152 The Cosmologkal Argument
relevance criterion P(7t|e&£) > P(h\k), and so the argument from the existence of a complex physical universe to God is a good C-induetive argument
The argument of the last few pages can now be put in simple words as follows. There is quite a chance that, if there is a God, he will make something of the finitude and complexity of a universe. It is very unlikely that a universe would exist uncaused, but rather more likely that God would exist uncaused. Hence the argument from the existence of the universe to the existence of God is a good C-inductive argument.
8
Teleological Arguments
I understand by an argument from design one that argues from some general pattern of order in the universe or provision for the needs of conscious beings to a God responsible for these phenomena. An argument from a general pattern of order I shall call a teleological argument. (The name 'teleological argument' has usually been used to characterize much the same arguments as 'argument from design'. I am giving 'teleological argument' a narrower use.) I shall deal with teleological arguments in this chapter. I shall deal in Chapter 10 with the argument from the occurrence of provision for the needs of conscious beings, and I shall call such an argument an argument from providence. In the definition of 'teleological argument' I emphasize the words 'general pattern'; I shall not count an argument to the existence of God from some particular pattern of order manifested on a unique occasion as a teleological argument
I begin with the distinction between spatial order and temporal order, between what I shall call regularities of co-presence and regularities of succession. An example of a regularity of co-presence would be a town with all its roads at right angles to each other, or a section of books in a library arranged in alphabetical order of authors. Regularities of succession are simple patterns of behaviour of objects such as someone moving his or her legs in accord with the standard movements of a dance. In both these cases the regularities are produced by humans. The universe is characterized by regularities of both kinds not produced by humans or other embodied beings. There is first the temporal order of the regular successions of events, codified in laws of nature. In books of physics, chemistry, and biology we can learn how almost everything in the world behaves. The laws of their behaviour can be set out by relatively simple formulae that humans can understand and by means of which they
154 Teleologkal Arguments
can successfully predict the future. The orderliness of the universe to which I draw attention here is its conformity to formulae, to simple, formulable, scientific laws. The orderliness of the universe in this respect is a very striking fact about it. The universe might so naturally have been chaotic, but it is not—it is very orderly. And then there is the spatial order of the intricate arrangement of parts in human (and animal) bodies. We have limbs, liver, heart, kidneys, stomach, sense organs, etc. of such a kind that, given the regularities of temporal order, our bodies are suitable vehicles to provide us with an enormous amount of knowledge of the world and to execute an enormous variety of purposes in it (as described more fully in Chapter 6). This is similar to the way in which parts of machines are arranged so as to produce an overall result from the operation of the machine; though—so far—machines intentionally constructed by humans are far less intricate than human bodies.
A ideological argument, whether from temporal or spatial order, is, I believe, a codification by philosophers of a reaction to the world deeply embedded in the human consciousness. Humans see the comprehensibility of the world as evidence of a comprehending creator. The prophet Jeremiah lived in an age in which the existence of a creator-god or gods of some sort was taken for granted. What was at stake was whether there was only one god, and the extent of his goodness, knowledge, and power. Jeremiah argued from the order of the world that there was one powerful and reliable god, and that god was God. He argued to the power of the creator from the extent of the creation: 'The host of heaven cannot be numbered, neither the sand of the sea measured'; he argued that its regular behaviour showed the rehabihty of the creator, and he spoke of the 'covenant of the day and night' whereby they follow each other regularly, and 'the ordinances of heaven and earth',1 and he used their existence as an argument for the trustworthiness of the Jewish God. The argument from temporal order has been with us ever since.
The Datum of Temporal Order
We find the argument from temporal order also in Aquinas's fifth way, which runs as follows:
1 Jer. 33:20 and 25.
Teleologkal Arguments 155
The fifth way is based on the guidedness of nature. Some things lacking awareness seek a goal—which is apparent from the fact that always or most usually they behave in the same way which leads to the best result. From this it is evident that it is not by chance but by intention that they reach their goal. Nothing, however, that lacks awareness tends to a goal, except under the direction of someone with awareness and with understanding; the arrow, for example, requires an archer. Everything in nature, therefore is directed to its goal by someone with understanding and this we call 'God'.2
Aquinas argues that the regular behaviour of each inanimate thing shows that some animate being is directing it (making it move, so as to achieve some purpose, attain some goal); and from that he comes—rather quickly—to the conclusion that one 'being with understanding' is responsible for the regular behaviour of all inanimate things (apart, maybe, from the behaviour for which humans and animals are responsible).
It seems to me fairly clear that no argument from temporal order, whether Aquinas's fifth way or any other argument, can be a good deductive argument For, although the premiss is undoubtedly correct—a vast pervasive order characterizes the world—the step from premiss to conclusion is not a valid deductive one. Although the existence of order may be good evidence of a designer, it is surely compatible with the non-existence of one—it is hardly a logically necessary truth that all order is brought about by a person. And although, as I have urged, the supposition that one person is responsible for the orderliness of the world is much simpler and so more probable than the supposition that many persons are thus responsible, nevertheless the latter supposition seems logically compatible with the data. So we must turn to the more substantial issue of whether the argument from the temporal order of the world to God is a good (C- or P-)inductive argument. Since just the same kind of considerations apply to all other claims that some argument from an observable feature of the world to the existence of God is a valid deductive argument, I shall not in future bother to repeat them where we come to new arguments. I shall assume that no such argument is deductively valid.
But, before considering whether the argument from temporal order is a good inductive argument, I must deal with three preliminary matters. First, there is an objection that this temporal order is
2 St Thomas Aquinas, Summa Theologiae, Ia2.3, my translation.
156
Teleological Arguments
Ideological Arguments
157
not an objective feature of the world but a mere human artefact; the order that we seem to see in the world is order that we impose on it, and is not there independently of our imposition. Put another way, all that this temporal order amounts to, it might be said, is a coincidence between how things have been so far in the world and the patterns that humans can recognize and describe. In tact, however, the temporal order of the world is something deeper than that. We rightly explain our observations so far in terms of laws of nature involving a physical necessity in nature (as I analysed on pp. 28-9), which determine how things behave and allow us to predict how they will behave in future. It is the operation of these simple natural laws that this argument seeks to explain.
Then there is the objection that there is nothing to be explained in the fact that we find an orderly universe—for we could not possibly find anything else. For, unless the universe were an orderly place, we would not be here to comment on the fact (If there were no natural laws, there would be no regularly functioning organisms, and so no Human beings.) Hence there is nothing surprising in the fact that humans find order—we could not possibly find anything else. This conclusion is clearly a little too strong. There would need to be quite a bit of order in and around our bodies if we are to exist and observe the world, but there could be chaos outside the earth, so long as the earth was largely unaffected by that chaos. There is a great deal more order in the world than is necessary for the existence of humans. So humans could still be here to comment on the fact of order even if the world were a much less orderly place than it is. But, quite apart from this minor consideration, the argument still fails totally for a different reason, which can be brought out by an analogy. Suppose that a madman kidnaps a victim and shuts him in a room with a card-shuffling machine. The machine shuffles ten packs of cards simultaneously and then draws a card from each pack and exhibits simultaneously the ten cards. The kidnapper tells the victim that he will shortly set the machine to work and it will exhibit its first draw, but that, unless the draw consists of an ace of hearts from each pack, the machine will simultaneously set off an explosion that will kill the victim, in consequence of which he will not see which cards the machine drew. The machine is then set to work, and to the amazement and relief of the victim the machine exhibits an ace of hearts drawn from each pack The victim thinks that this extraordinary fact needs an explanation in terms of the machine having been
rigged in some way. But the kidnapper, who now reappears, casts doubt on this suggestion. 'You ought not to be surprised', he says, 'that the machine draws only aces of hearts. You could not possibly see anything else. For you would not be here to see mything at all, if any other cards had been drawn.' But of course the victim is right and the kidnapper is wrong. There is indeed something extraordinary in need of explanation in ten aces of hearts being drawn. The fact that this peculiar order is a necessary condition of the draw being perceived at all makes what is perceived no less extraordinary and in need of explanation. The teleologisťs starting point is not that we perceive order rather than disorder, but that order rather than disorder is there. Maybe only if order is there can we know what is there, but that makes what is there no less extraordinary and in need of explanation.
The third preliminary matter is to note the kinds of regularities to which the argument appeals. The regularities of temporal succession in our universe are of two kinds. There are the phenomenal regularities that are very rough probabilistic laws about what happens on perhaps 97 per cent of occasions; and there are the fundamental regularities that explain these. The phenomenal regularities are the macroscopic ones by which humans (and the higher animals) guide their dairy life, ones evident to people without much scientific education. They include such regularities as that seeds when watered often grow into plants, people who do not eat or drink for a month or two die, mushrooms nourish but toadstools poison, arrows shot quickly penetrate human skin, day is followed by night and night by day at approximately similar intervals (as measured by pendulum clocks), and so on But scientists have discovered that these phenomenal regularities are brought about by more fundamental regularities. The phenomenal regularities arise from laws of chemistry about possible combinations of atoms into molecules and the resulting stability of solid objects; and these are brought about by laws of physics governing the electrons, protons, and neutrons of which atoms are made; and these are brought about by the laws governing the quarks of which the protons and neutrons are made and so on. These laws at the latter level include the laws of the four forces (gravity, electromagnetism, strong force, and weak force) constrained by the general requirements of Quantum Theory and the General Theory of Relativity. Probably the laws of electromagnetism and the weak force derive from the more general laws of an 'electroweak1
158 Teleological Arguments
theory, and there is some reason to suppose that, in due course, physicists will discover a 'theory of everything', whose laws have (within science) no more complete explanation; and which explain all physical phenomena. So the physical world is such that it is governed by relatively simple fundamental laws (detenninisric, °r—more likely—probabilistic), concerning the tiny unobservable building blocks of the world, of such a kind that they often lead to laws about the observable behaviour of medium-sized entities. Not all behaviour of physical objects at the phenomenal level is governed by simple regularities—the behaviour of the pendulum is, the behaviour of the weather is not And phenomenal regularities do not concern what always happens, only what happens almost always; and in that way they are very reliable, but not totally reliable. Buildings, bridges, and trees normally remain immovable, but just occasionally they collapse. Peanuts normally nourish, but very occasionally they poison.
It is the (almost entirely reliable) phenomenal regularities that we observe and then use in order to bring about our chosen goals. Regularities (fundamental or phenomenal) have to exist if we are to be able to make a difference to things beyond our bodies. If we are to grow plants, it has to be the case that certain basic actions of ours will have this result But, unless we are to be non-rational creatures, we need to be able to observe phenomenal regularities and learn from them. Humans (and often too higher animals) can
3 It may be urged that we have no reason to suppose that there is a most fundamental law of nature. Maybe law X operates in circumstances C because it follows from L' that it does; and L' operates in circumstances C (which include Q because it folows from L" that it does; lm operates in circumstances C" (which include C") because it follows from X" that it does, and so on ad infinitum. This difficulty can, however, be avoided as follows. Either such a series ends with a most fundamental law that holds in all circumstances, or it does not. In the latter case let us represent the more fundamental laws as conjunctions of laws that hold without exception under specifiable circumstances such as C Thus to say X' holds in C will be to say that X holds in C and Xj in Ci; to say that X" holds in C will be to say as well as this that X, holds in Q. Then the claim that there is an infinite series, X, X', X", etc. is the claim that there exists an infinite series of non-fundamental laws X, Lj, Xi, etc. that hold without exception in circumstances Q Q.Cj, etc; and that, although there is an explanation of the operation of any finite subseries, there is no explanation of the operation of the whole series. That the whole series operates will then be the starting point for a teleological argument; that it operates shows a conformity of the world to order similar to that shown by the'conformity of the world to statable most fundamental laws of nature, which form the starring point for the simpler argument I shall henceforward deal with the simpler argument on the not implausible assumption that there are most fundamental laws of nature.
Teleological Arguments 159
observe seeds being watered, or day being followed by night, and extrapolate to the regularities described as the simplest account of what they observe (that is, they can infer that what holds for the seeds that they have observed holds for seeds generally, and so on.) They can then rely on these regularities to produce effects. They can water seeds and grow plants. Wanting to travel a long distance easily, they can travel by day and not by night (since they know that day will come again soon). Knowledge of such regularities gives to humans choices. Discovering that toadstools poison, they can choose to poison someone by encouraging him or her to eat toadstools; or they can prevent accidental poisoning by toadstools by uprooting toadstools and telling people that they are poisonous; or they can choose not to bother.
Now in order for any of its inhabitants to attain their goals, the universe needs to evince regularity at some level or other. Then those inhabitants will need to have sense organs sensitive to how things are at that level if they are to detect and use the regularities. In our universe both the fundamental regularities and many less fundamental regularities are relatively simple. I have called the latter regularities the phenomenal regularities, because—as far as we know—these are the only regularities to which the rational creatures of our universe are sensitive. But if there are creatures whose senses inform them without the help of apparatus or inference of the location of individual atoms, then they can use the regularities in the behaviour of these atoms to attain their goals. There is, however, no guarantee that regularity at a fundamental level (in the behaviour of fundamental particles) will ensure usable regularity at a less fundamental level. Whether it does will depend on what the less fundamental laws are and on the boundary conditions of the universe. Even in our universe, if the temperature never became low enough for atoms to combine to form medium-sized solid objects, there would be only clouds of gas or liquid that do not as such behave in very simple ways. It is also possible to have laws of nature fairly complex on the small scale, producing many fairly simple regularities on the large scale. For it might be that the boundary conditions of the universe were such that fundamental particles were normally found only in states that easily allowed them to combine into larger objects, and the behaviour of the particles in combination was restricted to some very simple patterns. But clearly total chaos at a fundamental level will lead to chaos at any other leveL
160
Teleological Arguments
TeUological Arguments
161
The argument from temporal order is an argument from regularity at some level or other. And, while the operation of non-fundamental laws may be explained by the operation of fundamental laws, that these are fundamental laws of nature is, like the very existence of a complex physical universe, where science starts from in order to explain other things. It is something 'too big" for science itself to explain.
The Probability of Temporal Order in a Godless Universe
So how probable is it mtrinsically that in a Godless universe there will be laws of nature at some level guaranteeing that things behave in very largely predictable ways? The answer to this question depends to some extent on what laws of nature are. I discussed in Chapter 2 three theories of this. There is first the immensely implausible Humean account developed by Lewis—that the conformity of all objects to laws of nature is just the fact that they do so conform; there is no more fundamental explanation of this conformity. It is just a brute fact that (both at a fundamental level and at a phenomenal level) objects (substances) fall into kinds (electrons, positrons, pendula, seeds) in such a way that the simplest extrapolation from their past behaviour leads to generalizations that predict their future behaviour more or less correcdy. In the near past, as in the more remote past, every positron has continued to attract every electron with exacdy the same force inversely proportional to the square of their distances apart. There are innumerable other ways in which objects could have behaved, almost all of them such that the simplest extrapolation from their past behaviour would not have correctly predicted their subsequent behaviour. It is only if there is a common explanatory cause of the behaviour of objects that there is any reason to suppose that they will behave in the same way. And in a Godless universe on the Humean theory of laws of nature there is no more fundamental explanation of the coincidence in the ways in which objects behave. On this view 'laws' do notTeally explain the behaviour of objects, they merely describe it.
Alternative accounts of laws of nature represent talk of laws* as talk about a feature of the world additional to the mere succession of events, a feature of physical necessity that is part of the world. As we saw in Chapter 2, this feature of physical necessity may be thought of
either as separate from the objects (substances) that are governed by it, or as a constitutive aspect of those objects. The former approach leads to a picture of the world as consisting of events (constituted perhaps by substances with their properties), on the one hand, and laws of nature, on the other hand; the most common version of this view claims that laws of nature are logically contingent relations between universals. The conformity of all objects to simple laws of nature consists on this account of the instantiation of quite a few universals each connected in simple ways to one or two other universals. If, despite the difficulties raised in Chapter 2, we adopt this account, the first question is why should there be universals connected to each other before they are instantiated, and why—if there is a universe, and so some universals must be instantiated—should quite a few universals be instantiated in such a way as to form a whole system of laws of nature. There might be many universals that were instantiated without bringing any other universals with them, so that there was no predictable effect of the instantiation. But on this account virtually all universals are connected to other universals. And there might be universals, but only ones of kinds instantiated once or twice in the history of the universe, rather than ones like 'photon1 or copper' that are instantiated often and so can be used for useful prediction. And, again, the mathematical connections between the universals—for example, between the masses of bodies, their distance apart, and the gravitational attraction between them— might be of such complexity as never to be inferable from the past behaviour of objects.
Now I suggest that a universe without connections between universals would be simpler than one with connections; and one with simpler patterns of connection would be simpler than one with such complicated patterns of connection that rational beings would not be able to infer the future behaviour of objects by means of the simplest extrapolation from their past behaviour. Among theories of the universe as a whole (which will thus have equal scope), simplicity is the sole indicator of mtrinsic probability. It then follows that, if we give it the weight that I have urged that we should (so that a very simple theory is more probable than a disjunction of many more complex theories), it would be very probable that there would be no connections between universals at all—that the universe would be chaotic But note that, if we give simplicity much less weight and suppose that a simpler theory is merely somewhat more probable
162
Teleological Arguments
Teleological Arguments
163
than a more complex theory, it might be that it is more probable that one of a disjunction of alternative sets of fairly simple connections between universals holds rather than no connections at alL But in that case, since there are a very large number of complex ways in which universals could be associated, and we are giving simplicity only a moderate weight, then it will be at least as probable that one of the complex connections between universals will hold as that one of the simple connections will hold—there being so many more (infinitely many more) of the former. Either way, it is going to be improbable that in a Godless universe there will be simple connections between universals, and so simple laws of nature.
The same issues arise on the substances-powers-and-habuities account of laws of nature. On this account, powers and liabilities are among the properties of substances. Laws of nature are then just contingent regularities—not of mere temporal succession (as with Hume), but of causal succession, regularities in the causal powers (manifested and unmanifested) of substances of various kinds. The conformity of all objects to simple laws of nature consists on this account in all substances falling into very few kinds with the same powers and liabilities as each other. Why does this happen? The S-P-L model has an initial answer to this question; it can provide an explanation of this fact in a way that the other two models cannot provide an explanation of the corresponding fact. The S-P-L model's answer is an answer in terms of the causal ancestry of substances. A substance has the powers and liabilities it does because it was produced by another substance exercising (in virtue of some liability to do so) its power to produce a substance with just those powers and liabilities. If a proton is produced (together with an electron and an neutrino) by the decay of a neutron, then the proton's powers and liabilities are caused by the neutron, in virtue of its powers and liabilities. There are then different ways in which it could have come about that all substances fall into a small number of kinds in the way described, according to whether this process had a beginning and of what kind that beginning was.
Suppose, first, that the universe did have a beginning. There are two different kinds of theories of a beginning. The first state might have been a spatially extended state, or a spatially pointlike state. In the first case, we would still have a lot of substances, but perhaps crammed into a very small space. In terms of the Big Bang model,
there would not have been literally a singularity; it would just have been that, as you approach the first instant in the temporally backward direction, you would find denser and denser states; but it really all started in a very but not infinitely dense state. If that state was to give rise to our present universe of very few kinds of substance, it must itself have consisted of a very large number of substances of very few kinds. The alternative first state would be a literally pointlike one. At the first instant of the universe's history, on this theory, there was an unextended point, endowed with the power to decay into innumerable substances of very few kinds, and a liability to exercise that power at some time or other. Suppose, secondly, that the universe had an infinite age. The properties (of powers and liabilities) of every substance are then caused by those of a preceding substance. So there can be substances with exactly the same such properties (including the powers to produce substances of the existing kinds) only if there always have been.
Study of the present data of physics and cosmology will allow us to say roughly how probable on those data are the three different theories—on the basis of how probable it is that we would find those data, given each of the theories, and of how simple are the different theories. My assessment of the present state of cosmology is that a beginning is more probable than an infinite age; and that evolution from a very dense state is more probable than evolution from an infinitely dense state. (All matter-energy occupying an unextended point is, as I suggested in the previous chapter, not a possibility allowed by the current theory of matter-energy, which would require considerable complication in order to allow for this while continuing to make the present data probable.) But, of course, new data could change the probabilities.
The issue for us, however, is not what are posterior probabilities on the physical data that the different theories are true, but how probable is it a priori if there is no God that the true theory will be such as to lead to substances of a very few kinds. This will depend solely on the relative simplicity of the three theories (since they are of equal scope), and the probability on each of these theories that substances of such kinds would result Simplicity is the sole relevant a priori criterion. There is no doubt that the theory that the universe began at a point is simpler and so intrinsically more probable than any particular theory that it began with many substances, so much simpler that I suggest that it is more probable than the disjunction of all
164
Teleological Arguments
Teleologkal Arguments
165
theories claiming that it began from or always contained many substances.
But, if it did begin from an unextended point, the simplest theory of such a beginning would seem to be that that point would have had no power to produce extended substances. If it did have such a power, it would seem simpler to suppose that it would have the power to produce just one extended substance. The theory that it would have the power to produce extended substances falling into few kinds, themselves having the power to produce more such substances, all with the liability to exercise these powers from time to time, seems just one of a number of equally simple theories, less simple than that the theory that the unextended point had no power or only the power to produce just one extended substance. But any theory that at a beguuiing or always there were many substances, which fall into kinds with identical powers and liabilities, is again a theory of a very improbable coincidence. Such a coincidence cries out for explanation in terms of some single common source with the power to produce it Just as we would seek to explain all the coins of the realm having an identical pattern in terms of their origin from a common mould, or all of many pictures having a common style in terms of their being painted by the same painter, so we should seek to explain all physical objects having the same powers in terms of their deriving them from a common source. So again, on the S-P-L account of laws of nature, as on the universals account (and a fortiori on the Humean account), it is very improbable that there would be in a Godless universe laws of nature sufficiently simple for rational beings to extrapolate from past to future with normal success.
The Probability of Temporal Order, given Theism
Theism leads us to expect (with significant probability) that God will bring about humanly free agents, as described in Chapter 6. They will be embodied creatures, and will start with limited power and knowledge. If they are to extend their power, they must discover which of their basic actions will have which more remote effects in which circumstances—for example, which sequence of basic actions done in which circumstances will lead to a house being built, and which sequence will lead to a bomb being built Only with this knowledge will they have a choice of whether to build houses or bombs. For
there to be such recipes for producing different effects that creatures can discover, there need to be simple regularities in the behaviour of things that creatures can note and use. It has to be the case that this brick put on top of cement that is on top of another brick will stay there and resist much pressure, and so on and so forth. Theism leads us to expect a world at some phenomenal level, simple and reliable. It leads us to expect that God will bring about an initial singularity of the right kind, or an initial arrangement of substances with the same powers and liabilities of the right kind and keep them in existence, or that he will always have kept in existence such substances. (Or, on the universals model of laws of nature, theism leads us to expect that God will produce the right kinds of connections between universals. On the Humean model of laws of nature, theism leads us to expect that God will make things behave in simple regular ways.)
I have been assuming so for that there is only one universe. But there may be many universes. If there were actually existing all possible universes, some of them will be law governed and it might be expected that we would find ourselves in such a universe. However, it would be the height of irrationality to postulate innumerable universes just to explain the particular features of our universe, when we can do so by postulating just one additional entity—God. Science requires us to postulate the simplest explanation of the data, and one entity is simpler than a trillion. In order rationally to postulate other universes we would need to find new data in our universe best explained by postulating that there are also other universes. In particular, we would need to have data such that extrapolating back from the present state of our universe in accord with the mathematically simplest supposition about what are its laws that would explain these data would lead us to a state at which there was a universe split, a state in which those laws will have dictated that another universe would T>ud off' from our universe. But in that case the other universe would be governed by the same fundamental laws as govern our universe, and so we can consider the two universes (or however many universes we learn about) as one multiverse, and the whole preceding structure of argument gives the same results as before. So it does not affect the issue of why things are law-governed if we suppose (on good evidence) that there is more than one universe.4 And I have
* Hume, Diabgues concerning Natural Religion (first published 1779, ed. H. D. Aiken (Hafher, 1948), 53-5) suggested a temporal version of many universes—that perhaps this ordered universe is a mere accident among the chance arrangements of eternal matter. In
....--------..... ------— ••.
...s._v------ - -      - . .—
166
Teleological Arguments
argued that whether talk of 'laws' is talk of regular successions of events, of concrete entities determining the behaviour of substances, or of the powers and liabilities of substances, it is a priori improbable that a Godless universe would be governed by simple laws, but there is quite a significant probability that a God-created universe would be governed by simple laws. Hence the operation of laws of nature is evidence—one strand of a cumulative argument—for the existence of God.
Let us represent by e this conformity of the world to order, and let h be the hypothesis of theism It is not possible to treat a teleological argument in complete isolation from the cosmological argument We cannot ask how probable the premiss of the teleological argument makes theism, independently of the premiss of the cosmological argument, for the premiss of the teleological argument entails the premiss of the cosmological argument. That there is order of the kind described entails that there is a complex physical universe. So let k be now, not mere tautological evidence, but the existence of a complex physical universe (the premiss of the version of the cosmological argument to which I devoted attention). Let us ask how much more probable does the orderliness of such a universe make the existence of God than does the mere existence of the universe. As we have seem P(h\e8tk) will exceed P{h\k) if and only if P{e\h&k) > P{e\ ~h&1c). Put in words with our current fillings for k, e, and fc, the existence of order in the world confirms the existence of God if and only if the existence of this order in the world is more probable if there is a God than if there is not. The arguments of the previous pages have sought to show just this; and indeed that the probability of order of the right kind is very much greater if there is a God, and so that the existence of such order adds gready to the probability that there is a God.
the course of eternity matter arranges itself in all kinds of ways. We just happen to live in a period when it is characterized by order, and mistakenly conclude that natter is always ordered. Hume is certainly right in claiming that this is a logical possibility, but the point made above remains—that it is irrational to postulate other universes, unless features of this universe give support to laws that have the consequences that there are other universes; from which h would follow that they are governed by the same fundamental laws as our universe.
Teleological Arguments The Argument from Spatial Order
167
Those who marvel at the order of the universe may marvel at either or both of the regularities of co-presence and of succession. The thinkers of the eighteenth century to whom a teleological argument appealed so strongly were struck almost exclusively by the regularities of co-presence. They took the regularities of succession largely for granted. What struck them was the subde and coherent arrangement of parts in human and animal bodies and in plants, enabling humans and animals to acquire an enormous amount of knowledge and to execute an enormous variety of purposes, mcluding to reproduce their kind, and enabling plants to grow and multiply (without choosing to do so). Paley's Natural Theology dwells mainly on details of comparative anatomy, on eyes and ears and muscles and bones arranged with minute precision so as to operate with high efficiency, and in the Dialogues Hume's Cleanthes produces the same kind of examples: 'Consider, anatomize the eye, survey its structure and contrivance, and tell me from your own feeling, if the idea of a contriver does not immediately flow in upon you with a force like that of sensation."
The eighteenth-century argument from spatial order seems to go as follows. Humans, animals, and plants have the power to reproduce their kind, and so, given their past existence, their present existence is to be expected. But what is vastly surprising is the existence ofhumans, animals, and plants at all. By natural processes they can come into being only through generation from organisms of the same species. But, it was claimed, the world hasnot been going onfor ever, and so the great puzzle is the existence of the first humans, animals, and plants in 4004 bc or whenever exactly it was that they began to exist. Since they could not have come into existence by natural scientific processes, and since they are very similar to the machines that certain rational agents—namely, humans—make, it is very probable that they were
5 Hume, Dtakgues, 28.
* Even if they had supposed that the world had been going on for ever and had contained humans, animals, and plants for ever, the thinkers of the eighteenth century could still have constructed an argument from the eternal existence of a universe containing humans, animals, and plants rather than a universe not containing humans, animals, and plants; but the argument would have to have been a more subtle one than the one that we are considering.
Teleological Arguments
Teleological Arguments
169
made by a rational agent—only clearly one much more powerful and knowledgeable than humans.
Hume's objections in the Dialogues (through the mouth of Philo) to a teleological argument are directed against the argument from spatial order, although—if they worked—some of them would also have had force against the argument from temporal order and so I have considered them in that connection. Despite Hume's objections, the argument is, I think, a very plausible one— given its premisses. But one of its premisses was shown by Darwin and his successors to be clearly false. Humans can be produced through generation from complex animals, and complex animals and plants can be produced through generation from less complex animals and plants—species are not eternally distinct; and simple animals and plants can be produced by natural processes from inorganic matter. This discovery led to the virtual disappearance of the teleological argument from popular apologetic—mistakenly, I think, since it can easily be reconstructed in a form that does not rely on the premisses shown to be false by Darwin. The basic mistake of those who regarded Darwin's discoveries as destructive of the argument from spatial order is that they ignored the fact that only certain processes acting on a certain initial kind of inorganic matter would have produced human bodies (and animals and plants); and that it is a priori improbable that the processes and initial matter would be of the right kind, but that this is to be expected if theism is true.
The argument is, I think, best treated not as an argument from analogy (the way typical of the eighteenth century) but in the way in which other arguments are treated in this book, as an argument from evidence that it would he probable would occur if theism is true, but not otherwise. The argument must be construed as an argument from human bodies, not humans. The argument from human bodies being connected to a mental life is a separate argument, to be analysed in the next chapter. We also have the evidence of animal bodies and plants. To make the„exposition simpler I shall largely ignore these latter until towards the end of this chapter.
The argument from human bodies must be construed as an argument from the existence ofbodies having certain features characteristic of human bodies. They are features that the body of a humanly free agent, as denned in Chapter 6, would need to have. To be the body of a humanly free agent, a body needs to be suited for the acquisition of
true beliefs about the environment, the formation of purposes in the light of desires, and the expression of them via chosen basic actions designed to affect the agent, others, and the world for good or ill. To do this job a body needs: (1) sense organs with an enormous variety of possible states varying with an enormous variety of different inputs caused by different distant world states; (2) an information processor that can turn the states of sense organs into brain states that give rise to beliefs of moral or prudential importance; (3) a memory bank, to file states correlated with past experiences (we could not consciously reason about anything unless we could recall our past experiences and what others have told us); (4) brain states that give rise to desires, good and evil (desires to eat and drink, to care for others or to hurt them; and to discover whether or not there is a God); (5) brain states caused by many different purposes that we have; (6) a processor to turn these states into limb and other voluntary movements (to turn, for example, my purpose of telling you that today is Friday into those twists of tongue and hp that will produce an English sentence with that meaning); and (7) brain states that are not fully determined by other physical states. (As far as physical laws are concerned, there needs to be a certain amount of indeterminism in the brain if undetermined human choices are to determine what happens in the brain.)
Clearly human bodies have characteristics (1) to (6). Fairly clearly too there is a small amount of ^determinism in the brain, for, if the laws of Quantum Theory that govern matter on the smaller scale have no deeper detenrunistic explanation (as most physicists claim), then the behaviour of objects on the small scale is not fully determined. The laws of Quantum Theory merely tell us the physical probabilities of various outcomes. In general, ^deterministic behaviour on the small scale averages out to produce virtually deterministic behaviour on the large scale. If each coin had a physical probability of 1/2 of landing heads and 1/2 of landing tails, there will be a very large probability close to 1 that the number of coins landing heads in 1,000 tosses will not differ very much from 500. So, even if there is a significant probability that individual atoms will behave in ways different from the norm, bricks and billiard balls are most unlikely to do so. But it is possible to have devices that multiply small-scale indeterminacies, so that a small variation in the behaviour of one atom can have a large-scale effect The brain is an
170
Teleological Arguments
T
Teleological Arguments 171
extremely complicated system in which small differences cause large differences. But we do not yet know enough about the brain to know whether very small differences of the kind that Quantum Theory seems to tell us are not physically determined are such as to ensure that it is not physically determined which actions humans do. It is, however, evident that the brain is a physical system quite unlike any other physical system, in that it causes conscious events and its states are caused by conscious events; and so clearly laws of a very different kind govern the brain from those that govern all other physical states (as I argue more fully in Chapter 9). So it may well be that there is brain indeterminacy, sufficient for human free choices to produce physical effects, because of some feature of the brain other than the multiplication of indeterminacies within the Quantum limit, I conclude that, on present evidence, there is no good reason to suppose that the brain does not have characteristic (7).
In that case, human intentions will cause human behaviour without being caused to do so by physical events. Humans will therefore have libertarian free will unless something non-physical causes them to form the intentions they do. The most plausible such something is God. But, if I am right in my claim in Chapter 11 that a perfecdy good God will allow humans to suffer at the hands of other humans to the extent to which they do only if they have libertarian free will, God will not form their intentions for them. So, if there is a God, humans will have libertarian free will, and so humans will be humanly free agents. So, whether or not there is a God, there seems no reason at present to deny that humans have free will. Yet it seems to each of us at the moment of choice that we are making our choice independendy of the forces acting upon us (if we yield to some desire, we are choosing to yield to that desire; and, if we resist the desire, we are choosing to resist it), and we are justified in so believing in the absence of counter evidence7 (This follows from the Principle of Credulity that I discuss in Chapter 13.) Hence (in the absence of some new evidence from neurophysiology) humans are probably humanly free agents, and I shall so assume in future.
We know that human bodies have evolved by natural processes from inorganic matter. But dearly the evolution can have taken place only given certain special physical laws. These are, first, the
For fuller argument that humans do have libertarian free will that makes a difference to thdr behaviour in the physical world, see my The Evolution of the Said (2nd edn., Clarendon Press, 1997), ch. 13.
chemical laws stating how under certain circumstances inorganic molecules combine to make organic ones, and organic ones combine to make organisms. And, secondly, there are the biological laws of evolution stating how complex organisms evolve from simpler organisms. I have no wish to challenge Darwin's account of how this happens. Organisms have many offspring, some of which differ from their parents in small ways in respect of one or more characteristics—some offspring are a littie taller, some a litde shorter, some fatter, some thinner than their parents, some have a small extra growth, some have lost a small part and so on. The new characteristics in turn are passed on to the offspring's own offspring, which in turn also exhibit some variations of characteristics from those of their parents. Given predators and shortage of food, there will be competition for survival, and those organisms whose characteristics give them an advantage in the struggle for survival will survive. Among organisms very well fitted for survival (should they evolve) will be organisms who can see how their environment is changing in crucial ways (by the presence of predators, absence of prey and other food, etc.), and work out (using the criteria of what is evidence for what in more sophisticated ways than non-human animals) in the fight of past experience how to survive and to help their young to survive. These organisms will have human bodies of the kind described. It was shown by work subsequent to Darwin that the main mechanism by which these small variations are caused are 'mutations' in genes on chromosomes; there is no regular pattern in mutations—they may occur at any time on any gene affecting any characteristic.8
So the question arises why the inorganic matter of which the Earth was made was of a kind that under the operation of laws of chemistry and biology could be converted into human bodies. As we have noted, physics tells us that there was a Big Bang some fifteen billion years ago, which produced matter-energy that condensed into the fundamental particles that came together to form the chemical elements that eventually condensed to form the inorganic matter at the begiiming of the history of planet Earth. But why were there laws of physics that brought this about, and the laws of chemistry and
8 See Additional Note 2 for recent challenges to this Darwinian account, which—it is claimed—are evidence for various steps of the evolutionary process requiring the intervention of a 'designer'.
172 Teleologkal Arguments
biology that led to the inorganic matter being formed into human bodies? Presumably because these laws followed from the fundamental laws of physics. So our question becomes—why are there not just any laws of nature, but laws of a particular kind such that together with the initial matter-energy at the time of the 'Big Bang' would lead to the evolution of human bodies. That there are the laws of nature that there are, and the boundary conditions of the universe were as they were, is again where scientific explanation starts: it is something 'too big' for science itself to explain. I shall argue that the laws and initial conditions being such as to lead to the evolution of human bodies is very improbable a priori, but fairly probable if there is a God who brought it about, and so we have a further substantial C-inductive argument for the existence of God.
Fine-Tuning
Not all initial conditions or laws of nature would lead to, or even permit, the existence of human bodies at some place or other at some time or other in the universe. So we may say that the universe is 'tuned' for the evolution of human bodies if the laws and initial conditions allow this to occur (in the sense that they fully cause this evolution if the laws are deterministic, or make it significantly probable if the laws are probabilistic}. If only a very narrow range of laws and initial conditions allow such evolution, then we may say that the universe is 'fine-timed' for this evolution. If the fundamental laws and initial conditions are, as we suppose, the laws of Quantum Theory and Relativity Theory with the four forces (strong force, weak force, electromagnetic force, and gravity) governing the basic array of fundamental particles (photons; leptons, uicluding electrons; mesons; and baryons, mcmding protons and neutrons) (what I shall call the standard theory), and the initial conditions are such conditions as the velocity, density, and degree of isotropy of the matter-energy of the universe immediately after the time of the Big Bang; and these are measured in normal ways, then—recent work has shown—the universe is fine-tuned. The constants of its laws and the variables of its initial conditions needed to He widiin very narrow ranges if human bodies were ever to exist. One such set of narrow ranges are those centred on the actual values (as we believe them to be) of the constants of laws and variables of initial conditions. It is
Teleologkal Arguments 173
worthwhile in this context to bring out how this happens,9 why any one constant or variable lying outside the range (while the others lie within the range) would prevent the evolution of human bodies. This section may not be fully comprehensible to those without some scientific background. I suggest, nevertheless, that such readers read through these pages; they will get the main message.
A life based on carbon, in combination with certain other elements, especially hydrogen, oxygen, and nitrogen, is well suited for the formation of bodies with the seven features listed above. With a valence of 4, carbon can enter into many different chemical combinations. Carbon compounds are stable over long periods of time; but are also metastable in that in certain particular situations they can easily be induced to interact with other compounds to produce new compounds. Hence, 'more information can be stored in carbon compounds than in those of any other elements'.1 Together with hydrogen, nitrogen, and oxygen, carbon can form long, complex chain molecules; and, together with calcium giving skeletal rigidity, such an information-processing system can be made a continuing independent component of the universe. Carbon-based life requires for its stability a moderate range of temperature and pressure, and— if the purposes of organisms are to make much difference to things— a solid planet on which to live.
Given the standard theory with constants and variables of initial conditions having their actual values, it is highly doubtful whether there could be any other kind of intelligent life. It has sometimes been suggested that silicon could replace carbon in its central role, but this seems doubtful in that silicon compounds do not have the stability of carbon compounds.11 Another recent suggestion has been that intelligent systems of particles relying on the 'strong' interaction for their organization might exist inside neutron stars; but it seems doubtful whether they could have nearly as much information-processing capacity as does carbon-based life on Earth.12 So let us
9 The original classical physical analysis of the extent of fine-tuning in the universe is J. D. Barrow and E J. Tipler, TheAnthropic Cosmological Principle (Clarendon Press, 1986). This has been carefully reanalysed and updated in Robin Collins, 'Evidence for Fine-Tuning', in N. A. Manson (ed.), Cod and Design (Routledge, 2003). I am much indebted to the latter article for its presentation of the latest state of relevant physics.
10 Barrow and Tipler, Anthropic Cosmologkal Principle, 547.
11 Ibid. 545-«.
12 Barrow and Tipler, Anthropic Cosmologkal Principle, 343-6.
174 Teleological Arguments
suppose, plausibly enough, that carbon-based life is the only possible kind of life (given standard theory and the actual values of its constants and variables of initial conditions). If silicon-based life is possible, the argument below would not need much alteration (for the conditions necessary for the evolution of silicon-based life are very similar to those necessary for the evolution of carbon-based life), and neutron-star life is too speculative a suggestion to be taken into account. Given the four fundamental forces and the basic array of fundamental particles, the strengths of the forces and the masses of particles have to have ratios to each other within certain narrow bands if the larger chemical elements, including carbon and oxygen (needed for carbon-based life), are to occur at all; and the Pauli exclusion principle has to hold. This principle (applying to all fer-mions—for example, electrons and protons) says that in any one system (for example, one atom) only one particle of the same kind. can be in a given quantum state. In consequence there are only a small number of possible energy states for the electrons of an atom, and only a small number of electrons can be in each energy state. While the basic laws of Quantum Theory ensure the stability of the atom—electrons do not collapse onto the nucleus—the Pauli principle leads to the electrons being arranged in 'shells'. Hence atoms of a finite number of different kinds can be formed by different numbers of electrons surrounding the nucleus, and molecules can be formed by bonds between the electrons of different atoms. No exclusion principle, no chemistry. But not much chemistry unless there is plenty of possibility for different structures to be built up, to be relatively stable, to interact, and to form new structures. For that we need atoms to be large structures with plenty of empty space between well-defined central nuclei and electrons.
The build-up of the atoms required for carbon-based life requires the four forces to have certain strengths, relative to each other. If there are to be stable nuclei, the strong force that keeps the protons and neutrons together in the nucleus has to be strong enough to overcome the electromagnetic repulsion between the protons. A 50 per cent decrease in the strong force 'would undercut the stability of all elements essential for carbon-based life, with a slightly larger decrease ehmihating all elements except hydrogen'.13 But the process by which carbon and oxygen are built up from the actual initial
13 Coffins, 'Evidence for Fine-Tuning', 183.
Teleologkal Arguments 175
conditions of the universe requires a far greater degree of this tuning to lead to their production. An increase or decrease of more than 0.5 per cent in the strength of the strong force or more than 4 per cent in the strength of the electromagnetic force would lead to such small amounts of carbon and oxygen being produced as to make the production of intelligent life very unlikely.14 A tliirty-fold decrease in the weak force would lead to stars being made almost entirely of helium and so having a short life (of about 300 million years) in no way conducive to the evolution of intelligent life.15 An increase in the strength of the gravitational force by 3,000 would lead to stars with lives of no more than a billion years (compared to the ten billion years of our sun's lifetime), which would make the development of intelligent life much less probable.16 The weakest of the four forces is the gravitational force, whose effects are significant only over large distances. Over the small distances when the strong force has significant influence, its strength is of the order of 1040 times that of the force of gravity. It follows that the kinds of increases or decreases in the strengths of the forces mentioned above (50 per cent, 4 per cent, etc) compatible with the production of carbon-based life represent a very small range indeed of the values of the strengths of the forces involved within the range of actual values of any of the forces, and an infinitesimal range within the range of logically possible values of the forces. For example, G has to he between 0 and 3000G, which represents one part in 1036 of the range of values of the force constants. And so on for the other constants.17 The expansion of the universe is governed by the strength of the initial Big Bang, and the restraining effect of gravity possibly diminished or increased by the value (positive or negative) of the cosmological constant (A), which latter may be regarded as determining a fifth force. This needs to he extremely close to zero if space is not to expand so rapidly that every object in the universe flies apart, or to collapse so rapidly that every object is the universe is crushed.18
Further, given the actual laws of nature or laws at all similar thereto, boundary conditions will have to he within a narrow range of the present conditions if intelligent fife is to evolve (or else they will have to He well outside that range; this point will be discussed later). If the
14 Ibid 183-6. 15 Ibid. 188-9.
15 Ibid. 189-90,192-t.
17 Ibid. 190. 18 Ibid 180-2.
176
Teleological Arguments
Teleological Arguments
177
iiniverse had a beginning, the boundary conditions are the arrangements and properties of the matter-energy of the universe at the time that the universe started off. Present evidence suggests, as I have written earlier, that the universe began in a very dense state some fifteen billion years ago. For the formation of intelligent life in a universe expanding from such a state, conditions at the time of the Big Bang have to be (within the narrow ranges) just right The initial rate of expansion is critical. If (for the actual value of the gravitational and cosmological constants) the initial velocity of expansion had been slighdy greater than the actual initial velocity, the effect would have been the same as would result from a significant positive cosmological constant—stars and so the heavier elements would not form; if it had been slighdy less, the effect would have been the same as would have been produced by a significant negative cosmological constant—the universe would collapse before it was cool enough for the elements to form.19 It has been calculated that (barring a possible qualification from 'inflation theory' to which we shall come shortly) a reduction in the rate of expansion of one part in a million would lead to premature collapse, and an increase by one part in a million would have prevented the evolution of stars and heavier elements.20 Some initial inhomo-geneity in the distribution of matter-energy is needed if galaxies, and so stars, are to be produced; too much would lead to black holes being formed before stars could form.21 In the beginning there was a slight excess of baryons over anti-baryons; all but the excess baryons became matter-energy. If the excess number had been even slighter, there would not have been enough matter for galaxies or stars to form. If it had been much greater, there would have been too much radiation for planets to form.22 And so on. The universe has to start with the right density and amount of inhomogeneity of radiation and velocity of expansion, and that means (within a very narrow range) the actual amount
I have been setting out the generally agreed point that, if any one of the constants of the laws and variables of the initial conditions were to he outside a narrow range, human bodies would not evolve.
15 Barrow and Tipler, Anthropic Cosmological Principle, 410-12.
20 Papers by S. W. Hawking and by R. H. Dicke and P. J. E. Peebles cited in J. Leslie, Universes (Routledge, 1989), 29.
21 Barrow and Tipler, Anthropic Cosmological Principle, 414-19. 12 Ibid. 401-8.
Recent work has suggested 3 that, if a number of the constants and variables were all significandy clifferent, each having a value within a different small range, human bodies could still evolve. That is, there are several small islands within the space of possible values of constants and variables within which human life could evolve. But this does not alter significantly the point that such islands are the exceptions and the tuning needs to be fine-tuning for this evolution.
If standard theory provides the ultimate explanation of the universe (and so God does not bring it about that standard theory operates), such fine-tuning is a priori very improbable. For the form in which any theory, including the standard theory, is stated by scientists in their books and articles is its simplest form— scientists do not try to complicate things for themselves and their readers unnecessarily. This form involves variables and constants being measured in the normal way. It is the form in which we judge the simplicity of the theory that determines (for theories of equal scope) the intrinsic probability of its truth. Versions of standard theory expressed in its simplest form will differ only in respect of the values of constants of laws and of variables of boundary conditions therein. Given all that, a version that claims that a constant or variable lies within one range will not differ greatly24 in simplicity from theories that claim that it lies within another range of equal size; and so each such version will be approximately equally probable a priori. But since only a few versions of standard theory in which constants vary over a very small range are tuned for the evolution of human bodies, such evolution is a priori very improbable. In shghtly more technical terms, the claim is that the probability density for constants and variables measured in the normal way is roughly constant (that is, the probability that these will lie close to a given
23 For one example, see Collins, 'Evidence for Fine-Tuning', 185.
24 It will differ a little if the simplest formulation of the theory yields a unique zero point for measurement of some variable or constant (a unique point at which some quantity has its lowest value), as, for example, does the Kelvin scale for temperature measurement (0°K being that temperature at which an ideal gas would exert no pressure, and no lower temperature is possible). For then it will be a non-arbitrary matter whether the value of the constant or variable lies within a lower range or a higher range of possible values. It will be a bit more probable that it lies within the former, for the reason that laws containing small integers are simpler than ones containing larger integers (see P-54).
178
Teleological Arguments
Teleobgical Arguments
179
value is roughly constant for all values of the constants and variables of standard theory).25
It is worth noting the effect of not choosing the simplest formulation of a theory, on the probability density of different constants and variables. I take a very easy example. Newton's law of gravitational attraction F = could be expressed as F = ^ where d is defined as G-1/3. A constant probability density distribution for d (that is, the assumption that it is equally probable that d lies within any range of given size) will not yield a constant probability density distribution for G, and conversely. A constant probability distribution for d will yield the result that d is equally likely to he between 1 and 0.5 as between 0.5 and 0, and so that G is equally likely to he between 1 and 8 as between 8 and iiuinity (i.e. to have any value whatsoever above 8). Expressing the laws of our standard theory in very complicated forms, logically equivalent to their simplest forms, and assuming a constant probability density for the constants and variables of these forms, could have the consequence that much greater variation of these (far less 'fine-tuning') would be compatible with the universe being hospitable to human bodies. But laws are judged simpler and so to have greater prior probability in virtue of the features of their simplest forms. Since a constant is simpler than a constant to the power ( - 1/3), the traditional form of Newton's law is the simplest and so most fundamental form. And, more generally, insistence on the simplest form of a law should yield a unique probability density distribution for the constants and variables of laws of that kind (or, at most, if there are a number of equally simple forms of a law, a few different probability density distributions that
25 The constants and variables of standard theory with which we are concerned do in general have unique zero points (see n. 24 above). In measuring the density of matter-energy, or the velocity of relative recession of the galaxies, for example, velocity and density have unique zero points on the simplest way of measuring them. Hence lower values of these are somewhat more probable than higher values. This has the consequence that, although there is an infinite range of possible values of these constants and variables, there can be a finite probability that some such constant will have a value lying within any given range. But, if the constant or variable having a value within a range of given size was the same throughout the infinite range (as would be the case for constants and variables without a unique zero point), the probability of it lying within any finite range would be infinitesimal. (B5r the need to use iiifinitesimals in assessing probabilities, see my Epistank Justification (Clarendon Press, 2001), Additional Note GJ. So, either by ascribing higher mrrinsic probabilities to lower values of the constants and variables, or by using infinitesimals, I avoid what is known as the 'normalizabaity problem'. (See e.g. Timothy McGrew et «1, 'Probabilities and the Fine-Tuning Argumenf, in N. A. Manson, God and Design.)
are not likely to make much difference to the extent of need for fine-tuning). So, given the standard theory, and no more fundamental explanation thereof (physical or theistic), hrning is a priori immensely improbable.
Physical cosmology is a very unstable branch of physics. New theories are produced each year. Changes are possible that would have the consequence that constants and variables can vary within a much wider range and yet life still evolve. One possible change, though in my amateur judgement a fairly unlikely one, is that it may become established that the boundary conditions are significantly different from what has been supposed—for example, that the universe evolved not from an initial singularity, but from a very dense state resulting, perhaps, from a prior collapse, or perhaps from a quantum mechanical fluctuation of the 'vacuum'.2 Such a change, probably going with the adoption of the view that the universe was infinitely old, would have the consequence that a far wider range of boundary conditions would give rise to life.
The role of Tjoundary conditions' in a backwardly eternal (that is, infinitely old) universe may need clarification. Imagine a billiard table sealed under a glass cover in which the balls move in a vacuum (and that any energy transfer to or from the outside can be discounted). The laws of collision govern the interaction of the balls, which bounce off each other and off the walls for the indefinite future. It could have been that this process was started off by someone arranging the balls and giving them an initial push before the table was sealed. In that case the boundary conditions would be the initial conditions (arrangement and velocity of balls), and they together with the laws of collision would determine all the subsequent behaviour of the balls. Some initial conditions would allow the balls to arrange themselves in all the (logically) possible arrangements during the course of a subsequent infinite time. Yet some initial conditions (for example, the balls moving initially with velocities parallel to each other and to the walls) will ensure that the balls occupy only a few of the possible arrangements even in the course of infinite time. Suppose, now, that the process has been going on for ever (that is, is not merely forwardly but backwardly eternal). Then the set-up may still have certain features at some given time that would occur only if a narrow set of possible arrangements either ever have been or ever wfll
24 Barrow and Tipler, Anthropic Cosmologies! Principle, 440-1.
180
Tekologkal Arguments
Teleological Arguments
181
be occupied (for example, this might also be the case if at a certain time the balls are moving parallel to each other and to the walls); or, much more likely, features that would occur only if in the course of infinite time backward and forward all possible arrangements of those balls occur. However, the sealing of the table still ensures that the only possible arrangements are arrangements of those balls— there cannot be more or fewer balls in the past or future. The 'boundary conditions' of an infinite universe are those features of its condition at any one time (for example, in a Newtonian universe, the quantity of energy) that (together with the laws governing it) restrict its possible future and past states.
Now, if the universe is backwardly and forwardly eternal, its present state may be such that we can infer that it must pass through such and such a range of states in the course of infinite time. These might include all the logically possible states of matter-energy, but that is not very likely, for some kind of principle of conservation of energy (within quantum limits) will ensure that past (and future) states are limited to rearrangements of the existing amount of energy. However, although all this would have to be worked out, it is highly plausible to suppose that (for given scientific laws) life is much more likely to evolve at some time in the course of the history of our universe if it has an infinite past than if it has only a finite past. There is more time for more possible arrangements of the constituents of the universe. Nevertheless, the present evidence suggests a finite age of some fifteen billion years.
An alternative change in physics might be a discovery that the laws are other than previously supposed, again in such a way that they bring forth intelligent life out of a much wider range of boundary conditions than had hitherto been supposed. 'Inflation theory' suggests just that Inflation theory tells us that regions of the universe with certain features may have been subject, soon after the Big Bang, to a vast faster-than-light expansion, leading to them very quickly becoming cool homogeneous and isotropic regions.27 So features such as homogeneity and isotrepy for which a narrow range of initial conditions were thought vital are—according to inflation theory—to be expected, given certain laws, to arise from a wider range of initial conditions. Yet it may well be that inflation theory can be successful in any of its many variants in removing the need for fine-tuning from
27 Barrow and Tipler, Anthropic Cosmologksl Principle, 430-40.
the initial conditions only by putting more fme-tuning into the laws.28
There remains, however, a consensus among physicists that the values of the constants in the laws of standard theory (as opposed to the variables of initial conditions) must fie within very narrow ranges if life is to evolve anywhere in the universe—ranges that include the actual values of the constants and probably a few other small ranges in which the values of several of the constants are different from their actual ones. And there is also a consensus that, given an initial Big Bang, variables such as the initial velocity of recession have (even on inflation theory) to lie within a narrow range. There may, however, be a more fundamental physical theory that explains the standard theory, and a constant probability density for the constants and variables of the boundary conditions of the simplest form of that fundamental theory may have somewhat different consequences for the intrinsic probability of tuning (for example, that the more readily observable variables can take only certain values).29 But, more generally, there are innumerable possible scientific theories differing in their form from each other, and innumerable different kinds of boundary conditions differing in the number of entities that they postulate (big and small universes), each allowing many different sets of constants and boundary condition variables. A constant probability density over the latter (when each theory is expressed in its simplest form) will yield for each theory a different probability that a universe conforming to it will be tuned. The theories (although of equal scope—telling us about everything) themselves will differ in their simplicity, and so in their intrinsic probability. Hence, given a precise way of measuring simplicity, there will be a true value for the mtxinsic probability, the probability if there is no God, that any
28 See J. Earman and J. Mostevin, 'A Critical Look at Inflationary Cosmology', Philosophy of Science, 66 (1999), l-£9.
It is possible that the derivation of the fundamental laws of nature from string theory would greatly reduce the need for fine-tuning. This has heen argued in G. L. Kane et al, "The Beginning and End of the Anthropic Principle', astro-ph/0001197. They suggest that all string theories are equivalent; and that different possible 'vacua' uniquely determine all the constants and initial values of variables of laws of nature. They acknowledge that much work needs to be done before (if ever) string theory is established and their result can be demonstrated. But, even granted all this tentative speculation, they acknowledge that 'there wiH be a large number of possible vacua'; and that means both having string theory rather than many other alternative fundamental laws and requiring special variables of initial conditions if human bodies are to evolve.
182
Teleobgical Arguments
Teleobgical Arguments
183
universe will be tuned. It will be (loosely) that proportion of logically possible universes that are conducive to the evolution of human bodies, each weighted by the simplicity of the laws that govern it and the fewness and simplicity of entities in its boundary conditions. And, given the rough way we have of measuring simphcity, we could still give a rough value to this. So it does not matter—for the purposes of an argument from fme-tuning—whether we have the correct theory of our universe, or whether there is a more ultimate physical explanation of the forces that govern it; and whether only a small proportion of the versions of the correct theory lead to a tuned universe. For the prior probability (in a Godless universe), that a universe will be tuned is a function not of the true physical theory and actual kinds of boundary conditions that govern our universe, but of all the possible theories and boundary conditions there could be for any universe at all. It is not, however, within my ability to calculate this value, nor—I suspect—within the ability of any present-day mathematician.
What, however, I suggest, is fairly obvious is that no relatively simple universe would be tuned. For consider the seven features required by a human body listed earlier in this chapter. Such a body has parts. But the parts have to form one body distinct from other bodies and from the inanimate world. In our world this is secured by a chemistry whereby only some bits of matter link to other bits of matter—if I put my hand into a sandpit, my hand, will not absorb the sand; but, if I eat some bread, it will become part of my body. Sense organs require an enormous variety of stimuli impinging on a place, which vary with their distant source. In our universe the best of all such stimuli are light waves—an enormous variety of different light waves arrive every second at our eyes, which vary with the states of objects many metres away. The sense organs respond differendy to each very small range of incoming stimuli But we humans are interested only in certain aspects of the states of distant objects—whether they are the bodies of predators, or prey, or mates, and so on for a million possible differences. The stimuli have to cause brain states that give us information of moral or prudential importance. Our information processor will utilize states caused by past experiences to turn the states of sense organs into useful brain states. And, if we are not to be just automata but to reason consciously from past experiences, we need a memory bank to file those states in recoverable form. This requires a chemistry of stable
states (so that memories remain the same as time passes) and meta-stable states so that certain kinds of input will move a brain element from one state to another (as we learn that some previous belief was erroneous). And for output we need again an enormous variety of brain states comsponding to the different purposes we could form, a processor to turn these into the relevant limb movements (for example, if I want to tell you that today is Friday, to produce the twists of tongue and Hp that will cause the appropriate sounds of the English language). And we need a stable inorganic world to which we can make a difference that remains; there is no point in trying to build a house if the bricks immediately Uquidify.
One way in which all this could be achieved would be by bodies composed of only a few particles, each capable of existing in a trillion trillion trillion different states. But a physics that allowed such particles would be of incredible complexity. The other way, the way operative in our universe, is to have extended bodies, each composed of many fundamental particles of a number of different kinds, each particle capable of existing in a few different discrete states; the differences between bodies being a matter of the number and arrangement of the units and the discrete states of each. To do the job in this way you need a universe with a very large number of particles to compose many bodies and an inanimate environment through which people may influence each other. Change has to be affected through a particle (or group of particles) changing their states, causing other particles to change their states. To secure stable bodies that are nevertheless capable of existing in many different states, you need more than one simple force. One simple force of attraction would lead to crushed lumps of matter incapable of sensitive reaction; and one simple force of repulsion would lead to there being no extended bodies at all. Minimally a combination of two different simple forces (possibly both derivable from one more complicated force) is required. A force of attraction between particles inversely proportional to the square of the distance apart of the particles would be required to be balanced, for example, by a force of repulsion inversely proportional to the cube of their distance apart Forces of these kinds of the right strength would lead to particles coming together but not collapsing on top of each other. But to preserve states (of the brain correlates of belief, for example) intact we have to rule out small variations. We need metastability—systems that remain unchanged under forces of a certain strength but that change
184
Teleologkal Arguments
Teleological Arguments
185
from one discrete state to another discrete state when the strength of the force exceeds a certain amount. This is ensured in our universe by the laws of Quantum Theory, which guarantee the stability of the atom. And, to have distinct bodies that do not merge with each other, and distinct brain states that are open to change only under certain kinds of input, we need something like a chemistry allowing substances to combine easily with some substances but not with other substances. This is secured in our universe by chemical substances different from each other by the charge on their nucleus and the arrangements of charge-balancing electrons in shells around the nucleus—in other words, protons, neutrons, and the Pauli principle. And so on.
So we need large numbers of particles of a few different kinds and forces of some complexity acting between them. But universes are simpler, the fewer objects (for example, particles) they contain and the fewer kinds of mathematically simple forces that operate between them. No very simple universe could be tuned, whatever its boundary conditions. Clearly more complicated kinds of possible universes (for example, ours) can be tuned, and maybe normally the tuning needs to be fine-timing. Maybe, too, some very complicated kinds of universe would produce human bodies for most values of constants and variables of boundary conditions. But the considerable a priori weight of simplicity suggests that in a Godless universe it is a priori improbable that any one universe will be tuned so as to yield human bodies.30 With e as the existence of human bodies, h as theism, and k as the evidence of a universe conforming to natural laws, P(e| ~ h & k) is very low.
50 In order to show the improbability of tuning, it is not enough to show that tuning is improbable given the standard theory—that is, given the local area of possible words. John Leslie has compared this fine-tuning to a dart hitting a cherry on a wall when there are no other cherries in that area of the wall. He claims that (on the assumption that hitting a cherry is something a dart-thrower might wish to do) the fact that the dart hit the cherry is evidence that it was thrown intentionally by a dart-thrower—even though there are many cherries on other areas of the walL (See p. ,143 of his 'Anthropic Principle, World Ensemble, Design', American Philosophical Quarterly, 19 (1982), 141-52.) His claim seems to depend on a feature of his analogy, to which there is no parallel in the universe fine-tuning case. A dart-thrower would naturally try to tut a cherry in an area of a wall a long way from other cherries, his aim being to hit a cherry when that would be difficult for the average human dart-thrower. Hence he aims at the isolated cherry, rather than at cherries close to other cherries. A God tuning a universe seeks to produce human bodies; he has no particular concern to produce them in a possible world where all close possible worlds (except the very closest ones) would not allow their existence. He has no concern to show his universe-
Of course, if there was an infinite number of universes, each with different laws and different boundary conditions, one might expect at least one to be tuned. (Recall my earlier—see p. 133—definition of a universe as a collection of physical objects, all spatially related to each other. A universe other than our own would be a collection of physical objects spatially related to each other, but not to our Earth.) I have already in this chapter made the point that it is the height of irrationahty to postulate an infinite number of universes never causally connected with each other, merely to avoid the hypothesis of theism. Given that simplicity makes for prior probability, and a theory is simpler the fewer entities it postulates, it is far simpler to postulate one God than an infinite number of universes, each differing from each other in accord with a regular formula, uncaused by anything else.31 There might, however, be particular features of our universe (other than its taring) that are most simply explained by supposing that it Tmdded off from another universe in consequence
tuning skills, only to bring about an end product. So, if the fact that there is a tuned universe is to be evidence for God being its creator, what has to be shown improbable a priori is not that there be a tuned universe in our local area of possible worlds, but that there be a tuned universe among all possible worlds. I have given some argument for this— from the impossibility of any very simple universe (and so any universe intrinsically probable) being tuned.
1 Max Tegmark has, however, claimed that it is simpler to postulate an infinite number of universes than to postulate just one. See Max Tegmark, 'Is "The Theory of Everything* merely the Ultimate Ensemble Theory?', Annals of Physics, 270 (1998), 1-51, at 38. 'Our TOE [Theory of Everything] . . . postulates that all structures that exist in the mathematical sense exist in the physical sense as welL The elegance of this theory lies in its extreme simplicity, since it contains neither any free parameters nor any arbitrary assumptions about which of all mathematical equations are assumed to be "the real ones". He explicitly (his p. 44) assumes an account of simplicity, according to which a theory is simpler the fewer the number of computational symbols needed to express that theory. This 'algorithmic' account has the consequence that, for example (p. 44), the 'set of all perfect fluid solutions to the Einstein field equations has a smaller algorithmic complexity than a generic particular solution, since the former is specified simply by giving a few equations and the latter requires the specification of vast amounts of initial data on some hyper-surface'. So it is simplest of all to postulate that every possible universe exists, since that needs very few computational symbols indeed to state!
Tegmark's account of simplicity seems to me to yield in this case a bizarre result, totally out of line with all our inductive practice. If we are postulating entities to explain phenomena, we postulate the fewest number of entities needed to do the job. If we did adopt Tegmark's approach, we would need to amend and amplify his theory in two crucial respects. First, we would need to amend it to deal with the problem that the supposition that all possible entities exist is incoherent For the existence of some entities rules out the existence of others. Thus, the existence of an omnipotent all-good God rules out
186 Teleologkal Arguments
of a law whereby universes produce daughter universes differing from them in boundary conditions and laws; and so our universe is explained as one of a collection of an infinite number of universes (originally causally connected with each other) differing from each other in boundary conditions and laws. But that is tantamount to postulating a multiverse that has laws and boundary conditions such that it will contain at some time or other a tuned universe. But then there are an infinite number of logically possible multiverses that do not have this characteristic, and the shape of the problem has in no way changed. For the problem that concerns us is not really why is there one (in my sense) universe that is tuned for life, but why among all the universes there are (one or many) is there a universe tuned for life. One way in which this could come about is by there being only one such universe. But another way is by there being a universe-generating mechanism that produces universes of various lands, including a universe tuned for life. But, although the existence of this possibility does not change the shape of the problem, it draws our attention to a way in which a universe tuned for life could have come into existence. And so, in order to assess the intrinsic probability that there be a universe tuned for hfe, we need to assess the probability that that would come about by one or other route. And taking this into account may lead us to reassess the value of that probability.
It might seem that the value would turn out to be much higher than we originally supposed. Let us individuate universe-generating mechanisms by the multiverse (the collection of universes) that they generate (at some time or another). Then, if we consider all the possible multiverses, each consisting of r universes, chosen from n logically possible lands of universe only one of which is fine-tuned for life, it follows mathematically that a proportion        of these
the existence of an omnipotent Devil (in any actual universe at all). And, given God, he will certainly not choose to bring about the existence of certain other states—for example, endless suffering unchosen by the sufferer. And, secondly, the account of the simplicity of a theory in terms of the fewness of computational symbols needed to express it, which led to the claim that it is simple to suppose that every possible universe exists, needs considerable amplification. For how many symbols you need to express something depends on the language you use. AH theories can be expressed in the form 'a = V, when a and b represent some highly complicated multidimensional tensors. But, of course, it needs a language far removed from the language of observation to express the theory in that way. Tegmark's account of simplicity is not a dear one and its consequence in the case of current interest to us is bizarre and contains a contradiction.
Teleologkal Arguments 187
multiverses will contain a universe fine-tuned for life. For any r > l(r = 1 being the case where there is only universe), this will exceed - (the proportion of universes fine-tuned for hfe). And the more universes in a multiverse (the larger is r), the closer this value will be to 1. So it might seem that, as we consider more and more possible universe-generating mechanisms (generating more and more universes) (r getting larger and larger), the total proportion of universe-generating mechanisms that will generate a universe fine-tuned for life will approach 1. So if it were equally probable that there exist any possible universe-generating mechanism (most of them generating far more universes than the number of logically possible kinds of universe), it would seem to be very probable that there would occur at least one universe fine-tuned for life.
However, we cannot calculate the intrinsic probability (in a Godless world) of a universe-generating mechanism being such as to produce a universe tuned for life merely by counting the proportion of mechanisms that have this characteristic among the total number of possible such mechanisms. To start with, there will be an infinite number of possible mechanisms of which an infinite number will have the required feature. And infinity divided by infinity has no definite value. We have to divide up mechanisms into a finite number of kinds of mechanism, and then weight each kind by the prior probability of a mechanism being of that kind—which will be a function of the simplicity of the laws involved in the mechanism. Now, clearly mechanisms that yield universes varying from each other only in the constants involved in their laws will be much simpler than mechanisms that yield universes differing in the kinds of laws they have. A mechanism that produced universes with laws of totally different kinds from each other would need itself to be governed by some very complicated laws. Yet, if we are confined to mechanisms that yield only laws of one kind, my earlier arguments suggest that very few such mechanisms yielding only laws of relatively simple kinds (that is laws no more complex than are those of our universe) will yield a universe fine-tuned for life. Secondly, mechanisms that produce universes with simple laws are simpler and so intrinsically more probable than mechanisms that produce universes with more and more complex laws as welL And, thirdly, the existence of a multiverse with a universe-generating mechanism is a more complex supposition than the existence of one universe without such a mechanism.
188
Teleological Arguments
Teleological Arguments
189
So, even if there is a large range of possible multiverses tuned for life (in the sense of producing a universe tuned for life), and the proportion of the range of possible multiverses tuned for life is vastly greater than the proportion of the range of single universes so tuned, this holds only because the former range includes very complex multiverses that are mtrinsically very improbable. So I stick by my point that it is intrinsically very improbable that there be a universe tuned for life (whether it is a sole universe, or a universe produced by a universe-generating mechanism). Yet it may well be that this improbability is less than the improbability that a single universe would be tuned for life.
The Probability of Spatial Order, given Theism
A God, however, I argued earlier, has good reason for bringing about embodied humanly free agents, such as human beings appear to be; and so, on the hypothesis of theism, it is moderately probable that the universe will be tuned—that is, such as to allow and indeed make significantly probable the existence of human bodies. God could achieve this either by creating such bodies entire, or by creating and keeping in existence a universe designed to bring them about by a long evolutionary process, or even a multiverse designed to bring about such a universe.
What reason would God have for taking an evolutionary route? If his only aim in creating a universe was to populate it with human beings, there would indeed be no point in producing them by a long evolutionary process. But there are other good features of the universe that God has good reason to bring about I have commented already on the beauty of the manimate universe shown in the development of galaxies, stars, and planets. God has every reason for bringing about this process of development from the Big Bang for its beauty—even if he were the only person to observe it But, of course, he is not the only person to observe it. We can observe it through our telescopes reaching further and further back into the earliest stages of the universe. And God has the same reason for bringing about plants and animals—their beauty. And animals are good also, I have argued, in virtue not only of their beauty but also of their ability to have pleasant sensations and true beliefs and spontaneously to do good actions (even if not ones freely chosen). In view of all this, it is not too
surprising that God should take the long (by our timescale) evolutionary route to produce human bodies. And similar, though weaker arguments would show it to be unsurprising if God produced human bodies by an even longer route of going through more than universe to achieve this goaL
It may be that even given the initial conditions of the universe in all their detail, the laws of nature as such do not necessitate the evolution of human bodies, only make it quite probable. As I wrote earlier, it may be that the way by which God ensures that human free choices make differences to the world is by bringing it about that the fundamental laws of nature are probabilistic not fully deterministic And clearly God can guide the way in which the probabilistic laws operate so as to ensure that human bodies do evolve, without in anyway preventing their operation, simply by ensuring that the most probable outcome does occur. Yet there will be an argument from the existence of human (and animal) bodies to the existence of God of any great strength, via the route of 'fme-tuning', only if it follows that a fine-tuned universe will (not merely possibly but with significant probability) lead to embodied humans and animals. For fme-tuning as such is merely a necessary, and not a sufficient, condition for the evolution of humans and animals. There is, however, a very considerable, but not unanimous, scientific view that the laws and initial conditions of our universe make it very probable that human life will evolve in more than one place in the universe, and animal life will evolve in quite a number of places. And that is enough to make the argument a cogent one.
So it is quite likely that, if there is a God, the laws and boundary conditions of the universe will be such as to make probable the evolution of human bodies. It is otherwise very improbable that they will have this feature. I represent this evidence of the nature of the laws and boundary conditions as e, with h as the hypothesis of theism, and k as the background knowledge that formed the evidence of the two arguments considered previously—that there is a universe governed by simple laws of nature. The probability then, if there is no God, that the laws and boundary conditions will be such as to have this further feature of bringing about human bodies is P(e| ~/j&fc). The probability that this will happen if there is a God is P(e\h8*k). I have argued that P(e\h&k) » P(e| ~h&it), and so—by Bayes's theorem—P(ft|e& h) » P(h\k). We have here a powerful C-inductive argument for the existence of God.
190
Teleological Arguments The Argument from Beauty
The strength of the argument from the universe and its spatial and temporal order to God is increased when we take into account the beauty of that universe. As we have noted, the universe is beautiful in the plants, rocks, and rivers, and animal and human bodies on Earth, and also in the swirl of the galaxies and the birth and death of stars. Mark Wynn comments that nature is urnformly beautiful whereas the products of human beings are rarely beautiful in the absence of artistic intent'. I argued in Chapter 6 that, if God creates a universe, as a good workman he will create a beautiful universe. On the other hand, if the universe came into existence without being created by God, there is no reason to suppose that it would be a beautiful universe. The argument has force on the assumption with which I am happy and commend to my readers that beauty is an objective matter, that there are truths about what is beautiful and what is not. If this is denied and beauty is regarded as something that we project onto nature or artefacts, then the argument could be rephrased as an argument from human beings having aesthetic sensibilities that allow them to see the universe as beautiful. In the latter case, there is certainly no particular reason why, if the universe originated uncaused, psycho-physical laws (of the kind that I shall consider in the next chapter) would bring about aesthetic sensibilities in human beings. But, good though it is that humans should have these sensibilities, it would need to be shown that it would be involved in the equal best kind of act that constituted the creation of humanly free agents to endow them with aesthetic sensibilities.
For not to do would not deprive the universe of a kind of sensibd-ity, since God could himself have it whereas the ability to make significant choices between good and evil is not a kind of goodness that God himself could have. Because the argument from beauty needs, I suspect, an objectivist understanding of the aesthetic value of the universe, in order to have significant strength, and the establishment of such an understanding would require very considerable argument, I shall omit further discussion for reasons of space.32
52 An argument to God from the beauty of the world was presented by F. R. Tennant in his Philosophical Theory, vol Z, The World, the Soul and God (Cambridge University Press, 1930). There is a good short presentation of this argument and response to objections to řt in Mark Wynn, God and Goodness (Roirdedge, 1999), ch 1. For the quotation from Wynn, see ibid. p. 20.
r
Teleological Arguments
191
I should add that this point does not undermine the earlier point that the beauty of the physical universe (whether objective, or subjective in its perception by persons) provides a good reason for God to produce human bodies by the evolutionary route; my point here is simply that it needs much further discussion to show that the beauty of the physical universe provides a positive argument of significant strength for the existence of God.