Created 1996/12/05, Last updated 2000/09/05
Written for the 1997 UT Austin conference on Naturalism, Theism, and the Scientific Enterprise
There is a distinction to be made between descriptive and prescriptive philosophical treatments of science. For example, while some take Popperian falsificationism as a recipe for the practice of science, Popper's discussion of various sciences so-called makes it clear that he intended science to be recognized because it included falsifiability, not because falsifiability might be added, like some rare spice, to an existing field of study. Kuhn also wrote no cookbooks, rather, he reappraised what actual scientific practice looked like. It is no accident that both of these philosophers came to use natural selection as a metaphor for the operation of science. The modern practice of science is premised upon the radical assumption that the physical universe is comprehensible to humans. That this assumption is radical is supported by the fact that it has not always historically been accepted, that it remains largely unassimilated even today, and that many explicitly reject it since they believe that it denies any reality to theism, mysticism, or even mystery. The modern practice of science also requires that objectivity be approximated, even if it cannot in principle be completely achieved. The practice of science is a pragmatic endeavor whose principle product is the conversion of subjective personal experience into an approximation of objective knowledge concerning physical phenomena. While the subjective appreciation of a role for supernatural causation may be important to personal fulfillment, it does not afford a basis for objective knowledge, nor can it be counted as a means of comprehending the universe in a scientific manner. I will connote "naturalism" as "proposing only natural mechanisms for physical phenomena" rather than "asserting that only natural mechanisms have existence". The debate over naturalism and its relationship to scientific practice most often arises today in discussion of education in biology. There is concern in education that discussion of the scientific method should include information not only about how science ought to be conducted, but also about how science actually is conducted. My answer to the question, "Does the scientific method exclude appeals to supernatural causation?" has to be yes, since I consider naturalism to be a corollary to the assumption that the universe is comprehensible by humans. Rejection of naturalism amounts to an assertion that some parts of the universe are not comprehensible by humans, which even if true is a sterile stance. I entertain the possibility that the founding assumption of science may be literally false, but even if it is sometimes false, it is often true and retains much value thereby. Science has as its domain physical phenomena and its range is the set of scientific explanations of those phenomena. Natural explanations are the only known variety that produce an increase in scientific comprehension. In a particular case study, the claim made by A.E. Wilder-Smith that simulations of evolutionary processes demonstrated the failure of natural explanations is demonstrated to have been false, as evidenced by the success of the fields of artificial life, genetic algorithms, and evolutionary programming.
Science is "what scientists do," or so opined Judge Overton in his 1982 ruling in 'McLean vs. Arkansas' (Overton 1984, p.380). As a definition, it has the apparent drawback of recursiveness. However, this formulation is not necessarily recursive, if one avoids defining "scientist" as "someone who does science". It still retains some of the inscrutable character of fully recursive definitions, but provides at least one important insight. That insight is that science, as a method of study, is defined by its practice, and not by edicts from on high.
Edicts from on high concerning science include those prescriptive statements forwarded by Bacon and Descartes. Bacon believed science could only come from an inductive review of copious evidence collected from empirical study, while Descartes advocated the viewpoint that a rigorous application of logic should yield correct theories of science.
Newton, in producing his 'Principia Mathematica', did not strictly adhere to the recommendations of either Bacon or Descartes, but utilized elements from each. The emphasis in philosophical treatment of science shifted thereafter from the prescriptive to the descriptive. The question was not, "How should Newton have done it?", but rather, "How did Newton do it?" After all, Newton's immense success and regard would impose a heavy burden upon the prospective prescriptive philosopher of science, simply to avoid appearing embarrassingly arrogant.
Some have mistaken modern descriptive treatments for prescriptive treatments. This is especially true of Sir Karl Popper's famous demarcation criterion of falsification. Within Popper's framework, a field of study can be recognized as being scientific if it produces theories that can be falsified, not if it can add falsifiable statements to an existing body of conjecture. In particular, Popper establishes why he believes Freudian psychology and Marxian historicism to be beyond the pale of science.
Kuhn also made an attempt to describe rather than prescribe the course of scientific study in his 'Structure of Scientific Revolutions'. Unsatisfied with other popular descriptions of science as steadily progressing toward greater knowledge, Kuhn produced a different view of the scientific process. This view introduced the notion of the 'paradigm shift', where the basic postulates of an older paradigm are overthrown in favor of a more capable set of postulates that comprises a new paradigm.
But what about the possibility of supernatural causation for natural phenomena? Is science competent to examine the case where a supernatural agency acts in the physical world? In modern times, many have argued that for various reasons, diagnosed insufficiencies in current theories establish the necessity of invoking supernatural causation. I find various problems in this kind of endeavor, both on procedural and historical grounds. The best way of examining how argument for the serious treatment of supernatural causation in science works is to look at an example, so I will consider one below.
A. E. Wilder-Smith was an author of several books which expressed a critical view of what he termed "scientific materialism". He advanced various arguments to this end. The one that I will focus on here comes from his 1970 book, "The Creation of Life: A Cybernetic Approach to Evolution". In this argument, Wilder-Smith says that Paley's "argument from design" must be regarded as valid because the Darwinian theory of evolution which displaced it has been found to be actually false.
There are several reasons to suspect that Wilder-Smith's rehabilitation of Paley's argument was premature, including the fact that he ignored other critiques, but the one reason of primary concern to me here is his claim that Darwinian theory has been falsified by the only legitimate experimental evidence available. The nature of this claim and the subsequent developments in the field of computer science which are relevant say, quite eloquently, that the appeal to supernatural causation in the scientific method is always premature.
A prerequisite for establishing supernatural causation would be to find an area of natural phenomena without known natural mechanisms. Wilder-Smith attempts this early on in 'The Creation of Life'. He declares that the 1965 Wistar conference, 'Mathematical Challenges to the Neo-Darwinian Interpretation of Evolution', exposed "missing factors" in Neo-Darwinian theory, including ones arising from the study of cybernetics. By clearing away Darwinian theories, a playing field is opened for supernatural action.
"In addition Schutzenberger pointed out that recent developments in computer science have shown that the spontaneous evolution of a self-replicating organism is a phenomenon which has never been duplicated or simulated successfully even on the largest and most rapid computers available to date." (Wilder-Smith 1970, p.39.)
Above, Wilder-Smith identifies computer science as the field in which experiments relevant to abiogenesis and evolution can be tested.
"[...] Order is of two kinds in our present discussion. There is the kind of order which is truly a pattern -- like ripples on the seashore -- but which bears no code meaning. This order can be compared to letters in an ordered sequence which conveys no particular meaning. Then there is the other kind of sequenced order which hides a meaningful code -- like a section of Goethe's poetry. We know of only one way in which the latter can arise and that is by the exercise of intelligence. The first kind can arise either with or without the direct intervention of intelligence." (Wilder-Smith 1970, p.73.)
Wilder-Smith makes an assertion concerning how new information can arise, one which excludes any agency except that of the "direct intervention of intelligence." Wilder-Smith again relies upon the reports from the Wistar conference.
Wilder-Smith infers that the lack of algorithms in computer science to produce information from evolutionary principles indicates that Neo-Darwinian theories are deficient in content.
"The point is, of course, that, as things stand today, random and adaptive evolutionary theories have not yet supplied the programming ground rules for extracting order spontaneously from random processes or for constraining such processes. Surely this fact can only mean that some fundamental gaps still exist in current Neo-Darwinian theories which allegedly account for evolution as a result of random processes followed by competitive selection." (Wilder-Smith 1970, pp.110-111.)
Wilder-Smith asserts that simulations reveal an inability to produce new sequences via Darwinian mechanisms.
"Darwinians and Neo-Darwinians have long maintained that randomness, plus long time spans, plus natural selection would, in combination, do the synthetic trick and deliver specific codes and molecules. However, recent progress in cybernetics has shown by simulation experiments that order sequences, specificity and coding cannot be extracted from randomness on the basis of the Darwinian postulates." (Wilder-Smith 1970, p.116.)
The criterion of testing by experience is one which Wilder- Smith claims to have gone by the boards in the case of Darwinian mechanisms.
"The unwillingness to abandon the foundering ship of Darwinian chance hypothesis came out quite surprisingly in the symposium mentioned in chapter 1. The following citation shows the fundamental unwillingness to submit to the only sound experimental evidence available on the subject:
'Dr. Schutzenberger: I want to know how I can build on computers, programs which....'
The chairman, Dr. Waddington: 'We are not interested in your computers!'" (Wilder-Smith 1970, p.130.)
Wilder-Smith goes on to explicate how Darwinian theories displaced the argument from design.
"The great advantage of the randomness theory of Darwin with its accompanying natural selection and long time spans was that it destroyed the abhorred necessity of divine intelligent activity behind nature. Today, those in progressive circles in mathematics and physics conclude that cybernetic simulation experiments establish the fact that the principles of randomness plus selection plus long time spans cannot and do not replace the earlier concept of extramaterial constraint acting on matter to produce order, including the order of life. One hundred years ago Darwin's hypotheses were not susceptible of experimental and theoretical disproof. Now they are." (Wilder-Smith 1970, p.219.)
This gets to the nub of the argument. Wilder-Smith asserts again that computer simulations are the place where experimental test of Darwinian hypotheses can be made. It should be noted here that Darwin did not hypothesize concerning abiogenesis, which was a major topic in Wilder-Smith's book. Therefore, this passage above can be seen to apply directly to natural selection rather than abiogenesis.
Eventually, Wilder-Smith writes a rousing, emotive section to reiterate his claim:
"It is only in recent years, with the advent of the "super-computer" which could automatically, swiftly and surely deal with the astronomical numbers in which Darwin enshrouded his theory, that the denouement of the grand scheme becomes possible. The astronomical numbers of random changes, the long time spans and the alleged evolutionary "trends" in the midst of randomness have been programmed and fed into super-computers. The result has been dramatic, for the machines jam in their efforts to unravel such tangled masses of informational "noise." No wonder that the mathematical experts have crowded around the site of these experiments just as physicians crowd around the bed of a patient sick of a rare disease, to ascertain the cause of the excitement. The biologists have mocked from a distance and denied the result proclaimed by the mathematicians -- that the theory will not work but merely jams the best machines." (Wilder-Smith 1970, pp.232- 233.)
And here Wilder-Smith summarizes, but does not cite, the evidence of experimental test. Attempted simulation of biological evolution justs "jams" computers, thus Darwinian hypotheses are debarred, and thus Paley's argument from design is rehabilitated.
"It has been well pointed out by Robert Bernhard that a basic assumption of evolutionary theory is that "increasing complexity is an essential feature of evolution, but there is no explanation for that phenomenon in the theory." This very factor is the crux of the whole question of the missing factor in Neo-Darwinian theory. Information theory requires a programmer to account for the increasing complexity of the whole program of evolution. The theory as it stands provides for no information source to account for the increasing complexity. Yet it is perfectly clear today that life shows the most complex programs conceivable. Darwinians dare no longer close their eyes to this basic fact which will require explanation in terms of information theory -- the more so as knowledge in this area becomes more generally available." (Wilder-Smith 1970, p.244, emphasis in original.)
Wilder-Smith gets specific about what he sees as missing: a mechanism for the production of new information and complexity, indicated by the principles of information theory.
One consequence of a designer for the universe is explicated by Wilder-Smith in the following.
"Exactly the same problem would be expected to beset the relationship between the designer behind nature and the intelligently designed part of nature known as man. Obviously the great designer behind the universe speaks a huge number of languages in the expression of his huge intellectual capacity. He speaks, as Jeans said, among others, a mathematical language. But, on top of this, he speaks the chemical language of the elements as well as the languages of physics, geometry, algebra, philosophy and so on. The language of chemistry which he speaks in designing his thought according to DNA coding sequences is a subject in itself. The average human has all he can cope with in maintaining one language with which to communicate. Thus he is likely to be able to absorb only very small amounts of the designer's multilanguage. No one today can be familiar with all the languages of all the sciences. Once more we have the old difficulty of establishing communications between the designer and the designed on account of language barriers." (Wilder-Smith 1970, p.248.)
Wilder-Smith explicitly denies the axiom that the universe is comprehensible by humans in the above.
"[...] My point is that modern biology has made the use of aeons a necessity and a cardinal point of its dogma to overcome the inherent clumsiness of the trial-and-error mechanism it postulates. The supreme coding and programming of all nature should open our minds to the consequences of the factor of intelligence. For intelligence does things differently -- and more quickly!" (Wilder-Smith 1970, p.254)
Intelligence does do things differently, as I will explain in more detail later on. This does not turn out to be a point in favor of Wilder-Smith's argument.
Wilder-Smith asserts the invalidity of Neo-Darwinian and Darwinian hypotheses to satisfactorily explain abiogenesis, ontogenesis, or phylogenesis. Since the first two are not the subject of Neo-Darwinian or Darwinian hypotheses, we will exclude discussion of those claims. He relates the experience of Marcel-Paul Schutzenberger, who apparently failed in attempts to program simulations of biological evolution on super-computers. The approach of Schutzenberger is one which Wilder-Smith characterizes as the only sound means of gathering experimental evidence concerning these topics. This is an important point, since Wilder-Smith stakes his entire argument upon the conclusiveness of Schutzenberger's experiments.
In short, Wilder-Smith's logical argument may be stated as follows. Paley's "argument from design" was rebutted by Darwinian hypotheses. Computer simulations are the only sound means of accumulating experimental evidence to test Darwinian hypotheses. Computer simulations of biological evolution fail to accord with the stated results of Darwinian hypotheses, that sequence, specificity, or coding information can arise from random processes coupled with competitive selection. The failure of experimental tests of Darwinian hypotheses invalidate those hypotheses. Other hypotheses formerly abandoned due to the existence of the previously untested, and now falsified, Darwinian hypotheses must now be re-evaluated. Specifically, Paley's "argument from design" is rehabilitated and must be viewed as being valid. Information theory and the argument from design imply that humans can accrue only a limited and partial knowledge of the universe.
Because this point was crucial to Wilder-Smith's argument, it is useful to spend some effort in documenting how it is known to have failed. John Holland coined the term "genetic algorithm" for one component of his larger work on "classifier systems". The "genetic algorithm", or "GA", in Holland's formulation, was a process which operated upon a population of initially randomized fixed-length bit strings, each of which was evaluated at each iteration for fitness, and which were copied with variation to form a new population of bit strings. The whole process was premised on a fairly simple reading of Darwinian natural selection and basic genetics. Holland saw the bit string as analogous to the chromosomal content of an organism. The evaluation function performed the service of environmental constraint. Holland stressed the role of "crossover" in production of useful variation over mutation, though mutation still was part of the copying process. Because the original bit-strings forming the candidate population were randomly set, it can readily be seen that the operation of Darwinian processes was able to produce new and useful information.
The absolutely critical fact to be apprehended is that Holland's GAs were successful. These GAs not only did not "jam" computers, they were found to be capable of resolving very difficult optimization problems. Some of the applications for which GAs have been found useful include oilfield pipeline layout, Hubble telescope job scheduling, and contaminated ground-water remediation well placement.
Another insight revealed by GAs is that the utility of Darwinian processes is not dependent on logical necessity. That is, a GA does not necessarily produce an optimal or near-optimal result. Phillip Johnson critiqued one non-tautologous formulation of natural selection for not entailing a necessary outcome of success (1993, p.23). However, the common success of GAs points to general utility rather than the possible complete lack of utility which Johnson intimates in his critique.
One might well ask what the solutions that GAs produce look like upon human examination. The answer is that GAs produce near- optimal solutions that look very different from solutions that human engineers produce for the same problems. The literature on the use of GAs for parameter estimation and structure in artificial neural systems shows that GAs produce networks whose topology pretty often does not match the expectations of the humans who set up these problems. This gets back to Wilder-Smith's assertion that "intelligence does things differently". Wilder-Smith was right that differences in the products of intelligence and Darwinian process exist, but wrong in the inference that living organisms demonstrate the pedigree of intelligence. The quirky, odd, and just plain weird ways in which organisms are put together speaks much more clearly of a process like GAs than it does of intelligent planning.
This assertion is inconsistent with the axiom that the phenomena of the universe can be comprehended by humans. This is fundamental to scientific research. No researcher can approach phenomena under study with the belief that the phenomena cannot be comprehended. Since all phenomena are appropriate for study, the universal nature of this axiom is established. Note that this axiom is revealed in a bottom-up rather than a top-down fashion, in accordance with the principle that science is defined by its practice.
I should clarify what I mean by the axiom of comprehensibility. In many ways, it is a precursor to what some have termed "the scientific attitude". Briefly, the axiom of comprehensibility holds that for any particular phenomenon approached as a topic of scientific research, those engaged in the research hold an expectation that the research will yield comprehension of the phenomenon, either in whole or in part. This is simple to state, and as with most axioms, appears trivial when plainly stated. Yet the rejection of this axiom leads to a variety of non-scientific and even anti-scientific stances.
In summary, Wilder-Smith's premise that Darwinian hypotheses failed to provide adequate "rules" for producing algorithms capable of producing new information and were factually in contradiction to established tenets of information theory was false, and his conclusions drawn therefrom are invalid. Wilder-Smith's repudiation of the stubbornness of biologists confronted by mathematicians at the Wistar conference turns out to be merely embarrassing, for the biologists were absolutely correct and the mathematicians wrong in their assessments.
In this case, we find that a forcefully worded claim of falsification of Darwinian mechanism coupled with concomitant rehabilitation of supernatural causation failed. The failure of the claim stemmed from the acceptance of a certain few failed experiments as demonstrating conclusively that all experiments in the same vein would fail. Because of the universal nature of Wilder-Smith's claim, its downfall was demonstrated by the success of other computer simulations of Darwinian processes.
The philosophical underpinnings of science are, for the most part, invisible to its practitioners. While being schooled in scientific disciplines, it is relatively uncommon for students to be explicitly exposed to the philosophy of science. The practice of science is mostly conducted by people who have neither a grounding in or an appreciation for epistemology, and who may even find consideration of the topic unworthy of their attention. Even the much vaunted "scientific method" rarely receives a cogent explanation to the student at the secondary school level, and may be entirely absent from the curriculum of graduating college students. A mystery worthy of exploring is how science continues to perpetuate itself without an efficient and explicit pedagogy.
The solution is that many successful scientists learn by example and by doing science. In most cases, explicit lessons in the scientific method become superfluous because the neophyte scientist must undertake research directed by others, whether during their course work in school or in entry-level research positions. The structure of current scientific research holds many parallels to the medieval guild system, with the apprentice (read "undergraduate", "graduate", or "intern"), journeyman (read "graduate" or "postdoc"), and master (read "faculty", "researcher", or "scientist") levels. It is uncommon that one may find a master level scientist (one who obtains grants as a principal investigator in his or her own name) who has not completed one or both of the lower levels. The course work of science curricula is, generally speaking, an insufficient basis for the actual practice of science. That comes from getting involved in the practice of science. Actually practicing science is unlikely to enhance the practitioner's grounding in the philosophy of science.
Still, this persistent and pervasive gap in the knowledge of scientists concerning the philosophical basis of scientific endeavor cannot be viewed with equanimity. A proper understanding of what science is and does should be part of every citizen's education, especially as our society becomes more and more dependent upon technology. Miscomprehension of what scientists do hampers scientists in the long run, as public funding administration may then evaluate proposals using a skewed or even anti-scientific viewpoint. The objections of the late Senator Proxmire to various funded research studies in basic science demonstrate this point nicely.
In education as well as practice, the axiom of comprehensibility should be emphasized as a wellspring of scientific endeavor. Education in science should be well-rounded, to include not only the philosophical basis of scientific methodology, but also the flavor of how actual scientific research is conducted. An important part of this would include the subject of error and fraud in science, which illustrates the importance of inter-subjective experience. The history of science, including examples such as Wilder-Smith's failed rehabilitation of Paley, is also important to demonstrate why the axiom of comprehensibility is critical to doing science.
It is possible that the axiom of comprehensibility is false. There may exist some phenomenon which will never be comprehended by humans even after intense study. However, no such counterexample has yet been found, and we have no good reason to suppose that such a counterexample is likely to be found. (Some might interject here that Godel's incompleteness theorem as an indication that comprehensibility is limited. However, this does not bear upon the issue. Science can only deal with phenomena which exist. What the incompleteness theorem states is that certain concepts, such as a set of axioms sufficient to establish algebra, will not be both complete and consistent. In other words, the theorem tells us what phenomena will not exist, and not that some existing phenomenon is incapable of comprehension.) In many cases, we have examples where previously baffling phenomena have been rendered comprehensible. Asserting that the axiom of comprehensibility is false is a sterile stance. Such an assertion represents an abandonment of further study on some set of physical phenomena, rather than setting up another avenue of inquiry. Even if some few counterexamples to the axiom were to be found, its general utility would indicate that it should be tested by scientific inquiry in each case, not merely assumed to also be false for other cases.
Wilder-Smith rejected the axiom of comprehensibility based upon his view that the fact of supernatural causation of the phenomena of the universe implied a limitation on human comprehension of those phenomena. In a curious inversion of Wilder-Smith's logical argument for rejection of the axiom of comprehensibility, Phillip Johnson stated the following:
"[...] To theists, on the other hand, the concept of a supernatural Mind in whose image we are created is the essential metaphysical basis for our confidence that the cosmos is rational and to some extent understandable. Scientific naturalists insist, paradoxically, that the cosmos can be understood by a rational mind only if it was not created by a rational mind. (By such reasoning, a computer ought to be an impenetrable black box.) [...]" (Johnson 1993, p.164)
David Livingstone's description of various and sundry theists of the nineteenth century supports the notion that the metaphysical principle of creation yielded a bounded expectation of comprehension of phenomena, not an unlimited or universal expectation (1987). William Henry Dallinger, for example, is quoted to the effect that God's Creation implies a "'final purpose' too great for man to see" (Livingstone 1987, p.97). Johnson's "to some extent understandable" is an unnecessary metaphysical crippling of scientific endeavor, not an essential grant of license as he implies. Also, Johnson does not actually establish paradoxical inference on the part of "scientific naturalists", as reference to Wilder-Smith's argument reveals. Wilder-Smith argued that a superhuman intelligence was required for design of the universe and its components, and that of necessity the lesser intellectual capacity of humans would fail to comprehend some of its works. Johnson conveniently leaves out the critical modifiers in his characterization of his "scientific naturalist"'s stance: Scientific naturalists insist that the cosmos can be understood by a rational human mind only if it was not created by a rational superhuman mind. And as we have seen in Wilder-Smith's writings, it isn't even the scientific naturalists who are making the assertion of a limitation on comprehension; it is instead the theists who have promulgated this doctrine and who have insisted that supernatural causation of phenomena precludes complete comprehension of those phenomena by humans. Scientific naturalists need not do more than note that the argument is properly credited to theists who reject the axiom of comprehensibility. Johnson's parenthetic comment in the quote above is a cheap shot based upon an mischaracterization: computers are designed by human minds, not superhuman ones, and are thus comprehensible to human minds. The targets of Johnson's critique, scientific naturalists, have done nothing "by such reasoning" as was attributed to them.
While the subjective appreciation of a role for supernatural causation may be important to personal fulfillment, it does not afford a basis for objective knowledge, nor can it be counted as a means of comprehending the universe in a scientific manner. As Popper noted, objectivity can only be approximated by inter-subjective experience (Popper 1959, p.44). However, supernatural action is not considered to be amenable to test in the Judeo-Christian tradition. Similar initial conditions do not lead to similar outcomes across observers, nor to reliable repetition in the same observer.
I will connote "naturalism" as "proposing only natural mechanisms for physical phenomena" rather than "asserting that only natural mechanisms have existence". It is easy enough to define terms such that they become useless to anyone, which is how I view those who would make "naturalism", "scientism", and "scientific materialism" all synonymous. Science is incompetent to examine those conjectures which cannot be tested in the light of inter-subjective experience or criticism. The assertion that "only natural mechanisms have existence" is equivalent to the claim that "no supernatural causes exist". That is an example of proving a negative, and can only be regarded as a statement of faith, since it requires omniscience on the part of the claimant.
The term "supernatural mechanism" is an oxymoron. Humans are not privy to the mechanics of supernatural action, nor can they be. The most that can be hoped for is to demonstrate "supernatural causation". Even this appears chimerical, for the following two reasons. First, humans cannot establish a supernatural cause by experimental reproduction of that cause. No human is capable of producing a supernatural cause. Second, natural and supernatural causation are confounding: suspected supernatural causation may simply be due to currently indiscernible natural causes. Because of the confounding nature of the interaction, the only way to establish supernatural causation is through the elimination of all natural alternatives. This is simply another case of proving a negative, which is an intractable problem. That is, asserted supernatural causation logically requires an exhaustive study of possible natural causes of the phenomenon in question, which is counter to the usual desired outcome of such assertions.
Assertions of supernatural causation do nothing to aid human comprehension of physical phenomena. If Wilder-Smith had had his way, genetic algorithms would never have been attempted again, and a fruitful technique would have been denied to researchers and engineers everywhere. Only proposed mechanisms speak to comprehension, and theists do not box in deities with mechanism. Even attempting to assert supernatural causation for some phenomenon necessitates more examination of that phenomenon in light of natural explanation, not less. Enterprising science needs naturalism if it is to be considered a means of advancing knowledge of physical phenomena.
Holland, J.H. 1992. Adaptation in natural and artificial systems. Boston, MA: MIT Press.
Holland, J.H. & Reitman, J.S. 1978. Cognitive systems based on adaptive algorithms. In D.A. Waterman & F. Hayes-Roth, (eds) Pattern- directed inference systems. NY: Academic Press.
Johnson, Phillip E. 1993. Darwin On Trial. InterVarsity Press.
Kuhn, Thomas S. 1970. The Structure of Scientific Revolutions: Second Edition, Enlarged. Chicago: The University of Chicago Press.
Livingstone, David N. 1987. Darwin's Forgotten Defenders: The Encounter Between Evangelical Theology and Evolutionary Thought. Edinburgh, Scotland: Scottish Academic Press.
Medawar, P. 1967. Mathematical Challenges to the Neo-Darwinian Interpretation of Evolution. Philadelphia: Wistar Institute Press.
Overton, William R. 1984. Decision of the court: McLean vs. Arkansas. In: Ashley Montagu (ed.), Science and Creationism, pp.365-397. Oxford: Oxford University Press.
Popper, Karl R. 1959. The Logic of Scientific Discovery. New York: Harper Torchbooks, The Science Library.
Wilder-Smith, A. E. 1970. The Creation of Life: A Cybernetic Approach to Evolution. Wheaton, Illinois: Harold Shaw Publishers.