What exists is constantly changing. This fact requires that we take tense seriously.

(Chisholm and Zimmerman, 1997, 262)

Change is obviously real. But just how this real change might be related to a real time continues to puzzle philosophers. In this paper I argue that time, though in one sense unreal, nevertheless supervenes on change, which is real. I thereby support a theory of time that takes as its starting point the brute fact of change noted by Chisholm and Zimmerman.

The embrace is indeed reluctant in many quarters. There has always been an uneasiness with tenseless time arising from its awkwardness in dealing with our intuitions that we all view the world from roughly the same time and that this shared time is constantly changing. Michael Tooley (1997, 380) provides a fascinating quote from Rudolf Carnap revealing that none other than Albert Einstein was troubled by the anomaly that "the experience of the Now means something special for man, something essentially different from the past and the future, but that this important difference does not and cannot occur within physics." Even Mellor seems to abjure on the issue of tense in his most recent defense of tenseless time:

Both Carnap and Mellor suggest that the science of psychology might explain the anomalies associated with tenseless space-time. Carnap reassures Einstein that "the peculiarities of man’s experiences with respect to time...can be described and (in principle) explained in psychology." (Tooley, 1997, 380) But why should it be necessary, or for that matter even possible, to shore up the tenseless theory of time with human psychology?

Perhaps out of this uneasiness with tenseless time philosophers have shown an increasing openness to theories aimed at providing a coherent "third way" with respect to the metaphysics of time, a new approach that might improve upon the static, tenseless theory of space-time while avoiding the contradictions demonstrated early in this century by McTaggart with regard to changing tense attributes. This trend has been evidenced in recent years by the publication of two important works that seek to lay the foundations for a "dynamic" theory of time and/or tense. The first was Quentin Smith’s Language and Time, followed more recently by the appearance of Michael Tooley’s Time, Tense, and Causation. Smith presents a theory in which tensed facts "inhere in the present" (1993, 159). He thus avoids McTaggart’s paradox by allowing tensed facts about past/future events to change relative to the successive "presents" in which they inhere. By contrast, Tooley attempts a tenseless ontology based on the concept of a dynamic "accretion" of times, as events progress towards the future. Tooley claims to avoid McTaggart’s paradox by his stipulation that the emerging relations between present facts and their future successors are external relationships that do not result in contradictory properties for the events thus related. (1997, 378) While we may grant, for the sake of argument, that Smith and Tooley provide formally coherent theories of time, each theory requires significant leaps of faith with regard to certain key concepts and features. One wonders, for example, why Smith’s Present needs the peculiar tensed facts to inhere in its changing ontological fabric. Is it merely to provide the "semantic relata" for tensed propositions? Tooley’s source of dynamism, while perhaps skirting McTaggart’s paradox of contradictory events, seems to run afoul of J. J. C. Smart’s reductio with respect to "temporal motion" (1967, 126). Tooley’s universe tenselessly comprehends all existing times, but nevertheless changes as new events come to be, resulting in a puzzling, seemingly tensed, "truth at a time."

We think of time as a physical feature of the world, an objective phenomenon as real as the spatial extension of the common objects we take for granted. We assume that times are things we can simply see in everyday events, be it in the swift flow of a river, or in the snail-like course of the hands of a clock. "Seeing time" in this way, it’s quite natural to think of the times of events as forming a changing sequence, a "progress of time" somehow underlying the changing properties of the objects that we see as "progressing through time." We thus picture time as flowing swiftly with the river current, or inching ever so slowly with the advancing clock hands. Try as we may to abstract our view of time from these concrete images, it’s hard for us to imagine the normal course of times as anything other than, at the very least, a subtle kind of change.

Of course, the paradoxes tied to the view of time as a "changing" thing are by now quite familiar to philosophers. A brief review will serve to remind us of the difficulties. In the first place, how are we to imagine the advance, or "flow" of time? Is it best seen in the swift rush of the river or in the snail’s pace of the clock? A strange question, to be sure, but what do we see in these unlike spatial motions that could stand as a common "flow of time" that both could share? The answer would seem at first to be found in the familiar picture of time as a space-like dimension, a physical "metric" which seems to allow us to posit a special dynamic "moment" of time motion that dissimilar spatial motions might share. But this approach to time is far from unproblematic. For one thing, the progress of time along the space-like time dimension cannot be conceived as having a space-like direction or speed, at least not in any way that we can easily picture. First, consider time’s "direction." Whatever the spatial orientation of the physical time dimension that we may imagine, there will always be some spatial motions that cannot share a positive "moment" of motion along the chosen time dimension. Next, consider the matter of time’s "speed." Smart and others have noted that our normal concept of movement presupposes time. Thus, trying to measure the progress of time in the way we ordinarily measure movements results in an infinite regress of meta-time dimensions.(1967, 126) For this reason, the progress of time along the time dimension cannot be conceived to have a "speed," at least not in the ordinary way we conceive of the speed of a natural motion.

There are more problems for the space-like time dimension. The progress of time is found not just in rushing rivers and creeping clock hands but also in the turning colors of autumn leaves and the ascending notes of a piano scale. Why should we assume that these non-spatial changes, in order to "progress" through time, must participate in "motions" along a quasi-spatial time dimension? Isn’t it just a fluke of our language that we do not think of time as "reddening" with the maple leaves or "ascending" with the piano notes? One last, and perhaps most puzzling, problem for the notion of a physical dimension of changing time is that changing times do not seem to require physical changes at all. One would think that conceiving of time as a "moment" of motion along a special time dimension would require a "dynamics of time motion" as do the physical "moments" of motion we measure along more familiar spatial dimensions. But this does not seem to be so with the "changingness" of time. There is a strangely preternatural quality to our concept of time that has led some scholars to treat the change of time as a kind of philosophers’ change not clearly tied in a natural way to more familiar types of physical change. We think, for example, of the flow of time as marking a wind-up clock even if that clock has long-since wound down. What prevents us from saying that such a "stopped clock" points to the winding down of (its) time? We do, of course, continue to see changes outside the clock. Do we imagine that these external changes somehow insinuate themselves inside the stopped clock? Or, more likely, do we imagine subtle, unseen changes within the stopped clock itself? Imagine for the moment that there were no physical changes to be found within the stopped clock. Would we still suspect it of undergoing an ineluctable, "temporal" change? It seems that we would, but why would we?

Why can’t a Stopped Clock chronicle it’s own progress through time? One conceivable explanation, though one that does not work, is that change in time, like change in spatial location, can be measured only in relation to a frame of reference. Thus, since everything is constantly in motion at the same rate with respect to time, there is no temporal frame of reference in relation to which the change of time of the Stopped Clock can be directly perceived. On the one hand this argument proves too much. The moving hands of a working clock progress with respect to time at the same rate as any other object in the universe, including our Stopped Clock. By this argument, then, there should be no temporal frame of reference in relation to which the change of time of a working clock could be observed, any more than in the case of the Stopped Clock. How can it be, then, that a working clock can tell the time while a Stopped Clock cannot? On the other hand the frame of reference argument proves too little. It suggests a merely contingent relationship between a clock’s motions and the passage of time. But we know with far more than inductive certainty that a normal, moving, clock can "tell the time." If there were a merely contingent relationship involved, we might imagine a clock "defect" wherein moving clockhands falsely indicate the passage of time, much like a broken odometer whose wildly spinning dial falsely indicates the traversal of space. But we do not accept this notion, finding it odd that a clock’s hand movements should occur "in no time."

Summing up, the Riddle of the Stopped Clock shows two things: first, that the passage of time is tightly linked, some might say intrinsically linked, to the occurrence of qualitative change, and second, that the presumed passage of time for changeless objects is a singularly elusive commodity. Change of time for changeless objects must, it appears, be inferred on the basis of a theory of time from the proximate existence of changing objects. Assuming, for example, that objects can interact only if they are both at the same time, and observing that changeless objects are capable of interacting with changing objects at various phases of their change, we reason that time must change for such changeless objects. How else could they "keep up in time" with their changing counterparts? We thus have what we might label an "atmospheric" view of time according to which changing objects impart their temporality to nearby objects, even to objects that exhibit no temporality of their own. This "atmospheric" view of time has become less tenable with the advent of modern physics. Consider a variant of the "twin paradox," a fable of relativity where astronauts returning to Earth appear younger than their associates who have remained at home. Despite the measured differences in their "times," the astronauts and their earthbound colleagues succeed in shaking hands. Strictly speaking, we cannot say that the handshakers are all "at the same time." All are characterized by their own clocks and their own times. Only an arbitrary decision can determine which of their clocks is the "true clock." It appears that contiguity and interaction in space can no longer serve as unambiguous criteria for contiguity and interaction in time.

We have remarked on the fact that, faced with these paradoxes, many philosophers and scientists in our era have abandoned the idea that time is something that changes. It is important to document this point somewhat, for it may well come as a shock to the uninitiated that a significant segment of the scientific community, perhaps the dominant segment, does not think of the normal course of events as representing a "flow of time." This viewpoint is made clear in a revealing characterization of space-time by physicist Robert Geroch:

Geroch is forced to this static, "non-dynamic" view of space-time as a direct result of the paradoxes we have recounted with regard to the change of time. But it seems that more is lost in this popular metaphysics of space-time than merely the problematic changingness of time. The doctrine of the tenseless space-time continuum abandons, along with changes of times, changes of other sorts as well.

Supervenience, as that notion has emerged in contemporary philosophy, is a term-of-art with no "natural" meaning that we might usefully drawn upon. For this reason, I must take some care to mark, among the various uses of the term, that specific sense that can frame precisely what I mean when I say that "time supervenes on change." Indeed, there is a well-known association of the term that I want in particular to avoid. I do not want to suggest that the existence of time belongs to our "mental" phenomenology. This might suggest that the supervenience involved in Prior’s theory is akin to a kind of idealism that would deny the possibility of coming to an objective understanding of temporal facts. This view is encouraged by the use of such labels as "temporal solipsism" (Le Poidevin, 1991, 36-44) and "instrumentalism" (Earman, 1989,166) to characterize theories of the type that Prior has advanced. The kind of time supervenience that I wish to advocate is far from the kind of anti-realism suggested by Le Poidevin and Earman. If time supervenes on change, the core of objective reality that is left to us is nothing other than the changing world we see before us, a brute fact of such power that we may well wonder some day how twentieth century metaphysicians could have strayed so far from it in their theories of a tenseless space-time.

In framing a concept of time supervenience that can preserve this robust reality of change there are two useful sources to draw upon. The first is taken from tenseless theorists themselves in their characterizations of the unreality of tense. The second derives from the work of Jaegwon Kim on the notion of natural supervenience.

Jaegwon Kim is a good source to draw upon in characterizing this kind of "natural" relationship between our representations of time and the underlying grounds for those representations. In an analysis of phenomenal properties, Kim distinguishes between "nomological" supervenience (which seems akin to a "weak" supervenience), and "metaphysical" supervenience (which seems more like a "strong" supervenience). Kim’s account of nomological supervenience is as follows:

Although Kim raises some doubts about the adequacy of nomological supervenience in accounting for phenomenal properties, this kind of natural supervenience seems to be just what we need to characterize the lawlike relationship between an objectively changing world and our objective representations of that changing world. This is still, admittedly, a "weak" supervenience, which leaves open the logical possibility that a "Martian" observer might have a representation of "time" quite unlike our own. This is an interesting consequence in its own right, and in any case is a small price to pay for the avoidance of the anti-realism that might otherwise be suggested by the theory that time supervenes on change.

Prior’s notion of the present as the real, as distinguished from the unreal but remembered past, and the equally unreal but anticipated future, captures quite nicely the tense theorist’s demand for sharp tense distinctions. But it does so without postulating events with changing tense properties or relationships, concepts which McTaggart, and others including Dummett (1960) and Mellor (1981), have found to be incoherent. Prior’s theory provides a simple and elegant solution to the Riddle of the Stopped Clock. The Stopped Clock, seemingly fixed in time, has no distinguishable events that would allow the employment of our usual tense grammar. In the ideal case the Stopped Clock is truly timeless. By contrast, the normal clock, in virtue of its simple motions, has everything it needs to "tell the time," but without physical time properties or relationships.

The Stopped Clock is useless in telling the time precisely because it has no discernible properties that might be tied to one time rather than another. But the normal, moving clock is no more forthcoming in disclosing a changing, physical property related to its changing times. The successive times indicated by a clock’s moving hands cannot be identified with any discernible properties of the moving clock hands. If times could be so identified, then we would be forced to accept as literally true the old saw that a stopped clock "tells the time twice a day." A stopped clock, whatever its physical properties, cannot tell the time. A moving clock does not tell the time by exhibiting specifiable physical time properties.

Once again, the Stopped Clock presents an anomaly: no perceptible traces of the presumed physical time relationships are revealed to us in our successive "looks" at the Stopped Clock. We assume that our own "successive looks," or the motions of a normal, moving clock overcome this difficulty by revealing to us the physical relationships that comprise time. But is this really the case? We are perhaps fooled by our usual, spatial metaphor for time. We often think of the change of time in terms of a change of spatial position. We see a bus start at the corner, move to mid-block, and then come to a stop at the next corner. It seems natural to associate the various spatial positions of the bus with its presumed "temporal positions." Imagined in this way, it seems almost trivial to conclude that the temporal relationships of the bus at various times somehow mirror the progressive spatial positions that we remember. The problems with this assumption of space-like time relations can be brought out with a consideration of a change that does not involve spatial motion. Consider, for example, change of color. In experiencing a change of color we see time as vividly as in the case of the moving bus. But it is more difficult to imagine time as a kind of space-like relationship. Change has surely occurred, but the temporal relationships between the changing colors seem far more abstract than with spatial motions. Because the only real change that has occurred is a change of color, we cannot talk of the relationship of the before and after colors except by a re-imagining of the change itself (it was green, and then it was blue). There is no "temporal space" apparent in the transition between the green and the blue. Although it may be difficult for us to transpose these results to the case of spatial motion, they are exactly analogous. Although we may imagine that we see the successive temporal positions of the moving bus as a space-like progression, we in fact see, and remember, only its successive spatial positions. And, as in the case of change of color, we can only make sense of the "successive" nature of these spatial positions by re-imagining the spatial change itself (it was here, and then it was there).

"Times" naturally come into play in our concepts of changing objects. We have seen that, insofar as our experience of time is concerned, changeless objects are "timeless." This is part of the explanation of the Riddle of the Stopped Clock. A timeless object cannot serve as an empirical standard of time. Is there nevertheless a need for "times" in our consideration of changeless objects? Could changeless objects be literally timeless? That we do not readily accept this paradox is due to our unstated belief that we must assume "hidden" times of changeless objects in order to be assured of their "synchronization" with changing objects. But without physical time properties or relationships, "synchronization" reduces to coexistence. And the coexistence of a changeless object is in no way threatened by the "changingness" of its neighbors. Changing objects, though they exhibit successive properties through their changes, don’t "get anywhere in time" by their changes. They simply change. Thus, changeless objects, although timeless, have no trouble "keeping up in time" with changing objects. Le Poidevin notes this timelessness of changeless objects as a consequence of Prior’s theory. If we take Prior as identifying instants with the conjunction of propositions true at a certain time, then,

This thesis can be seen as a further corollary of Smart’s argument. Because temporal relationships in general are not verifiable as physical relationships, lawlike behaviors, if conceived as a necessary subspecies of such temporal relationships, are themselves incapable of verification as physical relationships. Lawlikeness, insofar as we can verify it, must therefore be conceived as a monadic property of objects rather than a dyadic, or polyadic, relationship between objects.

Critics of Prior’s "presentism" have claimed that it requires a concept of absolute simultaneity that is foreign to modern physics. However, if time supervenes on change, there are no physical times that objects can share. Thus, being "at the same time" is meaningless, even for co-existing objects or events. With the evidence of modern physics it seems just as well that this be the case. For example, in our fable from the theory of relativity, we imagine astronauts returning to Earth younger than their associates who had remained at home. Despite their claimed differences in times, the astronauts succeed in shaking hands. It is only in an informal sense, therefore, that we can say that we compare times by noting the physical coexistence of changing physical objects.

However, we may speculate as to whether Prior really ought to have held an absolutist view of the present given the informal nature of the simultaneity that seems to follow from his view of the present. Although Le Poidevin has characterized Prior’s present as a conjunction of contemporaries (1991,37), Prior himself defines the present in terms of the real, that is, all that coexists. To avoid circularity, it ought to follow that coexistence itself cannot be defined in terms of a singular time, but rather in terms of the various procedures, real or imagined, that we might use to establish the possibility of physical interaction between coexisting elements. It would seem that the relativity, or frame-dependence, of the descriptions of "coexisting" events that results from such procedures ought to place Prior’s view of the Present squarely on the side of the special theory of relativity.

Nevertheless, we can appreciate Prior’s resistance to such a relative view of his Present, violating as it does what for Prior is most likely a logical concept of coexistence. Peter Geach expresses a similar objection to the view of simultaneity as a physical, as opposed to logical, matter (1980, 312). It appears that both Geach and Prior would reject the view that the relativity of our procedures for establishing coexistence entails the relativity of coexistence itself.

This matter is of course not closed with these brief remarks, and in fact remains one of the chief issues with presentism when confronted with modern physics. I shall have further remarks on the possible resolution of this issue in a section dealing with so-called "time dilation."

One criticism of Prior’s modal tense ontology has been the claimed "solipsism" resulting from Prior’s insistence on the "unreality" of the past and the future. Smith (1993,166) and Tooley (1997,235) point out the lack of "semantic relata" or "truth-makers" for past-tensed propositions. But the lack of such "relata" in no way lessens the practical validity of our memories of the past or our anticipations of the future. We can attribute as much practical weight as ever to lessons from the past and form as easily as ever prudent anticipations of the future. The roles of tensed statements in our "language games" are unchanged by Prior’s ontology of time.

Le Poidevin notes a problem with Prior’s anti-realist position concerning the past having to do with the identity of objects across times in the cases of object fission and/or fusion (1991,41-4). With a four-dimensional domain, it seems easy to account for such cases through the lawlike relationships between an object and its subdivided progeny, or between separate objects and their merged descendent. But without such "causal relationships" It may seem mysterious how fission and fusion are to be conceived at all. In response, we must first distinguish cases of "mereological" fission and fusion from those that involve what we might call "metaphysical" fission and fusion. In mereological cases, an object is in reality a group of more basic "atoms" which in separating or converging give the impression of object fission and fusion. There does not seem to be any problem for Prior with this kind of case. Fission and fusion here constitute variations in the spatial "topology" of a set of atoms, a kind of change that Prior can easily countenance. However, cases of "metaphysical" fission and fusion do seem to be a problem for a theory where changes in times supervene on changes. In these metaphysical cases there seems to be a "coming to be" of new atoms, or a "ceasing to be" of existing atoms. Prior’s theory does have a problem with such metaphysical changes, as I attempt to explain in the following section on Ramsey’s problem of Cardinality.

Extending an argument from Frank Ramsey, Le Poidevin notes that it is a consequence of Prior’s ontology that the number of individuals in the his universe is necessarily fixed over time. (1991, 48-53) This does not seem to be a problem in a four-dimensional domain because the fixity of the domain applies across all of time as well as space. Thus, Earman has no difficulty in postulating for both classical and (certain) relativistic space-times objects which exist at some times but not at others (1989, 177-9). But for Prior, for whom the real exists only Now, this paradox of cardinality seems to fix the number of individuals that Now exist to all future and past "presents" as well. I shall not comment directly on Ramsey’s or Le Poidevin’s proof in this regard, other than to note that it rests on an important axiom of modal logic, namely, that:

Nevertheless, I do want to discuss a way in which Prior’s ontology of time lends itself to this kind of paradox. The problem for Prior is to provide a satisfying account of the metaphysical flavor of "coming to be" and "ceasing to be." Consider "ceasing to be." Immediately ruled out would seem to be the "ceasing to be" of a changeless object, since it has no times at which it might cease to be. So "ceasing to be" would appear to involve change. But how are we to understand this kind of "metaphysical" change in Prior’s ontology? Normal change we can attribute to an object that continues to exist (it was green, and then it was blue). But an object that ceases to exist results in a contradiction if understood in this sense (it was green, and then "it" was not). Normally, we would answer this by referring to a natural void, a continuing object to be sure, existing where the "it" used to be. But how do we know this natural void is "where" the object used to be? Only, it would seem, by observations of lawlike traces of the disappearing object. But this then negates the "metaphysical" nature of the presumed "ceasing to be." Of course, we might view this "problem" for Prior as an advantage. It puts him in the respectable company of Kant, for example, on the eternity of time. And is it just an accident that our natural laws seem to rule out the kind of "metaphysical" change that Prior’s theory fails to comprehend?

Le Poidevin claims that the temporal solipsist is committed to an untenable "theory of propositional instants." (1991, 53-7) He makes two points in this regard. The first is that an ontology of instants as "events" is foreclosed by the solipsist’s anti-realism with regard to past and future events. The second is that an acceptable theory of propositions must include the time of reference as part of its content. Responding to the first point, I need only two types of entities for a world in which time supervenes on change. I need substances and properties capable of change. Events might be thought of as a way of describing the changes of such substances and properties. Denying that such events might exist, or insisting that such changes do not support an adequate ontology of events is question-begging.

The second point may be related to a point by Zimmerman (1997, 83) and others about the difficulties in individuating events that have no "time" attribute. Zimmerman imagines an oscillating Universe wherein the attributes of events during each cycle are identical. Maintaining the distinctness of the otherwise identical events without physical time attributes amounts, in Zimmerman’s view, to a kind of "event-existentialism." In response, there are two possible courses open to the proponent of events without times. The first is simply to embrace event-existentialism, the irreducible "haecceity" of events. This is, by the way, a familiar recourse in the example of the absolute vs. relative views of space, where the same problem arises in individuating spatially identical objects in a symmetrical universe without stipulating absolute spatial locations. But there is another option for a theory of supervenient time. We might admit a truly reversible change, from A to B to A, that would nevertheless be represented in a supervening B-Series as "first A, then B, then A," but of course the "knower" is then conceived as being outside the perceived "world."

John Earman voices a kind of objection that amounts to a methodological dismissal of certain theories that do not seem to provide a satisfactory semantic grounding for space-time theories. Thus, Earman claims that a Machian space-time, which contains no physical time metric, is incapable of supporting a physical theory that can express what we mean when we say that an object is moving at a certain speed, or is accelerating at a certain rate (1989,30). Using Newtonian formulas to express principles of moving bodies in such a Machian space-time is for Earman akin to a kind of "instrumentalism" (1989,128). Note that the anti-realism suggested by Earman is not really a fair assessment of the situation. We can simply perceive such facts as the relative "speed" of a change and sense the change in motion involved in an "acceleration." Although the universe might be more satisfying if we could provide the graphical analysis afforded by the imagined time metric, we do not really need a physical time metric for events like velocity and acceleration to be recognized as objective facts.

If time weakly supervenes on change, the issues surrounding the "arrow" of time are changed in several respects. In one sense, the "arrow" seems tied to the idea of time as a physical dimension. In this sense we must conclude that, if the time dimension itself is unreal, so must be the "direction" of time evidenced by any particular change. However, the "asymmetry" of time often seems to refer to the asymmetry or irreversibility of change itself. Regarding asymmetry, though changes are "necessarily" asymmetric, they are so only in the trivial sense that a change is always from some A to some B, a change from A to A not counting as a change at all. Further, because time weakly supervenes on change, the perceived "direction" of the change is determined by our natural faculty of perceiving change. Although hard to imagine, what we perceive as a change from A to B might be perceived by a "Martian" as a change from B to A. Regarding reversibility, in a theory of supervenient time it is not clear that a reversible change can mark a reversible time. If changing objects don’t "get anywhere in time" by their changes, they cannot be conceived to "get back in time," however reversible their changes may be. Finally, as our theory shows, even with a truly reversible change from A-to-B-to-A, our representation of the B-series might remain as "first A, then B, then A."

The astronaut fable shows the problem with time travel. From the point of view of the homebound astronauts, the laws of physics have slowed the times of the returning astronauts. But it makes no more sense to say that the returning astronauts have traveled to the future than it does to say that the homebound astronauts have traveled to the past in order to participate in the same handshake. If times supervene on changes, there are no times one can "be at" or "travel to." With supervenient time, the times of co-existing true clocks may differ, but there is no sense in which such clocks can "travel in time."

Lawlikeness in a world of supervenient time presents a very different face from that imagined in a world of relational time. Because of the nonrelational, monadic character of lawlikeness in this view, the past (or the future) cannot be thought to present a credible threat to freedom in any way analogous to those physical constraints familiar to us in dealing with the present. Actual constraints achieve their constraining character by their continuing physical relationships to us in the present, that is, as coexisting objects. In a world of supervenient time, lawlikeness has no special role to play in determining which coexisting physical properties we come to see as constraints on our freedom. While it is true that many of our constraints are lawlike phenomena, constraints can just as easily be random, or anomalous, phenomena. One is reminded of Wittgenstein’s imagined railroad tracks that "suddenly change their shape" (Smythies, 1989, 85). An indeterminate railroad, though less predictable, might still constrain. We cannot even say with confidence that the elimination of our lawlike environment would result in freedom. Real, palpable, constraints on freedom can arise from the absence of a lawlike environment that would normally enable free action (think of the hapless astronaut on a space walk).

Constraints on human free will have also been claimed to follow as logically necessary consequences of past facts and the existence of natural laws. However, within Prior’s tense logic it is difficult even to express coherently the idea of a logically necessary consequence spanning time. Thus, Peter van Inwagen’s Consequence Argument comes into question. (1983, 16) In Prior’s tense grammar, the idea that temporally prior conditions plus natural laws can logically entail a temporally advanced consequence comes out something like the following:

P (p & (p -> Fq ) ) -> P (Fq).

The meaning of the modal consequence "P(Fq)" is open to question in a way that a simple, non-modal consequence "q" would not be.

Although I have suggested that a coherent alternative to Prior’s supervenient time is provided by the Standard tenseless theory of space-time, I want to conclude with an argument against that theory. The problem for the Standard theory of time, as is evident in the remarks of Mellor and Einstein, is to account for our undoubted experience of change in such a changeless, tenseless, space-time continuum. Mellor, and Carnap before him, suggest that this experience of the "moving Now" is due to our human psychology. Our psychology accounts for the changing succession of our experience. But how, within the Standard theory, are we to give an objective account of this "psychology" of changing perceptions? Only, it would seem, through a picture of a succession of experiences of the changeless space-time continuum from a succession of positions in this continuum. But, according to the Standard theory, this succession itself can be nothing other than a series of "experience events" corresponding to the occurrence of our experiences at the times claimed. This series of experience events can no more involve change than can the events that we claim to experience. Moreover, no number of apperceptive layers of experience can remedy this essential changelessness of the standard view of space-time. Ultimately, our conclusion must be that Prior’s theory of supervenient time is superior to the Standard theory, not on the basis of bald coherence, but rather on the basis of our undoubted experience of change. Prior’s theory starts from that experience, while the Standard theory struggles, and ultimately fails, to explain it.