Phillip Johnson on Trial:
A Critique of His Critique of Darwin

NANCEY MURPHY

Fuller Theological Seminary

School of Theology
Pasadena, CA 91182

From: PSCF 45 (March 1993): 26-36                                                                 Response: Claassen

Phillip Johnson's recent book, Darwin on Trial, claims to show that the reasoning presented in favor of evolutionary biology is defective. Such a book, being one of so many, would excite little attention were it not for the fact that the author is an expert in legal reasoning, and has contributed his particular skills to the debate. However, the canons of scientific argument are quite different from those of the courtroom, and it can be shown that Johnson's critique of Darwinian thought falls far short of the mark in that it does not fully appreciate the special requirements of scientific argumentation.

 Introduction

Phillip E. Johnson's recent book, Darwin on Trial1 has attracted a fair amount of attention among conservative Christians. Yet it may create an inaccurate impression of the status of evolutionary biologyˇan impression that I hope to correct in this article. On the book's dust jacket it is said that Johnson, a professor of law at the University of California at Berkeley, took up the study of Darwinism because he judged the books defending it to be dogmatic and unconvincing. I, at least, find Johnson's own arguments dogmatic and unconvincing. The main reason is that he does not adequately understand scientific reasoning.

Many readers will be impressed, even overawed, by Professor Johnson's credentials. He is not a scientist but a lawyer, who claims that his law career, with "a specialty in analyzing the logic of arguments and identifying the assumptions that lie behind those arguments" well qualifies him for the task (p. 13). The fact that he is a professor at U.C. Berkeley certainly adds to his credibility in the eyes of many. But I wish you would bear with me in a little foolishness (cf. 2 Cor. 11:1). Is he from Berkeley? So am I. One of my doctorates was earned in the philosophy department at U.C. Berkeley, where I specialized in philosophy of science. Is he an expert in critical reasoning? So am I. I teach critical reasoning to seminary students (now at Fuller Theological Seminary) and have just completed a textbook on the subject. Most of my other research and writing deals with methodological issues in theology, science, and the relations between the two.2

My plan is to describe some of the basic moves in scientific reasoning, and then examine in detail an important (and typical) passage in Johnson's book, explaining why it appears fallacious to one trained in scientific reasoning. Next, I shall describe some recent refinements in philosophers' understanding of scientific reasoning, and use them to describe the sort of study that would be required to make a fair assessment of the scientific standing of evolutionary biology.

Another issue that needs to be addressed is the very nature of science, and how it relates to religion. A bit of history will help us understand some of the positions taken by evolutionary biologists and excuse them from some of Johnson's criticisms.

I shall end with a few remarks on what I take to be the proper attitude for Christians toward evolutionary biology.

Before I proceed to the attack, however, I must say that Johnson's book has many good features. Johnson describes some of the failures and problems faced by evolutionary biology, and provides a valuable critique of popular writings that turn the science of evolutionary biology into an atheistic metaphysical system with many of the trappings of religion.

Basic Scientific Reasoning

Francis Bacon's description of scientific reasoning has been influential for many years. In brief, he claimed that scientists must first rid their minds of all prejudice and preconceptions, then collect all the facts relevant to the issue at hand, and finally draw inductive inferences from the facts.3 This view of scientific reasoning is inadequate, however, since it only accounts for our knowledge of observed regularities. An important advance in the philosophy of science of this century was the recognition of what has been called "hypothetico-deductive" reasoning.4 This kind of reasoning frees science from dependence on direct observation, and accounts for all of our theoretical knowledge. It is called "hypothetical" because it relies on the formation of hypotheses to explain a given set of data or observations. It is called "deductive" because hypotheses must be tested by drawing conclusions from them and seeing if they are corroborated by further observation or experiment. So the test of a hypothesis is not by direct observation (most scientific hypotheses postulate unobservable entities or processes), but by asking what observable consequences follow from the hypothesis, and by testing these instead. Another way of putting the matter: a hypothesis is accepted on the basis of its ability to explain observations and results of experiments.

Consider the following analogy drawn from everyday experience. You come home from work and find the front door ajar and muddy tracks leading into the kitchen. You form a hypothesis: the kids were here. But of course, there could be other explanations, such as a prowler. To test the hypothesis, you make predictions based on your knowledge of the children's behavior. For example, you check to see if anyone has been into the cookies, or if their school clothes are on the floor upstairs. If your predictions are confirmed you do not need to see the children to know that your original hypothesis was correct.

So the form of hypothetico-deductive reasoning is as follows:

 ´We observe O1.

´We formulate a hypothesis (H), which, if true, would explain O1.

´Then we ask, if H is true, what additional observations (O2.... On) ought we be able to make?

´ Finally, if O2 through On are observed, H is confirmed.

It is important to note that O2 through On are not equivalent to H; they are observable consequences that we deduce from H with the aid of additional assumptionsˇnibbled cookies and strewn clothing are not children.

Because the hypothesized entities or processes are unobservable, scientists often make use of modelsˇobservable entities or processes that are similar in important respects to the theoretical entities. A famous example is the billiard-ball model used to understand and account for the behavior of gasses in a closed container. Models are often helpful in deriving testable predictions from hypotheses (theories).

It is also important to note that hypothetical reasoning (like all reasoning about matters of fact) can never amount to proof. The best that can be hoped for is a high degree of confirmation. Much of what philosophy of science is about is examination of the conditions under which a scientific theory can be said to be well-confirmed. So objecting that any scientific theory is "not proved" is emptyˇnone can be.

The foregoing provides enough terminology to analyze some of Johnson's arguments, so we turn now to these.

Johnson on Natural Selection

Chapter Two of Darwin on Trial is an examination of the thesis that natural selection, or survival of the fittest, (when combined with natural variation) provides an adequate account of macroevolutionˇthat is, the evolution of all known species of living things from one or a few primitive ancestors. A crucial step in Johnson's overall criticism of evolutionary biology is his assessment of evidence for the efficacy of natural selection, so we must examine this short passage (pp. 17-20) with care. Johnson begins by noting that Darwin could not point to examples of natural selection in action, and so he had to rely heavily on an argument by analogy with artificial selection by breeders of domestic plants and animals.

However, Johnson replies to Darwin's argument as follows:

Artificial selection is not basically the same sort of thing as natural selection, but rather is something fundamentally different. Human breeders produce variations among sheep or pigeons for purposes absent in nature, including sheer delight in seeing how much variation can be achieved. If the breeders were interested only in having animals capable of surviving in the wild, the extremes of variation would not exist... .

What artificial selection actually shows is that there are definite limits to the amount of variation that even the most highly skilled breeders can achieve. Breeding of domestic animals has produced no new species, in the commonly accepted sense of new breeding communities that are infertile when crossed with the parent group.. .

In other words, the reason dogs don't become as big as elephants, much less change into elephants, is not that we just haven't been breeding them long enough. Dogs do not have the genetic capacity for that degree of change, and they stop getting bigger when the genetic limit is reached (p. 18).

Next, Johnson turns to evidence cited by contemporary evolutionists:

Darwinists disagree with that judgment, and they have some points to make. They point with pride to experiments with laboratory fruitflies. These have not produced anything but fruitflies, but they have produced changes in a multitude of characteristics. Plant hybrids have been developed which can breed with each other, but not with the parent species, and which therefore meet the accepted standard for new species. With respect to animals, Darwinists attribute the inability to produce new species to a lack of sufficient time.. . In some cases, convincing circumstantial evidence exists of evolution that has produced new species in nature. Familiar examples include the hundreds of fruitfly species in Hawaii and the famous variations among "Darwin's Finches" on the Galapagos Islands... .

Lack of time would be a reasonable excuse if there were no other known factor limiting the change that can be produced by selection, but in fact selective change is limited by the inherent variability in the gene pool. After a number of generations the capacity for variation runs out. It might conceivably be renewed by mutation, but whether (and how often) this happens is not known (p. 19).

And now Johnson's conclusion, drawn from the above considerations:

Whether selection has ever accomplished speciation (i.e., the production of a new species) is not the point. A biological species is simply a group capable of interbreeding. Success in dividing a fruitfly population into two or more separate populations that cannot interbreed would not constitute evidence that a similar process could in time produce a fruitfly from a bacterium. If breeders one day did succeed in producing a group of dogs that can reproduce with each other but not with other dogs, they would still have made only the tiniest step towards proving Darwinism's important claims.

 That the analogy to artificial selection is defective does not necessarily mean that Darwin's theory is wrong, but it does mean that we will have to look for more direct evidence to see if natural selection really does have a creative effect (pp. 19-20).

Analysis

What are we to make of this set of arguments? Before I begin a serious analysis, permit me another bit of foolishness: The series of steps in Johnson's argument recalls an old lawyer's joke about a defendant in a murder trial: "Your honor, I didn't kill him, and besides, it was an accident, and on top of that he really had it coming!" Similarly: artificial selection is not analogous to natural selection, and besides, selective breeders have not produced any new species, and on top of that they have only produced new plant species, but no new animal species.


We must ask what observations or results
are required to confirm (not prove) the scientific hypothesis
that natural selection is capable of producing
radically different new species


More seriously, we must ask what observations or results are required to confirm (not prove) the scientific hypothesis that natural selection is capable of producing radically different new species.5 Since we cannot directly observe natural selection at work, we need an observable model. Selective breeding has been proposed. (We will come back to the issue of the suitability of this model below.) What is at stake in testing the power of natural selection, then, is that our analogue to natural selection be shown to accomplish two things: First, we need to see that selection can produce radical differences within a population. Second, we need to see that selection can result in speciationˇthe development of one species out of another. The criterion here is incapacity to breed with the parent species.

Johnson seems to believe that both of these effects need to be observed in the same instance. He would have a point if there were something about one effect that precluded the other or made it less likely; for example, from the fact that you can pat your head and can also rub your stomach, I cannot infer with much confidence that you can do both at once. However, this does not appear to be such a case. The splitting of a population into two species isolates the gene pools, allowing them to diverge, and ultimately to manifest different physical characteristics. We can also imagine that a wide enough physical variation within a species would tend to isolate two or more gene pools and provide a necessary though not sufficient condition for speciation. Johnson notes, for example, that while dogs are all theoretically capable of interbreeding, size differences make it practically impossible.

Now, Johnson admits that we have examples of both of these changes as a result of intentional selective breeding. Regarding the first, he would like to see dogs as big as elephants, but the difference between a toy poodle and a great dane seems adequate to me. Regarding the second, there are instances from plant breeding and, he says, circumstantial evidence that many species of fruitflies have developed from one or a few species originally introduced to Hawaii. Yet his conclusion is that none of this is adequate evidence for the Darwinian thesis. In effect, he is claiming that because plant speciation and intra-species variation are not equivalent to macroevolution they provide no evidence for the power of natural selection. But recall that we never hoped to observe a case of macroevolution by means of natural selection. We were about the more modest task (and the only realistic task) of providing confirmatory evidence by means of a modelˇan analogous processˇthat macro-evolution by means of natural selection is possible (given sufficient time and enough environmental pressure).

The form of the Darwinian reasoning is as follows:

´ O1 is observed (here, the patterns of speciation in existence today).

´ A hypothesis (H) is formulated which, if true, would explain O1 (here, the correlative hypotheses of variation, natural selection, and geographical isolation).

 ´ If H is true, what additional observations (O2... On) ought we be able to make? (here, O1: radical change within a population, and O2: speciation).

´ Finally, O2 and O3 have been observed, so H is confirmed.

Again, O2 and O3 are not equivalent to H; they are observable consequences that can be deduced from H with the aid of additional assumptions.


In effect, Johnson is claiming that because plant speciation
and intra-species variation are not equivalent to macroevolution
they provide no evidence for the power of natural selection.


One of the crucial assumptions here is that selective breeding is like natural selection in relevant respects. It is like natural selection in that it operates by means of differential reproduction rates and within the variation that nature supplies. These seem to me to be the relevant factors. Johnson's claim that the characteristics breeders look for are different from the ones for which "Nature" selects seems to me beside the point. The issue is whether selective breeding can produce radical changes, including speciation; not the particular nature of those changes.

I believe it could be shown by examining other arguments that Johnson consistently fails to distinguish between evidence confirmatory of a hypothesis and a set of observations that together are equivalent to the hypothesis. For example, on pp. 25-7 he first lists six pieces of evidence that have been offered in support of the power of natural selection, then concludes without explanation that "none of these `proofs' provides any persuasive reason for believing that natural selection can produce new species... (p. 27, emphasis mine). In Chapter 8, discussing theories about the origin of life, he concludes that because the synthesis of some of the components of living organisms does not actually amount to the production of life in the laboratory there is "no reason to believe that life has a tendency to emerge when the right chemicals are sloshing about in a soup" (p. 103, emphasis mine).

Recent Philosophy of Science

I shall introduce in this section some of the refinements contributed by recent philosophers of science by commenting on further aspects of Johnson's arguments.

In the section quoted above, Johnson has said that there are definite limits to the amount of variation that even the most highly skilled breeders can achieve; that dogs do not get as big as elephants because they do not have the genetic capacity for that degree of change (p. 18); and that after a number of generations the capacity for variation runs out (p. 19). He then admits that mutation might renew the capacity for change, but claims that whether and how often this happens is not known (p. 19).

When Darwin proposed his theory of evolution, he speculated that there must be a mechanism that works predominantly to maintain the characteristics of a population from one generation to another, but that also allowed for some degree of fluctuation and for genuine novelty. At that time, of course, he did not know what that mechanism was. A great triumph for evolutionary theory, but one Johnson does not mention, came from the discovery of the role of genes in reproduction. The gene pool provides for variation within overall stability in most instances, but mutations allow for genuine novelty.

Johnson mentions mutation as though it is scarcely important at all, but in fact it is an essential "auxiliary hypothesis" for the evolutionary program, and it is simply not possible to draw Johnson's strong conclusions about the limits of variation without considering the frequency and kinds of mutations, and their potential contribution to viable changes in a population.6


Johnson mentions mutation as though
it is scarcely important at all, but in fact it is an essential "auxiliary hypothesis"
for the evolutionary program.


This fact illustrates an important point stressed by philosophers of science. Theories (hypotheses) rarely or never face the test of experience standing alone. We are (almost) always faced with the testing of whole networks of theories and auxiliary hypotheses. This makes the falsification of a major theory very difficultˇwhen negative evidence comes along, it can often be reconciled with the central theory by adding or changing lower-level (auxiliary) theories. If positive evidence is lacking, its absence can often be explained by the same strategy.

Johnson's book is full of examples of changes of this sort to make evolutionary theory consistent with the evidence (or the absence of evidence). For example, Darwin expected that the fossil record would soon provide evidence of species intermediate between known species and their ancestors (the "missing links"). When few such intermediates were found, later theorists proposed auxiliary hypotheses to explain their absence: for example, the fossil record is still only a small sample of all of the creatures that have existed, and it is to be expected that the intermediate species, being in between well-adapted forms, would not last long and would therefore leave little evidence behind in the form of fossils.

Theorizing of this sort is extremely common in science. Since major theories come along only rarely, most of scientific advance consists in the careful elaboration and qualification of major theories, fine-tuning them to fit the evidence. Several philosophers of science have noted, though, that there is a kind of fine-tuning that represents genuine improvement and growth in scientific knowledge, and another kind that is a mere face-saving deviceˇlinguistic tricks to protect a theory from falsification. So the important question is how to tell the difference.

Imre Lakatos (d. 1974) made a major contribution to philosophy of science by providing a criterion for distinguishing "progressive" from "degenerative" or "ad hoc" refinements of a network of theories.7 The essence of his criterion is this: if a hypothesis that is added to the network not only explains the problems for which it was designed, but also leads to the prediction and corroboration of new facts of a different sort, then the modification is progressive. On the other hand, if it only takes care of the problem and is not independently confirmed by the successful prediction of novel facts, then it is a degenerative move. Lakatos made a double claim about this criterion. First, he claimed that it could account for the history of science better than other views,8 in that history would show that scientists generally abandon research programs that are making mostly degenerative moves in favor of a more progressive rival. His second claim is that scientists should accept progressive programs and abandon degenerative onesˇthat this is what scientific rationality consists in.

Application to Darwinism

Now, what consequences does this criterion of "progressiveness" have for evaluating evolutionary theory? It shows, first of all, that the only fair way to assess the program is by examining the auxiliary hypotheses that have been added to it to see whether each is a progressive or degenerative modification.

It is clear that Johnson is aware of the problem of ad hoc developments of a theory, as the following passage indicates:

Darwinists have evolved an array of subsidiary concepts capable of furnishing a plausible explanation for just about any conceivable eventuality. For example, the living fossils, which have remained basically unchanged for millions of years while their cousins were supposedly evolving into more advanced creatures like human beings, are no embarrassment to Darwinists. They failed to evolve because the necessary mutations didn't arrive, or because of "developmental constraints," or because they were already adequately adapted to their environment. In short, they didn't evolve because they didn't evolve.

Some animals give warning signals at the approach of predators, apparently reducing their own safety for the benefit of others in the herd. How does natural selection encourage the evolution of a trait for self-sacrifice? Some Darwinists attribute the apparent anomaly to "group selection." Human nations benefit if they contain individuals willing to die in battle for their country, and likewise animal groups containing self-sacrificing individuals may have an advantage over groups composed exclusively of selfish individuals.

Other Darwinists are scornful of group selection and prefer to explain altruism on the basis of "kinship selection." By sacrificing itself to preserve its offspring or near relations an individual promotes the survival of its genes. Selection may thus operate at the genetic level to encourage the perpetuation of genetic combinations that produce individuals capable of altruistic behavior. By moving the focus of selection either up (to the group level) or down (to the genetic level), Darwinists can easily account for traits that seem to contradict the selection hypothesis at the level of individual organisms.

Potentially the most powerful explanatory tool in the entire Darwinist armory is pleiotropy, the fact that a single gene has multiple effects. This means that any mutation which affects one functional characteristic is likely to change other features as well, and whether or not it is advantageous depends upon the net effect. Characteristics which on their face appear to be maladaptive may therefore be presumed to be linked genetically to more favorable characteristics, and natural selection can be credited with preserving the package.

I am not implying that there is anything inherently unreasonable in invoking pleiotropy, or kinship selection, or developmental constraints to explain why apparent anomalies are not necessarily inconsistent with Darwinism. If we assume that Darwinism is basically true than it is perfectly reasonable to adjust the theory as necessary to make it conform to the observed facts. The problem is that the adjusting devices are so flexible that in combination they make it difficult to conceive of a way to test the claims of Darwinism empirically (pp. 29-30).


This passage indicates that Johnson
sees no difference between auxiliary hypotheses that are testable
and those that are not.


However, this passage also indicates that Johnson sees no difference between auxiliary hypotheses that are testable and those that are not. It is difficult to conceive a test for the hypothesis that the living fossils failed to evolve because they were already adapted to their environmentˇor to be more precise, it is hard to conceive of a way of showing this claim false. This seems to be an instance of a "linguistic trick" to protect the theory from falsification. But not so with all of the examples Johnson has cited here. For example, kinship selection is testable: if it is true, then there should be a direct relationship between the percentage of genes shared with another individual and the degree of "altruism" exhibited toward that individualˇa prediction that has in fact been confirmed. In addition, genetic mapping makes the concept of pleiotropy empirically testable.


It must be emphasized, though, that the mere existence of problems
does not disqualify a theoryˇgood theories are always in process,
and the question is whether they are progressing, overall, or degenerating.


So it is clear that Johnson has failed to see the import of such cases. He does not understand their role in demonstrating that there are in fact ways "to test the claims of Darwinism empirically" (p. 30).

In general Johnson has given too little attention to the role genetic theory has played in the history of evolutionary biology. Genetics arose as a major new theory in complete independence of evolutionary biology. Initially there was strong antagonism between workers in the two fields. However, with the advent of population genetics under Fisher, Wright and Haldane, the two fields were reconciled. In Lakatos's terms, the entire genetic program came to function as an "auxiliary hypothesis" within the evolutionary program, providing a tremendous amount of fresh empirical evidenceˇevidence of exactly the sort that Lakatos has led us to expect from a progressive program. Another instance is "neutral allele" theory, with its associated phenomenon of molecular clocks.

Much remains to be done to provide an adequate assessment of the evolutionary program. There are a number of problems with the theory, but whether there are more than with other major theories, such as Big-Bang cosmology, remains to be seen. It must be emphasized, though, that the mere existence of problems does not disqualify a theoryˇgood theories are always in process, and the question is whether they are progressing, overall, or degenerating. So the important question is how the evolutionary program deals with its problems; whether the auxiliary hypotheses needed to account for anomaliesˇfor the absence of certain kinds of expected confirmatory evidenceˇcan be independently tested and confirmed. Johnson does not pursue this question; nor can I do so here. Adequate treatment would require another book. But this is where the battle must be joined if we are to have a fair assessment of the evolutionary program.

It has been noted9 that the kind of "novel facts" needed to provide independent confirmation of auxiliary hypotheses are usually rare, and get harder rather than easier to find as a program progresses. This suggests that the crucial evidence for evolutionary theory, if it can be produced, will not be massive. It will consist in a few confirmed predictions here and there. In this way, evolutionary biology will be entirely in line with many well-respected programs such as Big-Bang cosmology.

A major problem for anyone undertaking an assessment of the evolutionary program is that philosophy of science provides criteria for relative rather than absolute assessment. That is, the criteria we have been discussing are only capable of telling us which of two or more competing programs is the most acceptable. While there is competition within the evolutionary program, between punctuated equilibrium and gradualist theories of change, for instance, there is no major scientific competitor for the program as a whole. This being the case, there are limits to what critics of Darwinism can hope to accomplish. When a theory is the only one available, the burden of proof falls on those who wish to do away with it. It is simply a fact of the history of science that a theory is seldomˇperhaps neverˇabandoned when there is no competitor to take its place. If criteria for rational choice are necessarily comparative, then this is a rational way to proceed. Beyond that, there is the practical question: what would evolutionary biologists do if there is no other conception of the field to guide their research?

The Nature of Science

In this section I shall take up three issues raised by Johnson:

1. Evolutionary biology is not scientific because (according to Karl Popper) science is characterized by falsifiability, and the central ideas of Darwinism are held dogmatically.

2. Evolutionary theory is held dogmatically because it is the only account of life that fits with a naturalistic philosophy.

3. Evolutionary biologists ought to consider the possibility that life is the product of creative intelligence.

Science and Falsifiability

In his final chapter Johnson adopts Karl Popper's criterion for distinguishing science from pseudoscience. Popper argued that what made science scientific was not its subject matter but the willingness of its proponents to allow their theories to be falsified.10 In Johnson's words: "Progress is made not by searching the world for confirming examples, which can always be found, but by searching out the falsifying evidence that reveals the need for a new and better explanation" (p. 147).

Imre Lakatos was a colleague of Popper's at the London School of Economics. Lakatos treated Popper's claims about the nature of science as an empirical theory and argued that, as such, the history of science falsified Popper's account. His own theory, introduced above, was proposed as a "new and better explanation" of the course of the history of science. We have already seen his proposed criterion for distinguishing between acceptable and unacceptable (progressive and degenerating) research programs. Here it is relevant to introduce another feature of his account of science.

All scientific research programs, he concluded, include a central idea, called the hard core, which is usually too vague to be tested directly. In addition, there are the auxiliary hypotheses that mediate between the core theory and empirical data. Lakatos's study of the history of science convinced him that a certain amount of dogmatism with respect to the core of a program was both a regular feature of good science and a necessary strategy to allow for the development of scientific thought. His new version of falsificationism allows researchers to protect their core theory "dogmatically" so long as the program is progressive overall.11

From what has just been reported,12 it follows that Johnson's criticism in the following quotation shows not the unscientific character of evolutionary biology, but rather that Johnson approaches it with an inadequate understanding of the philosophy of science:

The central Darwinist concept that later came to be called the "fact of evolution"ˇdescent with modificationˇwas thus from the start protected from empirical testing. Darwin did leave some important questions open, including the relative importance of natural selection as a mechanism of change. The resulting arguments about the process, which continue to this day, distracted attention from the fact that the all-important central concept had become a dogma (p. 149).

That is, the usual strategy in science is to hold on to a central ideaˇhold it "dogmatically," if you willˇso long as the theoretical elaborations and additions that are necessary to reconcile it with the evidence lead to new discoveries rather than to blind alleys.

Evolution and Naturalism

Johnson explains the evolutionists' dogmatism by attributing it, not to the usual processes of scientific development, but to an atheistic philosophical naturalism. Johnson is quite right about this in some cases, and perhaps in most of the cases of popular books written in defense of evolution.

However, a subtle distinction needs to be made here. On the one hand there are the proponents of "a religion of scientific naturalism, with its own ethical agenda and plan for salvation through social and genetic engineering" (Johnson, p. 150). This religion is fair game for criticism by proponents of other religions, and ought not be allowed establishment in the curriculum of the public schools. On the other hand, there is what we might call methodological atheism, which is by definition common to all natural science. This is simply the principle that scientific explanations are to be in terms of natural (not supernatural) entities and processes.

Johnson is critical of biologists and philosophers who define science in this way. However, it is a fact of history (perhaps an accident of history) that this is how the institution of natural science is understood in our era. It is ironic, perhaps, that Isaac Newton and Robert Boyle, two of the scientists who led the move to exclude all natural theology from science (then called "natural philosophy") did so for theological reasons. Their Calvinist doctrine of God's transcendence led them to make a radical distinction between God the Creator and the operation of the created universe, and hence to seek to protect theology from contamination by science. The metaphysical mixing of science and religion, Boyle and Newton believed, corrupted true religion.13

So, for better or for worse, we have inherited a view of science as methodologically atheisticˇmeaning that science qua science, seeks naturalistic explanations for all natural processes. Christians and atheists alike must pursue scientific questions in our era without invoking a creator. The conflict between Christianity and evolutionary thought only arises when scientists conclude that if the only scientific explanation that can be given is a chance happening, then there is no other explanation at all. Such a conclusion constitutes an invalid inference from a statement expressing the limits of scientific knowledge to a metaphysical (or a-religious or anti-religious) claim about the ultimate nature of reality.

This is a subtle differenceˇone beyond the grasp of a fourth-grade science class (and perhaps beyond the grasp of some outspoken scientific naturalists as well?). For this reason I am sympathetic with Christians who object to the teaching of evolution in the public schools. But the answer is to help educators make the distinction, not to cooperate in blurring it as Johnson has done.14

Creative Intelligence as a Scientific Hypothesis

Johnson writes:

Why not consider the possibility that life is what it so evidently seems to be, the product of creative intelligence? Science would not come to an end, because the task would remain of deciphering the languages in which genetic information is communicated, and in general finding out how the whole system works. What scientists would lose is not an inspiring research program, but the illusion of total mastery of nature. They would have to face the possibility that beyond the natural world there is a further reality which transcends science (p. 110).

The answer to Johnson's question is that anyone who attributes the characteristics of living things to creative intelligence has by definition stepped into the arena of either metaphysics or theology. Some might reply that the definition of science, then, needs to be changed. And perhaps it would be better if science had not taken this particular turn in its history. Could the nature of science change again in the near future to admit theistic explanations of natural events? There are a number of reasons for thinking this unlikely. A practical reason is the fact that much of the funding for scientific research in this country comes from the federal government. The mixing of science and religion would raise issues of the separation of church and state.

A second reason for thinking such a change unlikely is that many Christians in science, philosophy, and theology are still haunted by the idea of a "God of the gaps." Newton postulated divine intervention to adjust the orbits of the planets. When Laplace provided better calculations, God was no longer needed. Many Christians are wary of invoking divine action in any way in science, especially in biology, fearing that science will advance, providing the naturalistic explanations that will make God appear once again to have been an unnecessary hypothesis.

Concluding Remarks

What, then, of the relation between Christianity and Darwinism? I hope I have made it clear that this question is ambiguous. One question is: How ought Christianity be related to evolutionary biologyˇthe pure science? The other is: How ought Christianity be related to evolutionary metaphysics? The latter system of thought involves the use of scientific theory to legitimate a metaphysical-religious point of view, and it has been so successful that many cannot imagine Christian thought making its own, different use of biology. Nonetheless it can be done, and it has been done by the likes of biochemist-theologian Arthur Peacocke.15

Peacocke notes that the sciences can be organized in a hierarchy, with higher sciences studying more complex levels of organization in reality. For example, chemistry studies more complex organizations of matter than does physics; biochemistry more so than inorganic chemistry; within biology alone there is a hierarchy as we move from biochemistry to the study of cells, to tissues, organs, and finally to the functioning of entire organisms within their environments.

Each science has its recognized domain, and concepts and theories appropriate to its own level of interest. Yet each science is conditioned by the levels above and below. Lower levels set limits on the behavior of entities at higher levelsˇfor example, chemical processes in nerves and muscles set limits on how high or fast an animal can jump. However, lower levels do not uniquely determine the behavior of entitles at higher levelsˇhere one also has to take account of the environment. Thus, the animal's particular movements within the range permitted by chemistry and physics will be to some extent conditioned by ecological factors as well.

So any science alone provides an incomplete account of reality; it finds limits above and below, beyond which its explanatory concepts cannot reach. But what about the limits of the highest (or lowest) science in the hierarchy? Peacocke proposes that at the top of the hierarchy of the sciences we reach theology, the science that studies the most complex system of allˇthe interaction of God and the whole of creation.16

Peacocke's suggestion provides the groundwork for an exciting account of the relations between the sciences and theology. We can examine the kinds of relations that hold between two hierarchically ordered sciences, and then look for analogous relations between theology and one or more sciences. One relation we may expect to find is that when a science reaches an inherent limit, there may be a role for theology to play at that point. For example, it may be inherently impossible for science to describe what happened "before" the Big Bang.

Peacocke's understanding of the relation between science and theology means that we need not turn biology into theology, but we may and must bear in mind that there is a discipline "above" biology that answers questions that biology alone cannot answer. Is this discipline to be an atheistic metaphysic that elevates "Chance" to the role of ultimate explanation, or is it to be a theology of benevolent Design? The question calls for a careful comparison of the explanatory force of these two competing accounts of reality. The former has to explain (or explain away) all appearances of order and purpose; the latter has to explain a number of features of the world that (as biologists correctly point out) appear inconsistent with intelligent design.17

It looks to many as though these two explanatory systems are at a stand-off. For every feature that appears to be the product of design, there is another that appears to be the product of chance. However, I suspect that the design hypothesis, as the core of the theological research program, could be shown to be more progressive (in Lakatos's sense) than a research program based on chance. My guess is that while the atheistic program could explain (or explain away) all the evidence for design, it will have to do so by means of an assortment of ad hoc hypotheses. Besides this, the Christian program has at its disposal additional supporting evidence from a variety of domains: religious experience, history, the human sciences.

So there are two issues before us, both of which cry out for much more extensive and careful treatment than I have given them here: First, what is the true standing of evolutionary biology as a science and measured against the best criteria that have so far been proposed for evaluating scientific acceptability (truth). I make two claims with regard to Johnson's book: first, that he has allowed the Evolutionary Naturalists to confuse evolutionary science with something else and, second, that he has used too primitive a view of scientific methodology for his evaluation. I do not claim to have definitively refuted his claims against evolutionary science, but I hope to have undermined them, and to have shown the direction a definitive evaluation of evolutionary biology would have to take.

The second big issue is the clash of world views: evolutionary naturalism versus Christianity; Chance with a capital "C" versus Design. Settling this controversy is well beyond the capability of any single scholar on either side, but we do educators, school children, and perhaps even evolutionary biologists a great favor by carefully distinguishing this issue from the first.

An important effect of separating the theological-metaphysical issue from the scientific one may be to lessen the anxiety and heat of controversy that surrounds the latter. We want scientists to stop their attacks on Christianity, but all Bible-readers should know that the cessation of hostilities is not to be left to our opponents. Better to turn away wrath with a gentle word.18

ę1993

NOTES

1 InterVarsity, 1991.

2 This sort of credential swapping is quite out of place in academic writing, but nonetheless it deserves a name. In practical reasoning, some arguments are called ad hominem (to or against the man); this argument I shall dub an ab femina argument (from the woman).

3 This view of science has been particularly influential in conservative American Christian circles. John Witherspoon promoted Bacon's views among the Princeton theologians, such as Charles Hodge, who have influenced American Fundamentalism. It is described and criticized at somewhat greater length by Johnson, pp. 146-47.

4 This term was coined by Carl Hempel. See his Philosophy of Natural Science (Prentice-Hall, 1966).

5 Actually, we are asking more of natural selection here than is required by the theory. Darwinian theory does not require that natural selection be directly responsible for reproductive isolation. The classical theory is that geographical isolation, followed by differential adaptation to different conditions, is the principal agent of speciation.

6 Johnson does take up this issue in the following chapter. My point is that the conclusions he draws in this chapter regarding the limits of variation are quite unwarranted because they cannot be made independently of the assessment of the possibilities for mutations.

7 See "Falsification and the Methodology of Scientific Research Programmes," in J. Worrall and G. Currie, eds., The Methodology of Scientific Research Programmes: Philosophical Papers, Volume 1 (Cambridge University Press, 1978), pp. 8-101.

8 Such as Karl Popper's falsificationism.

9 By Alan Musgrave, in "Logical vs. Historical Theories of Confirmation," British Journal for the Philosophy of Science 25 (1974): 1-23.

10 Popper first elaborated this thesis in Logik der Forschung (Vienna, 1935); English translation, The Logic of Scientific Discovery (Harper, 1965).

11 There is insufficient space here to show that Lakatos's understanding of science is superior to Popper's. See my Theology in the Age of Scientific Reasoning (Cornell University Press, 1990), chapter 3; as well as Lakatos's "Falsification and the Methodology of Scientific Research Programmes," op. cit.; and especially his "History of Science and Its Rational Reconstructions," also in The Methodology os Scientific Research Programmes, op. cit., pp. 102-138.

12 The same point is made by Thomas Kuhn in The Structure of Scientific Revolutions (University of Chicago, 1970); Paul Feyerabend in Against Method (New Left Books, 1975); and Larry Laudan in Progress and its Problems (University of California Press, 1977).

13 See Eugene Klaaren, Religious Origins of Modern Science (Eerdmans, 1977); and Frank Manuel, A Portrait of Isaac Newton (Harvard University Press, 1968).

14 For an excellent discussion of this and other issues, see Howard Van Till, Robert Snow, John Stek, and Davis Young, eds., Portraits of Creation (Eerdmans, 1990).

15 See Creation and the World of Science (Clarendon, 1979); Intimations of Reality (University of Notre Dame Press, 1985); or Theology for a Scientific Age (Basil Blackwell, 1990).

16 I elaborate and apply this view of the hierarchy of the sciences and their relation to theology in "Evidence of Design in the Fine-Tuning of the Universe," in Robert Russell, Nancey Murphy, and Chris Isham, eds., Quantum Cosmology and the Laws of Nature: Scientific Perspectives on Divine Action (The Vatican Observatory, forthcoming).

17 Peacocke's view is that God creates through exploration of the possibilities provided by chance as well as through law-governed design.

18 I wish to thank Philip Spieth at the University of California, Berkeley for helpful comments on an earlier draft of this paper.

      Response