Science in Christian Perspective
Prediction Sets and
Their Consequences for Applying Intelligent Design Theories
David F. Siemens*
2703 E. Kenwood
Mesa, AZ 85213
From Perspectives on Science and Christian Faith 51.2 (June 1999): 108-113. Response: Mills
Intelligent design, stripped to essentials, covers a broad range of theistic views. It includes Van Till's "functional integrity," which insists that God, in the original creation, provided both the causal principles and physical basis for the development both of the inanimate universe and terrestrial life.1 This view is not to be confused with deism or process theology, for it holds (1) that the rational Creator originally established the universe so that, under his continual providential care, it developed "naturally," and (2) that he is omniscient, omnipotent, and sovereign.2 Though Van Till is Reformed, his view is comparable to Luther's teaching that all natural principles are larvae dei, the masks of God, behind which he is at work.3 Although it is always God at work, we see only the masks, whether we look at the development of the inorganic, from the Big Bang on, or the total development of the organic world. This is why we declare with the psalmist, "The heavens declare the glory of God," while recognizing that "The fool hath said in his heart, There is no God."4 We recognize God's hand behind natural events while the fool does not. Indeed, the fool's attitude is the same as that which brought forth Christ's rebuke: "Except ye see signs and wonders, ye will not believe."5
The Common Views of Intelligent Design
At the other extreme, intelligent design includes immediate creationism, the notion that God, in the recent past, produced the universe in something like its present state. Since this does not produce empirically testable consequences, it will not be further considered here.
Common among the views to be examined is continuous creationism. The more obvious version holds that every species, or every genus, or at least every family, was created independently at the appropriate time.6 A more moderate version of this type is expressed by those who are developing the Intelligent Design Theory. Since some of its members object to such placement, I will present a brief justification of my claim.
Behe's "irreducible complexity" holds that life exhibits structures and functions that could not have developed gradually.7 Despite his lack of expressed commitment, his view implies either that the mechanisms were introduced by some nonnatural force, or that organisms had to accumulate useless structures or "mutations" until the functional organ or process emerged.8 It appears difficult to envision a natural mechanism for this latter condition, while the former fits Dembski's approach.9 Dembski claims that, in a closed natural system, the amount of information cannot increase. Thus, more complex organisms with more advanced functional repertoires cannot arise except by the input of information from without. This claim may require a little modification,10 but necessarily excludes production of the genetic information found in organisms by material forces.11
Consequently, we may divide the total range of positions espousing intelligent design into two parts. On the one side is Van Till's "functional integrity" (hereafter referred to as FI), that is, what may be studied scientifically involves no external inputs.12 The alternative is a group of views which insist that there must be, in addition to the miraculous introduction of life, multiple inputs over time for the biosphere to have become what we now observe. To be sure, there is a notable difference among the claims that no more than information has to be inserted from without; that entire organs, functions, and their genetic support must be produced by nonnatural means; and that whole organisms are created and introduced at appropriate times during Earth's development. However, for what follows, these distinctions have overlapping consequences. So this set of views will be referred to as multiple input theories or MI.
The Two Sets of Predictions
Two distinct groups of predictions spring from these alternate versions of intelligent design. The first prediction provides that in the genome of normal individuals, whatever their species, we should find no quiescent genes that are similar to active genes. A related prediction requires that at least some structurally similar genes will have specific functions that do not parallel or overlap those found among related species. This is consistent with MI. The second prediction specifies that we should find sets of genes in various species in which one gene of the set preserves a function while other members of the set produce different effects. This includes the possibility that some of these similar genes may be quiescent in some lineages. Further, such sets should be more common in closely related species than in markedly different species. This is primarily compatible with FI.
These predictions do not depend on, or derive from, the questions usually discussed by proponents of intelligent design: namely, the inclusion of notions growing out of Intelligent Design Theory in the practice of science. I have discussed some of these elsewhere.13 Here the problems spring solely from the consequences of these two versions of the theory. Also, these two sets are not contradictory, coming closer to being contraries, though they even fall short of this. Yet each prediction grows out of, and tends to confirm, only one of the two fundamentally distinct intelligent design views noted above.
First Consequence: On Efficiency
This last point needs explanation and qualification. As noted above, extreme MI holds that all the primordial creatures came into existence directly by divine fiat. If we assume that the deity knows as much about efficiency as human beings, then the creatures' functions should proceed without excess demands on their metabolism. It has been observed, for example, that at least some plants' syntheses of those chemicals that defend against infectious attack takes away from the synthesis of materials needed for growth and reproduction. Thus, the defensive chemicals are synthesized only when a plant is attacked.14 In the interests of efficiency, one may expect that God would provide only the nuclear DNA necessary to produce the proteins vital to the organism or needed to control protein synthesis and other basic functions. Production of unnecessary DNA takes energy that otherwise could be expended for the organism's welfare. If all introns can be shown to be essential to the control function or to meet some other essential need, then we can claim confirmation of the direct creation of the creatures. On the other hand, if part of an intron is unnecessary, unless it is demonstrably not original,15 or if some exons are redundant, then the evidence against their immediate production by an intelligent deity is strengthened. The alternative is that God inserted "garbage" sequences to mislead us.16 But dare we suggest this of the one who swore by himself because he could swear by no greater?17
In contrast, moderate MI is not as affected by this test. Further, proving that some part of a creature's genome is useless is very difficult. It will not do to produce a modern version of Weiderscheim's late nineteenth century list of vestigial organs, which included all of the endocrine glands. Nevertheless, the volume of introns in the genomes of advanced plants and animals raises questions about their efficiency. For example, three bacteria have about 1,100 bases per gene in their genome; a yeast, 2,000. This we may attribute to their relative simplicity. A roundworm, Caenorhabditis elegans, and a simple dicot, Arabidopsis thaliana, have about 7,500. Homo sapiens comes in at about 30,000 bases per gene; but four species of grasses range from about 14,000 to over 500,000.18 Why should wheat require thirty-seven times as many bases per gene as rice?
Second Consequence: On Descent
The second prediction in the first set (at least some structurally similar genes will have specific functions that do not parallel or overlap those found among related species), which is connected to the first prediction of the second set (we should find sets of genes in various species in which one gene of the set preserves a function while other members of the set produce different effects), also requires comment. If a distinct gene springs from duplication followed by mutation to modify its purpose, we assume that the original function must be maintained by one of the pair, while the other is free to acquire a different function. Since either or both genes may repeat this path, there may be several similar genes with distinct functions, but one must maintain the original function if it is vital.19 At least in the simple case, we can then compare various organisms to determine the developmental sequence. For example, let us imagine that species A is ancestral to sibling species B and C, with B the progenitor of D. We must expect that gene a, found in A, will be found in all four species. If a gave rise to b in B and to c in C, we can expect that D will carry a and b, but not c. This is not a total prohibition, for there can be convergent and parallel evolution. Gene a can presumably give rise to c in D as easily as to a different gene, d. Of course, if there is a gene d, it could arise from b rather than from a. This possibility must also be considered. Any such sequences lend more support to FI, but do not eliminate MI.
What appears to be an example of this sort of development is found in the visual pigments. The gene for rhodopsin, the widely distributed basic visual pigment, is found on chromosome 3 in human beings. The blue-sensitive pigment, found on chromosome 7, is more than 40% identical with it. The sex-linked red and green pigments are 98% identical with each other, but only 43% identical with the blue pigment. All monkeys, apes, and humans have rhodopsin and the blue-sensitive pigment, but only the Old World monkeys, apes, and humans have both the red- and green-sensitive ones on the X chromosome. Interestingly, in Homo sapiens the latter may occur twice or thrice. New World monkeys have only one long-wave pigment on their X chromosome.20 Hence, all the males are dichromats. However, because some species have developed multiple alleles of the long-wave pigment, females can be trichromats.21
In contrast to the scheme noted above, if A, B, C, and D were independently created, there is the probability that, in the context of special sets of proteins or other enzymes, which in principle do not even have to overlap, gene a may be missing in one or more of the organisms. Ex hypothesi, its place can be taken by a totally different gene producing an unrelated enzyme having essentially the same effect within a totally different context as a has in its environment. A homolog to a might then have a totally different function. This has been argued in favor of MI. There may be an example in bacterial genomes.22 However, in making this interpretation, one must be careful of redundant or interacting pathways.23
One must be aware of a further problem in pressing this claim, namely, that some genes have multiple functions. If one knows of one function of a gene in A and of a different function of its closest known homolog in E, it is not established that the genes have different functions in the two organisms. A closer investigation may reveal that both functions are present in both species. Note that in C. elegans, the roundworm that has been studied in remarkable depth, a single pathway is involved in vulva formation, cell migration, oogenesis, and male tail development.24 Knowing only some of the functions could lead to a claim of nonoverlap when functions are actually similar. However, a genuine total set of homologs without overlapping functions in a pair of species would be a strong support of MI. On the other hand, several copies of genes with the same function apparently can coexist in an organism.25 These matters are open to experimental analysis, but will require a great deal of work to pin down.
A challenge to some forms of MI is found in the horizontal transmission of hereditary elements. Their appearance in pathogenic organisms may be ascribed to the consequences of sin, making them more efficient producers of pain and death. Yet, if the progenitors of all creatures came perfect from the hand of God, then horizontally transmitted elements can hardly increase the fitness of an organism. It must be that these elements interfere with their God-given original fitness, for MI cannot tolerate the notion that God created sloppily. Thus, there is an opportunity for research to confirm MI. If every horizontally transmitted genetic element interferes with the competence of the organism receiving the element, the stricter versions of MI are confirmed. On the other hand, if all of these elements enhance an organism's fitness, FI is confirmed. A mix of benefits and detriments is compatible with both FI and moderate MIs.26
Third Consequence: Extinction vs. Overlap
There is, however, a version of FI that can be distinguished from MI, only with greater difficulty. If one assumes that the Creator determined the total set of "natural" principles and then "supernaturally" determined every twist in the sequence of the outworking of those principles, the paleontological record would probably not differ observably from that of the sequential production of the creatures by divine fiat. Nevertheless, the tests noted above might detect subtle differences in current genomes. But, since this version of FI calls for as much direct divine intervention as MI, it is not usually encountered. However, there is a possible observation that would make it less likely, for it posits that the development of new taxa should be linear. If, in contrast, the geologic record indicates numerous variants among sibling species with only a few persisting to produce later taxa, this version of FI is rendered much less likely. The same applies to strict MI.
In addition, the vast number of highly specialized creatures found only in the geologic record fit better with normal FI than with this specialized version or with MI. It seems much harder, under the presuppositions of any version of MI, to explain the development or creation of so many different creatures, only for them to be exterminated.
The normal version of FI expects continuity throughout living things. It has no problem absorbing the recent observation that the same gene produces eyes in fruit flies, squid, and mice, although their structure is very different.27 An evolutionary tree is not automatically rejected. It assumes that God not only had ages within which to work, but has reasons for developing the universe and everything within it during these long periods. It asks why God should create sequentially over long ages when he could have done it all instantaneously.28 At this point, the discussion is philosophical and theological rather than scientific. But the discovery of the fossils of whales and snakes with legs seems to fit FI better than most versions of MI.29
A further complication is the apparent ease with which presumed ancestral features may be restored. For example, perissodactyls apparently originated in the late Cretaceous, ca 60 million years ago (Mya). The artiodactyls probably split off about 50 Mya. Yet mule-footed pigs are known today.30
Final Consequences: On Life, Leanings, and Logic
On the other hand, the origin of life is a graver problem for FI, at least in the current state of science. The computation of difficulty with odds beyond the astronomical31 assumes that all life must be at least as complex as what we currently observe.32 But this has not been demonstrated, only assumed. As Fulton notes, the possibility of "scaffolding," that is, the likelihood of intermediate supporting structures or states that have not survived, has seldom been considered.33 This is because it is extremely difficult to examine the problem of "scaffolding" within current scientific knowledge. Indeed, should someone come up with the correct sequence of ancient events, it is exceedingly unlikely that he or she could confirm its correctness, and it is certain that many adherents of MI would dismiss it as unfounded speculation.34
The variety of body plans has been urged against FI, on the basis that no mechanism has been suggested for the Cambrian explosion. Certainly, some of the Pre-Cambrian and Cambrian fossils seem exceedingly strange. As more becomes known, however, some of these fossils appear to fit into more familiar patterns. Further, the known forms are too few to tell us much about whatever happened between the first living forms and those preserved as fossils.35 So this appears to be what logicians know as the fallacy ad ignorantiam. The argument may also be turned around. Why would an intelligent designer introduce information to produce such bizarre creatures destined for oblivion? Unless one adopts the limited deity of process theology, there seems no reasonable explanation from MI. But this option excludes theism.36
This last point underscores a fundamental problem in this area. Tacit commitments have more effect than we normally recognize. The problem is more with theological and philosophical commitments than with scientific investigations. As I noted previously, some arguments are based on what we do not know,37 the ad ignorantiam fallacy. In contrast, I am tentatively suggesting some experimental tests which bear on FI and MI and produce a factual basis for decision. I grant that these are both difficult and inconclusive, although scientific progress should make them possible. Additionally, working scientists need to think through what such experiments can show, and plan better ones. Still, I justify my comments in that, first, they should bring more clarity to the discussion; and second, science is not capable of strict demonstration. Additionally, I trust that these thoughts will lead to discussion of additional areas of investigation and to more relevant tests.38
I thank the referee whose analysis and comments clarified both the argument and several points. He is not responsible for remaining errors.
1See Howard J. Van Till, "Special Creationism in Designer Clothing: A Response to The Creation Hypothesis," PSCF 47 (June 1995): 123˝31; "Is Special Creationism a Heresy?" Christian Scholar's Review 22 (June 1993): 380˝95. In personal correspondence, George L. Murphy noted that discussions of design almost always neglect this view.
2See Siemens, "Don't Tar Van Till: A Response to Anderson and Mill," PSCF 49 (March 1997): 70.
3Another traditional term is Deus absconditus, God hidden in his works.
4Pss. 19:1 and 14:1; 53:1, respectively.
5John 4:48. Is it too strong to suggest that some adherents of the Intelligent Design Theory, specifically those holding to continuous creationism, are similar to those to whom Christ spoke? They seem to hold that, unless there is an overt miracle, God cannot be at work.
6This appears to be the view of Phillip E. Johnson, Darwin on Trial (Washington, DC: Regnery Gateway, 1991); Evolution as Dogma: The Establishment of Naturalism (Dallas, TX: Haughton, 1990); and of J. P. Moreland, "Complementarity, Agency Theory, and the God-of-the-Gaps," PSCF 49 (March 1997): 2˝14.
7Michael J. Behe, Darwin's Black Box: The Biochemical Challenge to Evolution (New York: Free Press, 1996).
8If this external force is denied, then his version of intelligent design collapses to FI.
9William A. Dembski, "Intelligent Design as a Theory of Information," PSCF 49 (September 1997): 180˝90.
10See Siemens, "Dembski's Dogma Denied: Natural Causes Produce Information," PSCF 49 (December 1997): 287f. Also relevant is the comment in Denis Duboule, "The Evolution of Genomics," Science 278 (24 October 1997): 555.
11It has been claimed that the Intelligent Design Theory merely claims that design can be recognized in nature. While this may be true of some, Dembski explicitly and Behe implicitly require an intelligent source for the structure of information or design.
12This does not proscribe miracles, for God is sovereign. Some miracles are required by orthodoxy, for example, the virgin birth and resurrection of the Lord. But these events are on a par with "In the beginning God created" (Gen. 1:1), for they initiate a totally new entity.
13See Siemens, "On Moreland: Spurious Freedom, Mangled Science, Muddled Philosophy," PSCF 49 (September 1997): 196˝9.
14See Barbara Baker, et al., "Signaling in Plant-Microbe Interactions," Science 276 (2 May 1997): 726˝33; P.<|>J. O'Donnell, et al., "Ethylene as a Signal Mediating the Wound Response in Tomato Plants," ibid. 274 (13 December 1996): 1914˝7; Ann Simon Moffat, "Improving Plant Disease Resistance," ibid. 257 (24 July 1992): 482f; Barry McGurl, et al., "Structure, Expression and Antisense Inhibition of the Systemin Precursor Gene," ibid. 255 (20 March 1992): 1570˝3; Carole L. Cramer, et al., "Rapid Switching of Plant Gene Expression Induced by Fungal Elicitor," ibid. 227 (8 March 1985): 1240˝3.
This is different from the plant-wide synthesis of a signal to a parasitic wasp when a plant is being eaten. This signal is useful only at the time of an attack. See H.<|>T. Alborn, et al., "An Elicitor of Plant Volatiles from Beet Armyworm Oral Secretion," Science (9 May 1997): 945˝9.
15This is a possibility since the types of MI considered here allow for at least millions of years of development since their original creation for most species or whatever other taxon was primitive, or since the information or structure was introduced. Human beings are a possible exception, if the species is restricted to tens of thousands or, at most, to a few hundred thousand years. This places tighter restrictions on what may be allowed in the human genome under these versions of MI. Nevertheless, as much as 1% of the DNA people carry is worse than useless, being made up of sequences of genetic code that have the potential to produce infectious viral particles. See Robert P. Lanza, David K.<|>C. Cooper and William L. Chick, "Dealing with Viral Stowaways," Scientific American (July 1997): 59.
16Objection has been raised against calling any part of the genome "garbage." At one time it was popular to divide the genome into coding and "junk" DNA. But more recently, at least some noncoding DNA has been found to have vital functions. Still, no one can now demonstrate either that all DNA has some utility or that some is useless junk. Much further study is needed.
17Heb. 6:13. See Titus 1:2.
18John Cohen, "Corn Genome Pops Out of the Pack," Science 276 (27 June 1997): 1962.
19The necessity of a specific organ may vary with the environment. For example, whatever genes govern the normal development of chordate and arthropod eyes may become nonfunctional in cave-dwelling species, where blindness is no hindrance. However, similar conditions of light have had an opposite effect among abyssal fish, whose visual sensitivity appears greater and who have light-producing organs. So "one-size-fits-all" theories are too simplistic. An additional complication is that some functions are redundant in certain organisms but essential in others. Knocking out a "vital" gene may be lethal in one organism and have little effect in another. Living creatures are too complex for snap judgments. See, for example, John Ralph, et al., "Abnormal Lignin in a Loblolly Pine Mutant," Science 277 (11 July 1997): 235˝9, summary p. 157.
20Jeremy Nathans, et al., "Molecular Genetics of Human Color Vision: The Genes Encoding Blue, Green, and Red Pigments," Science 232 (11 April 1986): 193˝202; Jeremy Nathans, et al., "Molecular Genetics of Inherited Variation in Human Color Vision," ibid.: 203˝10.
21Song-Kun Shyue, et al., "Adaptive Evolution of Color Vision Genes in Higher Primates," Science 269 (1 September 1995): 1265˝7.
22Frederick C. Neidhardt, review of E.<|>C.<|>C. Lin and A. Simon Lynch, "Regulation of Gene Expression in Escherichia coli," in Science 276 (20 June 1997): 1812.
23See, for example, Michael Balter, "New Clues to Brain Dopamine Control, Cocaine Addiction," Science 271 (16 February 1996): 909, and the reference in note 19.
24Robert K. Herman, review of C. elegans II, Science 276 (13 June 1997): 1656.
25Nigel Williams, "Yeast Genome Sequence Ferments New Research," Science 272 (26 April 1996): 481.
26See Denis Duboule, "The Evolution of Genomics," Science 278 (24 October 1997): 555.
27Maria Barinaga, "Focussing on the eyeless Gene," Science 267 (24 March 1995): 1766f; George Halder, et al., "Induction of Ectopic Eyes by Targeted Expression of the eyeless Gene in Drosophila," ibid.: 1788˝92; "On the Path of the Primordial Eye," ibid. 275 (28 March 1997): 1885. Walter J. Gehring, "Letters," ibid. 272 (26 April 1996): 468f.
28This point goes back to St. Augustine in the fourth century.
29Philip D. Gingerich, et al., "Hind Limbs of Eocene Basilosaurus: Evidence of Feet in Whales," Science 249 (13 July 1990): 154˝7. Annalisa Berta, "What Is a Whale?" ibid. 263 (14 January 1994): 180f. J.<|>G.<|>M. Thewissen, S.<|>T. Hussain, and M. Arif, "Fossil Evidence for the Origin of Aquatic Locomotion in Whales," ibid.: 210˝2. Both Gingerich, et al. (p. 154) and Berta (p. 180) note that some embryonic whales have hind limbs. Carl Zimmer, "How the Snake Lost Its Legs," Discover (July 1997): 32f.
30The investigation of the genetics may have interest to the various versions of intelligent design. Is the gene that produces a single hoof in Sus homologous with that of perissodactyla, or a deletion or mutation of a different sort? A related matter involves archaic genes. See "Reviving Old Mouse DNA." Science 264 (1 April 1994): 27.
31See, for example, Charles B. Thaxton, et al., The Mystery of Life's Origin: Reassessing Current Theories (New York: Philosophical Library, 1984), 127˝66, 218f.
32This assumption is not restricted to adherents of MI. See the letters in Science 276 (20 June 1997): 1776f.
33Alice Fulton, review of Michael J. Behe, Darwin's Black Box in PSCF 49 (June 1997): 122.
34This is already explicit in the recent-creation group of adherents to MI. They dismiss all science that they cannot twist to their purpose. For example, I recently was challenged to explain the existence of comets, whose lifetime is on the order of ten thousand or so years, in a solar system five billion years old. The challenger would not accept the Oort cloud or Kuyper belt, even though members of the latter have recently been observed. As Swift observed, "There's none so blind as they that won't see."
35See, for example, Richard A. Kerr, "Timing Evolution's Early Bursts," Science 267 (6 January 1995): 33f; Gregory A. Wray, et al., "Molecular Evidence for Deep Cambrian Divergences Among Metazoan Phyla," ibid. 274 (25 October 1996): 568˝73; Kenneth J. McNamara, "Dating the Origin of Animals," ibid. (20 December 1996): 1995f; Heinrich D. Holland, "Evidence for Life More Than 3850 Million Years Ago," ibid. 275 (3 January 1997): 38f.
36This view, currently much discussed, considers God to be time-bound and therefore restricted to some extent in what he can predict as he acts. See, for example, James O. Morse, "The Great Experimenter?" PSCF 49 (June 1997): 108˝10; Siemens, "Don't Tar Van Till: A Response to Anderson and Mills," ibid. 49 (March 1977): 70. Despite its acceptance by some who claim to be evangelicals, I hold that this viewpoint pantheistically places the deity in the universe rather than understanding him theistically (or even deistically) as the eternal Creator of the universe ex nihilo. It further seems to compel modalism rather than orthodox Trinitarianism. So it will not be further discussed here.
37See note 2.
38These ruminations grow out of obiter dicta in a pair of juxtaposed "Perspectives" on pages 1658˝60 of Science 276 (13 June 1997). Gerald F. Joyce, "Evolutionary Chemistry: Getting There from Here," noted that gene-duplication allows "[o]ne gene copy to maintain the original function, while the other is free to evolve a new function." Daniel L. Hartl, "Mariner Sails into Leishmania" commented on horizontal transmission of genetic elements between families, orders, and phyla.