INFERENCE TO THE BEST EXPLANATION AND THE

As different as their approaches and substantive conclusions are, both Richard Swinburne and J. L. Mackie agree that the most interesting and promising arguments in natural theology have the structure of inferences to the best explanation.

[These] arguments can be seen as resting on one general principle, or as sharing one basic form and purpose: they are arguments to the best explanation. The evidence supports the conclusion, it is suggested, because if we postulate that the conclusion is true--or better, perhaps, that it is at least an approximation to the truth--we get a more adequate overall explanation of that whole body of evidence, in light of whatever considerations are cited, than would be given by any available alternative hypothesis.^[1]

This simple structure covers a wide range of potential religious evidence. The theistic hypothesis is offered by a number of thinkers as the best -- the simplest, most complete, most plausible -- explanation of the relevant data. Different types of data or evidence give rise to the classical arguments of natural theology. From the most general and abstract, that the physical universe exists in the first place, to very specific reports of mystical experiences or observed miracles, the existence of God is offered as a general explanation of these contingent and empirically discovered facts. According to the model of inference to the best explanation this data counts as good religious evidence just in case the theistic hypothesis is, not just a possible, explanation, but substantially better than any competing account.

The structure works equally well for negative religious evidence -- what Plantinga has been calling natural atheology. Why is there so much pain, suffering and misery in the world? Because, the atheist responds, there is no omnipotent and morally perfect God who is in any position to do anything about it. The standard theodicies -- free will, Satan, soul-making -- can be conceded as logically compatible with evil, yet the natural theologian can still argue that the evidential case against God's existence remains strong because the atheist's hypothesis is a better explanation of pain and suffering.

Inference to the best explanation is an amazingly powerful heuristic device. It allows us to see the inferential structure of many cases of scientific reasoning; consider for a moment how well the explanatory schema fits the evidentiary case for natural selection, or plate tectonics, or the big bang. There have been many outspoken critics of inference to the best explanation, but they have generally attributed too much to the model. It is no magic formula for flawless inductive reasoning. The best the D.A. can do is to assume that on the basis of some shared experiences, competencies and values, the jury will agree with the state's theory that the defendant's guilt best explains the finger prints, reports from witnesses, and the obvious motive. Each juror will have to search his or her own conscience and make a judgment about whether the state's explanatory hypothesis, or the defendant's rival account, best explains the evidence presented at trial.^[2] Thus, the process is fundamentally subjective, but we hope for intersubjective agreement. It remains unclear whether religion is an area where we can realistically hope for this kind of explanatory consensus.

Some critics are concerned that such an approach allows for the possibility of good evidence for false theories.^[3] Indeed the model assumes a clear distinction between truth and good evidence. Evidence evaluation is time dependent. As things stand now, a given explanation is better than any rival. Things could change in the future. Additional data may become available which significantly alters our judgments about explanatory adequacy; a rival that originally looked weak may now seem the strongest. New evidence might even suggest a completely new account, one that we hadn't even considered in our original assessment of the evidence, but now must concede better explains everything we now know.

The teleological argument calls attention to the remarkable degree of structure, order and purpose we find in the physical universe.

Look round the world, contemplate the whole and every part of it: you will find it to be nothing but one great machine, subdivided into an infinite number of lessor machines which again admit of subdivisions to a degree beyond what human senses and faculties can trace and explain. All these various machines, and even their most minute parts, are adjusted to each other with an accuracy that ravishes into admiration all men who have contemplated them. The curious adapting of means to ends, throughout all nature, resembles exactly, though it much exceeds the productions of human contrivance.^[4]

We see a "curious adapting of means to ends, throughout all nature." "[P]arts are adjusted to each other with [great] accuracy." How is this possible? The theist's answers with a proposed inference to the best explanation.

EVIDENCE: The physical universe exhibits great structure, order and apparent purpose.

Here is a classic case of circumstantial evidence. Just like the finger prints on the dagger, or the telltale lipstick on the collar, we infer from things we can see -- features of the natural world -- to things we can't see -- the workings of an omnipotent and morally perfect creator.

The philosophical tradition has spent a lot of time and effort trying to diagnose alledged methodological fallacies within the teleological argument. Thus we here that it is too dependent on a shakey analogy, or that it tries to generalize from a single case -- the creation of the universe, or that at its heart it depends on the discredited ontological argument. At the time of Hume and Kant's sustained analysis and critique, the truth is that neither the philosophical or scientific communities had any serious candidate for a better explanation of biological order, structure and purpose. But, within well less than a century both communities were well aware of a very serious and challanging rival to the theistic hypothesis. The theory of natural selection provides a purely naturalistic explanatory structure for the gradual development of biological structure, order and purpose. Its first articulation by Darwin and Wallace only contained the barest outline of this rival account. By the close of this century, however, we have a unified bio-chemical paradigm which, in principle, gets us from increasingly complex chemical molecules, to chemical replication, to life, to evolution by natural selection, and to potential cognitive scientific accounts of consciousness.

Wise theists long ago conceded the basic truth of natural selection, arguing that the theory is perfectly compatible with the existence of God. When I describe natural selection as a rival account, I do not mean to imply that the theory logically rules out the existence of God. Natural selection might well have been the mechanism of choice for an infinitely intelligent creator. But, the theory in no way depends on the supernatural component. Darwin did such serious damage to traditional religion, not because he contradicted Genesis, or because he produced scientific evidence against God, but rather because he gave us the first, and still only, naturalistic explanation for biological structure, order, and purpose. He produced a better explanation, and in the process undermined the evidential value of the teleological argument.

The most eloquent and persuasive articulations of the teleological argument in the eighteen and nineteenth centuries always emphasized biological structure, order and purpose. Paley's extended discussion of comparative anatomy is only the most familiar example. The success of the Darwinian rival suggests an obvious counter strategy. Rather than focus on the biological, perhaps the natural theologian should stress the structure and order in physics, chemistry and astronomy. Indeed, the is ample precedent for exactly such an appeal. Newton, for example, insisted that:

This most beautiful system of the sun, planets and comets, could only precede from the counsel and dominion of an intelligent and powerful being.^[5]

Swinburne argued, first in 1968,^[6] and then in his 1979 book,^[7] that this is precisely the appropriate strategy. He suggests that if we change the focus from "regularities of co-presence" which stress spatial order, to "regularities of succession" which stress temporal order, the teleological argument can "be reconstructed in a form which does not rely on the premises show false by Darwin."^[8] The key to Swinburne's "new" teleological argument is the orderliness of which allows for the articulation the basic mathematical laws of the natural sciences.

The orderliness of the universe to which I draw attention here is conformity to formula, 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.^[4]

Swinburne's restatement of the argument from design has widely been acknowledged as provocative and important. It has not, however, led to any kind of basic reevaluation of the teleological argument. I would hazard the following guess as to why. The orderliness of the universe does seem to require some explanation, but we have been willing to accept a promissory note from physics that a purely naturalistic one shall be forthcoming. After all, any physical system must have some structure and order, regardless of how dynamic or fluid. When Paley focused on the human eye, there was not just orderliness, but intricacy and complexity, to say nothing of functionality; the demand for an explanatory outline could not be deferred for further developments in the sciences. Simple law-like regularity is interesting and important, but does not have the evidential punch of biological order. Still, Swinburne's strategy is amazingly prescient, both by de-emphasizing biology to the advantage of physics and chemistry, but also because the orderliness of nature seems to be a prerequisite for life and intelligence.

A growing number of physical scientists have felt compelled to comment on the surprising fact the physical universe, at least this region of it, seems exquisitely designed to permit the evolution of intelligent beings. Not all of these scientists, by any means, infer that the universe must have been designed by an intelligent creator, but they do remark on how the slightest readjustment of some of the most basic constants in nature would have given rise to a universe that was totally inhospitable to life, or perhaps would not have given rise to a universe at all. Hawking puts this point as follows.

The laws of science, as we know them at present, contain many fundamental numbers, like the size of the electric charge of the electron and the ratio of masses of the proton and the electron. . . . The remarkable fact is that the values of these numbers seem to have been finely adjusted to make possible the development of life. For example if the eclectic charge of the electron had only been slightly different, stars either would have been unable to hydrogen and helium, or else they would not have exploded.^[10]

Many of the discussions of what is sometimes called the "cosmic coincidences," or the "fine-tuning" of the physical universe, quickly exceed the scientific competencies of the average reader. We are well advised, therefore, to state the evidential case as generally as possible.

DATA: The physical universe exhibits exceedingly fine-tuned structure and order. This precise structure and order seems to be a prerequisite for the evolution of conscious intelligence.

Several things require comment. The data here, in contrast to Swinburne's regularities of succession, really does illustrate "an accuracy that ravishes into admiration all men who have contemplated them." As impressive as the structure of the eye or the gene is, one has to be astounded to learn that immediately following the big bang

[t]he cosmos threatened to recollapse within a fraction of a second or else to expand so fast that galaxy formation would be impossible. To avoid these disasters its rate of expansion at early instants needed to be fine-tuned to perhaps one part in 10⁵⁵.^[11]

Such accuracy does demand explanation. The second thing to note, however, is that most of us must be tutored to even appreciate this data. Hawking's point about relative masses of protons and electrons, or the so-called "flatness problem" just above can only truly be appreciated by those with years of mathematical and technical training; the rest of us must simply take the scientist's, or the popular writer's, word for the amazing intricacy in the physical universe. This is quite different from biological structure, order and purpose, where anyone with reasonable intelligence and sensitivity can immediately see the explanatory problem. The final general point about cosmic fine-tuning is that kind of examples we have been looking at can be multiplied many times over. John Leslie devote two and a half chapters simply to articulating what he calls "the evidence of fine-tuning."^[12]

All of the above talk about fine-tuning and adjustment of values, naturally enough suggests the possibility of an intelligent fine-tuner and adjustor. What I labeled above as the DATA can obviously be construed by the natural theologian as evidence. And given the model of inference to the best explanation, the following evidential case demands careful analysis.

EVIDENCE: The details from the physical sciences regarding cosmic coincidences and fine-tuning.

EXPLANATION: God exists, and designed the physical universe in such a way that intelligent life would eventually evolve.

This inference is what I am calling the new teleological argument. Whether the evidence that is presented here is good or not depends, at least partially on the rival explanations that challenge this interpretation of the data.

Why is the physical universe so fantastically fine-tuned to permit the evolution of intelligent creatures? One answer to this question that is receiving considerable recent discussion sounds almost like a joke. Why?, because if it were not, you wouldn't be here to ask the question in the first place. A little less flippant statement of the anthropic principle is provided by Barrow and Tippler.

Weak Anthropic Principle (WAP): The observed values of all physical and cosmological quantities are not equally probable but they take on values restricted by the requirement that there exists sites where carbon-based life can evolve and by the requirement that the Universe be old enough for it to have already done so.^[13]

The weak anthropic principle reminds us that there is a selection bias at work in our observation of the physical universe, after all carbon-based life is doing the observing and questioning. Noting such a selection bias can be genuinely explanatory. Why do all these cars have Oregon license plates? Oh yeah!, I'm in Oregon now.

Nevertheless, the anthropic principle is not a very satisfactory rival explanation for why the universe is so finely-tuned. We can be thankful that it is; we would be here if it weren't; but we are still puzzled as to why it is. Swinburne, in a very recent piece, presents a great example for showing the explanatory inadequacy of the anthropic principle.

On a certain occasion the firing squad aim their rifles at the prisoner to be executed. There are twelve expert marksmen in the firing squad, and they fire twelve round each. However, on this occasion all 144 shots miss. The prisoner laughs and comments that the event is not something requiring any explanation because if the marksmen had not missed, he would not be here to observe them having done so. But of course the prisoner's comment is absurd; the marksmen all having missed is indeed something requiring explanation; and so too is what goes with it-‑the prisoner's being alive to observe it.^[14]

The anthropic principle can be salvaged as a partial account of the fine tuning data, if it is coupled with an additional strategy. What is needed is a huge number of spatially and temporally distinct parts, or perhaps aspects, of physical reality, where vastly different physical laws and constants hold. One finds in this literature the cliche of a roomful of monkeys at keyboards eventually producing a Shakespearean sonnet. Swinburne's firing squad story is adequately explained if we learn that there have been twenty-nine trillion similar firing squad scenarios in the indefinite history of the Universe, and only in three were there survivors lucky enough to listen to anthropic accounts.

The anthropic principle, along with concept of multiple universes, is part of a purely naturalistic account of the physical order, structure, and fine tuning we find in this (part of) the universe. The story, which is winning wide consensus. is actually composed of at least four distinct parts.

The beginning, of course, is the widely acknowledged big bang account of the origin of the physical universe. The second component is the theory of cosmic inflation. It is important enough to warrant some extended discussion.

Alan Guth reminds us that three of the cosmic coincidences are left totally unexplained by the big bang hypothesis.

2: How did the universe become so homogeneous on large scales? Do we have to assume that it started out that way?

3: Why is the mass density of the early universe so extraordinarily close to the critical density?

4: Can one find a physical origin for the primordial density perturbations which lead to the evolution of galaxies and clusters of galaxies? Are there physical processes which determine the spectrum of these perturbations?^[15]

We have already mentioned question 3 as the "flatness" problem. If the density of the universe were just a little greater, the universe would be tightly closed and would have collapsed due to the pressures of gravity long before stars and galaxies, let alone intelligent life, could have evolved. Were it a little less, the universe would be open, and the matter created would have rapidly dispersed before atoms, to say nothing of galaxies or humans, could have formed. As it is, the density is just right, it is approximately Euclidian, or "flat." This allows a formed physical universe with the necessary longevity to allow the evolution of intelligent observers. Questions 2 and 4 focus on the "smoothness" problem. The universe seems remarkably homogenous, this seems a requirement for regular cosmic expansion and longevity of the universe. At the same time, there seems to be just the right amount of inhomogeneity to allow the forces of gravity to begin the process of star and galaxy formation. To get some sense of the incredible precision of these processes, consider the following estimate of the flatness accuracy.

Pushing back to . . . the nearest definition we have to a beginning, the flatness of the universe must have been precise to within 1 part in 10⁶⁰. This makes the flatness parameter the most accurately determined number in all of physics.^[16]

Cosmologists now believe that these coincidences can be explained by the hypothesis that the very early universe underwent a process of inflation.

During this supercooling phase, the calculations show, the Universe will be driven into a wild, exponential expansion -- inflation. It will double in size roughly once every 10^-34 seconds, which doesn't sound too impressive until you realise that this means a hundred doublings in 10^-32 seconds, enough to expand a tennis ball to the size of the entire observable universe.^[17]

We are told that fairly reliable accounts of smoothness and flatness are natural consequences of the inflationary account. Less satisfactory, but still reasonable, accounts of the universe's inhomogeneity also follow. Inflation does not account for everything, however. Thus, contemporary cosmology must supplement the inflationary hypothesis with the multiple universes hypothesis, and confidence in the ultimate development of a "theory of everything" which accounts for the four basic physical forces.

Such a secular-naturalistic account is breath taking in its ambition and partial theoretical success. There is wide agreement that its support comes almost entirely from theoretical considerations, indeed observational confirmation is often ruled out on both theoretical and practical grounds. At this stage, this explanatory strategy demands faith in a worked out account of the "primordial density perturbations in the universe,"^[18] or a satisfactory theory of everything. Nevertheless, whenever we find such apparent consensus and optimism within a respect scientific community, we are bound to treat the explanatory hypotheses with respect, and some presumption of plausibility. Thus, a challenging rival explanation for the physical fine tuning is in the process of being developed, and must be integrated into any final assessment of the new teleological explanation.

It would be nice to close this discussion with a rousing endorsement or rejection of the evidential value of the new teleological argument. Several factors make such a clean conclusion impossible. I will, therefore, be content with a numbers of very short and somewhat scattered remarks.

Perhaps most importantly, the new teleological argument counts as a significant improvement on the old version of the design argument. Natural selection is clearly a better explanation of biological order, structure and purpose than the theist's design hypothesis. Thus, according to the model of inference to the best explanation, these biological facts have little evidential value in support of theism. Physical structure and order, what we have been calling cosmic coincidences and evidence of fine-tuning, are not nearly so easily explained away in a purely naturalistic model. We have seen that the beginning of such a model are well underway toward articulation. But at this point, even within the community of physicists, this rival account does not enjoy anything like the clear consensus that we find for Darwin. Thus, those who simply reject the project of natural theology as methodologically moribund may be premature. The new teleological argument presents a challenging case, and one that cannot simply be dismissed with time honored arguments or yesterdays scientific knowledge.

Hume's discussion of the teleological argument continues to be relevant. In particular, his introduction of the problem of evil into the analysis of the argument from design is prescient. If we are trying to make a case for theism based on observed facts about the natural world, we must consider all the relevant evidence. The D.A. who presents her case based on witnesses, finger prints and motives, but neglects to mention that the defendant's alibi has been corroborated, simply cheats. Theists who appeal to the evidence of fine-tuning seem honor bound to mention the disconfirming evidence provided by pain, suffering and misery. I have argued elsewhere that the model of inference to the best explanation allows us to see just how powerful an evidential case can be mounted against the existence of the God of western theism.^[19] Evil is particularly relevant to our discussion because it reminds us that at least three explanatory candidates are vying for our attention.

C₁. Some other metaphysical being or presence, consciously designed the physical universe.

C₂. The physical structure and order we find in the universe, including the evidence of cosmic coincidences and fine-tuning, are all accounted for in a completely naturalist model.

In my judgment, because of the great amount of moral and physical evil that pervades this world, C₀, is much weaker than either of its two rivals. Indeed, two of the most articulate defenders of the new teleological argument, Paul Davies, a physicists, and John Leslie, a philosopher, both see the evidence as pointing toward a nonstandard conception of "God."

I hesitate to mention a last worry about inference to the best explanation and the new teleological argument, but it is so obvious it has to be treated in some fashion. Those of us who believe that reasoning -- scientific, theological, or common sense -- is influenced by cultural factors, have to worry about a kind of incommensurability between theists and natural scientists on the assessment of this kind of data. As Gerald Doppelt persuasively argued, the most invidious kind of incommensurability is not so much concerned with semantics and linguistic misunderstanding, as with a shift-in-standards, "one which relativizes scientific evaluation to the standards internal and specific to particular physical theories."^[20] It seems reasonable to treat physicists and theologians as members of different paradigms or interpretive communities. But, if because of this different background they are doomed to apply different criteria for simplicity, plausibility, and general explanatory adequacy, this clearly bodes ill for productive natural theology. The apparent success that these two communities have recently had in discussing cosmic fine-tuning, however, leaves room for hope that productive discussion, if not consensus, remains a possibility.

^[1] J. L. Mackie, The Miracle of Theism, Oxford: Oxford University Press, 1982, p.4.

^[2] Note about special context of criminal trials.

^[3] Note about van Fraassen and Cartwright.

^[4] David Hume, Dialogues on Natural Religion,

^[5] Newton, quoted in Davies, p. 201.

^[6] Swinburne's article on the teleological argument

^[7] Swinburne, The Existence of God

^[8] Swinburne, 1979, p. 135.

^[9] Swinburne, 1979, p. 136.

^[10] Hawking, p. 125.

^[11] Leslie, p. ??

^[12] John Leslie, Universes.

^[13] Barrow and Tippler, p. 16.

^[14] Swinburne, 1989, p. 165.

^[15] Guth, p. 11.

^[16] Gribbin and Rees, p. 26.

^[17] Gribbin and Rees, p. 278.

^[18] Guth, p. 34.

^[19] My paper, reference deleted.

^[20] Gerald Doppelt, pp. 129-30.