Interview by Richard Marshall.
Joshua Mozerskysteps out of the Tardis to discuss his philosophy of time, talking about ontology and semantics, about whether what exists is time dependent or not, about presentism and eternalism, about whether we can think and speak in an untensed way, about whether physics alone can answer all the questions, about Hilary Putnam's argument about time, about McTaggart, about 'A-series' and 'B-series'approaches to time, about the reality of time, about the difference Special Relativity makes, about objective becoming and the eternal NOW, about Bourne and presentism, about his own version of 'B-theory', about time's arrow, about why causation explain why we privilege the present and about why perdurantism fits his approach. It's 3:AM, it's time, read on...
3:AM:What made you become a philosopher?
Joshua Mozersky:In many ways, I have had a philosophical disposition since I was young. Even as a child, I would regularly lose myself in thought for long periods of time about, say, the nature of eternity, death, other worlds, what have you. In addition, I was raised in a family culture in which debate, questioning, and knowledge were highly prized. The idea that one might spend one’s life dedicated to asking questions and seeking answers was not only understood and tolerated, but encouraged. The life of the mind was considered an end in itself, and pretty much any intellectual pursuit was accorded due respect.
I saw no real or natural division between the kinds of question one might ask, or the kinds of understanding one might seek out but eventually I found myself particularly drawn to the kinds of questions asked by physicists and philosophers. Not having had any formal exposure to philosophy in high school, I decided to pursue a bachelor’s degree in astronomy and physics, subjects that were much more familiar to me. As an undergraduate, I did, however, enroll in as many philosophy electives as I could, and I encountered some truly outstanding professors who made the material come alive for me. What really stuck with me was how the great philosophers were able to take problems that appeared to be intractably vague, imprecise, or impenetrable and reframe and clarify them. It was no surprise to me that questions in physics and math were able be put in (relatively) precise form, but I was blown away at how such nebulous concerns as the existence of God, the nature of consciousness, or the difference between right and wrong, could be placed in a conceptual space that allowed for deepened understanding and movement toward an answer. I was hooked. After I completed my B.Sc., I spent a couple of years playing music, but then went back to school to pursue my M.A. and then Ph.D. in philosophy. I was fortunate enough to secure full-time academic employment and have been able to work on developing philosophical ideas ever since.
3:AM: You’re an expert in the philosophy of time. You say there are two great debates about time. One is ontological, and the other is semantic. Can you briefly sketch out the outlines of these two debates for us.
JM:In fact, I think there are numerous worthwhile debates about the nature of time, but there are two that I see as particularly important because solutions to them will impact, and even suggest solutions to, most of the others.
The first debate centers on the following question: is what exists time-dependent (is position in time existentially relevant)? To many thinkers—not just philosophers—it has seemed clear that what there is, in the most general, unrestricted sense of the term, constantly changes. The most prominent of these views is presentism, which is the theory that only what is present exists. For the presentist, the past, for example, is not just out of sight or out of mind, but out of reality, period. A related view is that there exists a concrete present and past, but the future is non-existent. Opposed to these sorts of view is what has come to be called eternalism. Eternalists argue that all times (and their contents) are real; all are equally part of the furniture of the universe. When something is in time makes no difference as to whether it should be included in our inventory of existing things. For the eternalist, Caesar, Napoleon, and Barrack Obama all exist, separated in time. This is similar to how we might insist that New York, London, and Paris all exist, but not at the same place. So that’s one debate, an ontological one.
I conceive of this first debate as a dispute over whether reality is temporally centered on the present. There are, broadly speaking, two ways of relating to time. First, in our everyday lives, the present looms as a unique moment that must be singled out in all thought and action: to act, is to act now; to think, is to think now; to experience, is to experience now; and so on. Secondly, we can step back and theorize about time as a whole, in which case the present becomes just one time among indefinitely many other, similar times. John Campbell (in his Past, Space and Self) draws a similar distinction with respect to space, and calls the latter perspective the ‘objective conception’. So, I see the first debate as one over whether the objective conception of time is the correct one.
We use the device of linguistic tense to mark the distinction between past, present, and future, and the second dispute concerns the logical structure of tensed language and thought. There is a train of argument in the philosophical literature that the truth of certain statements—for example, ‘it is now 12:00 p.m.’—logically commits us to objective, mind-independent, tense. In other words, such obvious truths—which can hardly be denied—entail a conception of reality that is itself divided along tense lines (not just divided by language). This is in tension with the objective conception, which does not see relation to the present as ontologically significant.
A related line of argument one encounters is that it is impossible for us to speak or think in an untensed mode; there simply does not exist an objective conception of time because we lack the linguistic and conceptual resources to frame one. If this were right, then there would be no way to defend the objective conception (though it may still be true), so even minor considerations against it would be pretty strong grounds for rejecting it. So, there are semantic disputes that have implications for the ontological one.
I defend the objective conception directly, on ontological grounds, but I also argue that the logical structure of tensed language is, in fact, tenseless. In other words, I argue that the objective conception is not only true, but that the resources of our language allow us to express it adequately. I think this is interesting in its own right. However, as I suggested earlier, I think this combination leads to satisfactory answers on a broader range of philosophical questions.
For example, I argue that the conception of time I defend leads to a powerful and satisfying understanding of change, temporal passage, and direction. In other words, there are non-centered accounts of time, change, passage, and direction that weave together into a seamless whole.
3:AM: Do you think that any theory that can’t explain why we experience it as tensed, as flowing like an arrow, is flawed simply because of that? Why wouldn’t it be ok to prove that time is tenseless and shrug and say the experience is just some psychological, evolved quirk – something for the cognitive scientists and not the responsibility of a metaphysical theorist of time?
JM:That’s a good question. Part, perhaps most, of the work on our experience of time will have to be done by cognitive scientists of various sorts. But there is an argument against the objective conception that turns on a conceptual or logical point. The argument is this: the very fact that we have the experience that we do—i.e. at least as if of a dynamic, changing, temporally directed world—is incompatible with the non-centered (i.e. tenseless, eternalist) viewpoint; i.e. the very existence of the experiences that we undeniably have falsifies the theory. This argument is run independently of any claims about the physiological, neurological, or psychological bases of such experience. Rather, it takes it as a given that we have experience with a certain character, and argues that the objective conception is incapable in principle of explaining that fact. So, that is one argument that I think the philosopher/metaphysiciancan and ought to tackle.
Furthermore, in the case of time and tense, as I note above, many philosophers have argued that the content of our language and thought commits us to a world that is tensed, and whether this entailment holds is, in my eyes, a philosophical question.
Having said that, I of course think that it is possible that in various cases the content of our experience simply conflicts with the way the world is in itself, in which case so much the worse for the content. That’s fine. However, in the case at hand, I happen to believe that time passes and that things in the world really do change: growth, decay, evolution, process, etc., are all genuine phenomena. So in my view, our understanding of time should be compatible with all of this being the case. Typically, however, those who defend the tenseless, eternalist conception of time, prefer to bite the bullet and deny that time passes or even that change is an illusion. I think this is a mistake, and lends unnecessary credence to the opposing, tensed views of time. If genuine change, temporal passage, and direction can be shown to be compatible with the non-centered conception, then we can in fact have the best of both worlds.
JM:As long ago as 1967, Hilary Putnam argued (in ‘Time and Physical Geometry’) that all the philosophical questions about time have been settled, not by philosophy but by physics(in particular Einstein’s Special Theory of Relativity). Putnam argues that the frame-dependence of simultaneity—the fact that different observers will project different ‘nows’ across the universe—demonstrates that any attempt to tie existence to what is present is refuted. The reason is that no observer can reasonably claim to stand in a privileged position or reference frame; so if everybody carves up the universe with a different temporal knife, then we have to either accept that different observers live in distinct realities, or else concede that what exists is not time-dependent after all (each reference frame merely offers a different perspective on the same, unchanging whole). Putnam thinks that the former option is absurd, so the latter is the only plausible position.
Of course, had Putnam imagined that his argument would put an end to philosophers debating the nature of time (which I assume he didn’t), then he would have been mistaken; the debate over the nature of time has, if anything, heated up since then. In the nearly fifty years since Putnam published that article, there has grown a cottage industry of philosophers defending or attacking his argument (not the only time Putnam has made such an impact, of course). For close to 50 years, some of the best philosophers of physics, and even physicists, have debated each other over whether Special Relativity is compatible with an ‘open’ or indeterminate future, whether there exists an objective direction to time, or even whether time is real. So, as concerns the nature of time, I think that there is plenty of evidence that contemporary physical theories of space-time do not logically determine all the questions we might reasonably raise.
Further, if you think of semantics as, roughly, the theory of what would have to be the case in order for our utterances or thoughts to be true, then the fact that the ontological questions—questions about what there is—remain incompletely determined by physics, entails that physics will not dictate our temporal semantics. So I think that the philosophical projects remain.
3:AM:So the semantic argument is about tenses. I guess before we get to understand your position we’re going to need to know about A-series and B-series positions regarding time. A’s about time being tensed and B tenseless but can you flesh this out so we get what you take to be the salient elements of the two views. You start with an analogy between space and time but point to an important disanalogy when we theorise ‘here’ and ‘now’. Can you tell us about this important difference between space and time?
JM:The terms ‘A-series’ and ‘B-series’ are inherited from McTaggart’s argument against the reality of time. McTaggart began with the observation that we have two sets of terms to describe positions in time. One set orders times with respect to whether they are ‘past’, ‘present’, or ‘future’. Another set orders times in terms of ‘is earlier than’. McTaggart called the ordering of events that results from the first set the ‘A-series’ and the order that results from the second the ‘B-series’. Now, both sets will produce a series of events that is in fact the same order. For example, both series will put the beginning of the First World War in 1914, 25 years before the beginning of the Second World War, which ended 24 years prior to the first moon landing, and so on. However, though the A-series and B-series agree on the order of events, the difference is that A-series attributions are constantly changing, while B-series relations are not. World War I begins 25 years before World War II, and this never changes, but each was once future, then present, and is now past. So, B-series relations don’t change, but A-series predications do change with time.
The A-series is described explicitly in terms of tense. Accordingly, views that take the A-series to be a fundamental part of the theory of time are sometimes called ‘tensed theories’ of time (often just the more generic ‘A-theories’). The B-series is defined in terms of temporally invariant relations, so theories that deny the objective reality of A-series properties are often called ‘tenseless’ or ‘B-theoretic’ accounts of time. So, as a result of McTaggart, philosophers tend to divide into A-/tense theorists or B-/tenseless theorists. McTaggart himself considered both the A-series and the B-series to be essential to time, but he argued that the A-series is incoherent—it harbors a hidden contradiction—and the B-series alone incompatible with genuine change. He therefore concluded that time is unreal.
So you can see how this connects to the semantic dispute I mention above. Part of what drives the philosophy of time is the question of whether the world itself has a tensed structure. There are at least two ways it could. First, what is real could depend on what is present. Secondly, events in the world could instantiate tensed properties. If, as I argue, the semantics of natural language is in fact tenseless—that is, includes only B-series relations—then we do not commit ourselves to the reality of mind-independent, A-series properties. For something to be ‘present’ or ‘now’ is just for it to stand in a simultaneity relation to a particular time or event; for it to be ‘past’ is for it to be earlier than a particular event of interest; and so on. In other words, ‘past’, ‘present’, and ‘future’ are relations.
One of the phenomena that has seemed to many to count against this tenseless account is the sheer utility and, indeed, indispensability of tensed language and thought in navigating our environment, communicating, achieving goals, and so on. Numerous philosophers have noted that we simply wouldn’t be able to get on with our lives were we to avoid thinking and speaking about the world in tensed terms. But, note that we must also distinguish what is ‘here’ from all other places as well. For example, suppose you want to take the train from Toronto to Montreal. You know when you set out to the train station, that Montreal is 541 kilometers to the east of Toronto. You know this when you board the train; you know this at the halfway mark; and you know this in Montreal. So, conceiving of Montreal as ‘541 km east of Toronto’ is not a form of thought that can prompt you to disembark at, and only at, Montreal (since it doesn’t change when your position changes). Rather, you need to conceive of Montreal first as ‘there’, and then eventually as ‘here’, in order to be prompted to leave the train at, and only at, the right place. In other words, you need a belief form that is spatially variable. Similarly, if you want to meet someone at noon, then you can’t conceive of the meeting time as ‘12 hours later than midnight’, for that form—which is B-theoretic—is unchanging and will be truthfully held not only at noon, but all other times. So, one needs to conceive of noon first as ‘future’, then as ‘now’, if one wants to successfully time one’s actions. So, we need terms or concepts that are subject to temporal and spatial variability if we are to act successfully in the world.
Nobody, however, takes the need for spatially variable terms—spatial indexicals—to have any implications about the nature of space itself. Everyone cheerfully accepts that though it is pragmatically necessary to divide space up into the ‘here’ and the ‘there’, space itself is perfectly isotropic; there is no such-thing as an observe-independent ‘here’. Space is not centered, even though our orientation toward space requires us (most of the time) to center our representation of it upon ourselves. This is just one of the ways in which, we all agree, our normal ways of thinking about space come apart from the way it is itself. On the other hand, many thinkers argue that we cannot do the same in the case of time. In the case of time, the need to center our thought, actions and goals on the present moment is taken to indicate that the present itself is distinguished from all other times—either in terms of being a unique point of existence or else in terms of instantiating a mind-independent property of absolute nowness. I find this curious because this argument in favour of a temporally centered universe is perfectly analogous to the one for a spatially centered one. So, something has gone wrong. That is why I consider it important to actually provide the tenseless semantics for tensed discourse; I am not content to simply say ‘well, the same argument could be made in the case of space, and since that argument is invalid, so is the temporal one’. I happen to believe that to be the case, but the details have to be worked out and presented, because it is obviously tempting to suppose that time is different from space in this way (of course time differs from space, just not in this way). This is another reason why the semantic project matters.
3:AM:Is McTaggart still useful when arguing about whether time exists?
JM:I think McTaggart’s arguments about time are somewhat misunderstood. This is largely due to his own presentation, which is convoluted and, at key moments, unclear; also, he makes a serious mistake about the nature of change. Furthermore, his association with Hegelian idealism renders him immediately suspect in some quarters. So, any appeal to McTaggart is essentially guaranteed to be controversial in philosophy.
Nonetheless, I think his argument against the reality of time contains an important insight into the logic of temporal predication. Consider the A-series predicates ‘x is past’, ‘x is present’, and ‘x is future’. They obviously are temporally variable: 2015 is present, but it won’t be forever. So, 2015 changes from being present to being past. But how can we represent this change? Suppose we represent it as follows: ‘2015 is present, but also 2015 is past’. This appears to be a contradiction. We can do better as follows: ‘2015 is present and 2015 will be past’. This is fine, but notice that we have replaced the predicate ‘x is past’ with the relation ‘x will be past’. The latter is only true with respect to some times (not 2016, for example). Similarly, ‘2015 is present’ must be a relation as well because it only holds with respect to some times, namely any moment in 2015. So, when we try to represent the variability of A-series properties in non-relational terms, we end up with a contradiction—this is something McTaggart noticed. But if we represent them in relational terms, then A-series properties end up equivalent to temporally invariant, B-series relations. For example, ‘with regard to 2014, 2015 will be past’ is just another way of saying that 2015 is later than 2014. If we try to recapture some non-relational content to ‘is past’, we end up with a semantics for tensed predicates that leads to absurdity.
I should point out that none of this entails that time is unreal unless one thinks that a model that includes all moments standing in temporally invariant temporal relations to each other couldn’t be a model of time. This is what McTaggart thought, but his argument for it is both unmotivated and undefended. His argument is this: (1) time entails change; (2) a model that includes all times standing in temporally invariant temporal relations to each other lacks change; so (3) the B-series is not time. McTaggart never defends the second assumption: he simply asserts that if the sum total of facts never changes with time, then there is no real change. Consider the following example: the poker is hot at noon, cold at midnight. This does not, for McTaggart, record genuine change because it is itself a temporally invariant fact: the poker’s being hot at noon never changes—if it is hot at noon, it will never be the case at some future time that it was in fact cold at noon; the same goes for its being cold at midnight; that will never change. So, McTaggart assumes that the conjunction of two facts that don’t change with time cannot ever be an instance of actual change in the world.
Not only does McTaggart merely assert this claim without defence, it is in fact indefensible. Change is variation with respect to time. For something to change is for it first to be one way at one time, then another way at another time. These facts themselves don’t have to change in order for the change in the object to be genuine. To require that the facts themselves change as well is simply to insist that change must itself change; that is, that the fact that the object is first one way then another must itself change for there to be real change. But why isn’t change in the object itself enough?
So, McTaggart does not give us any reason to think that time is unreal because he gives us no reason to suppose that a relational, B-theoretic account of change is unsatisfactory. Still, he does point us in a very interesting direction, namely, that tense, change, and the passage of time, can only be understood in relational terms. This idea is the heart of the B-theory, in my opinion.
3:AM:Does Einstein’s Special Relativity theory have a claim to being a reason for rejecting the A-series view, and do you think it is justified to think it still does?
JM:I do think that Einstein’s theory gives us reason to doubt the A-series view, but as I indicate above, it is not conclusive reason on its own; it needs to be supplemented with philosophical argumentation. Relativity Theory does inherently, I think, put a certain amount of pressure on the A-series view, simply because it does away with the notion of a universal now that divides all times into what was, is and will be. Since I think that a great deal of the appeal of A-theoretic views is that the world seems to be divided into the past, the present, and the future, the fact that such a division is perspectival does undercut some of the motivation for the A-theory. Thanks to Einstein, we now know that the world is not exactly as it seems to be.
Having said that, there are ways to recover A-theoretic or presentist-like features in the context of relativistic space-time. These pictures don’t look much like the classical presentism defended by people such as Arthur Prior but, on the other hand, they do not look quite like the eternalist picture defended by Putnam either. There are also certain assumptions underlying Putnam’s argument that might be challenged. For example, Putnam assumes that it would be absurd to suppose that different observers exist in distinct realities. However, Kit Fine has recently argued that, in order to make sense of the passage of time, we should suppose that reality is ‘fragmented’, i.e. that different moments in time are distinct realities that simply cannot be combined into a coherent whole. If Fine is right, then the idea that relative motion creates distinct realities in virtue of creating distinct temporal orderings is less implausible. I think that Fine’s position needs a philosophical rebuttal; appeal to Relativity Theory alone won’t do it.
So, as you can see, I believe that philosophy and science are both essential in coming to a full understanding of our world; they need each other. I happen to think that there are philosophical arguments that suffice to rebut non-eternalist ontologies and establish the non-centered conception of time. This picture happens to cohere nicely with Einstien’s theories, so we are, I conclude, in the fortunate position that our physics and metaphysics do not conflict. Still, metaphysics on its own would probably never drive us toward the Special Theory of Relativity and all the rich, interesting, geometrical structure of Einstein-Minkowski space-time, so, just to be repetitive, we need both disciplines to fill out our understanding of the world.
3:AM:You’d think that the ontological and semantic claims combine in predictable ways, so I’d guess that eternalists were tenseless and presentists were tensed. But it doesn’t always pan out like that does it? Can you tell us a little about the typical positions I noted above, and then say how someone like Bourne brings a different combination to the table. Does it work?
JM:In general, I think that semantic and ontological views are independent. I also think that we sometimes commit ourselves to the existence of certain things in virtue of the views we hold, but there is usually the possibility of reinterpreting the view to avoid the commitment. So one can often combine any number of ontological positions with any number of semantic theories.
For example, consider what is often called the ‘moving spotlight’ theory of time, investigated by Bradford Skow (in his recent book, Objective Becoming). This view is explicitly eternalist: it includes a B-series that contains all times. However, superimposed on top of this completed timeline is an objective present, a moving NOW, that ticks along from one moment to the next. Note that this NOW is absolute: when a certain time is present, it is present period, not in a relational sense. As I indicate above, I can’t make sense of such an absolute present, but the point here is that on the moving spotlight view, tense is fundamental and irreducible: the world is ontologically eternalist but also tensed, and any theory of the world that gets things right, on this view, must reflect the tensed nature of reality.
Other eternalists, such as D. H. Mellor, completely reject the absolute, moving NOW as incoherent. For them, tense distinctions don’t correspond to anything absolute or observer-independent: they represent relations between a thinker and the non-tensed world. So Mellor, and others, argue that our representations of time don’t need to represent objective tense, and tend toward a tenseless semantics, as I do.
Does a tensed—what I call ‘centered’—ontology require a tensed semantics? That’s the usual combination but, again, it seems possible to go against the grain here. For example, one could be a presentist but still insist that, for reasons of logical coherence, all tensed predicates are relations—to be past is to be earlier than the moment at which one is thinking, that kind of thing. This might require some revisions. One might say that, even though all talk of something being past is talk of it standing in some relation to the present, nevertheless the past doesn’t exist: the existence of a relation does not entail the existence of its relata. Or, one might suppose that to be past is to be earlier than the present, but because there is nothing earlier than the present, nothing is in fact past. However attractive such views may or may not be, they show that one can combine a presentist ontology with a tenseless semantics for natural language.
Bourne’sposition is interesting. He is a presentist but he accepts that there must exist something is in virtue of which propositions about the past and future are true; to use the jargon, the presentist must tell us what the truthmakers are for propositions about the past and future. He further accepts that there is a semantic need for times other than the present: a semantics that includes other times simplifies and systematizes temporal inference. However, he suggests that though the presentist must commit to the existence of non-present times, she needn’t concede that those non-present times are concrete. Bourne argues for a version of presentism in which all concrete entities exist in, and only in, the present; all other times exist as abstract, ersatz times. Accordingly, linguistic tense can be explained in terms of relations that hold between times, and even though reality is tensed, linguistic tense can be explained in terms of B-series relations (technically, they are analogues of B-series relations because at least one of the items in the relation is non-concrete).
So, there are certainly connections between semantics and ontology, and restrictions that the latter can place on the former, but overall, the two realms are largely independent and one can hold a wide range of combinations.
3:AM:You’ve also investigated whether tenseless presentism works. What does this position look like and can it account for enough of our experience of time?
JM:My position is that the presentist needs both a tensed and tenseless sense of existence for the view to make sense. The core expression of presentism is this: ‘only what is present exists (concretely)’. Suppose that ‘exists’ is understood in a tensed manner, as equivalent to ‘exists now’. In that case, the view is a simple truism: ‘only what is present exists now’. Suppose, on the other hand, that ‘exists’ is taken as equivalent to ‘existed, exists, or will exist’. In that case, ‘only what is present exists’ is an obvious falsehood because there clearly existed things that are not present. Hence, to make sense of the claim, one needs a tenseless sense of ‘to exist’, one that could encompass even non-spatiotemporal entities such as numbers. To exist in the tenseless sense is just to not fail to be; there is no implication as to when something exists. If we have this tenseless understanding, then presentism is not trivial; it becomes: ‘only what is present tenselessly exists’. I think this is untrue, but it is neither trivial nor absurd.
As Bourne’s arguments indicate, I think that presentism is compatible with tenseless (or ersatz tenseless) relations. Nontheless, I think that presentism is a fundamentally tensed theory that requires some mechanism to distinguish what is absolutely present from what is not. Hence, I doubt that one can have a purely tenseless presentism, unless one is considering solipsism of the present moment. On this view, the universe is frozen in a single moment, but that moment is loaded with memories and anticipations, so its occupants don’t realize that time doesn’t pass and that they have never existed, nor will they exist, at any other time. This is interesting, and I don’t think our experience obviously undermines it, though it is extremely difficult to motivate. It is much like Cartesian skepticism: logically possible, but with very little actual evidence in its favour. Someone might reply that all our evidence of the past and future lies in our current experience, so it is a leap to suppose there is something more than the present, and the burden is the other way around. However, there is need to consider whether the epistemic limitations we find ourselves in have ontological significance. In general, should we suppose that the limits of our evidence are the limits of the world? This sort of anthropocentrism seems to me to be impossible to justify, so if one is arguing from a lack of a certain kind of evidence to a lack of a certain kind of being, then I think one is making a dubious move that requires further support, which is why I think it is the solipsistic view that bears the burden here.
3:AM:You’re a B-series guy aren’t you? What are the advantages of the B-series approach to the A-series?
JM:For me, the B-series view falls out of other considerations, in particular the ontological arguments against a temporally centered view, and the semantic considerations I outline above. The B-series, as I understand and defend it, just is the combination of the eternalist idea that all times exist, and the semantic doctrine that the logical form of temporal predicates is that of temporally invariant relations. Since I consider the arguments for these two positions to be compelling, the primary advantage I see for the B-series view is that it is correct; it is where the arguments lead.
Nonetheless, I do think the B-series is an appealing view, with a number of interesting strengths. For example, it renders the content of propositions about the past and future both clear and familiar. On the B-series, when we claim that Napoleon lost at Waterloo, we are talking about is Napolean, the person, in the ordinary sense. What makes the proposition true is a complex set of relations that hold between Napoleon, Wellington, and their armies at a particular location and time. Compare that to, say, ersatz presentism, according to which propositions about the present are about, and made true by, concrete objects and events, while propositions about the past can at best be about ersatz entities. On the B-series, ordinary, familiar objects and events are available to complete the analysis. Both the familiarity of the objects appealed to as well as the theoretical unity offered by the B-series—the same kind of story is told for propositions about any region of time—is an advantage, in my view.
Another advantage is that the B-theory offers a successful example of a non-centered conception. Philosophical disputes over the viability of such conceptions are often carried out at a very high level of generality. I think it is worthwhile to consider specific problems, such as the nature of time, and see whether we can work out a perspective-free conception of just that. Can we actually model it and provide a semantics for it? That is worth doing, if it can be done. If it can’t, then we learn something important. If it can, then we can continue with the work of trying to provide similar models for other domains such as consciousness or the self.
In short, I see the B-series as serving both as an example of what to aim for in the larger project and as a piece of that larger project, to which further pieces can be added.
3:AM:Can you sketch some of the main arguments you find successful in showing that tenseless theories of time can account for linguistic facts?
JM:I defend what would probably be considered a fairly extreme or austere position, which could be called reductionism or even eliminativism about tense. On my view, any tensed utterance, predicate or thought can only express a relation between times or events. I can sketch out some of the reasons I favour this view.
As I write this sentence, it is September, 2015. Now suppose that while writing I mutter to myself, ‘2015 is present’ (to take a very simplified example). I argue that the content of this sentence is simply this: that 2015 is simultaneous with 2015. I defend this answer both positively, by showing that it works, and negatively, by demonstrating that the arguments against it fail.
To give you a sense of this, let me present a little story. Imagine that, at noon, you notice that a lecture you wish to attend begins at 3:00 and will be held at a venue that is an hour away from your current location. You settle in to enjoy your afternoon, occasionally checking a clock. Then, at 2:00 you look at your watch and exclaim, ‘the lecture begins in one hour!’, at which point you stop what you were doing and head out to catch a bus. I argue that, ‘the lecture begins in one hour’, uttered at 2:00, has the following content: that the start of the lecture, which is 3:00, is one hour later than now, which is 2:00 or, equivalently, 3:00 is one hour later than 2:00. As noted above, ‘3:00 is one hour later than 2:00’ is something you believed all afternoon long. For example, if you had put your book down at 1:00 and been asked, ‘is 3:00 one hour later than 2:00’, you would have said, ‘of course’. Assume, further, that at all times between noon and 2:00 you retain your desire to see the lecture. Accordingly, on my story, you believe all afternoon that 3:00 is one hour later than 2:00, and you desire, all afternoon, to attend the lecture at 3:00. Why, then, do you leave your house at, and only at, 2:00 rather than any other time? Why, in fact, do you change your behavior at all, at 2:00, given that your beliefs about the relation between 2:00 and the start of the lecture never change? My opponent insists that there must be more to the content of the utterance than the expression of a temporally invariant relation between two times, and this content is tensed: it involves the attribution of a non-relational property of ‘presentness’ to 2:00 and ‘futurity’ to 3:00, or something along those lines.
Let me outline my response. First, I argue that temporally invariant content is compatible with timely action. In brief, the story I propose is that we learn to respond to perceived cues in the environment in a temporally sensitive manner. For instance, we are taught not to say things such as ‘it is now 2:00’ or ‘3:00 is in one hour’, or to leave an hour in advance of 3:00, until we actually see a clock face that reads ‘2:00’, or until we smell the familiar scent of baking bread that comes through he window every day at 2:00, or until an alarm that we set to ring at 2:00 is heard, and so on. Knowing how to time our actions involves a perceptual skill that we need to develop. So, even if the content of one’s belief about the relation between the lecture and 2:00 p.m. doesn’t change, changes in one’s environment can still prompt changes in action. That’s the positive part of the defence.
The negative part, secondly, is to show that the alternative proposal of tensed content fails. For example, what could it mean to suppose that 2:00 has an absolute, non-relational property of presentness? How can we describe the process by which 2:00 comes to gain and lose such a property? If I am right, then all such attempts run into a roadblock: either they are impossible to coherently describe—2:00 is both absolutely present and absolutely not-present—or else they rely on an ontological model, such as presentism, which, I argue, we have good reason to reject.
3:AM:Though you argue that tenseless theories of time are not undermined by linguistic facts, neither are tensed ones. And tensed ones seem easier to defend – they need less machinery – than the tenseless ones. So what seems to be needed to decide between the two isn’t semantic facts but ontological ones about nature. But if that’s the case, why bother with the semantic debate? Or is it your point that we need a theory that can reconcile the semantic and the ontological as part of the project to reconcile the objective with the subjective points of view about time?
JM:In fact, I think it is the tenseless theories that require less machinery, are less complex, and even more intuitive. In part, this is because tensed accounts often rely upon models that can’t bear their ontological weight, or end up proposing implausible truth makers for claims about the future and past, such as abstract, ersatz entities. So, I do think that a combination of semantic and ontological arguments is required.
However, it is your last point that is the key one for me. My overarching goal is to defend the objective conception (not just as concerns time, but that is my focus here). In part, this involves demonstrating that we have the logical and semantic resources to form the conception in the first place, but it also involves showing how to integrate our perspective on the world into our account of that world. After all, even if the world itself is not centered on the here or now, it contains many points of view that must be explicable as part of reality. So, what makes a tenseless, temporally invariant semantics of philosophical as well as linguistic interest is that it indicates how to reconcile the existence of a non-centered world with indefinitely many perspectives on that world.
3:AM:Is yours a deflationary account of the passage of time and if so why do you think that’s the most fruitful way to approach the issue? Does time's arrow survive your treatment?
JM:Both objective change and temporal passage are compatible with the tenseless, eternalist worldview. Indeed, I think that only the B-theory has the resources for a satisfactory account of change and passage.
As I mention above, many defenders of the B-series deny that time passes. I argue that this stems from two errors. The first is the acceptance of McTaggart’s idea that temporal passage involves an absolute, non-relational now, or something equivalent. The second is the assumption that passage, whatever it is, must be understood as a kind of motion.
Against the first, I point out that since the B-theory involves the rejection of pretty much everything associated with McTaggart’s A-series, it is odd to defend the former while accepting the latter’s conception of passage. This is especially odd given that McTaggart himself never defends his conception of passage and change; he merely asserts it. So, the failure to articulate a satisfactory account of passage in terms of an absolute now, or something analogous to it, just has no bearing on whether time in fact passes, in my opinion.
Against the second, I point out that there are kinds of change that can’t be usefully integrated into the paradigm of an object in physical motion; for example, the change in colour of a leaf in the autumn, the change in one’s mood throughout the day, or the freezing of water in cold weather. Accordingly, there is no motivation to suppose that the passage of time, which is a concept that covers change in general, is itself a kind of motion. Hence, the failure to articulate a satisfactory account of the motion of time, or its rate, or what have you, similarly has no bearing on whether time in fact passes.
So, my account of passage is deflationary in the sense that it makes no appeal to anything such as a moving now, or a growing block, ersatz entities, or anything along those lines. As far as I can see, if first one thing happens, and then another, then time has passed (put more formally: if there exist events that are temporally ordered, then passage exists). Suppose you have a leaf that is green at one time, then after a while is red; time has passed. Suppose you once had a beard, and now you do not; time has passed. Suppose you once weighed less than you do now; time has passed. Suppose what was once a cup of water is now filled with ice; time has passed. And so on. The passage of time is nothing more than the fact that temporally ordered occurrences exist. So, you could say my account is deflationary. I insist, however, that such a minimal account captures everything we need the concept of temporal passage to capture: growth, decay, aging, evolution, motion, etc.
What makes this the most fruitful way to approach the problem is familiar. First, the alternatives depend on ontological models that can’t be sustained. Secondly, this account does all the work that we need it to do since it renders change a genuine, objective phenomenon, though one that is not tied to any particular temporal viewpoint: the passage of time does not require an objective now in order to be real.
Time’s arrow survives this account quite nicely. Usually, philosophers who are skeptical about the possibility of an objective arrow of time argue that though the world lacks an objective temporal direction it nonetheless contains an asymmetry, for example the thermodynamic asymmetry of entropy. To put it loosely, the world is more ordered at one temporal end, less ordered at the other, but this doesn’t entail that there is a unique direction to time—consider that a sign in the shape of an arrow could just as easily point toward its tail as toward its head; it is a convention which one to choose. For some reason, evolution latched onto this asymmetry so that we can only experience the world from the low entropy end toward the higher entropy end, but this no more reflects a deep fact about the world than does the fact that we can only see a small portion of the light spectrum; each case reflects a limitation in us.
I reject this kind of explaining away of temporal direction because it seems to me to conflict with objective probability. Here’s the argument.
There are certain kinds of correlation that we notice regularly. For example, an organized deck of cards that is dropped will become jumbled, but a jumbled deck will not become ordered when thrown upward; or a vase that is dropped will often shatter, but collect the glass pieces and drop them all you want, a new vase will not be composed. The concept of entropy is often appealed to in order to explain these sorts of correlations: disorder is more probable than order, so entropy always increases.
Now, the transition from the less probable to the more probable state is what is to be expected; the reverse is not. But, then, if there is no fact of the matter as to which way this interaction transpired—that is, if it is merely a subjective projection of ours to suppose the vase was first in the integrated state and later in the dispersed state—then there is no fact of the matter as to whether a really probable transition occurred or a very improbable transition occurred. But that can’t be right: it has to be the case that improbable transitions happen relatively infrequently. We know that our universe contains a certain kind of coherence here: we don’t observe vases reassembling themselves; this isn’t 50-50. So anybody who argues that it is just as correct to view the universe as unfolding from our future to our past as vice versa, is arguing that it is just as correct to view our universe as a long string of highly improbable transitions as it is to view it as a long string of probable transitions. It seems to me that this can’t be right.
Call the state of the universe at the Big Bang, B. Call the next state of the universe, C. Suppose that C has higher entropy, so that the transition B-C is probable, but the transition C-B is improbable. If there is no fact of the matter as to whether B occurs first or C occurs first—if there is no objective fact concerning the direction of time—then there is no fact of the matter as to whether a probable transition occurred or an improbable one occurred. But we shouldn’t accept a picture of the world in which commitment to the improbable is just as valid as commitment to the probable.
So, I hypothesize that the direction of time is a primitive feature of the universe, built in, as it were, to every position in space-time. We have reason to believe this, I think, because it enables us to explain why the coherence we observe exists; otherwise the asymmetries that we in fact see become mysterious: it should be closer to 50-50.
Of course, it may be the case that the universe lacks any temporal direction or asymmetry at the deepest, rock bottom, sub-temporal physical level. That’s okay as far as I’m concerned, so long as the direction of time exists as an objective feature of the universe, either as emergent from the more basic, sub-temporal level or as irreducibly higher level. What I object to is the idea that there is no objective direction of time; I think that simply can’t be right.
Either way, the key point is that all of this is consistent with tenseless eternalism. One needn’t reject objective change, temporal passage, or the direction of time in order to hold onto the ontological and semantic insights delivered by the B-theory.
3:AM:If you’re right then why is the present privileged in terms of perception and action if no such objective privileges come with time?
JM:In brief, it is causation that is the key to the story. Our perceptual experiences are, in the normal case, the effects of interactions with the environment. In the normal case, we believe there is a tiger in front of us because the presence of the tiger causes such a belief (the story must also involve the proper functioning of our organs, and so on, but I take that for granted here). Or, to return to the earlier case, we believe that it is now 2:00 o’clock because the light reflecting from a clock face that reads ‘2:00’ has impacted our nervous system and, because of our learning, triggered the belief.
It so happens, that while causal signals from our environment tend to be fast, they travel at finite speeds and tend to dissipate rapidly with distance. So, I notice the pop of the firecracker three centimeters from my feet almost instantly, but the sound of a firecracker on the other side of the planet has no noticeable impact on me. The result is that we are constantly being consciously impacted by—and forming beliefs about—nearby events, almost instantaneously, while rarely being made aware at all of distant events.
If I arrive where a firecracker exploded an hour earlier, there is no longer any sound or flash to witness; the signals have long since dissipated. I may conclude that a small explosion occurred earlier by noting some charring on the sidewalk, but I won’t see or hear anything from the earlier explosion. As I bend down to examine the charred sidewalk, I have new perceptual experiences: I believe, for example, that I am currently seeing a dark patch on the ground. Later, when the charring has been washed away, anybody who arrives at the same place will have no experience of that. So you see, there is no reason to suppose that we would, in general, perceive events other than when they occur.
This last point allows us to explain how we perceive the passage of time. Event E1 at time T1 will be perceived at, or very closely after, T1, but not at some significantly later time (or earlier time, given the absence of backward causation). Then, event E2 at T2 will be perceived at time T2, and so on. Hence, it might seem as though we are following or tracking a unique, absolute now, but in fact we just have a succession of experiences which adds up to an experience of succession (because short term memory allows us to keep the recent experiences in mind as we have the new ones); hence, we become aware of the sequence of events in time.
That is the basic story, in outline.
As for perception, so it is with action. Assuming that we act on the basis of our beliefs and desires, we must decide to do something when we believe the time is right. Since, as I mentioned earlier, our beliefs about what time it is are formed in response to triggers from our environment, we learn to wait for those triggers before acting, as in the case of the visit to the lecture one wants to catch. So once the relevant belief—e.g. ‘it is now 2:00’—is triggered, then the relevant action is as well and until that belief is triggered, no action is taken.
So, I don’t think the B-theorist has any principled difficulty in telling a story of perception and action that explains why it is that our perceptions are of the time of the perception itself, and not the past or future. Furthermore, the B-theorist can explain why timely action is possible. I don’t go into the psychological details. I am mostly interested in showing the coherence of the account. I do rely on the assumption that perceptions of our environment are at least partly caused, in the sense of being triggered, by events in that environment. If that turns out to be false, then my story can’t be right.
3:AM:Why isn’t a perdurantist position able to fit with your B-series theory?
JM:Perdurantism (four-dimensionalism) actually fits perfectly well with my version of the B-theory. So does exdurantism (stage theory) and endurantism (three-dimensionalism). My argument against temporal parts is mostly negative: they aren’t needed for a satisfactory theory of change, persistence, and identity, and since it is not entirely clear what temporal parts are, we can happily jettison them from our model.
What we are left with, I argue, is a relational account of change. As I note above, I think that ‘past’ is a relative term: tenses are relations. Similarly, I argue that ordinary predicates, such as ‘red’, ‘tall’, ‘heavy’, and so on, are relations: for something to be red is for it to be red at a time, and the best way to make sense of this, I argue, is to suppose that ‘red’ is relation between an object and a time. So, for something to change from green to red is for it to stand in a certain relation (‘greenness’) to one time and a different relation (‘redness’) to a later time.
There are, by the way, a number of reasons to accept this view. First, it fits perfectly with the tenseless view of time and eternalism. If we assume that relations entail the existence of their relata, then the claim that the leaf was once green entails that the leaf stands in the ‘greenness’ relation to an earlier time, which must exist. Eternalism has no problem with this entailment and, indeed, encourages it.
Secondly, consider the puzzle of change: how can some object change its properties, say going from green to red? After all, a red object is surely not the same thing as a green object, so what was green cannot be the same object as what is now red. Change, in other words, destroys identity. If, however, to be red or green is to stand in a certain relation to a time, then an object that changes in virtue of standing in distinct relations to distinct times can very well be the same object in both cases. So the relational view handles the puzzle of change nicely.
Finally, such a view is compatible with three-dimensionalism, the view that objects persist by enduring rather than by having distinct temporal parts located at different times. When an object, such as a leaf, stands in the greenness relation to T1 and the redness relation to T2, it is the leaf that stands in those relations, not a temporal part of it. (By the way; David Lewis objects to pretty much all of this, so I consider his views in a fair bit of detail in my book.)
So, one can combine the B-theory with a coherent story about the nature of change that does not involve temporal parts. While four-dimensionalism and stage theory are compatible with my version of the B-theory, they are unnecessary, so as far as I am concerned ought not to be a part of the view; things are simpler and more unified without them.
3:AM:And finally are there five books you could recommend that will take us further into your philosophical world?
JM:Saul Kripke, Naming and Neccessity: a marvelous disentangling of modal from epistemic concepts.
Heather Dyke, Metaphysics and the Representational Fallacy: a terrific antidote to the fallacy of conflating features of our representations with features of what we represent.
Tim Maudlin, The Metaphysics within Physics: simply fantastic. In fact, I recommend reading anything Maudlin has written.
Bertrand Russell, The Principles of Mathematics: it is more than 100 years old now, but I still consider it worth returning to regularly, and essentially correct on many matters.
Finally, Kurt Gödel is philosophically important to me for two reasons. First, he so forcefully reminds us that truth and provability are distinct. Secondly, he demonstrates that mathematics can be its own meta-language, which is an example I try to follow in arguing that our natural language can coherently contain the objective conception (we are not, in other words, trapped in our language but are in fact able to gain an external perspective, even on language itself). Both of Gödel’s lessons are guiding principles in my approach to philosophy. Ernest Nageland James Newman’s Gödel’s Proof is a good, introductory treatment of Gödel’s famous incompleteness theorem and some of its implications.
ABOUT THE INTERVIEWER
Richard Marshallis still biding his time.
Buy his book hereto keep him biding!