects are compounds of material objects and K-paths’ and to debate, as I shall in the section that follows, how good an account of ordinary objects this is and what some of its intended and unintended consequences may be. In fact, I have been making such claims repeatedly throughout this review and Sattig’s book makes such claims in almost every sentence! These claims are not to be understood as either formal or material predications. They would clearly come out as false if taken in either of those ways, rendering the theory self-defeating. Sattig, therefore, postulates not two modes of predication, but three: formal, material and absolute. Taking a judgement as expressing an absolute predication means, roughly, treating it in the (relatively speaking) straightforward way that all predication is taken if one does not subscribe to the complicated semantic revisionism introduced here. If taken as a formal predication, the sentence ‘S is P’ is true (or false) in virtue of the K-path that is part of S. If taken as a material predication, it is true (or false) in virtue of the material object that is part of S. If taken as an absolute predication, it is true (or false) in virtue of S.

The existence of absolute predication, alongside formal and material predication, threatens to upset the nice alignment between semantic perspectivalism and Spelke’s work on object representation. For Spelke, sortal-abstract representation is the infant’s take on the world. The infant, if she could speak, would, according to Sattig, use only material predication. To that is gradually added a different, sortal-specific take on the world, to express which, Sattig holds, we use formal predication. Absolute predication, and the ‘perspective’ on the world it expresses, is ‘off-limits’ to ordinary speakers. It is the province of the metaphysician, who now seems to stand to the adult as the adult stands to the infant! (Is there an analogous kind of object representation for it? Pure Kantian intuition, perhaps?) And in what sense does absolute predication offer a ‘perspective’ on reality? Is it not the means by which to express what that reality is really like? Formal and material predication now seem like vehicles of illusion – perspectives that creep into the simple syntax of absolute predication, passing themselves off as if they were really telling us about an object when, in fact, they are telling us only about a part of that object.

Despite, therefore, a promising and ingenious way of motivating the semantic innovations needed to bring q-hylomorphism into line with what we ordinarily say, I remain unpersuaded that the innovations are not ad hoc.

Metaphysical issues

I turn now to some questions about the metaphysics of q-hylomorphism that I find troubling. According to q-hylomorphism, ordinary objects are compounds of material objects and K-paths. Although the operation of compounding is distinct from the classical mereological operation of summing (the former being a restriction of the latter), still, a compound is a mereological sum of the things of which it is a compound. Thus, on q-hylomorphism, ordinary objects are mereological sums. This, it seems to me, is highly implausible. Mereological sums are mereologically rigid – they cannot change their parts over time and they could not have had different parts from those they have. But ordinary objects can change their parts and could have had different parts. True, given semantic perspectivalism, we can employ the formal mode of predication to yield sentences like ‘ordinary objects can change their parts, ’ etc., which are true. But those sentences do not represent the claim that I just asserted two sentences up (where I used the absolute mode of predication) and that claim, I believe, is true, and inconsistent with q-hylomorphism.

Setting this aside, a further consequence of taking ordinary objects as mereological sums of material objects and K-paths will be that ordinary objects, in their nature, are no different from random mereological sums of material objects and K-paths, that is, sums of such things that are not compounds of them because the material objects in question are not the subjects of any of the states in the K-paths in question. (For example, think of the mereological sum of the material object that is part of the chair I am sitting on and the K-path that is part of my dog.) Although ordinary objects will be distinctive, relative to other mereological sums of material objects and K-paths, in that their material object parts will be the subjects of some of the states in their K-path parts, this will not be an ontologically distinctive feature of them. It will be (in some sense) an accidental feature of the sums in question that leads them to be classified as ordinary objects (or indeed, as compounds at all).

Sattig is clear about both of these consequences of q-hylomorphism and so I am merely registering my dissent from something he affirms. Another problematic feature, in my opinion, is brought out less explicitly in Sattig’s discussion. I shall call a view about the nature of objects extensional if it makes the identity of an object entail the disposition of its matter (i.e. what its matter is) at all times at which it exists.Footnote ⁴ An obviously extensional view is the four-dimensionalism associated with David Lewis. But other views which depart radically from Lewis’s in many ways may also count as extensional. Judith Jarvis Thomson (Reference Thomson1983) vigorously rejects four-dimensionalism but adapts the classical mereology that usually accompanies four-dimensionalism into her cross-temporal calculus of individuals. According to Thomson, an object is a cross-temporal sum of its parts at different times so, again, the object’s identity fixes the disposition of its material parts at all times at which it exists. An even greater departure from Lewisian metaphysical territory is the view of Richard Grandy (Reference Grandy1975), who takes an ordinary object to be a function from times to quantities of matter. This view is equally extensionalist because Grandy takes a function to be a set of pairs, so the object (the function) fixes the disposition of its matter (the range of the function) for all times at which it exists (the domain of the function).Footnote ⁵ Extensional views are very implausible, in my opinion. In some sense, the disposition of its matter, gaining or losing parts, is adventitious to an object and it flows from the nature of that object what parts it has at any time. Ordinary objects have a ‘metabolism’ that draws in and expels material parts. In some cases, as with biological organisms, this is literally true with respect to some changes in material disposition (but not to all such changes, such as traumatic loss of a limb). In other, non-biological cases, and for some changes in biological objects, the expression ‘metabolism’ has only a figurative sense. The way in which an object, owing to the kind-dependent ‘metabolism’ it has, controls its parts at any time is quite different from the way in which, according to extensional views, the very identity of the object is bound up with its parts at different times. Perhaps one could put it like this: on the metabolic view, the way an object ‘selects’ its parts is dependent on what context it finds itself in (which sandwich it happens to eat, say). On the extensional view, because it is the object’s identity that fixes its parts at all times, that determination is quite independent of context.Footnote ⁶

It might seem as if, in his desire to accommodate both three-dimensional, four-dimensional and stage-theoretic conceptions of the material objects that form parts of ordinary objects that Sattig’s view is neutral with respect to extensionalism. If we took a four-dimensional view of these material objects then the material disposition of a compound through its entire career will be built into its identity. If we took a stage-theoretic or three-dimensional approach, then the material changes in compounds (as described with the resources of formal predication) would be adventitious to them. But of course, that is not how it works. For whatever the metaphysics of the material object parts of ordinary objects, their K-paths are described in such a way as to fix the material disposition of compounds (as described using formal predication) over their entire career. Hence, Sattig’s view is extensional in my sense, and at odds with the metabolic priority of the object to its matter.

In this connection, it is interesting to compare Sattig’s treatments of time and modality. I said above that the changes in an object’s material disposition should be treated as adventitious to it. I might have said contingent. And one might have responded by saying that the fact that an object’s material disposition at all times at which it exists are fixed by the identity of the object, as they are in Lewis’s four-dimensionalism or Grandy’s objects-as-functions view, is consistent with those differences in material disposition at different times being contingent. How this exchange might continue would depend on how one treated the relation between time and modality. One might push the temporal extensionalism into the modal dimension. So one might advocate five-dimensionalism and see objects as having modal parts as well as temporal parts; or one might take objects to be functions not from times to quantities of matter, but from world-time pairs to quantities of matter. In that case, although the material disposition of an object will be contingent, in the sense that it differs at different worlds, the objectionable aspect of the original extensionalism will not be mitigated. The identity of the object will still be such that it fixes its material disposition, not just at all times, now, but at all worlds too. On the other hand, one might be an extensionalist about time but give some other treatment of the modal properties of the object, a treatment which does not make the identity of the object automatically fix its material disposition in all possible worlds.

Sattig’s own treatment of modality turns out to be unexpectedly complex. One might have thought that he could have just taken the analogous approach to the first strategy described in the previous paragraph and extended the notion of a K-path. Assuming a three-dimensional or stage-theoretic view of material objects, let k be the K-path that is part of an object O and m its material object part. Formal temporal predications about O may be true in virtue of the properties of material objects other than m if the facts about those objects are themselves included in the K-path k. If the K-path now included facts about material objects in other possible worlds, formal de re modal predications about an object might be made true by the facts about material objects in other possible worlds. In that case, K-paths would be sets that consisted of facts in different worlds, so K-paths would be transworld entities. Sattig does not take this line and assumes that K-paths are world-bound. There is a problem that might be behind his rejection of the proposed extension, namely that the facts in a K-path have to be related in a law-like way; those law-like connections are what enable us to assemble facts about different material objects into a single coherent ‘history’ of an object as described in terms of formal predication. Since there are no law-like connections between facts in different worlds, we would have to find some other basis on which to include and exclude facts from a transworld K-path. But I think it is unlikely that this is Sattig’s reason for not developing a notion of transworld K-paths since his own approach to formal de re modal predications relies on a counterpart relation, which is the sort of thing that one might appeal to in order to theorize about transworld K-paths. Sattig’s solution is to invoke a counterpart relation among K-paths in different worlds, a relation similar to, but not quite the same as, Lewis’s counterpart relation, and to say that formal de re modal predications about an object are true just in case the object’s K-path has the right kind of counterpart in another world.

Alongside this account, the account of material de re modal predications works quite differently. Unlike K-paths, which are world-bound according to Sattig, material objects can exist in more than one world. Material de re modal predication of an object O, therefore, works without a counterpart relation by appealing to what is true of O’s material part itself in some other world(s). Sattig claims that both temporal and modal predications conform to the following principle: formal predications are based on qualitative resemblances, material predications are based on non-qualitative identities (of material objects across times and worlds). One can see the attraction of this, and Sattig’s point in claiming it, though it is slightly misleading in that the qualitative resemblances for formal de re modal predication do all the work, whereas, as we have seen, in the case of formal temporal predication, the notion of law-like connection between facts is also important.

Common sense

The third condition under which a package of revisionary metaphysical and semantic theories might be more than Rube Goldberg machine, I suggested above, is if our common sense thought and talk about ordinary objects appears to be in such bad shape internally that some kind of rescue operation must be mounted. As noted, many of the problems to which Sattig applies perspectival hylomorphism have the form of an apparent contradiction in common sense judgments between some general principle (e.g. two distinct objects cannot coincide) and intuitions about cases (e.g. the piece of paper and the aeroplane made out of it are distinct because they have different histories). Semantic perspectivalism, as we saw, offers a nice resolution of this kind of problem, showing that the appearance of contradiction is illusory. The general principles are true if taken as material predications while the intuitions about cases are true if taken as formal predications.

I would like to suggest that common sense is in some ways better off than Sattig allows and in some ways, worse. (Here I return to the theme of my opening paragraph.) Common sense is neither in need of the rescue operation Sattig mounts, nor is it rescuable in this way. My reason for thinking it is better off than he allows is that I am much less convinced that there is a distinct set of judgments that are mostly agreed upon by the common folk and that have the form of apparently conflicting general principles and particular intuitions. In other words, I don’t see common sense as having the shape or structure that Sattig finds in it. In the example we looked at above in some detail, the general principle involved in the apparent conflict was that distinct objects cannot coincide. But I don’t think there is a generally agreed upon principle of common sense to this effect at all. What there is, is a lot of disagreement among people all of whom, to a greater or lesser extent, are already philosophers in embryo. But of course, this is also why the rescue Sattig proposes won’t work. When opinions like this are argued about, it is because people have genuinely different takes on how things are. They don’t just want to assert a sentence expressing a general principle here and another expressing a particular judgement about a case there, hoping that somehow, they can be allowed to assert all those things without conflict. Rather, they are propounding inchoate (or sometimes, not so inchoate) metaphysics themselves. Often they are puzzled by their own contradictory inclinations. More often, they are puzzled by others’ inclinations to think differently from them! But there is real opposition and it is, to repeat, already on the way to being philosophy. That’s why, returning to a point I made earlier, I find it so jarring to have not just a distinction between formal and material predication, in which the folk are supposed to speak, but the addition of absolute predication in which the philosopher speaks, swooping in to soothe the waters of the confused folk, allowing them to utter the sentences they wanted to, but that have now been hollowed out and filled up with the philosopher’s meanings. The situation is a lot more confusing than Sattig’s view makes it seem, but it is also a lot less sharply delineated and hence the confusion is negotiable (and argued over) without needing to be tidied up. There is only absolute predication and it is in this common currency that we disagree, or sometimes agree, with each other.Footnote ⁷