The Use and Abuse of Philosophy of Science: Response to J.P. Moreland

Stephen C. Meyer, Ph.D.

Reprinted from Perspectives on Science & Christian Faith: The Journal of the American Scientific Affiliation 46, no. 1

J.P. Moreland's "Conceptual Problems and the Scientific Status of Creationism" argues against the notion that creationist theories are inherently unscientific. He suggests: (1) there are no good reasons to exclude postulations of intelligent design or special creative acts of God from science a priori and (2) there is at least one good reason to allow consideration of such postulations in science—namely, creationist theories attempt to solve conceptual problems which, following Laudan, he regards as a primary function of many scientific theories. Moreland's analysis does not address any of the specific empirical claims that the various creationist theories (old-earth, young earth, theistic macromutationalist, etc.) make, but instead seeks to counter the claim that such theories can not (i.e., in principle) be considered scientific because they invoke special divine action as part of their explanatory framework. Thus, unlike Ruse, Stent, Gould, Grizzle, Murphy and others, Moreland does not regard the possibility of a scientific theory of creation as "self-contradictory nonsense."

While Moreland's conclusions no doubt seem quite radical to many practicing scientists and longtime ASA members, his arguments are, in my opinion, quite sound. Philosophers of science have generally lost patience with attempts to discredit theories as "unscientific" by using philosophical or methodological litmus tests. Such so-called "demarcation criteria"—criteria that purport to distinguish true science from pseudo-science, metaphysics and religion—have inevitably fallen prey to death by a thousand counter examples. Well-established scientific theories often lack some of the allegedly necessary features of true science (e.g., falsifiability, observability, repeatability, use of law-like explanation, etc.), while many disreputable or "crank" ideas have often manifested some of these same features.

Consider, for example, falsifiability. As Imre Lakatos has shown some of the most powerful scientific theories have been constructed by those who stubbornly refused to reject their theories in the face of anomalous data. On the basis of his theory of Universal Gravitation, Newton, for example, made a number of predictions about the position of planets that did not materialize. Nevertheless, rather than rejecting the notion of universal gravitation he refined his auxiliary assumptions (e.g. the assumption that planets are perfectly spherical and influenced only by gravitational force) and left his core theory in place. As Lakatos has shown, the explanatory flexibility of Newton's theory in the face of apparently disconfirming evidence turned out to be one of its greatest strengths. Such flexibility was emphatically not a token of "non-scientific status" as the Popperian model would suggest.

Indeed, more careful study in the history of science has shown the falsificationist ideal to be extremely simplistic. Rarely are the core commitments of theories directly falsified via a single failed prediction. Instead, predictions occur when core theoretical commitments are conjoined with auxiliary hypotheses; thus, leaving open the possibility that auxiliary hypotheses, not core commitments, are responsible for deviations from prediction. On the other hand, the history of science is littered with the remains of failed theories that have been falsified, not by the air-tight disproof of a single anomaly, but by the judgment of the scientific community concerning the preponderance of data. Are such falsified, and therefore falsifiable, theories (e.g. the flat earth, phlogiston, heliocentricism, etc.) more scientific than successful theories (such as Newton's in, say, 1750) that are capable of wide-ranging explanatory power.

As the philosopher of science Larry Laudan has shown, such contradictions have plagued the demarcation enterprise from its inception. As a result, most contemporary philosophers of science regard the question 'what distinguishes science from non-science' as both intractable and uninteresting. Instead, philosophers of science have increasingly realized that the real issue is not whether a theory is scientific, but whether a theory is true, or warranted by the evidence. As Laudan puts it, "If we could stand up on the side of reason, we ought to drop terms like 'pseudo-science'. . .they do only emotive work for us." As Martin Eger has summarized, "[d]emarcation arguments have collapsed. Philosophers of science don't hold them anymore. They may still enjoy acceptance in the popular world, but that's a different world."

Despite having fallen into disrepute with philosophers of science, demarcation arguments remain especially popular with both creationist and evolutionary polemicists (and, alas, many contributors to this journal). Nevertheless, the use of demarcation arguments to distinguish the scientific status of competing programs of origins research is especially problematic. One of the reasons for this is that many origins theories, if true, have obvious metaphysical overtones or implications. Those wishing to separate the scientific from the religious in contemporary cosmogony, for example, may find themselves facing quite a conundrum. On what basis could one assert that the various secular anthropic principles, many-worlds scenarios or quantum-cosmologies are any more or less scientific (or more or less religious) than, for example, recent theistic interpretations of "delicately balanced" physical constants or the big bang singularity?

Attempts to distinguish the scientific from the pseudo-scientific in origins research fail for another reason. The demarcationist arguments used in the origins controversy almost inevitably presuppose a positivistic or neo-positivistic conception of science. Among their other deficiencies, such accounts of science fail to take into account the distinctive methodological character and limitations of the historical sciences. Theories of intelligent design or creation have been alleged to be necessarily unscientific because they: (a) do not explain by reference to natural law, (b) invoke unobservables, (c) are not testable, (d) do not make predictions, (e) are not falsifiable, (f) provide no mechanisms (g) are not tentative, (h) have no problem solving capability, etc. Evolutionary theories have been tarred with many of the same methodological brushes.

As I have argued elsewhere, however, none of these criteria provide grounds for distinguishing the a priori scientific status of either program of origins research over the other—unless, that is, the criteria are applied in a tendentious or question begging way. Indeed, my research has suggested that metaphysically neutral criteria do not exist that can define science narrowly enough to disqualify theories of intelligent design or creation a priori without also disqualifying theories of naturalistic descent or evolution on identical grounds. Either science will be defined so narrowly as to disqualify both types of theory, or science may be defined more broadly, in accord with appropriate desiderata for historical inquiry, and the initial reasons for excluding opposing theories will evaporate.

Consider the following example. Creationist theories have often been said to be unscientific because they make reference to an unobservable intelligence that can not be studied or tested empirically. Yet, if unobservability precluded testability neither evolutionary nor creationist theories could qualify as scientific. Indeed, a core commitment of evolutionary theory—namely, that present species are related by common ancestry—has a very similar epistemological character to present theories of intelligent design or creation. The transitional life forms that ostensibly occupy the nodes on Darwin's branching tree of life are unobservable, just as the postulated past activity of a designer is unobservable.

Origins theories generally must make assertions about what happened in the past to cause present features of the universe (or the universe itself) to arise. They necessarily must attempt to reconstruct unobservable past causal events from present clues or evidences. Methods of testing, therefore, that depend upon direct or repeated observation of cause/effect relationships have minimal relevance to origins theories of whatever type. Those who insist that testing must involve direct observation of causal antecedents or predicted consequents rather than explanation of data after the fact will find nothing scientific in any origins theory. If, however, one accepts the necessity of testing competing historical theories ex post facto by comparing their explanatory power, then the original reason for excluding creationist theories from consideration dissolves. My analysis of the other demarcation arguments enumerated above suggests they are similarly incapable of discriminating the a priori scientific status of creationist and evolutionary theories.

Thus, from the standpoint of contemporary philosophy of science and recent work on evolutionary demarcation arguments Moreland's first assertion seems to me unproblematic. There do not seem to be convincing arguments for disqualifying creationist theories as inherently unscientific. Nevertheless, without further demonstration, many practicing scientists may be forgiven for a certain suspicion of philosophers of science. What after all do philosophers really know about science? Yet as Moreland has pointed out, demarcation arguments do not make claims about nature itself, thus, reflecting the domain of scientists; rather, they make second-order assertions about the nature and method of scientific practice, the study of which does directly and legitimately concern philosophers of science. In this case, it is scientists, not philosophers who assert beyond their authority.

Moreover, given the recent trend within the philosophy of science to integrate philosophical analysis with historical study, there seems, to me at least, little reason to doubt philosopher's conclusions about the failure of demarcation arguments. Too many examples from the history of science itself show these arguments to depend upon oversimplifications and caricatures of scientific practice. Nor is this fact surprising when one considers the history of philosophy. Demarcation, historically, has been the special project of positivist and neo-positivistic philosophers whose claims were judged deficient in part because they misrepresented actual scientific practice—as Michael Polanyi a scientist turned philosopher so convincingly demonstrated.

Thus, the first part of Moreland's argument seems sound. There do not seem to be good reasons to exclude postulations of intelligent design or special creative acts of God from science a priori. Instead, most of the reasons for disqualifying such theories seem to be derived from discredited positivistic accounts of scientific rationality. Yet what about Moreland's second (and main) argument? Are there any good positive (though not positivistic) reasons to consider creationist theories scientific? Do recent non-postivistic accounts of scientific method and rationality suggest the possibility of a scientific theory of creation? Moreland invokes recent work by Larry Laudan to answer both questions in the affirmative.

And here again, I agree with Moreland's conclusion. His analysis illustrates persuasively to me one way that creationist theories might well conform to a general model of scientific practice. If one takes Laudan's work as a good descriptive account of what scientific theories do, then creationist theories seem to be as scientific as many other theories already regarded as such. Like other theories that already enjoy scientific status, creationist theories attempt to solve both internal and external conceptual problems.

Yet one might want to ask whether or not Laudan's account of the nature of science is accurate or complete. Perhaps true scientific theories do other things besides solve conceptual problems that creationist theories don't do. Perhaps there are other better (non-positivistic) accounts of scientific method and rationality that would not cast so favorable a light on the possibility of a scientific theory of creation as does Laudan's. Moreland, of course, does not address such possibilities, as he carefully limits the scope of his paper to analyzing the implications of Laudan's work for the scientific status of creationism.

Nevertheless, those hoping to find a post-positivist philosophy of science to assist them in defining creationist theories out of existence may well have to look long and hard. In my opinion, other recent accounts of scientific rationality offer little hope for a renewed program of demarcation. In fact, quite the reverse is the case. Paul Thagard and Peter Lipton's work on the use of inference to the best explanation has, for example, suggested an unexpected similarity between scientific reasoning and the reasoning used in religious, historical, philosophical and ordinary discourse. Knowledge simply does not appear to be divided into the neat isolated compartments assumed by many demarcationists and complementarians. Empirical data may have metaphysical implications, while unobservable (even metaphysical) entities may serve to explain observable data or their origins.

Furthermore, as Elliot Sober and I have argued (independently), both the argument for intelligent design and the Darwinian argument for naturalistic descent with modification can be understood as attempts to make retrodictive inferences to the best explanation. This logical similarity between the two theories raises an important question: What makes either intelligent design or naturalistic descent inherently more or less scientific than the other when both theories depend upon similar forms of inference and methods of empirical evaluation?

Recent work on the methods of the historical sciences has suggested that the methodological and logical similarity between creationist and evolutionary theories runs quite deep. Both programs of research attempt to answer characteristically historical questions; both may have metaphysical implications or overtones; both employ characteristically historical forms of inference, explanation and testing; and, finally, both are subject to similar epistemological limitations. Thus, intelligent design and naturalistic descent appear to be what I term "methodologically equivalent"—that is, both prove equally scientific or equally unscientific provided the same criteria are used to adjudicate their scientific status and provided metaphysically neutral criteria are selected to make such assessments. (Design and descent may not, of course, be equivalent in their ability to explain particular empirical data, but that is a separate issue).

Clearly, I can not demonstrate exhaustively the above arguments in the space available in this review essay. Nevertheless, I mention my work and other developments in the philosophy of science as a warning against the cavalier use of philosophy to make categorical judgments against ideas we would prefer not to engage. I readily understand the distaste that many ASA friends feel for the sloppy handling of data that has unfortunately characterized the work of some of our more fundamentalist brothers and sisters. I myself reject young-earth flood geology. Nevertheless, I do so, not because the supposition of a young earth is intrinsically any less scientific than its opposite. Rather, I do so because the empirical data seems to me to support strongly the supposition of great antiquity.

I wonder, however, if in our zeal to disassociate ourselves from our young-earth colleagues, we have too readily embraced an unnecessarily secularized vision of science that serves neither truth nor Christ. With the demarcation arguments that have underwritten methodological naturalism now exposed as yet another vestige of an untenable enlightenment view of rationality, the time seems ripe for ASA members especially to take the lead in probing the extent to which strict materialistic assumptions must govern all branches of science. Nowhere is such a re-evaluation more necessary, for the sake of science itself, than in the area of origins research.

Consider for the moment a radical possibility. It might well be the case that God acted in special way (i.e., in a way that differs from his ordinary supervenience over nature that we describe with laws). He might well have acted discretely or specially to create, for example, the universe, the first life, the major taxa and/or human consciousness. It might also be the case that unambiguous traces of His special creative activity remain by which such activity could be convincingly inferred. And, then again, it may not be so. Yet I see no reason, and philosophy of science currently provides no reason, to limit the inferences that scientists are allowed to draw in their attempts to explain the origin of the evidence they observe. Artificial limitations upon theory construction only leave open the possibility that the best explanations many not have been considered. Scientia so encumbered is unworthy of the name.

Indeed, the most important reason to question methodological naturalism is not that it undermines the claims of religion; the best reason to question the doctrine is that it limits the prerogatives of science. Methodological naturalism is not so much irreligious, as irrational. Hyperbole aside, strict naturalism functions (at least within origins research) to close off legitimate lines of inquiry and avenues of potential explanation. It, therefore, limits the ability of scientists to pursue the truth wherever, and perhaps, to Whomever, it might lead. Moreland correctly challenges ASA members to reassess this truncated and anti-intellectual view of science.