Metaphysical Foundations of Science

[Photo: Tuomas E. Tahko]

E.J. Loweinterviewed by Richard Marshall.

E.J. Lowe is a frost-cool deep fry who goes to the heavy core of the metaphysical lodestone and thinks about kinds of being all the time by building a system in the old style in order to get a grip on the very nature of reality itself. He thinks metaphysics is a slow business but we shouldn't be fooled into thinking slowness is stasis, doesn't think that common sense is riddled with confusions but there are some inconsistencies in it, thinks ontologies are expensive and in-car/out-car ones are too cheap, thinks there's a four category ontology, thinks Aristotle the king of the metaphysicians but prefers his own version of the ontological square, thinks hard about the nature of the laws of nature, thinks about universals and particulars, about powers and categories, can count tables but not red things, thinks empty sets can't be empty sets, thinks he has hands, thinks freewill can't be disproved by any empirical evidence, and thinks scientists should be more philosophical when entering important philosophical debates than they have tended to be recently. Which makes him hard-core.

3:AM:What made you become a philosopher? Were you always having these metaphysicalthoughts even when young? Or did they just emerge?

EJ Lowe:As a schoolboy, I was extremely interested in science and mathematics, but especially in astronomy and cosmology. I had my own telescope – a 6-inch Newtonian reflector – of which I was inordinately proud and fond, although light pollution in my home town was so bad as to make it pretty useless for observing the stars. Cosmology was in an exciting state of turbulence in those days, with the rivalry between the now accepted ‘Big Bang’ theory and Fred Hoyle’s elegant theory of the continuous creation of matter and energy, according to which the universe, even as it continuously expands, remains spatiotemporally homogeneous on the large scale. I might even go so far as to say that Hoyle was my scientific ‘hero’ at that time. I went up to Cambridge University in 1968 – where Hoyle was Professor of Astronomy – to read Natural Sciences, with the hope of eventually getting into cosmology. But I found the first-year course dull and tedious, especially the experimental side of it, and also began to realise that I probably didn’t possess the necessary mathematical aptitude to be really successful as a theoretical physicist. So, after one year, I changed to History, and took a BA in that subject two years later (1971). I’d always been interested in history anyway, and soon discovered that I could concentrate a good deal on the history of political thought, which especially attracted me. This introduced me to early modern political philosophy, including the works of Locke, which soon led me on to early modern metaphysics and epistemology. However, this wasn’t what first acquainted me with philosophy. My eldest brother, who is 11 years older than me, took a degree in physics at Oxford and then went on to do the BPhil in philosophy, eventually becoming a lecturer in that subject in London for a while, so even as a young teenager I had a pretty good idea of what philosophy is. As a teenager, I started to try to work out a metaphysical system of my own, partly drawing on my (admittedly rather limited) understanding of developments in modern physics, especially the idea of symmetry and equivalence principles of the sort that lie at the heart of Einstein’s theories of relativity – which I tried my best to understand.

My idea was that one might hope to frame some very general equivalence principle which enabled one to see that all the great metaphysical systems are ultimately just different ways of formulating the same basic facts about the ultimate structure of reality, one implication of this being that no empirical test could possibly choose between them. This, I thought, fitted in well with Karl Popper’s idea that what distinguishes science from metaphysics is that scientific theories are empirically falsifiable whereas metaphysical theories are not – although this doesn’t mean that metaphysical theories are meaningless or worthless, indeed, quite the contrary, since they provide indispensable ‘framework’ principles for scientific theorising. Actually, I still believe something like this to be the case. Anyway, in time I began to realise that my early interest in cosmology was really an interest in fundamental metaphysics and ontology, and this is what eventually took me away from history and political philosophy to pure philosophy. After completing my BA degree, I wanted to switch to do a PhD in philosophy, but Cambridge wouldn’t let me do that, so I departed to Oxford to do the BPhil in philosophy and subsequently a DPhil. At that time, I concentrated in metaphysics and the philosophy of science, having Rom Harré as my BPhil dissertation supervisor and Simon Blackburnas my DPhil supervisor, with a thesis entitled ‘Induction and Causal Inference’ (1975). In that thesis, I criticised Hume’s account of induction and causation (while recognising that he espoused what later came to be known as a ‘sceptical realist’ view of causation and causal powers) and developed a theory of ampliative reasoning which appeals to our knowledge of causal powers, linking the notion of such powers with a ‘dispositional’ account of natural laws according to which such laws primarily concern natural kinds and only derivatively their particular instances (a view which resembles the Armstrong-Dretske-Tooley view of laws as involving universals rather than regularities amongst particulars, but also differs importantly from that view with regard to the type of universals involved). These ideas eventually developed into the view that I defended in my first book, Kinds of BeingThe foregoing is, in as short as space as possible, a summary of my philosophical development and education.

3:AM:Armstrong, van Inwagen and Lewis (and Lewis in particular I guess) kind of give the general options for contemporary metaphysics. Are you cutting free from these and setting out a new option? You think Lewis’s approach tends to be engaged in an exciting battle between revisionary and merely descriptive metaphysics, which is a thrill but may not help us understand better the nature of the world don’t you? Is serious metaphysics dull?

EJL:I have always thought that metaphysics needs to be tackled systematically, rather than piecemeal. I liken the task to that of putting together the pieces of a gigantic jigsaw puzzle: it’s no use just trying to perfect many small but disconnected parts of the puzzle in the hope that these will eventually fit together, since it’s likely there are several different ways in which any small number of pieces will fit together, no more than one of which will be correct. Rather, you need to work simultaneously on the ‘big picture’ and on its many parts.

I don’t altogether accept P. F. Strawson’s distinction between ‘revisionary’ and ‘descriptive’ metaphysics, much though I admire his book Individuals, which was in fact one of the first serious books in metaphysics that I read, long before I became a student of philosophy. I do, however, follow Aristotle and Locke in taking it that ‘common sense’ is a reasonable – indeed, the only reasonable – starting point for philosophical thinking. But it is only a starting point, and some common sense ideas will inevitably have to be abandoned by the philosopher, since some of them, when pursued to their logical conclusions, give rise to puzzles and paradoxes, which it is the task of the philosopher to try to resolve – an idea, of course, that goes back at least as far as Socrates. The revisionary/descriptive distinction is best seen as marking the poles of a spectrum of positions, with the most sensible and defensible positions lying somewhere in the middle. Common sense cannot intelligibly be abandoned completely, but neither can it be defended from every charge of incoherence. The task of the philosopher is to strike the right balance between its rejection and its revision. Russell once said that common sense leads to science, and science shows common sense to be false. But that is too stark a judgement. Certainly, many common sense notions – for instance, those of so-called ‘folk physics’ – are shown to be false by modern science. But a physics which is so disconnected with common sense as to be nothing more than an abstract mathematical formalism can at best be of only instrumental value: it cannot help us to understand the fundamental nature of reality, which is the aspiration of metaphysics. In pursuing its task, metaphysics must take notice of developments in theoretical science, but should not be in servitude to them. It will need to deploy distinctive formal methods and tools of its own, but these too should not simply be carried over slavishly from logic and mathematics – for instance, in the shape of formal logic, set theory, and classical mereology – as though metaphysics were some kind of applied logic or mathematics. In my view, metaphysics, with ontology at its heart, is an autonomous and fundamental mode of inquiry, beholden neither to the empirical sciences nor to the a priori sciences of logic and mathematics. It really is, as Aristotle said, ‘first philosophy’, and as such an implicit pre-requisite for any more specific form of intellectual inquiry whatever. In that sense, I am not a ‘naturalistic’ metaphysician. But my kind of metaphysics is far from being ‘dull’, I would venture to say. It seeks to articulate a coherent system of ontological categories and a consistent account of the fundamental formal relations obtaining between entities belonging to these categories, in terms of which we may hope to understand the fundamental structure of reality as a whole. That is just about the most ambitious intellectual task that anyone could hope to undertake. And because it is so difficult, we should not be surprised that progress in it is slow – much slower than in theoretical physics or mathematics, for instance. We should not mistake its slowness for complete stasis. Genuine progress can be and has been made in metaphysics.  

3:AM:You say that common sense ontology is riddled with confusions – are these errors? If so, how come we manage to survive? Survival suggests either that they can’t be all that confused, or that common sense ontology isn’t that important. Both alternatives seem to be threats to the whole business of metaphysical examination of ontologies, especially revisionary metaphysics. What’s your push-back?

EJL:I’ve already implicitly answered this question. I don’t think that common sense ontology is ‘riddled’ with confusions, only that it harbours some inconsistencies which emerge in the form of various puzzles and paradoxes. These inconsistencies only manifest themselves when common sense notions are pushed to their limits, and that’s why common sense thinking serves us well enough for everyday purposes, or is ‘adaptive’, to use the jargon of evolutionary psychology. One task of philosophy in general and of metaphysics in particular is to tease out these hidden inconsistencies and consider how best to deal with them – either by replacing common sense notions with significantly different ones, or by revising common sense notions in certain ways. It is this process that eventually leads to the development of a comprehensive system of ontology, such as the one that I currently favour, the ‘four-category ontology’, as I call it, which has its historical roots in Aristotle’s early work, the Categories. A good example of the sort of puzzle or paradox that I have in mind is the ancient problem of the Ship of Theseus, which forces us to rethink certain aspects of our common sense understanding of the identity of material objects over time (their diachronic identity). Another equally ancient one is the problem of Dion and Theon (Theon being Dion except for, or ‘minus’, one of his feet, so that when Dion loses that foot he coincides with Theon, even though, it seems, Dion and Theon must remain numerically distinct). This problem forces us to rethink our understanding of the identity of material objects at a single time (their synchronic identity).

3:AM:I think we get a good handle on the way you come at metaphysics in your subtle argument against Hawthorne’s views about ‘parity’ and ‘plenitude lovers’ when discussing the in-car/out-car example of Eli Hirsch. You conclude that we shouldn’t be supplementing common sense ontology on the grounds of parity but instead ‘rather than either embrace and add to that ontology or simply reject it, we do better to reform or refine it.’ You worry that contemporary metaphysicians don’t handle common sense ontology carefully enough don’t you. You are kind of harsh: you say that if you’re a young metaphysician wanting to discover new kinds of objects you should retrain in physics! Could you explain why you disagree with the ‘plenitude lover’ and their notion of ‘parity’ – and just explain the Hirschexample which is a pretty cool and weird scenario?

EJL:The in-car/out-car example can be summarised this way. Consider your drive to work each morning. As we ordinarily think of it, you get into your car and drive it out of the garage. Here, a single object – your car – moves continuously across a borderline, marked by the door of your garage. But it seems that we might think of this situation differently, as follows. Inside your garage there is your ‘in-car’. As you drive out, this in-car shrinks until it eventually disappears, and at the same time another object, your ‘out-car’, grows outside your garage until it eventually reaches the same length that your in-car originally had. Here there are two adjacent objects, neither of which moves, but one grows as the other shrinks. (If you drive further than the door of your garage, the same process is repeated with a subsequent series of non-moving but alternately growing and shrinking objects.) The first way of thinking is the common sense way, but can we really charge the other way with being any less satisfactory as a way of thinking of the situation? Should we say that the two ways are just two of many ‘equivalent’ ways of thinking of the situation, no one of which should be privileged as being the ‘true’ way? My answers to these questions are ‘Yes’ and ‘No’ respectively. I think common sense is correct in supposing there to be objects which retain their identity through processes of movement across space – at least, if this is wrong, then it is only wrong for reasons to do with issues in fundamental physics. I don’t think that there are any such things as in-cars and out-cars. Just thinking about the example in isolation may not reveal any problem with the in-car/out-car way of describing the situation. But we have to remember that what we are looking for, as metaphysicians, is a comprehensive system of ontology, not just a piecemeal treatment of particular cases. To be consistent in the example under consideration, we have, for example, to apply the in-car/out-car mode of description also to you, the driver. We shall have to speak of ‘you’ not in terms of your being a single person who moves from being inside to being outside your garage, but in terms of a shrinking ‘in-you’ and a growing ‘out-you’, which are numerically distinct objects. So, suppose you start having a thought as you go through the door of your garage (as we would ordinarily describe it). On the in-you/out-you model, we’ll have to say that in-you begins the thought and out-you completes it. So one and the same thought must be attributed to two distinct subjects, and this appears to make no sense.

As I’ve just suggested, there might turn out to be reasons based in fundamental physics for thinking that the common sense notion of objects moving through space is problematic – for instance, that we do better to think of ‘movement’ as really consisting in variations of mass/energy density across tracts of spacetime. An analogy would be with the way in which we now understand that sea-waves don’t literally move in from the sea towards the beach, but consist in the regular rising and falling of the seawater’s height above the seabed, giving the illusion of real movement. But there’s no reason to abandon the common sense notion of moving objects, or to regard it as merely ‘conventional’, on purely philosophical grounds. It is deeply entrenched in common sense ontology, which is largely very successful in this regard, and it’s very difficult to see how it could be rejected without rendering other aspects of that ontology – such as its inclusion of ourselves as objects which both move and think – incoherent. So it should only be rejected as a last resort. The in-car/out-car fantasy might sound exotic and intriguing, but this is a cheap way to try to install a radically new ontology. Fundamental physics, with its invocation of such strange entities as superstrings, which have a genuinely explanatory role to play in physical science, is the place to look for interesting new and non-commonsensical types of entity. Such entities can genuinely earn their keep, but not so entities such as in-cars and out-cars. In fact, the latter are really just parasitic upon common-sense ontology, as their very names suggest: the only way to introduce them into philosophical discourse is by way of redescribing a common-sense scenario, such as your morning drive to work, in a bizarre new fashion. Our understanding of the in-car/out-car description of the scenario would be impossible without this reliance upon our prior common sense description of it. That’s not the case with entities like superstrings, where even the ‘string’ metaphor is not essential to understanding their nature.

3:AM:You’ve written a couple of books about Locke. Do you find Locke’smetaphysics of more than historical interest? And is he a good example of the way you think contemporary metaphysicians should operate?

EJL:It was more by accident than by design that I wrote the first of these books on Locke. A former student of mine – Tim Crane, in fact – asked me if I’d write it for the new ‘Guidebook’ series that Routledge was starting, in the mid-1990s. I lectured on modern philosophy for many years, and maybe Tim enjoyed my lectures on Locke – I do hope so! Anyway, I was happy to write on Locke, both because he is, like Aristotle, a philosopher with a healthy regard for common sense, combined with a respect for developments in empirical and theoretical science, and because I felt, and still feel, that Locke is often traduced by his commentators, who are too often willing to attribute indefensible and even downright silly views to him. Moreover, I share Locke’s general opinion about many key matters in metaphysics, epistemology, and the philosophy of mind and action. For instance, like him, I defend volitionism in the philosophy of action. So, in that first book, Locke on Human Understanding, I tried to offer sympathetic but faithful interpretations of Locke’s views and defences of modern versions of some of them. But one deficiency of Locke’s overall approach, I feel, is that it is not sufficiently systematic – it is too piecemeal. In this respect, as well as others, I consider Aristotle to be the vastly superior philosopher, indeed as being unsurpassed.

3:AM:You have defended what you call a ‘four-category ontology’, something that has an ancestry that traces back to Aristotle. It’s often useful to know what a philosopher sees as rival positions that require seeing off. So before telling us about this ontology, what are the rival positions that you are facing up to and why do you find them inadequate?

EJL:I am first and foremost opposed to those metaphysicians who see no need at all to frame a system of ontological categories (categories of being). I describe them as espousing a ‘no category ontology’. Typically, these philosophers think that we can just talk indiscriminately of ‘entities’ or ‘things’, distinguishing between them merely in terms of different descriptive predicates whose application is determined purely empirically. This is the ontology of Lewis Carroll’s walrus and the carpenter, who spoke of many things – of ships and shoes and sealing wax, cabbages and kings. In logico-philosophical terms, an ‘entity’, for such a theorist, is just any possible value of a bound variable, in line with W. V. Quine’s famous dictum, ‘to be is to be the value of a (bound) variable’. Even Quine himself didn’t really espouse a no category ontology, but rather one containing two basic categories: spacetime regions and sets (although at one point he thought that the regions could be reduced to sets). Such an ontology clearly has very little in common with common-sense ontology and that, fundamentally, is why I am not at all sympathetic to it. I don’t believe, for instance, that it can find a satisfactory way to accommodate ourselves, as thinking subjects, amongst the furniture of reality.

Other fashionable modern systems of categorial ontology include D. M. Armstrong’s ontology of states of affairs, with particulars and universals as their ‘constituents’, and the pure ‘trope’ ontology, which is a one-category ontology of property-instances. With regard to states of affairs, I find difficulty in understanding their supposedly ‘non-mereological’ mode of composition (as, notoriously, did David Lewis). With regard to pure trope ontology, I find difficulty in understanding how tropes are supposed to be individuated – that is, what determines their identity conditions. I think much better of the two-category ontology of C. B. Martin and John Heil, which includes not only tropes (or ‘modes’) but also, as an irreducible type of entity, objects or ‘individual substances’, conceived as bearers of such tropes. Where I disagree with this system is with its rejection of universals of any type, which I, like Armstrong, regard as indispensable for a satisfactory account of natural laws.

3:AM:So you defend your ‘ontological square’. How does this work?

EJL:The term ‘ontological square’ is not one that I coined myself, but I am very happy to use it. The square is a diagram depicting the formal ontological relationships between the entities in my four fundamental ontological categories: the categories of object (or individual substance), (substantial) kind, attribute, and mode. The bottom left corner of the square is occupied by objects, the bottom right by modes, the top left by kinds, and the top right by attributes. Kinds and attributes are universals, whereas objects and modes are particulars. Accordingly, objects instantiate (are particular instances of) kinds and modes instantiate (are particular instances of) attributes, so each vertical side of the square consists of an upward-directed arrow denoting the formal ontological relation of instantiation. Next, attributes and modes are properties, whereas objects and kinds are property-bearers, i.e. are entities which are characterized by properties. Accordingly, each horizontal side of the square consists of a right-to-left-directed arrow denoting the formal ontological relation of characterisation. Finally, there is a diagonal arrow leading from the bottom left (object) corner to the top right (attribute) corner, denoting the formal ontological relation of exemplification.

Kinds characterisation Attributes

instantiationexemplification instantiation

Objectscharacterisation Modes

My version of the ontological square is based on a four-fold distinction that Aristotle introduces very early in the Categories, utilising the notions of ‘being in a subject’ and ‘being said of a subject’. According to Aristotle, individual substances (my ‘objects’) are neither in a subject nor said of a subject, substantial kinds (in his terms, species and genera) are said of but not in a subject, attributes (as I call them) are both in a subject and said of a subject, and modes (individual accidents or particularised properties) are in a subject but not said of a subject. Naturally, I prefer my version to Aristotle’s, but the difference between them is not enormous. Both his and my version map rather nicely on to syntactical features of sentences in everyday natural language, and this is part of what makes them align with common-sense thinking. For instance, there is a clear difference between saying something like ‘Rover is a dog’, in which we assign Rover, a particular animal, to a certain natural kind or species, and saying something like ‘Rover is brown’, in which we attribute a certain property or quality to Rover. Modern first-order predicate logic completely obliterates this distinction, representing both sentences as having the logical form ‘Fa’. My view is that our formal logic should perspicuously reflect our fundamental categorial ontology, so that modern first-order predicate logic is, in my view, deficient and in need of revision in this respect. Too many modern metaphysicians uncritically accept the formalism of modern first-order predicate logic and allow it to influence their thoughts about ontology. Ontology should drive formal logic, not the reverse, in my opinion.

3:AM:This is a system that you think handles laws of nature and the power-categorical distinction better than rivals. Starting with laws of nature, how does your theory conceive of laws of nature and what does it do better than Armstrong’s view, which I guess is the most powerful contemporary rival view?

EJL:My view about laws of nature can be explained fairly simply in terms of the ontological square. There is an arrow denoting characterisation going from the top right (attribute) corner of the square to the top left (kind) corner. In other words, attributes (which are one type of universal) are said to characterise kinds (which are another type of universal). Sentences expressing such facts have the form ‘Ks are F’ – for example, ‘Dogs are carnivorous’ and ‘Planets move in elliptical orbits’. Linguists call such sentences ‘generics’. ‘Dogs are carnivorous’ doesn’t mean the same as ‘All dogs are carnivorous’, which attributes the property of being carnivorous to each and every individual dog. The truth of ‘Dogs are carnivorous’ is consistent with the truth of ‘Rover is a dog and Rover is not carnivorous’, because Rover might be an abnormal dog. Armstrong, like me, thinks that natural laws involve universals rather than particulars, but he recognises no distinction between substantial universals (kinds) and attributes. Thus, he takes the basic form of a natural law, in the simplest sort of case, to be ‘N(F, G)’, where F and G are first-order attributes (that is, attributes of particulars) and N is a second-order relational universal of ‘necessitation’, holding between first-order attributes. So, for example, he would regard Kepler’s first law of planetary motion, which I earlier expressed in the form ‘Planets move in elliptical orbits’, as having the logical form ‘Being a planet necessitates moving in an elliptical orbit’. My way of expressing the law is clearly much more in tune with everyday natural language and avoids any appeal to a ‘second-order’ relational universal. Armstrong’s account is particularly vulnerable to what is known as ‘the inference problem’, posed by critics such as Bas van Fraassen. According to Armstrong, ‘N(F, G)’ entails ‘For all x, if Fx then Gx’ (where the variable ‘x’ ranges over particulars), but it’s not clear what licenses this inference, which doesn’t appear to be formally valid. My account doesn’t have this problem because, according to it, ‘Planets move in elliptical orbits’ – for instance – doesn’t entail ‘Every planet is moving in an elliptical orbit’ (which is clearly not strictly true, given the gravitational interference between the planets and other disturbing factors) but only ‘Every planet is disposed to move in an elliptical orbit’, in which only a disposition or tendency so to move is ascribed to all individual planets: and my theory of dispositions or powers explains why this is so and why the corresponding inference is formally valid. (The technical details are rather too complex to be described here, but can be found in the last few chapters of my book More Kinds of Being; I also say more about these matters in my newest book, Forms of Thought.)

3:AM:How does your account explain the link between universals and particulars?

EJL:I regard the ‘link’ between universals and particulars as being the formal ontological relation of instantiation (depicted by the vertical, upward-directed arrows in the ontological square). I contend that every particular instantiates (is an instance of) at least one universal (in other words, there are no ‘bare’ particulars). I take instantiation to be a fundamental and therefore unanalysable relation. (In that sense, I don’t think that the link can be ‘explained’, because it is so basic – but none the worse for that, in my view, since some elements in any system of ontology must be taken to be basic.) In fact, I define the distinction between universals and particulars in terms of this relation, as follows. A particular is an entity which has, and can have, no instances (other than itself, if we take instantiation to be a reflexive relation). A universal is an entity which has, or at least can have, instances (other than itself, again if we take instantiation to be a reflexive relation). (It is a matter for debate whether we should take instantiation to be a reflexive relation, though if we do we should obviously not take it to be an asymmetric relation. However, we could instead take it to be an anti-symmetric relation: a relation R is asymmetric if ‘Rxy’ entails ‘not Ryx’ but anti-symmetric if ‘Rxy & Ryx’ entails ‘x = y’. For many purposes it doesn’t matter which option we take.) Because I am an ‘immanent’ realist concerning universals, I consider that there cannot be uninstantiated universals – a view associated with Aristotle and also held, for example, by Armstrong. However, in order to accommodate, at least for the purposes of argument, the ‘transcendent’ realist view (associated with Plato) that there can be uninstantiated universals, I define a universal only as an entity which can have instances (other than itself). To this it may be objected by some philosophers that there may be universals which cannot be instantiated, such as being both round and square. To accommodate, if need be, such philosophers, I am prepared to modify my definition of a universal in the following way: a universal is an entity which either can have instances of its own or is wholly composed by entities which can have instances. Thus, if we want to allow that there is such a universal as being both round and square, my modified definition accommodates this view because this universal, if it exists, is wholly composed by being round and being square, and each of these can have instances – in other words, it is a universal because it is wholly composed by other universals. It is hard to see how there could be a simple (non-composite) universal which could not have instances, so this modified definition appears to be unassailable.

3:AM:The power (e.g. fragility) and categorical (e.g. squareness) distinction is something that we might have thought about in reading Locke I guess, but the contemporary discussion is pretty interesting too at the moment isn’t it, with people like Mumford, Bird, Ladyman, van Inwagen et al all having their theories. You think there is a distinction but not at the level of properties but at the level of predicates. Does this mean you think they’re just two different ways of talking about the same thing? And are you denying the existence of uninstantiated universals? And what’s the role of tropes in all this?

EJL:Many philosophers draw a distinction between ‘dispositional’ properties (‘powers’) and ‘categorical’ properties (‘qualities’) and, indeed, fragility and squareness would be taken by many to be paradigm examples of, respectively, a dispositional and a categorical property. I don’t like using the term ‘categorical’ in this context because, historically, a distinction has commonly been drawn between ‘hypothetical’ and ‘categorical’ statements, the former being conditional in form and the latter unconditional. This has the unfortunate effect of linking the ascription of dispositional properties to the assertion of conditionals – and, indeed, it is still commonly supposed that a statement such as ‘This vase is fragile’ entails (or is even analysable in terms of) some such conditional as ‘If this vase were struck, it would shatter’ (a supposition that was roundly attacked in an important paper by C. B. Martin). I prefer to draw a distinction between ‘dispositional’ and (what I call) ‘occurrent’ predication. (This mirrors Aristotle’s distinction between ‘potency’ and ‘act’, or ‘potentiality’ and ‘actuality’.) In natural language, one way in which this distinction is registered is in terms of (what grammarians call) the aspect of verbs (not to be confused with their tense). For example, the two present-tensed statements ‘This liquid dissolves salt’ and ‘This liquid is dissolving salt’ differ in respect of the ‘aspect’ of their verbs, and the first involves dispositional predication while the second involves (what I call) occurrent predication. This can be seen from the fact that the first statement, but not the second, is equivalent to ‘Salt is soluble in (or by) this liquid’, in which a dispositional adjective is used. In my view, any property (attribute) can be predicated either dispositionally or occurrently of an object, even a so-called ‘categorical’ property like squareness – although English grammar may partially obscure this fact because such properties are standardly expressed in English only by means of adjectives, not by means of verbs. Suppose, however, that we were to introduce into English the (intransitive) verb ‘to square’: then, for example, we could explicitly distinguish between, for example, ‘This piece of rubber squares’ (dispositional) and ‘This piece of rubber is squaring’ (occurrent), the former saying that the piece of rubber is disposed to take on a square shape and the latter that it is actually taking on a square shape. Clearly, this is a distinction that we should be ready to make, since objects often have a ‘natural’ shape which can be distorted under stress, as when we bend or stretch a ‘naturally’ square piece of rubber. The ‘natural’ shape is the one that is actually taken on by the object when it is not subjected to forces of stress. However, even though the distinction that I make between dispositional and occurrent predication is one at the level of grammar or logic, this doesn’t mean that I think that the distinction has no underlying ontological ground. Quite the contrary, in fact. As I see it, the ontological ground of the dispositional/occurrent distinction lies in the fact that exemplification (as I call it) has two different forms – recalling that exemplification, depicted by a diagonal arrow in the ontological square, is a relation between objects and attributes. I hold that an object can either exemplify an attribute in virtue of being characterized by a mode which instantiates that attribute (occurrent exemplification) or it can exemplify an attribute in virtue of instantiating a kind which is characterized by that attribute (dispositional exemplification). These two forms of exemplification correspond to the two different ‘routes’ around the ontological square from bottom left (object) corner and the top right (attribute) corner, one of these routes going via the top left (kind) corner and the other via the bottom right (mode) corner. More details and a logical formalism for expressing statements of these types can be found in my More Kinds of Beingand Forms of Thought.

3:AM:You wrote Kinds of Beingand then twenty years later More Kinds of Beingwhich uses the four-category ontology work to bolster the original position. Is that right? This is your considered position on what there is isn’t it? So can you first give us an overview of the position you take and why you thought developing your theory of ontology was necessary to improve your initial presentation?

EJL:Basically, when I wrote Kinds of BeingI was not yet persuaded that tropes – or what I prefer to call modes, and what were traditionally called individual accidents – need to be included in our ontology. Effectively, then, Kinds of Beingrecognised only three corners of the ontological square: those representing objects, kinds and attributes. At that time, I was prey to the common mistake of thinking that an ontology should not include both property universals (attributes) and tropes, and that these categories are in rivalry with each other as ways to conceive of properties. I then saw that, just as we have objects as particulars instantiating kinds, so we can also have – and need to have – modes as particulars instantiating attributes. Once this piece of the ‘jigsaw’ fell into place, the ontological square and its structuring relations of instantiation, characterisation and exemplification suddenly became blindingly obvious, and the four-category ontology could emerge in its full form. This was, for me, one of those ‘eureka’ moments that rarely happen to one in philosophical thinking, but which make philosophy so rewarding. Of course, I can’t claim any deep originality in any of this, since Aristotle had got there well over 2,000 years before, and other modern philosophers appreciated the importance of his insights concerning the ‘being in a subject’/’being said of a subject’ distinction long before I did. Where I think I have contributed something original to the notion of the ontological square is specifically with regard to the relation of exemplification, and in particular the idea that it comes in two different forms, dispositional and occurrent, corresponding to the two different ‘routes’ around the square.

3:AM:The ideasyou have are prolific and rich so we can only look at a few of them, but there are some things you argue that are striking. So for example, why don’t you think that we can count all the red things that there are, but you do think we can count how many tables (for example) there are?

EJL:In order to count things – say, the things in a certain room – we have to be able to distinguish each of the things that we are supposed to be counting from all of the others, not least in order to avoid double-counting some things. That we means that we need to grasp the identity conditions of the things that we are supposed to be counting, i.e. we need to grasp their criterion of identity. But the term ‘thing’, although it is grammatically a count noun (inasmuch as it has a plural form, ‘things’), doesn’t have any criterion of identity associated with it. Anything whatever, of any kind whatever, is a ‘thing’, in the broadest sense of that term. Mountains, mice, and motorcycles are all ‘things’, but they are things of very different kinds and have very different identity conditions, conveyed by very different criteria of identity. So, provided we are told what kinds of things we are supposed to be counting in a room – for instance, all the tables, or all the tables and all the mice, or all the tables and all the mice and all the motorcycles – we can intelligibly undertake the task. But if we are just told to count all the things, or even just all the red things, we don’t really know where to begin, how to proceed, or where to end. Suppose, for instance, that one of the red things in the room is a red table. We could, I suppose, begin with that. But should we also count each of its red legs? And should we also count, say, each one-inch long (and each half-inch long, and each quarter-inch long, and each ...) red cross-section of each of those legs? Should we count the red paint on the table’s surface (indeed, should we count its red surface)? Does the red paint even qualify as a red thing at all (or is it just some red stuff)? And if so, what about the parts of that red paint? Should we count each square-shaped red part and each triangular-shaped red part, and so on – bearing in mind that we can divide the painted surface into squares and triangles (and infinitely many other shapes) of infinitely many different sizes? It should become readily apparent, in the light of these considerations, that the request to ‘count all the red things in the room’ is not just difficult to comply with, but doesn’t really make any sense at all. I say much more about these matters in the opening chapters of More Kinds of Being. The deeper lesson, however, is that we can only individuate things as things of this or that kind: there is nothing that is merely a ‘thing’, without being a thing of some specific kind. And, of course, we may be – indeed, surely must be – in complete ignorance of vastly many kinds of things (even within a single room). So I don’t think my stance on the question just raised is really so striking after all, once it is thought through carefully.

3:AM:You also don’t think zero is a number do you – doesn’t this mean you’re going up against a whole bunch of philosophers of math to hold your theory together – and isn’t that risky? Why do you hold this thought?

EJL:What I’m really opposed to is the idea that the so-called ‘empty set’ exists. Many mathematicians and philosophers of mathematics do indeed identify zero with this supposed set. My problem with the empty set is that I really don’t understand what it could be. We are told that it is a ‘set’. But all other sets have members, which stand in the formal set-theoretical membership relation to the sets to which they belong. Indeed, intuitively, a set is just a ‘collection’ of some things. But how could you have a ‘collection’ which didn’t collect anything? The empty set, then, is supposed to be set, and in that respect just like any other set, and yet it is supposed to be unique in having no members (in the set-theoretical sense of ‘member’). But lots of things have no members (in this sense of ‘member’) – for instance, Napoleon and my left foot, since neither is a set. So what makes the empty set different from any of these other non-membered things and distinctively a set, even though it is completely unlike all other sets in the only way, it seems, that has anything to do with the nature of sets as ‘collections’ of a certain kind? Here’s an analogy: in another (non-set-theoretical) sense of ‘member’, clubs have members, but it would be thought a poor joke at best if someone said that they had founded ‘the empty club’, which was unique in having no members (not just no current members, but no members at any time, by the very rules of its constitution). So, since I don’t even understand what the empty set could be, I see no reason to believe in its existence. It has an even worse ontological standing than something like the golden mountain, which could at least exist, and whose distinctive nature we can understand. As for mathematics, it’s true enough that the zero sign, ‘0’, is invaluable for mathematical purposes, but that’s no reason to suppose that we have to take it as denoting some distinctive mathematical object. (Indeed, there’s a good case for saying that ‘0’ denotes nothing: but ‘denoting nothing’ surely just means ‘having no denotation’, not denoting a weird kind of thing, nothing: that again looks a poor joke at best.) The ontology of mathematics is really a matter for the philosophy of mathematics, not for mathematics itself. I’m not proposing to deprive mathematicians of their zero sign and debar them from using it for all the purposes that they do. Philosophers of mathematics may object to me, but then it is incumbent upon them to explain what they do understand the empty set to be, in a way which doesn’t beg the question against me by just presupposing that ‘the empty set’ is a meaningful expression.

3:AM: Parts and wholes is always interesting, to me anyway. We’ve had have Olsonthe philosopher with no hands. So what’s your take on this – do you have hands?

EJL:Yes, I think I have hands. (I’m inclined to echo G. E. Moore and prove this by demonstration: here is a hand, and here is another! No philosopher’s reason to doubt that I have hands could outweigh the conviction that this demonstration licenses me to have.) However, I very much doubt that I have a ‘hand complement’, that is, a part of me which consists of all of me except, or ‘minus’, one of my hands. The assumption that I do is a presupposition of puzzles of the Dion/Theon type mentioned earlier. There’s a big difference (not just in size) between my left hand and my ‘left hand complement’, and it is this: an exact but unattached duplicate of my hand would not qualify as a human being, but an exact but unattached duplicate of my left hand complement would qualify as a human being (a human being lacking its left hand). I don’t believe that a human being can have, as one of its proper parts, something that would on its own qualify as a human being. Philosophers like Eric Olson often deny that living beings have ‘undetached parts’, such as hands, in order to escape paradoxes like that of Dion and Theon. But this is overkill. That sort of paradox can be overcome by denying that living beings of a certain kind have undetached parts which, were they to be detached (i.e. unattached), would qualify as living beings of that same kind. This still allows me to say, in line with common sense and biological science, that I have a certain living cell as a proper part, because even if that living cell were to be removed from me and continue to live and so still be a living being, it would not be a living being of the same kind as me – it would not be a human being, just a living human cell.

3:AM:A really fascinating and important argument you make is about persons. You say that persons are not to be identified with living organisms? You think that ‘person’ is unanalysable, and so we can’t reduce it to something else. Is that right? Does that mean that biological science is irrelevant to understanding what a person is – which seems counter-intuitive doesn’t it?

EJL:I do indeed believe that a person cannot be identified with a living organism, so that I am not identical with the living organism that is my biological body, and this is because it seems clear to me that persons and living organisms have different diachronic identity conditions (i.e. different persistence conditions). Very plausibly, there are changes that I could survive but which my present biological body could not survive. For example, it seems plausible that I could in principle survive the gradual replacement of every part of my biological body (even the neurons in my brain) by some non-biological substitute. At the end of that process, my current biological body would no longer exist, but I would still exist, and that implies that I am not identical with my current biological body. This is not to say that I think that ‘person’ is unanalysable, though. I more or less agree with Locke’s definition of the term, and hold that a person is a self-aware, rational agent and subject of thought and experience. I don’t agree, however, with Locke’s proposed criterion of personal identity, which is usually construed as being a psychological, memory-based one. In fact, I contend that there is no non-trivial, non-circular criterion of personal identity, i.e. that personal identity is ‘simple’ or ‘primitive’ (quite unlike the identity of a living organism). So, it’s not the concept of a person that I take to be unanalysable, just personal identity. As for the relevance of biological science to our understanding of persons, I certainly accept that it has relevance where human persons are concerned, that is, persons with human biological bodies. (The mere fact that a human person is not identical with his or her biological body provides, it seems to me, no reason whatever to suppose that facts about that body are irrelevant to facts about that person.) But I also think that there could be persons with non-human and even non-biological bodies and, obviously, biological science would not be relevant in the case of the latter. None of this implies, incidentally, that I think that there could be persons lacking bodies of any kind whatever – disembodied persons. I have no settled opinion about that matter, and don’t really know what considerations might be able to settle the matter one way or the other. Incidentally, a thesis that I have argued for concerning persons is that they are ‘simple substances’, in sense of having no proper parts. (This isn’t my primary ground for refusing to identify human persons with their bodies, I should stress, since one premise in my argument for the simplicity of persons is precisely that they are not identical with their bodies.) So when I implied, in answer to the previous question, that my left hand is a part of me, I was talking loosely (by my own lights), since what I really ought to say is that my left hand is a part of the living organism that is my human body. However, it would have been confusing to put matters that way prior to answering the present question. Descartes is also often interpreted as regarding persons, or ‘selves’, as being simple substances, and more specifically as being simple immaterial substances, lacking all physical attributes such as shape, size, mass and even location. That is not my view at all. I believe that I am located wherever my body is located and have the same shape, size and mass that it has. I just consider that I am not identical with that body and do not literally possess, as proper parts of me, any of its proper parts. I call this view ‘Non-Cartesian Substance Dualism’. It may look crazy at first sight, but in fact I think that there are powerful arguments in favour of it. I hope that those arguments will at least be allowed a hearing. (I develop them further in a number of places, including my Subjects of Experience, Personal Agency, and More Kinds of Being.)

3:AM:As always, the work of metaphysicians such as yourself raises the issue of why we should heed you guys when we have science. Take your thoughts about natural laws, for example. Why can’t we just leave it to the scientists to decide what they are and how they work? “How can analytical metaphysicians know anything from their armchairs?” is the question people ask isn’t it? What’s your answer?

EJL:I’m happy to leave it to the scientists to tell us what natural laws there are, but not happy for them to tell us, purely in their capacity as scientists, what a natural law is. There’s a very big difference between these two questions. The first can be settled on largely empirical grounds, that is, on the basis of observation and experiment. But the second isn’t an empirical question at all. In fact, some sort of answer to the second question is presupposed by any answer to the first. Unless we have some idea of what a natural law is supposed to be, we can’t really set about trying to establish which natural laws actually obtain. Scientists will, then, inevitably have at least an implicit idea of what they take a natural law to be, but that’s no guarantee, of course, that this implicit idea is a sound one. And different scientists may well have different implicit ideas. That’s why we need to engage in some explicit metaphysical thought and reasoning to work out what would be a cogent conception of natural law – and this has been a matter of intense dispute amongst metaphysicians of science in recent years. None of this implies that scientists themselves are debarred from entering this debate. On the contrary, their contribution should be most welcome. But it comes at a price: they should at least acknowledge that what they are contributing to is a metaphysical and thus a philosophical debate, not some discussion in which they are qualified, purely in their capacity as scientists, to assume that their opinions on the matter are authoritative and constitute the final word. Philosophical debate should be open to anyone, but one can only take part in such a debate if one recognises, as every rational person should, that there is such a thing as a philosophical debate, which differs in important ways from purely factual debates. Unfortunately, this very simple and, on reflection, very obvious fact seems to elude a number of well-known scientists who, in the course of publishing best-selling works of popular science, have taken the opportunity to pour scorn on philosophy. They should follow the lead of their wiser and greater forebears, including Newton and Einstein, who were far from being unphilosophical in their thinking, and whose philosophical cast of mind contributed in a major way to the originality and importance of their theories. At the same time, however, metaphysicians should not presume to think that they can fruitfully pursue their inquiries in complete ignorance of developments in scientific theory. When functioning properly, science and metaphysics complement and invigorate each other, and both stagnate when they ignore or are hostile to one another. That’s one reason, I believe, why the 17th century was such a fruitful period for the development of both science and metaphysics. The greatest scientists of the period were also philosophers and the greatest philosophers were also scientists. Now it’s difficult for any one person to master both fields of thought, as Descartes, Newton, and Leibniz could. But we can still foster constructive dialogue between those fields of thought. And we should.

3:AM:Has imaginative fiction – stories or films – helped you come up with ideas?

EJL:Not I great deal, I have to confess, although I was an avid reader of science fiction when I was in my teens and twenties, and I do think that some of the free imagination, inventiveness, and open-mindedness of the best science fiction can inspire productive philosophical reflection, and that it did so in my own case once in a while. That’s not to say, however, that I place much confidence in the ‘method of thought experiments’ where fundamental questions of metaphysics are concerned. I suppose I might be accused of using precisely this method when, earlier, I invoked the possibility of gradually replacing biological parts for non-biological ones as a reason for denying the identity of a human person with his or her body. However, in that case I wasn’t merely relying on free imagination, since there seem to be more solid scientific and metaphysical grounds for regarding the replacement hypothesis as tenable, at least in principle. It’s certainly not at all obvious that consciousness and reasoning require a specifically biological substrate, and optimists in the field of artificial intelligence clearly assume that they don’t.

3:AM:Your views have been formulating for decades. It would be hard for anyone to spend so long thinking on issues and then realise that they’re wrong but I wondered if there are things that over time you’ve had to give up? How hard is it for philosophers to change their minds?

EJL:Lots of very good philosophers change their minds, sometimes fairly frequently: think of Russell and Putnam, for instance. I often change my mind, but in recent years only regarding parts (though sometimes quite major ones) of my overall system – a prominent example being my conversion, mentioned earlier, to an acceptance of tropes or modes. However, I am perhaps a little unusual in having an overall system, since system-building nowadays tends to regarded as a thing of the past in philosophy. Sometimes, I feel it would be refreshing and exhilarating to throw over my entire system and start anew, and indeed I would have no hesitation in doing so if I could see a really fundamental flaw in it. The real excitement in philosophy, for me at any rate, consists in having new thoughts and discovering new arguments, not in the building of a great edifice. I’ve just found that thinking systematically is most productive of new thoughts and new arguments, since there are endless opportunities for forging new connections between parts of a system.

3:AM:You’ve written extensively about personal agency: what is your position on mental causation and how do you square your libertarian view of agency with the existence of physical causal determination?

EJL:Well, first of all, I don’t think there is physical causal determination, at least in this world, however it might be in other possible worlds. Quantum mechanics seems to require us to acknowledge that there is an ineliminable degree of causal indeterminacy in the physical world. This, then, at least makes room for the possibility of libertarian free agency. (And it won’t do to object here that quantum indeterminacy can only manifest itself on the atomic scale, since effects at the atomic level can readily be amplified to make an impact at the macroscopic level, as happens whenever a Geiger-counter records the decay of a single radium atom.) Furthermore, I reject the common assumption that all causation is event causation – the causation of one event by one or more other events. In fact, I hold that all causation is fundamentally ‘substance’ causation – the causation of events by individual substances, that is, by objects possessing causal powers (causal dispositions). Objects cause events by exerting (or ‘exercising’ or ‘manifesting’) their causal powers. As for events, they are just changes in the properties and/or relations of objects. So, for example, some water causes some salt to dissolve, by exerting its power to dissolve salt. The effect here is the change in the salt’s condition, from being crystalline to being dispersed in the water – this is what it is for salt to undergo dissolution in water. Water has this power because its molecules possess a dipole moment which overcomes the electrostatic forces between the sodium and chlorine ions forming the salt’s cubic crystal lattice structure. The event-causalist would say something like this: the salt’s being immersed in the water (event C) caused the salt’s dissolving in the water (event E), and would very probably add that this relation between C and E obtains because events like C are always followed by events like E (the ‘Humean’ or ‘regularity’ account of laws). It is much more illuminating, in my view, to say that the water, by exerting the power it has in virtue of the dipole moment of its molecules, causes the disruption of the salt’s lattice structure, dispersing the salt’s constituent ions amongst the water molecules. This account reveals the casual mechanism at work in the process of dissolution, by identifying the causal power whose operation is involved in the production of the given effect. Once we see causation as involving the multiple agency of many powerful particulars, there is less temptation to suppose that the physical world is governed by some sort of Laplacean determinism, according to which the total state of the physical universe at one moment of time somehow fixes its total state at any subsequent moment. I regard human agency as just a special case of the agency that is ubiquitous in the world at large. In the human case, the key power involved is the will – which, like Locke, I regard as a ‘two-way’ power to choose or not to choose to perform some specific action, such as the action of raising one’s arm. That we have a power to choose does not, of course, guarantee that our actual choice will be successful on any given occasion, since opposing causal powers may frustrate that choice – as when one tries to raise one’s arm but finds that it is impeded by an obstacle. But, as I see it, one always has a capacity to make a free choice, even if one’s actual choices may be prevented from being successful on some occasions. (An unsuccessful choice is still a choice.) Physical determinists will say, of course, that our choices are always causally determined by prior events, such as the onsets of certain beliefs and desires of ours and ultimately by certain physical events in our brains (since that is what they take those onsets to be). They may even appeal to the notorious psychoneural experiments which are taken by some to be evidence for this view. However, the experiments in question and their proper interpretation are highly contestable matters, and pending much more solid evidence than this I see no reason to abandon my view. In fact, I suspect that the notion of the power of free choice is so deeply embedded in our conception of what it is to be a rational agent that it must be questionable whether any merely empirical evidence could constitute adequate grounds for denying that we have such power. For we need to be able to exercise our capacity for reason in judging the merits of any purported evidence for or against the existence of some phenomenon. So, if the question is whether certain empirical data constitute good evidence for the claim that we lack a genuine power of free choice, and yet the possession of such a power is partly constitutive of what it is to be able to assess evidence rationally, we would be involved in at least some kind of pragmatic contradiction in accepting the evidence as supporting that claim. In other words, a ‘transcendental argument’ of a quasi-Kantian kind seems to be applicable against any empirically-based claim that we lack a genuine power of free choice. Of course, none of this explains how, by exercising one’s power of free choice, one can cause some physical effect to occur, such as the rising of one’s arm – in the sense of revealing the ‘mechanism’ at work in such cases. However, it would appear that the revelation of a ‘mechanism’ is only possible where non-fundamental powers are concerned – that is, powers whose efficacy is explicable in terms of more basic powers (as the power of water to dissolve salt is explicable in terms of the power of water molecules to disrupt ionic bonds in a crystal lattice). The powers of fundamental physical particles presumably cannot be explained in such a way. So, if the will is a fundamental mental power, as it seems likely to be, we should not expect such an explanation to be forthcoming in its case either.

3:AM:And finally, for the metaphysicians here at 3:AM, are there five books (other than your own which they’ll all be dashing away to read straight after this) you could recommend that will take us further into your metaphysical world?

EJL:It’s very difficult to choose just five, but the following five would certainly appear near the top of any more extended list that I would choose: P. F. Strawson, Individuals; David Wiggins, Sameness and Substance; David Lewis, On the Plurality of Worlds; Peter van Inwagen, Material Beings; and D. M. Armstrong, A World of States of Affairs. These are all philosophers and books that I admire immensely, even though I disagree with many aspects of every one of those books.

Richard Marshallis still biding his time.

This site was built using