The gradual separation of physics from philosophy over the last 400 years has had its advantages and disadvantages. On one hand, the autonomy of quantitative physics has enabled a robust development of mathematical theory which corresponds to repeatable empirical measurements to an astoundingly high degree of accuracy. On the other hand, a purely mathematical physics does little to impart real understanding of the ontology of physical substances or the nature of physical causation. Physicists of unquestionable mathematical genius are nonetheless capable of astonishing ignorance in even the most elementary aspects of the lost science of natural philosophy, sophistically conflating efficient and final cause, potentiality and actuality, mathematical form and corporeal matter. In the absence of a robust theory of natural philosophy, the philosophy of physics is a virtual free-for-all, with no theory too absurd to be accepted. While we can hardly hope to address all cases of philosophically misguided meta-theories of physics, we may address some common fallacies by analyzing a particular class of physical theorizing, that which makes use of an “anthropic principle.”
An “anthropic principle” makes use of the fact of human existence to explain the physical conditions of the universe. Physicists and biologists alike make use of such principles, which may be characterized in several forms. Several definitions proposed by John D. Barrow and Frank J. Tipler (1986) have come into common use:
Weak anthropic principle (WAP): "The observed values of all physical and cosmological quantities are not equally probable but they take on values restricted by the requirement that there exist sites where carbon-based life can evolve and by the requirements that the Universe be old enough for it to have already done so."
Strong anthropic principle (SAP): "The Universe must have those properties which allow life to develop within it at some stage in its history."
Final anthropic principle (FAP): "Intelligent information-processing must come into existence in the Universe, and, once it comes into existence, it will never die out."
The first two principles make use of the self-evident fact that intelligent human life does, in fact, exist, therefore the Universe “must” have properties that allow the development of human life. This is certainly true in some tautological sense, but the nature of the necessity underlying that “must” is a more delicate matter, which requires us to be very clear in our thinking about causality and conditional probability. If we are not careful philosophically, as many scientists are not, we will make the logical error of reversing cause and effect, and effectively say that the existence of human intelligence explains or causes the conditions of the primitive universe. The existence of human intelligence is certainly true a posteriori, but there is no a priori necessity for the existence of human intelligence. Thus “anthropic principles” are not explanatory principles, but practical principles for restricting our inquiry to areas that will lead to fruitful results. The final anthropic principle is entirely gratuitous, and rests upon the inverse causation fallacy we have just briefly described.
Anthropic principles are motivated by the observation that many fundamental physical constants appear to be finely tuned to the development of human life, meaning that if any one of these constants were slightly changed, human life, or more broadly, any animal life, could not exist. To many, this is strong evidence of an intelligent, even benevolent, design to the universe. For those who wish to resist such a conclusion, or at the very least, keep theism out of physics, it is better to resort to a pseudo-probabilistic theory that shows there is nothing remarkable about this fine tuning, but it is a mere consequence of the trivial fact that only universes that contain intelligent life can be observed by any one. More ambitiously, some theorists hope to use anthropic principles for the positive development of physical theory, ironically returning in practice to the anthropocentric world-view they have rejected in the context of theistic physics.
Most uses of the anthropic principle evince a failure to understand the meaning of probability. Probability is not an objective absolute, but relative to a determined set of conditions. When all conditions are specified, the only probabilities are one or zero. For example, consider what is the probability that a particular person will live to at least age eighty. That probability will increase or decrease depending on whether we are able to specify the person's gender, race, health, and physical circumstances. These revised computations do not actually alter the person's lifespan, but reflect our increasingly specific knowledge, so that our computed probability more closely corresponds to the particular person we are studying. Since we are studying a particular human being, and not human beings in general, or males in general, or smokers in general, our computed probabilities will only crudely correspond to the person's real state unless we know absolutely everything about the person's life, in which case the probability will be one if he lives past eighty, or zero if he does not. All probability is conditional probability; there is no “real” probability, save 1 if it happens, and 0 if it does not.
When an event occurs, we can say with the benefit of hindsight that the probability of its occurrence is one. This is true of all events, so this fact tells us nothing of the anterior probability of the event. If I have won the lottery, we may say that the probability of my success is now 1. Notwithstanding this fact, my success is still remarkable, because the anterior probability of winning was less than one in a million. This is the conditional probability determined from specifying only the fact that I had purchased a ticket in a lottery for which less than one in a million combinations were winners, and I had no prior knowledge of whether my ticket had a winning combination. There was no prior necessity that I or anyone else should win, but as things turned out, I did.
The weak anthropic principle is certainly true if we regard probability with the benefit of hindsight. We know that, in fact, human sentience has arisen in this universe, so we might restrict the possible values of the universe's fundamental constants to those which permit the sustaining of human life. This is an after-the-fact revision, based on our knowledge that human life does in fact exist. Our revised probabilities are thus based on the condition that human life does in fact exist. This revision does not abolish the fact that, without specifying the condition that human life exists, the anterior probability distribution of values for the fundamental constants might have been very different. Similarly, we have not established any a priori necessity that human life should exist, nor have we shown that there is nothing remarkable in the “fine tuning” of the universe's constants, since we have assumed that human life does exist (assigning it a probability of one), making it a condition of the probability of the selection of fundamental constants.
This exercise of retroactively adjusting the probabilities of different values of fundamental constants (and other “fine-tuned” aspects of the physical structure of the universe) in light of our a posteriori knowledge of human existence helps us determine what these values actually are, but not what they had to have been. There is no a priori necessity that human beings must exist, as we shall see when we examine the strong anthropic principle, but for now it suffices to show that no such necessity is established by the weak anthropic principle. The weak anthropic principle assumes that humans exist, based on a posteriori knowledge, and uses this condition to impose a constraint on allowable fundamental constants. It would be circular logic to invoke this restriction in probabilities as the condition determining human existence, since these restricted probabilities assume human existence as their condition. Returning to the lottery example, I know that I won the lottery, therefore I must have chosen a winning combination, with 100% certainty. It would be sophistry for me to conclude that there was nothing remarkable in my winning the lottery, being a consequence of the 100% necessity of my choosing a winning combination. I know that my combination was a winner only with the benefit of hindsight; at the time of purchase there was no necessity that I should win.
A more obvious criticism of this abuse of the weak anthropic principle is that it inverts cause and effect. The fundamental constants of the universe held their present values long before human beings existed. We know that physical substances acting in conformity with these values are in some ways an efficient cause of human existence (though not necessarily to the exclusion of other causes), since our bodies are made of matter that obeys the laws of physics. If anything, the physical structure of the universe (as represented by fundamental constants and various mathematical relations) are an efficient cause of human existence. Humans obviously cannot be an efficient cause of the physico-mathematical structure of the universe, since our existence arose later in time, and is in fact dependent upon the prior existence of that structure. Furthermore, there is no mechanism by which we could act upon the entire universe and change its structure. Abuse of the weak anthropic principle uses a flawed understanding of conditional probability to circumvent these obvious absurdities. Once we admit that such an argument cannot establish an a priori necessity for human existence, we have shown that the “weak anthropic principle” does not provide any insight into the question of why the universe was formed in a way that was amenable to human life.
It is, of course, possible that human existence is an indirect final cause of the universe's physical structure, but this is not the sort of causation in which modern scientists have interest. The Aristotelian concept of final cause is based on the analysis that an act is defined by its end-point, or the end toward which it is directed. For example, the act of dying ends in death or at least tends toward it, or the act of stopping ends in a stop, or at least tends toward it. That which is already dead does not die, nor would we say that something already at rest now stops, so the act is defined by a goal or end-point, or an inclination toward a goal or end-point. Thus all activity has teleology built into it at a metaphysical level. These “goals” are final causes, the forms toward which a substance tends when it is subjected to a particular act. Final causes can be rather potent arguments for the existence of divine purpose in nature, which is one reason they have been ignored by modern science. Another reason is the fallacy that efficient causes (the only type of cause recognized by modern thought) are sufficient to account for an effect's existence, making final causes superfluous. This way of thinking misunderstands the metaphysics of action and reduces causality to a disjointed succession of events rather than agents and their acts. A critique of this position is best reserved for a philosophical discussion, but for our present purposes it suffices to note that anthropic principles might be legitimately used as explanations of the universe's structure if we are speaking of final causes, not efficient causes.
Stronger statements of the anthropic principle assert that human life must exist, not merely in the simple a posteriori sense that it does exist, but because of some pre-existing requirement. We have already shown that it is circular reasoning to make this argument from probability conditioned by the post hoc observation of human existence. Further, it is absurd to make human existence an efficient cause of that which precedes it temporally and causally, as does the pseudo-probabilistic argument we have considered thus far. Proponents of the strong anthropic principle provide another rationale for the necessity of human existence: without intelligent life, there would be no one to contemplate the universe.
The idea that someone is needed to contemplate the universe can be interpreted in several ways. One is that there have been (potentially or actually) many universes, but only those with sentient life can be observed, thus “explaining” away the seemingly improbable coincidence of a universe fit for human life. Unlike the pseudo-probabilistic argument we considered earlier, this interpretation does not assert an a priori necessity for human existence. Instead, admitting that the anterior probability of a physically possible universe sustaining human life is low, we suggest that there have in fact been many universes, though naturally we can only observe one that has sentient life in it. This argument relies on the supposition that many universes actually exist or have existed; if it refers only to the potential existence of other universes, we encounter the same limitations as the pseudo-probabilistic argument. To expound this last point, suppose there has only been one (or several) actual universe, and the anterior probability that a universe will be able to support intelligent life is extremely low. If this universe cannot support sentient life, no one will observe it, but if by some incredibly good fortune it can support intelligent life, it will be observed. This truism does nothing to explain why our universe has had such incredibly good fortune, as there is no a priori necessity that there should be any universe that is observed. We have not established anything showing that it would be absurd for a physical universe to exist without life, and never be observed. Our own universe existed for eons without any humans to observe it.
Probabilistic arguments using the strong anthropic principle are on stronger ground if we suppose that many universes have in fact actually existed, and only those sustaining human life have actually been observed. This really would account for our seemingly incredible good fortune, and it is the only non-sophistical probability argument using a truly explanatory anthropic principle. Unfortunately, it is predicated on a highly speculative and empirically unverifiable “many-worlds” hypothesis, taken in the stronger form of many actual universes, rather than many possible universes.
Even if we accept the “many-worlds” scenario, there remain serious limitations to its explanatory power. We have no real insight into how a universe develops one structure rather than another, or how it may transition from one set of laws to another, if such a thing is indeed possible. If the many universes are causally connected, we would have to affirm that at some point a universe can change its laws. If they exist in independent realities, it is unclear why we should contemplate them in the same probability space. It is also unclear what prevents universes from disappearing permanently, or what causes them to arise in the first place. We could conjecture that there is some cosmological requirement that “all possibilities,” or infinitely many random possibilities, must manifest themselves, but this is very ill-defined, and there is no reason to exclude the possibility that there could only be one universe, or none. These limitations do not disprove the “many-worlds” hypothesis, but show that, in order for the “many-worlds” scenario to be applied to the strong anthropic principle, we must construct a specific type of “multiverse” with special constraints. Without these constraints, there is no guarantee that sentient life is likely, a priori, to come into existence, for there are many conceivable multiverse scenarios that allow for no sentient life. Then we must ask, why, a priori, should our multiverse have a form that does permit intelligent life? So the same logical difficulty arises as before, and the critique of the anthropic principle has not been circumvented.
Lastly, we may consider the final anthropic principle, which would make the existence of intelligent life or other forms of information-processing ontologically necessary. This is perhaps most easily refuted by the observation that the universe was able to manage itself quite nicely long before human beings came about, and presumably will be able to do so should we cease to exist. Moreover, most of the universe is not under continual contemplation by humans, yet it continues to exist. There are certain interpretations of quantum mechanics that would hold a thing does not exist in a definite state unless it is observed, but this is manifestly inapplicable to macroscopic objects, which was the intended point of Schrödinger's “cat paradox.” There may be deeper philosophical reasons why a thing must be contemplated by some intellect in order to exist, but that intellect need not be that of biological beings.
The various forms of the anthropic principle fail as an explanation of the existence of human life or the conditions of the early universe, with the possible exception of an argument founded upon a well-contrived “many-worlds” scenario. Still, the anthropic principle might be useful for discerning physical conditions conducive to the development of life. Even here, the principle has serious limitations. We barely understand our own mode of biological existence, and only after centuries of painstaking empirical research, so it is unlikely that we would be able to guess at the viability of other forms of life known only to our imagination. We cannot dismiss the possibility of beings of fantastically different bodily constitution, or even those which are utterly incorporeal. Hence, we have strong reason to doubt that the anthropic principle can impose verifiable constraints on the number of spatial dimensions and the mathematical forms of force laws and statistical mechanics which are conducive to sentient life. Conversely, it is equally difficult to show that there exist other scenarios where sentient life would be feasible, since there is fundamentally no way to test such a hypothesis in practice. The anthropic principle, therefore, is of extremely limited practical value.
In reality, we know very little about the anterior probability of human life coming into being, not only because we do not know which conditions are compatible with sentient life, but also because we have limited understanding of the processes by which life came into being. It is true that our particular form of life would have been impossible without “fine tuning” of the universe's structure, but we do not know if there could have been equally fine-tuned, yet fantastically different universes supporting biological life. Anthropic principles fail to explain why our universe is in one such fine-tuned state, when “all things being equal”, it was much more likely to develop into chaos. The many-worlds hypothesis is a contrived attempt to reconcile our elegant reality with blind chance, at odds with our universal observation that reality follows one path or another, but not all paths. A more obvious solution to the enigma is that all things are not equal, and that the path that results in life was preferred from the beginning. Anthropic principles attempt to mask this blatantly teleological principle with the language of probability, illogically “proving” the a priori necessity of the cause from the a posteriori necessity of the effect. When their faulty logic is removed, we are left with pure teleology: the universe was made for man. Anthropic principles, perhaps unsurprisingly, ultimately lead to the anthropocentric view of the universe they are intended to undermine.
© 2006 Daniel J. Castellano. All rights reserved. http://www.arcaneknowledge.org