John Heil interviewed by Richard Marshall.

John Heil is a big beast of analytical metaphysics who thinks all the time about the ontological point of view, about objecting to ontological conclusions coming from linguistic premises, on ontology not being an analytical enterprise, on Charlie Martin's gift, on why the picture theory is wrong, on why not every truth has a truthmaker, on rejecting the Special Composition Question, on accepting a level-free picture of the universe, on not accepting tropes, on the difference between higher order and higher level properties, on substances, on qualities and powers, on physics and substances and on what is fundamental. Freak out in a spaceage daydream, man ... and now enter this competitionto win tickets to see him in action.

3:AM:What made you become a philosopher?

John Heil:Becoming a philosopher is not easy. First, there is the Ph. D.. As an undergraduate I majored in political science aiming for graduate school. In the fall of my senior year I took a course in political theory that was built around a textbook on the subject by an eminent political scientist. The book seemed to me shallow and uncomprehending. I decided that, if this is what political scientists regard as scholarship, I wanted no part of it.

I am sure this was unfair, but I was in the thrall of philosophy. Reflecting on my own case, I tell students contemplating graduate school that they should do philosophy only if they can’t imagine doing anything else. I emerged from graduate school woefully under-prepared and certainly not a philosopher. I eventually developed into a competent professional, but it is only recently that I have begun to think that I might actually be approaching philosopherhood. (In my experience, few philosophers are philosophers.) My models have included David Armstrong, Keith Campbell, Donald Davidson, E. J. Lowe, and C. B. Martin. While not putting myself at their level, I find I am engaged in the same enterprise, an endeavor continuous with the early moderns and medievals.

3:AM: You defend what you call an ontological point of viewfrom a host of positions that have not taken this seriously enough. Metaphysics as it has traditionally been done is ontologically serious, and Australian philosophers that you have worked with have also been serious about ontology. Is it primarily linguistic approaches to philosophical problems that you’re dismissing? And what’s so important for you about CB Martin’s approach to metaphysics?

JH:C. B. Martin – Charlie Martin – gave me the greatest gift you could give a philosopher: confidence in my own instincts. I had always had strong realist leanings, but I tended to follow the path trod by many junior academics, taking for granted that, when I had doubts about a prevailing doctrine, I should accommodate the doctrine. Charlie made me see that, despite philosophy’s reputation as a discipline that questions all things, philosophers are too often happy to defer to received doctrines with the result that the substantive nature of those doctrines remains hidden.

In my case the watershed was a 1985 APA paper, ‘Are We Brains in A Vat? Top Philosopher says “No”’ on which Charlie commented. I was willing to let Putnam (the Top Philosopher) have the premises he wanted (or most of them) but suggested that they failed to imply what Putnam thought they did. Charlie would have none of it. He was unwilling to concede anything to Putnam. I eventually came to see that he was right. I subsequently learned from J. J. C. (Jack) Smart that this was a common reaction to Charlie.

What I object to is the unthinking move from linguistic premises to ontological conclusions, from the assumption, for instance, that if you have an ‘ineliminable’ predicate that features in an explanation of some phenomenon of interest, the predicate must name a property shared by everything to which it applies. (A predicate is ineliminable if it cannot be analyzed, paraphrased, or translated into less vexed predicates.)

Philosophers speak of ‘the pain predicate’. When you look at creatures plausibly regarded as being in pain, you do not see a single physical property they all share (and in virtue of which it would be true to say that they are in pain). Instead of thinking that the predicate, ‘is in pain’, designates a family of similar properties, philosophers (including Putnam in one of his moods) conclude that the predicate must name a ‘higher-level’ property possessed by a creature by virtue of that creature’s ‘lower-level’ physical properties. You have many different kinds of physical property supporting a single nonphysical property. This is the kind of ‘non-reductive physicalism’ you have in functionalism.

Non-reductive physicalism has become a default view, a heavyweight champ that retains its status until decisively defeated. Non-reductive physicalism acquired the crown, however, not by merit, but by a kind of linguistic subterfuge. If you read early anti-reductionist tracts – for instance, Jerry Fodor’s‘Special Sciences (Or: The Disunity of Science as a Working Hypothesis)’ (Synthese, 1974) – you will see that the arguments concern predicates, categories, taxonomies. Fodor’s point, a correct one in my judgment, is that there is no prospect of replacing taxonomies in the special sciences with one drawn from physics. But from this no ontological conclusions follow – unless you assume that every ‘irreducible’ predicate names a property.

This language-driven way of thinking is not one that would have occurred to the ancients, the medievals, or the early moderns – or to my aforementioned philosophical models. It is an invention of the 20th century, one that has led to the emasculation of serious ontology.

3:AM:You say that ontology is not an analytical enterprise. What do you mean by that?

JH:You don’t arrive at ontological conclusions by analyzing concepts. David Lewis hoped to provide an illuminating account of the concept of causation by way of a counterfactual analysis. But Lewis was careful to distinguish questions about what causation was from questions about the concept of causation we deploy in everyday life. As a Humean, Lewis had no room for a robust ontological conception of causation, so he, quite rightly, felt compelled to account for our use of causal concepts.

(Lewis’s Humeanism, like that of his Harvard teacher, D. C. Williams, is the doctrine that the universe comprises a four dimensional spread of qualities. These, and their mereological sums – their endless combinations – provide truthmakers for all the truths that have truthmakers. Such a picture admits of no ‘necessary connections’, no causal nexus, no bringings about.)

But suppose you want to know what causation is and you find Humeanism philosophically dodgy. You will then want to say something about what it is about the universe that makes causal claims true when they are true. Appeals to counterfactuals do not advance matters because you would need to say what it is about the universe in which those counterfactuals are true. This would be so even if, as Lewis thinks, you could produce an analysis of the causal concept in terms of counterfactuals.

The idea that metaphysics should start with Humeanism and adjust strikes me as pernicious. I prefer an Aristotelianism according to which objects do what they do because they are what they are: objects are equipped with powers that, in concert with powers of other objects, lead them to behave as they do. Powers provide truthmakers for counterfactuals and, ultimately, for causal claims.

3:AM: You argue that the picture theory of representation is wrong and a culprit in the pervasive shallowness and anti-metaphysics bias you suggest is inherent in philosophy. What’s wrong with this theory, or cluster of different theories, and have you something better to put in its place?

JH:It seems implausible on the face of it that you could infer ontology from language. Language can take us to items in the universe, ways the universe is, but language cannot reveal the nature of those items, the deep story. You can refer to, or pick up, or slice a tomato, but your capacity to do these things reveals nothing of the real nature of the tomato. For that, you need science and, in the long run, fundamental physics.
The key idea here – traceable to Charlie Martin in the 1950’s – is that of truthmaking. Truths about the universe are made true by ways the universe is. If it is true that this is a tomato, the truthmaker for this truth is the tomato. This sounds trivial, as well it should. What is not trivial is what the tomato is, and this could be almost anything. The tomato could be a cloud of particles, a disturbance in a field, or something stranger still. Whatever it is, the tomato’s nature is not going to be revealed by analysis of the tomato concept.

Some philosophers are skeptical about truthmaking. One reason for this has been the absence of plausible accounts of the truthmaking relation. Thus truthmaking is often characterized as a necessitation relation: a truthmaker – this tomato, for instance – necessitates the truth of ‘this is a tomato’. So you have ways the universe is necessitating truths.

But truths require truth bearers, something to be true. And it seems crazy to think that these could be necessitated by ways the universe is. (I consider appeals to propositions, mysterious timeless entities with built-in meanings, to be an evasion.)

I reject such accounts. I see the truthmaking relation as an internal relation between a truth bearer, a representation (that this is a tomato, for instance), and some way the world is (a particular tomato way). The relation is internal in the sense that if you have the relata as they are, you have the relation. The relation is not something in addition to the relata. If you have the tomato and a representation that this is a tomato, you have the representation’s being true. But it takes two to tango: a truth bearer as well as a truthmaker.
Truthmaking, then, is not an asymmetrical ‘upward’ relation of necessitation, but an internal relation that obtains when you have the right relata.

3:AM:So can you sketch the basic ideas that go into taking an ontological point of view in philosophy?

JH:In the simplest terms, the ontological question is always going to be, what might the truthmakers be for some truth or class of truths. If you want to know what causation is, for instance, ask what the truthmakers might be for causal claims. If you like counterfactual and subjunctive conditionals, ask what the truthmakers might be. If there is a role for conceptual analysis here, it is a very preliminary one of clarification: making explicit application conditions for terms of interest.

A grasp, implicit or explicit, of the application conditions of a term used to describe some way the universe is provides us with instances of the term in question. The ontological question is then, what is the nature of these instances? And this is not something that could be revealed via further analysis. In understanding how to apply ‘tomato’, we put ourselves in a position to identify tomatoes, instances of the tomato concept. It remains to be seen what tomatoes are, what the deep story about tomatoes might be.

3:AM:What does your position commit us to thinking has an ontology, and what kind of things does it deny has an ontology? Do non-fundamentals like tables and people have an ontology, for example?

JH:I am skeptical of the fundamental/nonfundamental distinction as it is typically deployed. If tomatoes turn out to be clouds of particles, I do not think the clouds are any less real, any less fundamental than the particles that make them up. The clouds are perfectly real. What does not follow, however, is that, in addition to the particles duly arranged, you have the cloud.

Here I am rejecting what philosophers call the Special Composition Question, or rather rejecting the picture that the question seems to presuppose. If tomatoes are clouds of particles, you do not have to be able to say what makes a cloud of particles a tomato to be able to think about, identify, pick, slice, eat, or throw tomatoes. Indeed, if Fodor is right, if there is no prospect of recasting talk of tomatoes into talk of particles, no prospect of providing an informative answer to the Special Composition Question for tomatoes and their ilk.

Incidentally, if tomatoes are clouds of particles, it is at least highly misleading to describe them as tomato-shaped aggregates. A wax tomato is a tomato-shaped aggregate of particles, but it is not a tomato. A cloud of particles making up a tomato includes massive complex interactions among the particles that make it up, and the particles will exhibit distinctive causal histories.

3:AM:Can you explain why you say that some propositions are true but don’t have a truth maker, and others do? Perhaps you could sketch out what truth-makers actually are in your theory – are they substances?

JH:I do not follow David Armstrong in thinking that every truth has a truthmaker. I doubt, for instance that truths of logic and mathematics have truthmakers. I agree with Armstrong, however, that ascertaining the truthmakers for truths that have truthmakers is a broadly empirical enterprise. If it is true that this is a tomato, then the nature of the truthmaker is a matter, ultimately, of fundamental physics.

Although I think that ontology is a matter of negotiation between philosophy and fundamental physics, I find it hard to see how an ontology could dispense with substances and properties. Substance and property are correlative notions: a substance is a something that is various ways, and the ways are properties.

What the substances and properties are is another matter. The substances might be particles, or fields, or superstrings, or space itself. The properties would be particular ways these things are.
Note that, if the substances turned out to be fields or space itself, then what we regard as objects would be properties – or, as I prefer, modes – particular ways the fields are, particular ways space is, as Jonathan Bennett has put it, ‘thickenings’ in space.

Some philosophers favor a more austere ontology encompassing just substances or just properties. If substances are various ways, however – and how could they not be? – you have properties. And if you think of properties as ways, they must be ways something is – the red way or the spherical way, or the negatively charged way – and this gets you substances.

One prominent competitor to this view is trope theory. Tropes, an invention of the aforementioned D. C. Williams, are particularized properties. What you might think of as a substance is at bottom a ‘bundle’ of tropes. On such a conception, properties are literally parts of objects to which they belong.

So you might think of a tomato as made up of a red trope, a spherical trope, a juicy trope. The tomato is red by virtue of having a red part, spherical by virtue of having a spherical part, and juicy by virtue of having a juicy part. But a red part, a spherical part, and a juicy part would be red, spherical, juicy somethings. somethings that are the red way, the spherical way, the juicy way. And these somethings look like substances.

You could also ask whether the red trope has a shape. Could you have a something that is just red and nothing more? How could there be the tomato’s redness without its being some shape or other? But then you have something red and spherical, and something red and spherical is a substance. (I shall qualify this point presently.)
The worry that I am raising for tropes is one that bothered Descartes and many medievals when they considered ‘real accidents’ trope-like entities that survived the conversion, or transubstantiation, of the substances that bore them. In the Eucharist, bread and wine are converted into Christ’s body and blood, but the accidents of the bread and wine remain, thereby accounting for appearances. These accidents are the most obvious precursors to tropes. But, as many medieval philosophers noted, if the accidents are white, or round, or sweet, they look like substances: somethings that are white, or round, or sweet.

So, following Descartes, I deny that properties are tropes. So are properties universals? Why think that? The idea that properties must be universals is of relatively recent vintage. Traditional properties – modes and accidents – were fully particular ways substances were. I prefer ‘mode’ to accident. You could no more subtract the substance and leave the modes than you could subtract the cat and leave the grin.

3:AM:You don’t think there can be levels of reality do you, although we can have levels of organization, description and explanation. How does this work?

JH:Start with the Lewis–Williams conception of the universe as comprising a four-dimensional distribution of qualities. (Qualities of what? Maybe space itself.) You have the qualities, but you also have mereological sums of qualities. These sums are not something in addition to the qualities: if you have these qualities, you have these sums.

Now, although all sums are created equal, not all sums are equally salient or equally stable, not all answer to time-tested predicates deployed in the sciences and in everyday life. The sum of qualities answering to ‘this is a tomato’, is similar and dissimilar to myriad other sums. There are lots of tomatoes and lots of non-tomatoes.

If this is your picture, you have one reality, many different ways of truly describing this reality. Some of these ways are orthogonal to others. One reality, many taxonomies. The big mistake would be to move from the idea that there are different ways of describing, different ways of ‘taxonomizing’ reality, to the thought that there are different realities or different levels of reality.

Although I do not accept their Humean ontology, I do accept the Lewis–Williams ‘flat’, level-free picture of the universe.

3:AM:Does this mean that you are arguing that the world is ontologically reductive but not analytically? Is this why mind and materialism can’t be analyzed reductively but ontologically there’s just one ontology in play. Does this mean you reject Ladyman and Ross’s claim that there isn’t a fundamental ontology?

JH:I am not sure what ‘ontological reduction’ would be. How could you reduce one thing to another? I understand reduction as a relation among categories, or predicates, or taxonomies, or theories. Can you take a true description formulated in biological terms, and paraphrase it into a description formulated in terms of quarks and leptons? If that seems unlikely, then the reduction of biology to physics is not in the cards.

My conception of ontology differs from that of Ladyman and Ross, so I am unmoved by their rhetoric. Most philosophers nowadays accept Quine’s observation that there is no sharp boundary between philosophy and science. And, as Donald Davidson puts it, ‘where there are no fixed boundaries, only the timid never risk trespass’. Measured reflection on what we know about the universe suggests that wherever physics leads us we will find substances and properties.

The substances could well fail to be thing-like. What the substances are – particles? fields? space itself? – is a question for fundamental physics, as is the question, what are the properties. And these are ontological questions that constrtain any inquiry into the nature of the universe.

3:AM:How can you account for property similarities if you deny the ontology of higher-order properties?

JH:I like to distinguish higher-order properties from higher-level properties. A higher-order property would be a property of a property. Higher-level properties, in contrast, are meant to be properties possessed by some object by virtue of that object’s possession of some distinct lower-level ‘realizing’ property.
I have said enough already to make it clear that I regard higher-level properties as artifacts of linguisticized ontology. What of higher-order properties? And, in particular, might we need higher-order properties to account for similarities among properties?

No. The idea that, if properties are similar, this must be because they possess similar properties is confused. Properties are particular ways substances are. Substances are similar (or dissimilar) by virtue of possessing similar (dissimilar) properties. Two complex objects might be similar in some respects, dissimilar in others. The respects are properties. But properties are similar or dissimilar tout court, not by virtue of anything other than themselves.

There is nothing mysterious here, nothing ad hoc. Property similarity is an excellent example of an internal relation: if you have the properties you have their being similar (or not). God can make similar objects by making the objects and giving them similar properties. Once He has made the properties, however, God need do nothing more to make them similar (or not). If the red of this tomato is similar to the red of your shirt, all God needs to do to make the two properties similar is to make the properties.

Appeals to higher-order properties to account for property similarity are not only confused and unnecessary, but deeply unhelpful. If two properties are similar by virtue of respectively possessing similar higher-order properties, what makes these higher-order properties similar? Still higher-order properties?

Do we need higher-order properties at all? Ask yourself, what a higher-order property would be. If a property is a way its possessor is, then a higher-order property would be a way a property is. But a way a property is, is just the property itself.

Nonchalant acceptance of higher-order properties is another unfortunate product of linguisticized ontology. The redness of this tomato is a color, the redness is a property, so the tomato’s redness has two properties: the property of being a color, the property of being a property (a property possessed by every property). But in reality what we have is a single property, the tomato’s redness, that answers to many predicates. The tomato’s redness is the truthmaker for
(a) The tomato is red
(b) The tomato is colored
(c) The tomato has a property
Three truths, one truthmaker.

3:AM:You argue that substances are things without parts; substances are simple and that only substances bear properties. Does that mean that there are more than one substance? Can you spell out the ontology you argue is our universe?

JH:Yes, I follow a long tradition in regarding substances as simple, at least in the sense that substances lack parts that are themselves substances. A substance can have spatial parts – a right and left half, for instance – and, if there are such things, temporal parts, but not substantial parts.

Why think this? First, it is important to see that ‘substance’ is a term of art. To deny that a tomato is a substance, is not to deny that tomatoes exist or that tomatoes are, in a perfectly ordinary sense, objects.

Second, substances have traditionally been characterized as non-dependent entities. The best way to understand what this means is to consider examples of dependent entities. Socrates’s paleness depends on Socrates. A triangle made of matchsticks depends on the three matchsticks that make it up. Properties (Socrates’s paleness) and complex objects (triangles) are dependent. But on what? On substances. Properties are ways substances are; complex objects are particular kinds of arrangement of substances. If there are any dependent entities, then, there must be substances.

Third, I claim that only substances are bearers of properties. Strictly speaking, then, Socrates’s paleness and the redness of this tomato are not properties of Socrates and the tomato, respectively. It is true that Socrates is pale and that the tomato is red, but the truthmakers for these truths are not Socrates’s possession of a property, paleness, and the tomato’s possession of a property, redness. The tomato’s being red is (let us say) a matter of myriad particles being interactively arranged in a particular way. If you organize particles in this way, you get something correctly describable as red. What you do not have is, in addition to the particles duly organized, redness.

This gives us truthmakers for the application of ordinary predicates, truthmakers for ordinary property claims. But just as ordinary objects need not be substances, what we ordinarily call properties are not, in the strict sense, properties. The tomato’s redness is a property-by-courtesy, what Keith Campbell calls a quasi-property, what I have called an Episcopalian property.

Why think this? In the first place, you do not need properties of complexes as something in addition to the properties of the substances that make them up together with the interrelations of the substances with one another and with their surroundings. These provide all the truthmakers we need for ordinary property ascriptions.

In the second place, if you start with the idea that properties are ways substances are, it would appear that, for a complex to be some way, would be for its constituents to be interactively arranged in a particular way. I believe that this thought bears a direct relation to the initial thought that substances are non-dependent entities. The non-dependent status of substances and their status as property bearers require simplicity, the absence of substantial parts.

So we have a metaphysically strict notion of substance and property, and an ordinary conception of objects and their Episcopalian properties-by-courtesy. So long as this is understood, there is no harm treating ordinary objects as property bearers. Thus, I will continue to describe the tomato’s redness and sphericality as properties of the tomato, counting on you, the reader, to play along. You can speak with the vulgar, provided you do not then think with the vulgar.

3:AM:What are the powers of properties? Are all qualities powers?

JH:Substances, not properties, have powers. Objects do what they do because they are what they are (and their circumstances are what they are, a qualification I shall henceforth omit for ease of exposition). The tomato rolls, makes a circular concave impression in the carpet, looks spherical, feels spherical because it is spherical. The tomato’s sphericality is a quality, and it is by virtue of possessing this quality that the ball does or would do what it does. If the ball’s sphericality is a property of the ball, this property is a powerful quality.

Most of us are trained to start with Humeanism and adjust. Humeanism is the default. You need a good reason move from a default position. If you start with the idea that properties are qualities, and that objects do what they do because they are governed by laws of nature, the addition of powers will seem gratuitous. We have laws. Laws explain why objects behave as they do. Who needs powers?

So you begin with (i) a sharp distinction between powers and qualities, and (ii) the explanation of objects’ behavior external to the objects. These two ideas are closely connected.

Until the 17th century objects were thought to do what they did because they were as they were. This is the Aristotelian picture. God creates the objects and endows them with powers. God can intervene in the course of nature in either of two ways: by miraculously modifying the powers possessed by objects, or by directly manipulating them. With Descartes this picture alters dramatically. The source of motion is not to be found in material objects, but in mental substances, finite or infinite. At best, material objects can transmit motion via impact on other material objects. But only a mental substance can initiate motion. Laws of nature are expressions of principles in accord with which God acts.

On a view of this kind, you can see what it would be for a law to govern the behavior of objects. But now subtract God from the picture. What is left of the laws? You have equations and principles, but in what sense could these be said to govern objects’ behavior? What are candidate non-transcendent truthmakers for claims concerning laws of nature?

Descartes stripped the objects of powers and placed them in God. You might think that the natural move to make after removing God from the picture would be to return powers to the objects. But somehow philosophers persevered in the belief that objects are governed by external laws while evincing little curiosity in how this was supposed to work. (David Armstrong is an notable exception, although Armstrong’s laws are still extrinsic to what they govern.)

By returning powers to the objects, you ensure that the source of objects’ behavior lies in the objects themselves. Objects’ powers would then be truthmakers for laws, where laws are regarded, not as entities, but as what scientists seem to say they are: equations, principles, formulae. Take Newton’s laws of motion. These might be thought to capture the contribution mass makes to the behavior of objects that have mass. Because massive objects typically have various other properties, how objects actually behave is going to depend on many factors other than their mass.

Unlike many contemporary philosophers who take powers seriously, I want to locate objects’ powers in their qualities. Properties are powerful qualities. Although some philosophers regard such a conception as verging on incoherence, I regard it as both commonsensical and inevitable. I say more in response to the next question.

3:AM:Are laws of nature necessary in your theory – and how do Unger’s yellow and blue spheres help us here?

JH:In responding to the previous question, I have addressed the topic of laws. Now I will say more about the relation qualities bear to powers.

The 20th century gave birth to the idea that powers (or dispositions; I use the terms interchangeably) and qualities are mutually exclusive kinds of property. Qualities are ‘categorical’ properties, properties possessed here and now by objects, properties possessed unconditionally. Powers, in contrast, are non-categorical, purely if-then, purely conditional. Ryle insisted that dispositional predicates – is fragile, is soluble – held true of objects, not by virtue of any categorical feature of those objects. A fragile object is one that would shatter if struck, but there is nothing about a fragile object by virtue of which this is so. The counterfactual, ‘if the glass were struck, it would shatter’ is ‘barely true’ of the glass, the counterfactual lacks a truthmaker.

This strange Oxfordian idea was supplanted by the thesis that dispositional properties were ‘grounded in’ non-dispositional, presumably qualitative, features of objects. Dispositional properties, although perfectly real, perfectly here and now, were ‘higher-level’ properties possessed by objects by virtue of those objects’ possessing distinct ‘lower-level’ categorical properties.

Arguments for this thesis resemble functionalist arguments for ‘multiple realizability’. Many different kinds of object are fragile. If you look for a single categorical property possessed by all and only the fragile things (and by virtue of which it is true that they are fragile) you will find only a variety of distinct, and largely unrelated properties.

Need I point out that this is the 'Picture Theory' at work?
An alternative approach would be to suppose that ‘is fragile’ designates a large, possibly open-ended family of cases. A particular object is fragile because it is qualitatively a particular way. The fact that another object is fragile owing to its possession of a very different qualitative character is neither here nor there. Both would manifest their fragility in similar ways.

A tomato’s sphericality is a paradigmatic quality of the tomato. Owing to its sphericality the tomato would roll down an incline, would make a circular concave impression in the carpet, would look spherical, would sound spherical to a creature equipped with a capacity for echolocation. The tomato’s sphericality is a powerful quality. This idea would seem odd only to a philosopher with a theory.

Peter Unger is a philosopher who has appreciated the need for qualities (see his All the Power in the World). Unger decries what he calls the ‘scientificalist’ world view, according to which properties possessed by physical objects are exclusively powers and qualities are all non-physical. He imagines a variety of toy universes populated by objects possessing various qualities – shapes and colors are his favorites – with powers to affect one another. If you try to subtract the qualities, however, the picture collapses into incoherence.

Imagine a universepopulated by blue spheres and yellow spheres. Blue spheres repel one another and attract yellow spheres. Colors and shapes, admitted qualities, play an ineliminable role in individuating the powers. This simply reflects a more basic point, namely that a power’s identity depends on how it would manifest itself, what it is a power for. (I would add, how it would manifest itself with particular kinds of reciprocal power.)

At any rate, that is how I would put it. Unger central argument is that our ability to distinguish a universe of mutually attracting and repelling spheres from a universe comprising bubbles moving in a liquid plenum, requires an appeal to qualities, in this case colors and shapes. Such ideas are variants on arguments advanced by Berkeley, by Charlie Martin, by Keith Campbell, and by David Armstrong that are aimed at exposing the incoherence of a universe of ‘pure powers’. Powers are always powers for something qualitative, some qualitative change, for instance. Powers and qualities go hand in hand.

Despite his attraction to powers and to qualities, Unger’s approach bespeaks his Humean sympathies. We need both qualities and powers, he thinks, but their relation is purely contingent. In some imaginary universes blue spheres attract yellow spheres, in others blue spheres repel yellow spheres.

The idea that qualities and powers could vary independently is sustainable only so long as you start with a Humean picture and graft on powers. Think again of the tomato’s sphericality. This is a quality of the tomato, but it is also a power: by virtue of being spherical the tomato does or would do various things under various conditions. What things? The tomato would
(a) roll (rather than slide or tumble) down an incline
(b) make a circular (rather than square) concave impression in the carpet
(c) look spherical (rather than cubical)
(d) feel spherical (rather than cubical)
(e) pass smoothly through a circular hole of the same diameter
(f) …
Notice, first, that properties considered as powers contribute to the behavior of objects possessing them, but what an object does will depend on its full compliment of powers. A steel ball resting on a magnetic incline would not roll. Second, a power’s identity depends on how it would manifest itself in concert with particular kinds of reciprocal partner. This means that one power could manifest itself differently with different kinds of partner.
The second is particularly important here. Unger can describe pretend universes in which blue spheres attract yellow spheres and universes in which they do not. It is much harder to imagine a universe in which a spherical object tumbles down an incline, makes a square concave impression in the carpet, looks and feels cubical, passes smoothly through a circular hole the diameter of which matches the length of an edge.

If the identity of a power depends on what it is a power for with particular kinds of reciprocal partner the identity of which depends on their qualitative natures, attempts to divorce powers from qualities are bound to fail. Unger’s imaginary spheres could have any powers at all. They do Unger’s bidding. But there is no reason to think that reality is like this and every reason to think that it is not. (Reality determines not only what is actual, but what is possible as well.)

3:AM:Given that physics doesn’t theorise a fundamental substance neither does it discuss powers how does your fundamental ontology approach agree with physics? I guess I’m interested in how the metaphysics of the ontological point of view relates to science. Why does it survive science? What does it add? Or do you think that science is committed to finding the fundamental level of reality?

JH:Physicists might not speak of substances, might use the word, but this does not mean that physics does not concern substances and their properties. Fundamental physics has just not settled on what the substances are. Newton thought the substances were corpuscles, Descartes held that there was but one material substance, space itself. Perhaps the substances are fields, or superstrings, or something even stranger. From the fact that physics has given up the simple corpuscular model, it does not follow that it has given up on substances. What would it be to do that? I have no idea.

What of powers? Are these absent in fundamental physics? Again, it is hard to see how this could be possible. By virtue of possessing a particular charge an electron would repel other electrons, attract positrons, and interact in particular ways with protons. Space, we are told, has energy some of which spontaneously yields short-lived particles that subsequently revert to energy. (Notice, by the way, that this makes space substantival.)

Here is the cosmologist, John Barrow:
As one digs deeper to the roots of scientific theories one finds that there is a foundation of a sort that we call laws of Nature, which govern the behavior of the most elementary particles of Nature. The identities of these particles, the things that they are able to do, and the ways in which they can combine are like axioms whose consequences we can test against the facts of experience. To some extent we may find that it is very difficult to imagine how things could be otherwise because the properties become so closely bound up with the nature of the populations of identical elementary particles that they govern. (The Book of Nothing: Vacuums, Voids, and the Latest Ideas about the Origins of the Universe. )

Does physics, or fundamental physics, lead us to a fundamental level of reality? Again, I think talk of fundamentality here is slippery and best avoided. If ordinary objects are dynamic clouds on massively interacting particles, or disturbances in fields, or, as Descartes would have it, modes of space, they are no less real than the particles, or fields, or space itself. Complexes depend on their parts and properties depend on their bearers, but complexes and properties are fully real.

3:AM:And are there five books you could recommend for us here at 3:AM that would take us further into your philosophical world?

JH:Here are five titles, although I could just as well have included dozens of references to my medieval and early modern heroes.
Keith Campbell, Abstract Particulars(Oxford: Basil Blackwell, 1990).
Edward Grant, Much Ado About Nothing: Theories of Space and Vacuum from the Middle Ages to the Scientific Revolution(Cambridge: Cambridge University Press, 1981).
C. B. Martin, The Mind in Nature.(Oxford: Clarendon Press, 2008).
Galen Strawson, Real Materialism and Other Essays(Oxford: Clarendon Press, 2008).
D. C. Williams, Principles of Empirical Realism(Springfield, IL: Charles C. Thomas, 1966).

Richard Marshallis still biding his time.