Ronald E. Merrill copyright 1997

There is a game that is played by young children and by scientists. When a child learns the use of the word "why", he soon advances to the game of asking endless "why"s. To each answer or explanation offered by his exasperated parent, the child asks again, "Why?" Eventually, of course, the answer must be either, "I don't know," or "It just is, that's all."

Scientists play this same game, in a somewhat more sophisticated way. As Aristotle pointed out, the question "why?" is the way we ask for an explanation in terms of causality. The answer to "why?" generally begins with "because." But science is simply the systematic investigation of causality. And so scientists also generate chains of "why"s.

For instance, we may observe that sodium metal reacts violently with water. Why? Because sodium is a very electropositive metal. Why? We may explain this in terms of the electronic shell structure of atoms, which in turn may be explained in terms of the rules of quantum mechanics and the behavior of elementary particles such as electrons.

And then? What conclusion--if any--may we expect for such a causal chain? Might it end in a primary or first cause, which explains, but itself has no explanation? This is an old and fundamental issue in metaphysics. But it is also a quite practical problem in science. When we begin to search for a deeper level of explanation than we have yet achieved, how do we know that there is such a deeper level? Perhaps there is no reason for the laws of quantum mechanics; perhaps they just are, that's all. But how can we know? After all, we could have said there was no reason for sodium to react with water, it just does, that's all. If we had done so, we would have learned very little about chemistry. Philosophers refer to a truth which has no explanation as a "brute fact." But must we submit to being bullied by brute facts?

The problem for scientists (and children), then, may be expressed as follows: Given any true statement about reality, we may ask why it is true. The answer to this question must also be a true statement about reality, to which we may address the same question. Is there any way to avoid an infinite regress? If so, how? Must we at some point inevitably encounter a brute fact? If so, by what means do we recognize it as a brute fact?

Another way of looking at this is to invoke the Principle of Sufficient Reason (PSR). This states that for every fact there is a sufficient reason, that is, an explanation that "is both correct and fully satisfying and informative." (van Inwagen 1993, 100-101) This is to say, for every "why" there must be an answer. But is this true?

What areas are problematic for the PSR? What sorts of things are candidates to be brute facts? "Accidental" facts--for instance, that my car was not running right last month--are not. We believe that in general the "ordinary" phenomena we encounter always can be explained. We seem to invoke brute fact for only three kinds of phenomena:

(1) An occurrence which is due to metaphysical chance must, almost by definition, be a brute fact. And metaphysical chance seems to be demanded by the laws of quantum mechanics. For instance, if we have a sample of radium, the decay of a specific radium atom within that sample appears to be a purely chance event. There is an explanation for why it decays, but no explanation for why it decays at some particular time. There have been attempts to remove metaphysical chance from quantum mechanical phenomena by developing "hidden variable" theories (Jammer 1974, 252-339). Currently these alternatives seem to have little credibility among physicists, but it is possible that they could be put on firmer foundation.

(2) When a rational being commits a volitional action, we might be inclined to consider it, at least in a sense, a brute fact. It could be argued that if it could be explained, it would be deterministic, not volitional. (For a discussion of a physical model of volition, and how it relates to the next problem, see Merrill 1993.)

(3) The strongest candidates for brute facts are the physical laws of nature. Thus a physicist could explain the slow rotation of the sun as being due to the absorbtion of angular momentum by the planets. In doing so, he would invoke conservation of angular momentum as part of the explanation. But he would most likely consider the Law of Conservation of Angular Momentum itself to be a brute fact.

For the scientist, as for the little child, there is a strong inclination to believe in the Principle of Sufficient Reason as an absolute: For every "why" there is an answer. Philosophers, on the other hand, have tended to deny the PSR; indeed, it has even been asserted that the PSR can be disproved. The argument is made that if the PSR is true, there can be no contingent facts, yet we believe some facts are contingent. However, since "contingent fact" is defined in a way that makes it equivalent to "brute fact," a certain circularity seems to creep into this proof (van Inwagen 1993, 104-107).

In this essay I will argue that the PSR can be maintained without exception, provided we are willing to make two concessions: First, we must accept that "explanations" will sometimes, though not always, be weaker than we are accustomed to in science. Second, we must accept certain radical revisions in our notions of causality. The results may seem a bit strange; but, as Robert Nozick says, "Someone who proposes a non-strange answer shows he didn't understand this question." (Nozick 1981, 116)

In order to set the context, I will give a very abbreviated overview of some of the issues in causality that are relevant to the problem in hand. First, however, let us set out the alternatives that are conceivable solutions to the problem.

First Causes and Their Alternatives

Without, at this stage, specifying what kind of causality we are talking about, we may consider what causal chains might look like. Let us adopt the symbolism "A --> B" to signify "A causes B." Of course we must take into account that something may have more than one cause. For instance, the taste of lemonade is caused (explained) by both the lemon juice and the sugar; we could write such as situation as A + B --> C. Similarly, one cause might have more than one effect, which we could symbolize as A --> B + C. Now, our problem arises from the fact that in reality there are causal chains, where we encounter a situation such as . . . --> E --> F --> G --> H --> . . . and we ask ourselves, how do, or how can, such causal chains begin?

The first possibility is that each causal chain begins with a "first cause" which is a brute fact. At the start of the chain we find the situation: A --> B --> C --> . . . In this model A is a "first cause." It does not itself have any cause; or, alternatively, it may be regarded as being its own cause: A <--> A --> B --> C --> . . . It is not clear that there is any real distinction between these two ways of looking at a first cause.

This is the approach taken by Aristotle; we must always come to a first cause. In this, Ayn Rand, it is asserted, follows Aristotle: "Rand fully accepts the view that explanations come to an end somewhere, and the somewhere is the nature, the what-it-is-to-be, of a thing." (Den Uyl and Rasmussen 1984, 4-5) In this view, entities themselves are brute facts. But in fact Rand diverges from Aristotle here. For Aristotle does indeed say that it is meaningless to ask, eg, "why man is man." But he goes on to point out that one can sensibly ask what it is that makes a certain type of animal a man (Aristotle 1963, 101-103). He therefore finds himself invoking an "Immovable Mover" as first cause (Aristotle 1963, 120-132).

Rand seems to simply refuse to take the final step with Aristotle. Entities are what they are because of the nature of existence, and the nature of existence does not need explanation. Existence exists, and it is meaningless to ask why it exists, or why existence takes the form that it does. There is, therefore, no need to invoke any Immovable Mover. (Rand did not write extensively on metaphysics and it is difficult to pin down her exact position on causality. For some further evidence see Rand 1995, 19-21.)

A second expedient is what we might call the necessary universe argument. Here we invoke first causes, but we do not accept them as brute facts. Instead, we assert that A, the first cause, is necessarily true; that is, ~A would involve a logical contradiction.

This is the implied metaphysic of some physicists who seek the grail of the "theory of everything." They would ultimately like to find a physics which is logically necessary. This approach has little support among scientists, let alone philosophers. Still, it is interesting to note that in recent years basic theories in physics have been more likely to be discarded because an internal logical contradiction was found than because of conflict with experimental evidence.

A third alternative is simply to accept an infinite regress. The chain of causes, in this view, has no beginning; everything has a cause, which in turn has a cause. It's turtles all the way down, so to speak.

This solution was repugnant to Aristotle, who indeed considered it quite impossible; that is why he invoked a first cause. The idea of infinite regress is not so automatically unacceptable to the modern mind, since the calculus has made it familiar to us. However, an infinite regress of causes requires an actual infinity of phenomena in the causal chain. This is certainly problematic for moderns who, like Aristotle, consider that infinities can be only potential and not actual. Ayn Rand, for instance, falls into this category (Peikoff 1991, 31-32).

Finally, there is a fourth expedient. We might invoke circular causality, in which causal loops are allowed to appear. For instance, we might say A --> B --> C --> A. This has commonly been considered too absurd to be even considered; it can allow causality to operate backward in time and result in "bilking paradoxes" (Nahin 1993, 176-201). Moreover, one might question whether a causal loop really provides an explanation at all. If we do not admit circular arguments, how can we admit circular explanations? (Although Nozick has argued that the two can be distinguished; Nozick 1981, 116-121.) On the other hand, the same objection could be raised to invoking either brute fact or infinite regress; neither can be considered a valid explanation. Further, as pointed out above, a brute fact may be considered as a causal loop with one unit. It turns out, as we shall see below, that the idea of causal loops may yet be viable.

The Development of Causality as a Concept

It was Aristotle who first developed a systematic concept of causality, and he took the common-sense approach that causality is that which we are inquiring about when we ask the question "why?" He considered that there were four ways in which this question could be answered. This has come to be called the doctrine of the four causes. However, Aristotle did not really consider them separate causes, but rather different aspects by which we can describe causality.

Suppose, for instance, we are looking at a billiard table and observe that the eight-ball goes into the corner pocket. We may ask, why did the eight-ball go into the corner pocket?

To our modern mind, the most obvious cause is what Aristotle called the efficient cause. The eight-ball headed for the corner pocket because it had just been hit by the cue-ball. The later event was caused by an earlier event. (Aristotle did not conceive of the efficient cause in exactly this modern way. Lear 1988, 30-33)

But why did the eight-ball, on being struck by the cue-ball, go "click" and head across the table? Why did it not, say, shatter into a thousand pieces, or collapse into mush, or vaporize? Obviously because it was made out of hard plastic rather than some other substance such as gelatin. This too is an explanation of its behavior, the material cause.

Again, why did the eight-ball smoothly roll straight into the pocket, instead of sliding, or wavering around the table? Here we must invoke the formal cause. It rolled because it was round, a nearly perfect sphere, rather than, say, a cube, which surely would have behaved differently. Note that the formal cause includes more than the shape of the entity. For instance, if the eight-ball had contained a metal weight inside, sitting off-center, it would have rolled, not straight, but with a wobble. So the formal cause includes everything about the way an entity is organized (Adler 1978, 38-39).

Finally, why did the eight-ball go into the corner pocket instead of heading into some other direction? The obvious answer is that the player had just announced, "eight-ball in the corner pocket." The eight-ball went where it did and when it did ultimately because that was the intention of the person playing, whose end was to win the game. We have not fully understood why the eight-ball behaved as it did until we take into account this, the final cause.

Aristotle's identification of causality and explanation was taken over by the medieval Scholastics and put to religious uses. In particular, the final cause underwent a sort of rococo elaboration in which virtually all phenomena were explained by God's intent. Actual causality was replaced by a series of "just-so stories."

The development of modern science began with the expulsion of the final cause from "natural philosophy," and the rejection of religious explanations for observed phenomena. This principle became gradually more strict, reaching its culmination with the Behaviorism of Watson and Skinner, which rejected teleological explanations even for human action.

As a corollary to this process, the material and formal causes also came under attack. The material cause invoked "substance" as a cause; but, science asked, what is substance? Since substance was conceived as, so to speak, that part of an entity which is left when all the attributes (form) are removed, substance has no attributes. But in that case, how could it be at all, let alone cause phenomena? As for the formal cause, how can we assign causal powers to universals unless we ascribe some sort of independent reality to universals, as in Plato's theory of the forms? But this seemed much less plausible to scientists than the nominalist conception that universals represent merely our way of classifying phenomena.

That left the efficient cause, and in modern terminology "causality" is limited to this type. Of course the efficient cause cannot fully "explain" phenomena, so that in the end the identification of causality and explanation was quietly abandoned. The way in which scientists explain phenomena by invoking deeper layers of structure in fact bears a strong relation to the way Aristotle used the formal cause, but this is no longer referred to as causality.

One result of this has been some confusion over what we mean by a theory. When we say that we have a "theory" about something, what we mean is that we have an explanation. A theory is an assertion of a causal connection between observed facts. That smokers are more likely to get lung cancer is a fact. That lung cancer is caused by smoking is a theory. But when we have a restricted notion of causality, it is not always clear where the distinction between fact and theory should be made.

The culmination of this process was the now dominant scientific epistemology of Thomas Kuhn. In this view, the search for true explanations ("theories") is a delusion. Science can never offer anything better than an increasingly efficient toolkit of "puzzle solutions." We may use equations or other methods to make predictions, but these do not explain anything; our explanations are merely "paradigms" which will one day be replaced with different but no more valid "paradigms" (Kuhn 1970).

But the question "why?" remains. Human beings invented science because they wanted explanations. Is there no chance that science once again could become "natural philosophy" and satisfy our need to know why things are as they are?

Causality as the Logic of Reality

In the foregoing discussion, I have made an implicit assumption that there is such a thing as causality. Our certainty on this point has, however, been challenged. We normally take it for granted that the same cause under the same circumstances will always result in the same effect; this is what scientists call the principle of "uniformity of nature." But how do we know this is true? Many philosophers assert that our belief in the uniformity of nature is simply a matter of faith. (A prominent modern example is Karl Popper. For a critique, see Dykes 1996, 5-11.) Following the argument of David Hume, they point out that "inductive reasoning" from past events can prove nothing. Even though, in our previous experience, the impact of the cue ball always results in motion of the eight-ball, we have no grounds for assuming that will happen next time. It is just as possible that the eight-ball will do nothing, or explode, or turn into a toucan.

Of course, Hume himself freely conceded that one would have to be a madman not to believe in the uniformity of nature. He was merely pointing out that he could see no way to prove what he, and every sensible person, knew was true. This ought to suggest to us that causality is axiomatic.

All phenomena of causality reflect the logic that controls the behavior of entities. This is reflected in the very way we speak of it. We easily recognize that the statement, "What was the cause of that event?" and the statement, "What was the reason that event happened?" are equivalent questions. Causality is identity as applied to action. Causality is what makes the universe make sense; it is the way we describe the underlying logic of reality. We cannot prove causation from any observation, for both proof and observation can exist only because causality is valid. There is no standpoint outside of causal reality from which we can address such issues; for the very fact that we seek proof, or make observations, shows that we have already accepted causality as axiomatic.

Causality as Axiomatic

Causality is axiomatic. I have previously argued that an axiom is a statement that is both undeniable and inescapable (Merrill 1995). If we examine how this applies to causality, we may be able to clarify our understanding of the concept.

We begin by reviewing two other axioms. The axiom of existence states that there are entities ("existence exists"). Something is; it cannot half-exist, or exist for one person but not for another. The axiom of identity states that each entity has a specific and unique identity. Something is what it is; it cannot be everything, or nothing in particular, it cannot be something to one person and something else to another.

We can see that this is axiomatic because it is inescapable. Any statement we make about reality must be a statement ascribing some characteristic ("predicate") to something that exists. This is to say, making a statement about the identity of the entity. To contradict this, even to say something like, "everything is changing," assumes identity; for "change" identifies a difference between what something is at one moment and what it is at a different moment, but if it had no identity there would be nothing to change. Another way of looking at it is that identity is the denial of subjectivity. The subjectivist cannot say anything meaningful without assuming identity. For instance, if he says, "Reality is one thing to you, another to me," he is already assigning separate, objective identities to "you" and "me".

The axiom of causality states that the identity of each entity depends on logical relationships to other entities. When we see an apple, we know that it was once an apple blossom, and not, say, a cherry blossom. When we see the cue ball headed toward the eight ball, we know it will strike it and that the eight ball will move in response; the cue ball will not pass through it like a ghost.

We are constantly faced with the need to identify connections in the behavior of entities that are separated either by space or by time. Thus you and I are different entities, and what I do may affect you--for instance, based on what I say, you may decide to study philosophy. But also, if I say, "I am a different person from what I was 30 years ago when I was in college," you will find this quite coherent. And you can easily believe that something I did then--such as choosing to major in chemistry--affects the way I am now.

Identity is axiomatic over the domain of all statements about existence. Causality is axiomatic over the domain of all statements about relationships. A better way to put it, perhaps, is that whereas existence and identity are restrictions on individual propositions, causality is a restriction on all propositions about reality taken together, viz, that they must be logically consistent.

Causality, then, is both undeniable and inescapable. Suppose, for instance, that a Humean skeptic says, "You believe in causality because you observe regularities in nature, but your belief is not justified." But this statement includes an ascription of causality, viz, that observation of regularities in nature causes people to tend to believe in causality. The skeptic might attempt to escape by revising his assertion: "I do not believe in causality because there is no justification for it." Once again, though, he is invoking causality; he has a belief (or does not have a belief) "because." And indeed, an assumption, explicit or implicit, of causality is quite inescapable as long as he provides any reason for what he believes. The only way to avoid the presumption of causality is to assert only the arbitrary--ie, that which, concededly, one has no reason to believe. (On the arbitrary, see Peikoff 1991, 64-67.)

Causality and Time

One of the reasons modern science abolished Aristotle's final cause is that it has the peculiar property of working backward in time. When something happens right now, the final cause may be something that is in the future. For instance, I turn off my computer so that I can go have dinner. How can my dinner, an event half an hour from now, cause the event now? Note that, for one reason or another, I might change my mind and not even eat dinner. In that case, the turning-off of the computer would be caused by something that didn't even happen!

Problems of this sort impelled scientists to discard the notion of final causation, even if it meant completely abandoning teleology. In fact, teleology became somewhat of a dirty word in the sciences by the mid-Twentieth century. Of course, one may attempt to restore teleology by accounting for it without using the final cause, an exercise demonstrated by Harry Binswanger (Binswanger 1990).

For a while it seemed that restricting all causation to the efficient cause would eliminate the awkward problems that grew out of the traditional four causes. The development of modern physics, however, revived the whole problem of time-reversed causality. This first showed up in the Einstein-Podolsky-Rosen (EPR) paradox. A simple thought experiment (later confirmed in the laboratory) showed that quantum mechanical phenomena resulted in "causality violations" (that is, in effect, time-reversed causation) in the context of special relativity (Jammer 1974, 159-251).

The problem became even more forceful and explicit when another thought experiment was proposed by physicist John Wheeler. (This is sometimes referred to as the Wheeler "delayed choice" experiment and is described in Wheeler 1978. For a simplified account for laymen, see Penrose 1989, 286-287. The experiment has actually been carried out in a modified form, as reported in Helmuth 1987.) We will consider here a slightly revised and simplified form of the thought experiment, which we might call the Baseball Diamond Thought Experiment (BDTE).

Here's how it goes. It's a well-established fact that quantum-mechanical entities can have either particle or wave properties. A photon (light particle), for instance, will act like a particle if it is measured with a particle-measuring device such as a photon multiplier. It will act like a wave if it is measured with a wave-measuring device such as an interferometer. This is peculiar enough, but now consider the following thought experiment.

We have a large baseball diamond. Photons are fired, one by one, at home plate from a point way out in center field. At second base they encounter a half-silvered mirror. There are ordinary mirrors at first and third bases, and a detector at home plate.

Now, if we have a particle detector at home plate--a dual photon multiplier, one arm on the first base side, one on the third base side--here's what happens. The photon (presumably--we don't actually see it till it gets to home) hits the half-silvered mirror at second base. It bounces randomly toward either first or third, hits the mirror there, and goes to home. Sure enough, half the time we get a "hit" on one arm of the photon multiplier, and half the time on the other.

If we put in the wave detector at home plate, something different happens. The photon, behaving as a wave packet, is split in half at second base. The two waves are reflected off the first and third base mirrors, and combine at home plate in the interferometer to give a classic wave interference pattern.

Now, here's the gimmick. The photon in this experiment has to commit itself to particle-like or wave-like behavior; it can't do both. And it has to make that commitment when it is at second base. If it acts like a particle it goes either to first or to third; if it acts like a wave it goes to both first and third. But it takes time to get from second to home. What if we switch the detector after the photon has passed second?

What is found experimentally (the actual experiment has been done, though in a modified form) is that the photon cannot be fooled. It always behaves the right way for the detector that's in place--or rather, that will be in place--when it reaches home plate.

This thought experiment floodlights the really serious issue in quantum mechanics: the breakdown of classical notions of efficient causality. What is to be done? There are basically three ways to explain the Baseball Diamond Thought Experiment:

1. Existents do not have any identity until they are "measured" (ie, perceived by a conscious mind). The photon simply didn't have either wave or particle identity until it was measured at home plate. This is the Copenhagen Interpretation or, we might say, Umpire Theory ("It ain't a ball, it ain't a strike, it ain't nothing till I calls it.")

2. Every time a quantum mechanical chance phenomenon occurs, the entire universe splits into two or more universes, in each of which one of the possibilities actually occurs. This is the so-called Multiple Universes Interpretation.

2. Causality can operate backwards in time, at least under certain circumstances. Our decision to use a certain type of detector causes the photon to behave differently at an earlier time.

The Copenhagen Interpretation relies on a Positivistic principle: If something cannot be observed, even in principle, it does not exist. In the BDTE, the behavior of the photon at second base cannot be directly observed (any attempt to do so will change the experiment and its outcome). So it's pointless to pretend that the photon is "really" there.

Note that the Copenhagen Interpretation is not just a denial of objectivity. It raises the issue of what we precisely mean by "objectivity." Objectivity, we usually think, means that the world exists independent of our knowledge of it. The Copenhagen argument is that we never have knowledge without observation (perception, broadly defined); that observation is interaction with the object being observed; and that all interactions are mutual. We therefore change reality by the act of observing it; so "objectivity" in the traditional sense of the word is impossible.

This runs into problems, though, with the BDTE. We know the photon really does do something at second base (since if we change the equipment there, the results of the experiment change). It really is there, it really does act there, and the action it takes there affects subsequent events, so how can we say that its identity is not fixed until it gets to home plate?

The Multiple Universes Interpretation also has problems with the BDTE. Suppose we have the particle detector in place. Then the Multiple Universes model says that when the photon reaches second base the universe splits into two new universes, one in which the photon goes to first base, and one in which it goes to third. We happen, let's say, to live in the universe in which it goes to first, and we therefore find that the photon approaches from that direction. Our duplicates in the other universe find it comes in from third. (If we have the wave detector in, no splitting of the universe is necessary.)

But how does this model account for the choice of wave vs. particle? Let us say that the universe splits, when the photon is at second, into three new universes: One in which the photon behaves as a wave, and the two in which it behaves as a particle. How is it that in the wave-choice universe we inevitably choose to have the wave detector in, and in the two particle-choice universes we inevitably choose to have the particle detector in? The Multiple Universes model cannot account for this.

This is why we might consider the desperate expedient of accepting the third alternative. Huw Price argues for biting the bullet and accepting backward causation (Price 1996). I cannot agree with his justification for this--which is based on an a priori reverence for the principle of microscopic reversibility in physics--but I agree with his argument that this is the best way to escape from the paradoxes of quantum mechanics.

If we permit backward causation (subject to certain restrictions), all the conceptual detritus of quantum mechanics can be cleaned up. The EPR Paradox simply dissolves. Schroedinger's Cat is no longer in limbo. The whole problem of the "collapse of the wave function" simply disappears.

We ought also to note that even outside the realm of quantum mechanics, physics finds itself encountering backward causation. The basic equations of electrodynamics allow for time-reversed solutions; these are usually simply ignored, but it has been proposed that they can have physical consequences (Nahin 1993, 217-220, 224-231). It has been known since at least 1921 that the equations of General Relativity allow for backward causation under certain circumstances, and physicists have proposed designs for actual time machines (Nahin 1993, 62-64).

If backward causation can occur, then we can encounter causal loops. This raises the spectre of "bilking paradoxes." We normally think of the past as fixed; but suppose we used backward causation to change the past?

The simple answer is that backward causation can be restricted so that this cannot occur. As we saw in the BDTE, we cannot bilk the photon. The same seems to be true of other "time travel" thought experiments from General Relativity. It is notable that relativistic singularities, which can in principle generate any phenomenon no matter how absurd, are always hidden behind an event horizon which separates them from the visible universe. Penrose has called this "cosmic censorship."

Backward causation requires us, however, to think somewhat differently about time. In the classical view, the present is a knife-edge separating a past, which is absolutely fixed, from the future. We now will have to see the past as partly but not entirely fixed, and the present as a wide, fluid boundary between the past and future. This is compatible with relativity theory, which, by restricting the idea of simultaneity, forces the "present" to be different for observers in different frames of reference.

The Neo-Design Argument and the Anthropic Principle

We often find intimate connections between the quantum world of the very small, and the cosmological world of the very large. In current models of the Big Bang, basic physical laws are attributed to quantum-mechanical fluctuations that occurred during the very earliest instants of the universe. In recent years new arguments have appeared to explain things that once were thought to be brute facts. I want to indicate how these arguments relate to the issue of time-reversed causation. First, however, we must summarize the state of the argument.

In 1974, an article by Brandon Carter brought into focus what is now called the Anthropic Principle (Carter 1990). The basic issue may be set out as follows. In the fundamental laws of physics, a number of important constants, such as the speed of light, the gravitational force constant, and so on, appear. These quantities appear to be "written in by hand," as physicists say. There is no evident reason why they ought to have the specific values they have, and so they have traditionally been regarded as brute facts.

However, the question may be asked: If one or more of these constants were to have a different value, how would physics be different? It turns out that even very slight changes in these fundamental constants would dramatically alter the nature of the physical world. Some examples (Leslie 1989, 25-56):

The initial rate of expansion out of the Big Bang must be exactly chosen to within not more than one part in a million; a little slower and the universe would immediately collapse again, a little faster and matter would be so dispersed that no stars could form. (The "inflationary scenario" has been claimed to solve this problem. However, it appears that inflation itself must be equally fine-tuned to produce a universe with galaxies and stars. See discussion and references provided by Leslie 1989, 29-32.)

Were the weak nuclear force just slightly stronger all matter in the universe would quickly convert to its most stable form, iron, precluding the formation of fusion-powered stars. On the other hand, a slight weakening of the weak nuclear force would have resulted in an absence of hydrogen coming out of the Big Bang; only helium would have been formed.

The strong nuclear force is also fine-tuned. If it were even as much as one percent stronger or weaker, a peculiar nuclear resonance which allows fusion of helium into carbon would no longer exist, and carbon would not be formed in the universe.

The mass of the neutron is larger than that of the proton by about one part in a thousand. If it were just slightly heavier, neutrons would be too unstable to exist within nuclei and hydrogen would be the only chemical element. If it were just slightly lighter, protons would decay to neutrons and the universe would consist only of neutron stars and black holes.

The interesting thing is that only certain very tightly constrained values of the fundamental constants are compatible with a universe in which life could exist. And these are precisely the values that they have. In fact, the universe seems to be "fine-tuned" to permit life. At a still deeper level, one can consider what might be a "physics" in the most abstract possible terms. It turns out that a parameter "lambda" can be defined which identifies, roughly speaking, the mobility of information which results from a given set of laws of physics. It can be shown that only a physics for which lambda has a tightly constrained value can result in a universe that will support life (Levy 1992, 108-111; see also Dennett 1995, 175). Surely these "cosmic coincidences" require some sort of explanation.

Some have opted to simply assert that the universe was fine-tuned to permit life. This is a revival, in new and stronger form, of the Argument from Design that was once used by thinkers such as William Paley to assert the existence of God. Paley used this analogy: Suppose, walking through the country, we find a watch lying on the ground. Observing how its complex mechanism is organized to serve a purpose, we would conclude that the watch didn't just accidentally occur; it must have been made by a purposeful consciousness. Since the living organisms in the world are even more complicated and organized than a watch, we must assume that they also had a Maker, that is, God.

Paley's argument collapsed when Darwin's theory of evolution by natural selection showed how organization can arise spontaneously through random biological processes (Dawkins 1986). Now, though, what might be called the "neo-Design Argument" switches from biology to physics, and responds, "You account for the watch by saying that it was made in this factory [evolution]. Very well. Now, who built the factory?" For the process of evolution can develop life only if the underlying physical laws allow for the necessary complex phenomena, and very long times in which they can evolve. But, the neo-Design proponents argue, the necessary laws of physics required for this are far more complex, logically organized, and delicately adjusted than any watch, or any living organism. Surely this is a powerful argument for intelligent design.

The problem with the neo-Design Argument is not a lack of logical coherence, but the fact that it does not provide a real solution. An analogy may be helpful. Suppose we are faced with a differential equation that is difficult to solve. A mathematician offers to help, and he shows how, by a series of manipulations, we can tranform the equation into an integral equation. His mathematical logic is perfectly sound; unfortunately, the integral equation is as difficult or perhaps even more difficult to solve than the original differential equation. The neo-Designers have done much the same thing; they have explained why our universe is, and is as it is. But now they leave us with no explanation for why God is, and is as He is. After all, suppose that, like Stapledon's character, we could come face-to-face with the Star Maker and ask Him the "why" of the universe He has created (Stapledon 1968). We still would not know the "why" of the Star Maker.

The neo-Design Argument invokes a first cause. If we reject it, what are our alternatives? So far, the idea of assuming an infinite regress of layers of physical laws has not received much support. More popular have been appeals to a logically necessary universe, which come in two flavors.

As we previously mentioned, some physicists seek a Theory of Everything which will show that the laws of physics are logically necessary. The universe is as it is because if it were any other way, that would lead to a contradiction. This program does not currently appear promising. It is true that physics is now understood to be more constrained by the requirement of logical consistency than was once believed. However, currently there is no good theoretical basis for a priori assignment of all physical constants. On the contrary, in the current model, crucial constants are set by random fluctuations as forces "freeze out" of the cooling Big Bang.

An alternative called the Weak Anthropic Principle has attracted more favor. The argument is simple: Since we, conscious beings, are here to observe the universe, the universe must be such that conscious beings can exist. Thus the cosmic coincidences are not surprising; for if they had not occurred, there would be no consciousness to comment on how ordinary the universe was.

The difficulty with the Weak Anthropic Principle is that it requires us to believe in the Multiple Universes Theory. This is argued forcefully by John Leslie, who gives us the following analogy (Leslie 1989, 9-11): Suppose we go fishing in a lake, using a net that can only catch fish that are exactly 23.2576 inches long (though it catches those very efficiently). We indeed catch a fish, which of course is 23.2576 inches long. Now, if we do not believe that some intelligence deliberately designed the fish to fit the net, or the net to fit the fish, we must believe that the lake contains many fish, of which at least one happened to be the right length. We could not plausibly believe that the lake contained only one fish, which just by coincidence happened to be 23.2576 inches long.

The necessary multiple universes can in principle be generated. In this view, the quantum fluctuations in the early stages of the Big Bang result in a very large, if not infinite, number of universes, each with different physical constants. We, presumably, inhabit one of the very tiny fraction of these universes that are suitable for life.

The Multiple Universes Interpretation, though, has always seemed rather contrived, and it is vulnerable to a number of cogent criticisms. For one thing, since the other universes are held to be unobservable from our situation, the theory seems to be intrinsically neither verifiable nor falsifiable. Furthermore, as we have seen, the Multiple Universes Interpretation cannot even account for microscopic quantum phenomena, such as the BDTE, in the first place. Even so, it has been proposed that multiple universes not only can exist, but can reproduce and even evolve (Smolin 1997).

There is a final alternative that has been proposed: the Strong Anthropic Principle. What if we simply come out and say that the cosmic coincidences took place because they make it possible for the universe to hold life and intelligence? If we set aside the neo-Design Argument (which comes to the same conclusion), the only way we can hold this position is to assert that causality can work backwards in time. That is, the presence of life or intelligence in the present-day universe must somehow cause the quantum fluctuations during the Big Bang to "freeze out" into an appropriate set of physical laws--which in turn make the presence of life possible.

There are three strong objections to the Strong Anthropic Principle. First, it invokes backward causation. Second, it relies on a causality loop. And third, it requires additional explanatory efforts. Let us take these in turn.

Since conditions suitable for life did not develop until long after the Big Bang, if the cosmic coincidences were caused by life the causation must have worked backward in time. But we are not necessarily entitled to reject this notion out of hand. As we saw from the Baseball Diamond Thought Experiment, we may have reason to conclude that causality can work backward in time, at least under certain restrictions. Furthermore, it is clear from the BDTE that reversed causation can operate over very long time spans. In fact, by making the baseball diamond large enough, we can cause action in the photon billions of years in the past.

We also may feel some discomfort because we are postulating a causal loop: The laws of physics take on a specific form (A); which leads to formation of a very specific kind of universe (B); in which life and consciousness are possible and perhaps inevitable (C); and the existence of life and/or consciousness causes (backward in time) the Big Bang to result in certain physical laws (A). So, in the notation we previously adopted, A --> B --> C --> A.

It might be argued that a causal loop does not really explain why an event happened. It introduces a circularity into our logic of reality. So it is no better than a brute fact; but then again, it is no worse!

Naturally we must also be concerned about bilking paradoxes. Could we somehow interfere with the causation that resulted in our universe, and retroactively change it into one that had different physical laws? At present this question is unanswerable, but our experience with quantum mechanics and relativistic physics suggests that time-reversed causality is always subject to restrictions that will prevent bilking paradoxes.

A final objection to the Strong Anthropic Principle is that it fails to give us any hint of how the necessary time-reversed causal effect is produced. It could be argued from a subjectivistic, primacy-of-consciousness viewpoint. However, it will be seen that I am doing nothing of the sort. The Strong Anthropic Principle is quite compatible with the notion that the outcome of the Big Bang, and the quantitative nature of the laws of physics, are caused by the life/consciousness-bearing nature of our present universe. And, specifically, there is nothing outlandish in the hypothesis that the detailed process of this time-reversed causality is as objective, physical, and real as photons bouncing off mirrors.

Of course the actual mechanism must be vastly more complicated than that of the BDTE. We have not even a starting point on a specific theory of how such a process could operate. That may be considered a drawback of the hypothesis, but it may also be considered an opportunity. For one thing we seek from a hypothesis is that it should generate interesting questions to investigate.

The Limits of Causality

Let us conclude, then, by returning to the question with which we began: What would, or could, we consider a satisfactory explanation for for allegedly brute facts? Could we defend the proposition that the universe is logically necessary--that everything is as it is because if it weren't that would lead to a logical contradiction? This would be going Liebniz one better; we live not just in the best of all possible worlds, but in the only possible world. This has the attraction that it fully defends the PSR. Unfortunately, the current state of physics provides no support for it. So far as we can see, physics is "underconstrained." We could only defend the logically necessary universe by finding some novel logical law to serve as an additional constraint. (For instance, that we logically must live in the most interesting of all possible universes, or that consciousness must be possible.) If we cannot carry out such a program, we must turn to the alternatives: first causes; infinite regress of causes; or circular causation.

All three of these descriptions are profoundly unsatisfactory, in that they do not give us the conviction that we have a "real" explanation. In fact, they all lead to this problem for much the same reason. As we've seen, a first cause may be regarded as just a very small causal loop: A --> A --> A or A <--> A. Again, we may ask whether there is any fundamental distinction between a causal loop and an infinite regress of causes. For, after all, the standard objection to circular reasoning is that it leads to an infinite regress. If we write a causal loop as A --> B --> C --> A --> B --> C --> A --> . . .

we can see how it can also be regarded as an infinite regress.

However, there are excellent arguments for accepting the validity of time-reversed causation. Its drawbacks are tolerable. True, it can lead to causal loops (though it does not always have to do so). However, as we have seen, causal loops are no more intolerable than alternatives such as brute facts. Time-reversed causality can, and apparently does, operate under restrictions that prevent paradoxes even when causal loops occur.

On the positive side, time-reversed causation helps us to explain all three of the causality gaps that we identified at the beginning of this essay. The probabilistic character of quantum mechanics can be removed by "hidden variable" theories, as has long been known. These approaches have been rejected because it can be shown that hidden variable theories inevitably must lead to "causality violations," that is, time-reversed causality. But if we accept this, the problem disappears. Similarly, we have now the option of returning to some version of the final cause. This enables us to explain, at least at some level, volitional choice. Finally, time-reversed causation gives us, via the Strong Anthropic Principle, a potential explanation, even if it is not as robust as we might like, for an otherwise utterly intractable problem, the laws of physics.

At a very fundamental level, this view of causality acknowledges an existential role for consciousness. Physical scientists have tended to have, so to speak, a Deistic view of consciousness. It stands outside the universe of action, looking, but doing nothing. It observes causation, but causes nothing itself. But we need to take a stronger focus on how consciousness acts to cause events.

It is because consciousness exists in the universe that the question "why" is asked. Paradoxically, it appears that just because consciousness exists, some of its questions cannot be fully answered. There is something irreducible about conscious intention which requires us to accept explanations that are ultimately limited. This is the price we pay for being the kind of creatures who can care about the Principle of Sufficient Reason. Once we understand its limits, though, perhaps we need not be disturbed by them. As usual, Aristotle says it best:

Our treatment will be adequate if we make it as precise as the subject matter allows. (Aristotle 1963, 287)

Adler, M. J. 1978. Aristotle for Everybody. New York: Bantam Books.

Aristotle 1963. The Philosophy of Aristotle. R. Bambrough, editor. New York: Mentor Books.

Binswanger, Harry 1990. The Biological Basis of Teleological Concepts. Los Angeles: Ayn Rand Institute Press. (First published 1976.)

Carter, Brandon 1990. Large Number Coincidences and the Anthropic Principle in Cosmology. In John Leslie (ed.), Physical Cosmology and Philosophy. New York: Macmillan. (First published 1974.)

Davies, Paul 1982. The Accidental Universe. Cambridge: Cambridge University Press.

Davies, Paul 1992. The Mind of God: The Scientific Basis for a Rational World. New York: Simon & Schuster.

Dawkins, Richard 1986. The Blind Watchmaker. New York: W. W. Norton.

Dennett, Daniel C. 1995. Darwin's Dangerous Idea: Evolution and the Meanings of Life. New York: Simon & Schuster.

Den Uyl, Douglas J., and Rasmussen, Douglas B. 1984. Ayn Rand's Realism. In Den Uyl, Douglas J., and Rasmussen, Douglas B. (eds.), The Philosophic Thought of Ayn Rand. Chicago: University of Illinois Press.

Dykes, Nicholas 1996. A Tangled Web of Guesses: A Critical Assessment of the Philosophy of Karl Popper. London: Libertarian Alliance.

Helmuth, T., Walther, H., Zajoonc, A., and Schlich, W. 1987. Physical Review A 35, 2532.

Jammer, Max 1974. The Philosophy of Quantum Mechanics. New York: Wiley.

Kuhn, Thomas S. 1970. The Structure of Scientific Revolutions. Chicago: University of Chicago Press.

Lear, J. 1988. Aristotle: The Desire to Understand. Cambridge: Cambridge University Press.

Leslie, John 1989. Universes. London: Routledge.

Levy, Steven 1992. Artificial Life. New York: Pantheon Books.

Merrill, Ronald E. 1993. "On the Physical Meaning of Volition." Objectivity 1(5), 69.

----- 1995. "Axioms: The Eightfold Way." Objectivity 2(2), 1.

Nahin, Paul J. 1993. Time Machines: Time Travel in Physics, Metaphysics, and Science Fiction. New York: American Institute of Physics.

Nozick, Robert 1981. Philosophical Explanations. Cambridge, Massachusetts: Harvard University Press.

Peikoff, Leonard 1991. Objectivism: The Philosophy of Ayn Rand. New York: Dutton.

Penrose, Roger 1989. The Emperor's New Mind: Concerning Computers, Minds, and the Laws of Physics. New York: Oxford University Press.

Price, Huw 1996. Time's Arrow and Archimedes Point. New York: Oxford University Press.

Rand, Ayn 1995 (ed. Robert Mayhew). Ayn Rand's Marginalia. New Milford, Connecticut: Second Renaissance Books.

Smolin, Lee 1997. The Life of the Cosmos. Oxford: Oxford University Press.

Stapledon, Olaf 1968. Last and First Men and Star Maker. New York: Dover. (Star Maker first published 1937.)

van Inwagen, Peter 1993. Metaphysics. Boulder, Colorado: Westview Press.

Wheeler, J. A. 1978. In A. R. Marlow (ed.), Mathematical Foundations of Quantum Theory. New York: Academic Press.

Whitrow, G. J. 1980. The Natural Philosophy of Time. Oxford: Clarendon Press.