News and Commentary
Origins & Design 20:1
Issue 38

Blaming the Handyman--An Exchange

By Walter R. Hearn

In “Blaming the Handyman” (O&D, Winter 1999, p. 8), Lydia McGrew related the true story of a couple investigating the disappearance of a gun from their apartment. After eliminating many possible alternative explanations, the couple concluded that a handyman had taken it—which turned out to be correct.

McGrew offered the Tale of the Dishonest Handyman as a parable to counter the charge that theists are prone to abandon scientific investigation. To a philosophical naturalist, theism appears to have a built-in anti-scientific bias. But is it true that science is threatened by allowing “a divine foot in the door”? Faced with unexplained phenomena, do theistic scientists simply give up and say, “God did it,” rather than continue to search for scientific explanations?

As McGrew’s parable illustrates, admitting a supernatural cause in one instance does not commit one to invoking divine activity indiscriminately. After the couple discovered that the handyman had taken the gun, they did not continue to assert, when anything else was missing, “the handyman did it.” McGrew’s point was that Christians working in the sciences are not “forced to bring in supernatural explanations for what otherwise seem to be natural events.”

To me, however, the parable illustrates another point equally well: namely, that at the level of investigation into the causes of particular events, most scientific methodology is independent of ultimate beliefs or worldviews. In the story, the only beliefs that played a role were those unspoken assumptions of intelligibility underlying any rational inquiry (e.g., that real events have real causes, and that we can make sense of their relationships). Clearly, the couple believed that they should be able to figure out what had happened to the missing gun.

On the other hand, the handyman’s existence and access to the apartment were not matters of belief. He was one of those entities “already known to be present in the system.” That is, the couple did not come to believe in the possibility of an intentional causative agent as a result of anything uncovered in their search for the missing gun. Rather, that was information they had prior to their investigation. The wife “remembered the handyman” who “had been left alone in the apartment not long ago.” Considering what they knew of his character (he had “always seemed to be an honest person”), the couple was skeptical that he had committed the deed. Yet as they eliminated other possibilities, the hypothesis of the handyman as the culprit seemed increasingly likely.

In fact, any rational person would have carried out the investigation in the same way. Whatever a scientist’s personal beliefs, science is done by considering the merits of various explanations and eliminating them one by one. At some point, a scientist may draw on prior knowledge based on non-scientific experience—just as prior knowledge led the couple to consider the handyman as a suspect in the theft of the gun. For the theist, the possibility that a supernatural cause may be directly responsible for some event cannot automatically be excluded. Until a scientist runs out of plausible naturalistic explanations, however, a supernaturalist behaves methodologically, just the way a naturalist does.

To acknowledge that supernaturalists use the same research protocols as naturalists is not to champion “methodological naturalism.” To pin a such pejorative label on the way all scientists proceed seems inaccurate, even unfair. It might be best to seek some neutral term for the empirical methods of science—for the “bottom-up analysis” or “data-driven reasoning” in which all scientists engage. Ignoring any ultimate biases toward naturalism or intelligent design, most working scientists begin “from the bottom up,” examining data pertinent to their immediate problem and reasoning from the data toward a satisfying yet tentative solution.

To some, this reading of the parable will indicate a hopelessly “Baconian” view of science. As pointed out by management scientist David Tyler in the same issue of O&D (p. 12), Francis Bacon tried to eliminate dogma and superstition from science by replacing them with “a blank sheet of paper” on which to write data. According to Tyler, this 17th-century Baconian view lingers on as a “myth” despite evidence that “all scientists start with a (personal) conceptual framework or template.” The question I am raising here is not whether scientists have such a framework, but rather what they do with it. In my career as a biochemist, I did start each day’s work by recording data on a page of my laboratory notebook.

In another article in the same issue (p. 30) legal scholar Phil Johnson commented that, “There is no intellectually viable midpoint between naturalism and intelligent design.” That may be true at the level of philosophical or theological debate about science, but the Tale of the Handyman illustrates how scientists actually do their work. At that level, research takes place in a broad territory common to both naturalists and theists. The fact that most scientific work can be carried out in exactly the same way by practitioners across a wide spectrum of basic beliefs shows that such beliefs are not necessarily controlling.

The reason they are not controlling is that even those with a biblical commitment to the Creator’s agency cannot specify ahead of time how divine design might be exercised in a given situation—say, to introduce new information into a macromolecular structure. Going back to the thieving Handyman, I wonder what the couple’s next move would have been if the handyman had not confessed and produced the gun. Without such a revelation, they might still have suspected him of being the culprit. Yet, lacking evidence of how the deed was done, no claim that “he did it” would stand up in court. Similarly, in scientific investigations, lacking revelation about how divine agency has acted, theists should be cautious in their claims.

If scientific claims about God’s action must be tentative, how do theists justify the belief that He does in fact act in the created world? We come to that belief on biblical grounds and on the basis of personal and communal experience, but is it justifiable? Let’s turn to other parables that have been offered, both challenging and supporting theistic belief. More than 50 years ago, John Wisdom offered the parable of the invisible gardener.1 In this parable, two men return to a long-neglected garden to find signs that it has been cultivated in their absence. The first man says a gardener must still be working there. The other denies it. They design various tests to detect the gardener, but without success. The first man continually adjusts his claim, insisting that the gardener exists but must be invisible and intangible. The point of the parable is, in the words of atheist philosopher Antony Flew, that “Theism is a victim of a thousand qualifications.” In other words, attempts to make empirical tests of claims about divine action in the world tend to come up empty.2

In response, Basil Mitchell, Nolloth Professor of the philosophy of the Christian religion at Oxford, offered various parables to illustrate why it is reasonable for Christians to trust their knowledge of God even when some evidence appears to contradict it. In Mitchell’s parable of the lighthouse, for example, a ship is at sea in stormy weather during wartime. The officer of the watch reports seeing a lighthouse, but the navigating officer’s own careful reckoning says they are a hundred miles from land, so he discounts the sighting as a mistake. Soon afterward, however, a lookout reports seeing land. What should the navigating officer now conclude? Given that the coast may be hostile, he should probably trust the sightings despite his own calculations and get out of there quickly. “It would be nice to be surer,” Mitchell concludes, “but in the circumstances it would be prudent to act on the hypothesis which gives him the stronger reason for action.”3 The point is that when we have to make any decision with life-and-death consequences, on the basis of limited or even contradictory evidence, it is reasonable to trust the most compelling evidence.

Mitchell made the same point in the parable of the partisan.4 In this parable, the protagonist is living in enemy-occupied territory and meets a stranger in the resistance movement whose character impresses him as trustworthy. The stranger warns him that at times he will be in disguise in order to obtain information from the enemy. And indeed, later the partisan sees the stranger wearing an enemy uniform and apparently working for the enemy. Nevertheless, based on the evidence garnered through that initial encounter, the partisan trusts the stranger’s word that he is in fact on the same side. The parable suggests that belief in God rests on more than one line of evidence, and that God may intentionally keep some of his activities covert.

The conclusion is that lack of scientific evidence of how God works in the world does not render our faith untenable, but it should make us cautious in making specific claims about how and where God acts. We may have good reason to believe that “the handyman did it,” but in many cases we remain without a clue as to exactly how he pulled it off.


  1. Wisdom, J. “Gods,” Proceedings of the Aristotelian Society, vol. 45, 1944, p. 194e; reprinted in A. Flew, ed., Logic and Language, vol. 1. Blackwell, 1951. See also B. Mitchell, The Justification of Religious Belief. Seabury, 1973, p. 114.
  2. Flew, A. “Theology and Falsification,” in A. Flew and A. MacIntyre, eds., New Essays in Philosophical Theology. SCM Press, 1955, pp. 98-99.
  3. Mitchell, B. op cit, pp. 112-113.
  4. Mitchell, B. “Theology and Falsification,” in A. Flew and A. MacIntyre, eds., New Essays in Philosophical Theology, p. 105. See also, I.B. Barbour, Myths, Models, and Paradigms. Harper & Row, 1974, p. 128.

Blaming the Designer

By Lydia McGrew

“No, no, my good sir,” said Holmes. “There is a master hand here. It is no case of sawed-off shotguns and clumsy six-shooters. You can tell an old master by the sweep of his brush. I can tell a Moriarty when I see one.”

-- Sir Arthur Conan Doyle, The Valley of Fear

Walter Hearn uses my Tale of the Handyman to illustrate a rather different point from mine regard ing the relationship between science and the supernatural, namely, “that at the level of investigation into the causes of particular events, most scientific methodology is independent of ultimate beliefs or world views.” In elucidating this claim, Hearn draws a contrast between “ultimate beliefs” or “personal beliefs,” on the one hand, and “natural explanations” on the other. He implies that in order to consider either of these kinds of hypotheses as an explanation for a physical fact, one must have antecedent reasons to believe in the existence of the entity postulated by the explanation. And in the case of “ultimate” explanations, he suggests that a Christian’s antecedent belief in the supernatural is based on “non-scientific experience,” “personal and communal experience” or “biblical grounds.”

I believe that Hearn has drawn a false dichotomy here. It is arbitrary to treat hypotheses about supernatural beings in a radically different fashion from the way we treat hypotheses about mundane beings. Belief in the supernatural is not necessarily based on non-scientific grounds, nor does it require an antecedent ultimate commitment. On the contrary, it is possible to infer the existence of a supernatural or super-powerful being on purely scientific grounds–indeed, on the basis of a single scientific argument.

For example, in contemplating the origin of life, we can argue for the action of a designer from the phenomenon of irreducible complexity, as Michael Behe has done in Darwin’s Black Box. Given that we ourselves are not (yet) capable of designing life from the ground up, and that scientific data indicate that complex life forms existed before beings such as ourselves were around, it makes sense to conclude that the designer was a being different from, and more powerful than, ourselves. This conclusion does not require a prior ultimate commitment, nor does it directly entail such a commitment; it merely involves inferring the existence of an otherwise unknown being from a single line of scientific argument.

Hearn illustrates the contrast he draws between supernatural and natural hypotheses by appealing to the original story, noting that “the couple did not come to believe in the possibility of a cognitive causative agent because of information they uncovered in the course of their investigation of the missing gun.” Of course, the sheer logical possibility of the existence of such an agent is self-evident. And it’s true that in the story, the couple did not come to accept the plausibility of such an agent’s existence from their investigation of the missing gun, since they already knew that there was a handyman. Moreover, it certainly would have been much harder to convince them that the gun had disappeared by divine, rather than human, intervention. What this shows, however, is not that there is a radical difference between the way we ought to treat supernatural and natural hypotheses but rather that supernatural explanations have a lower prior probability than explanations involving only mundane entities. However, sufficiently strong empirical evidence can overcome even a low prior probability for a hypothesis.

Hearn notes that “basic beliefs” need not be epistemically “controlling” in scientific investigation, and he advocates what he calls “data-driven” scientific investigation, in which committed theists and committed naturalists usually behave indistinguishably. While I strongly agree with him in rejecting the notion of “controlling” presuppositions, I cannot agree with the suggestion that scientists may invoke a divine agent only after exhausting all naturalistic explanations (and then only tentatively and cautiously). Hearn seems to assume that the broad overlap in scientific practice among theists and naturalists is essential evidence that objectivity is possible, and that scientific neutrality is best maintained by treating supernatural hypotheses as personal religious beliefs and largely ignoring them. But I do not believe that this model characterizes truly objective science. Certainly, the ideal investigator should not be biased by theistic (or other) preconceptions. But neither should he exclude supernatural hypotheses a priori (as a committed naturalist would do), nor artificially defer and segregate them. A scientist can be data-driven when he confidently invokes the action of a supernatural agent as the best explanation of the empirical evidence.

Because Hearn believes we are confronted with ambiguous and even contradictory evidence regarding the existence of God, he urges us to fall back on prudential arguments such as those given by Basil Mitchell in order to justify theistic belief. Though he does not specify what the contrary evidence might be, the quotations from Flew and Mitchell suggest that he is thinking of the problem of evil. But it is one thing to ask whether there is unambiguous evidence that life was designed and quite another to ask why God permits evil in the world. In its simplest form, ID does not postulate a designer of a particular character, still less the God of Abraham, Isaac, and Jacob. Since the argument that life is the product of design rests on highly specific and compelling evidence regarding the nature of organisms (even organisms hostile to man), it simply is not true that ID proponents must be “cautious” in making this claim.

Finally, Hearn stresses that theists should be cautious and tentative in making “scientific claims about God’s action” because “even those with a biblical commitment to the Creator’s agency cannot specify ahead of time how divine design might be exercised in a given situation.” The insistence that design theorists propose a highly specific description of the designer’s modus operandi has dogged the ID movement from the beginning. Both from Darwinists and from some who are in a broad sense sympathetic, such as Hearn, we hear that we must say exactly how design was carried out in order for ID to be taken seriously as science. But this is incorrect. The hypothesis of agent involvement has predictive content in itself and thus can be scientifically supported. For example, since we know that agents make irreducibly complex entities, but that such objects are highly improbable if only non-intelligent causes are operating, the presence of such an object is predicted more strongly and explained better by agency than by naturalistic hypotheses. This is a matter of predictions and probabilities, even without a more specific theory as to how the designer has acted.

Hearn argues that without an admission of guilt from the handyman in the original story, an accusation would not have stood up in court, because without a confession the couple would have lacked “evidence of how the deed was done.” His point is, once again, that operational details are required for a high degree of confidence regarding agent intervention. But an admission of guilt in itself would not tell the couple whether, for example, the thief picked the lock on their gun case or used a copy of the key. In fact, a confession supplies no further details as to how the deed was done, and such details were not a necessary part of the argument for the handyman’s guilt.

To take another real example, some years ago in the state of Arkansas a man’s body was discovered decapitated. There was no machinery in the vicinity to indicate that death resulted from an accident, and the coroner’s ruling was that the man died of natural causes! That some agent was involved in the man’s death and that the coroner was corrupt is patently clear, even in the absence of details such as the exact weapon used, whether the killer was right-handed or left-handed, and so forth. Once the coroner’s faulty ruling was revised, it was possible for investigators to search for the killer under the description “the perpetrator of this beheading,” even though they would have had no reason to believe that there was any such person had the body never been discovered. Similarly, even if our only evidence for a designer is the discovery of objects which appear to have been designed, the nature of the objects themselves allows us to conclude confidently that a designer exists, regardless of whether we have any other reason to suspect his existence or independent evidence as to exactly how and when he intervened.

Suppose, however, that we wished to “pin” an event on the Christian God specifically, rather than merely conjecturing the involvement of some intelligent agent. Would Hearn then be correct that we need independent “revelation about how divine agency has acted” or that we must be able to “specify ahead of time how divine design might be exercised in a given situation” if we are to make more than “tentative” and “cautious” claims that God has acted at a particular place and time? To answer that question, we should consider what such specific theories could possibly look like: Must we have some sort of inductive generalization to the effect that God only performs miracles a certain proportion of the time and otherwise leaves things to natural law? Must we have an independently plausible conjecture at the quantum-physical level as to why Christ’s body did not decay in the tomb? I cannot see that any such information would be either helpful or necessary. It would not be helpful, because it would not distinguish between the action of the Christian God and any other supernatural or super-powerful being. And it would not be necessary, since God is by definition an omnipotent being; therefore, He is capable of intervening in the physical world in an infinite number of different ways and at any time. We have no reason to expect, and need not look for, some particular underlying mechanism. If God has acted, we should not expect to find the mark of His chisel and mallet, but rather effects best explained as the work of His hand.

Copyright 2000 Access Research Network. All rights reserved. International copyright secured.
File Date: 6.27.01