On The Integrity of Science

If Kansas insists on introducing supernaturalism into science, then science equally insists on showing supernaturalism the door.

Key propositions

  • Science generates the natural-supernatural distinction, it doesn’t presume it.
  • The supernatural is just that which can’t play a role in scientific explanations.
  • Kansas is powerless to change scientific practice to incorporate supernaturalism.

The debate about teaching intelligent design (ID) in public school science classes raises the fundamental issue of what constitutes science. Are there agreed-upon features of the scientific method which can help decide the question of whether ID is science or not? And who decides the nature of science? Is it up for vote at school board and curriculum committee meetings, or might science have its own agenda, independent of the democratic process?

Such questions are very much on the front burner thanks to recent (2005) revisions to science standards adopted by the Kansas Board of Education, including the redefinition of science itself to permit consideration of supernatural causes. Is it even logically possible for science to traffic in the supernatural? It seems obvious that science is restricted to natural phenomena, both in its purview and in its explanations.

There are of course esteemed governing bodies such as the National Academy of Sciences, the American Association for the Advancement of Science and the National Science Teachers Association, that explicitly define science and good scientific practice. There’s also the National Center for Science Education led by the intrepid Eugenie Scott, who makes the rounds defending science against ID, most recently and successfully in the Kitzmiller trial.

These groups indeed promulgate a definition of science that bars consideration of the supernatural, and have taken Kansas to task for crossing that line. For instance, the National Academy of Science says in Teaching About Evolution and the Nature of Science, p. 58:

Science is a way of knowing about the natural world. It is limited to explaining the natural world through natural causes. Science can say nothing about the supernatural. Whether God exists or not is a question about which science is neutral.

And the National Science Teachers Association says (also in Teaching About Evolution..., p. 124):

Because science is limited to explaining the natural world by means of natural processes, it cannot use supernatural causation in its explanations. Similarly, science is precluded from making statements about supernatural forces because these are outside its provenance. Science has increased our knowledge because of this insistence on the search for natural causes.

But these definitions raise a further question: from whence comes our conception of the natural? What I’d suggest is that the very idea of the natural has been generated by science itself. The importance of this point for the science-religion debate is that science does not, as proponents of ID and religion often argue, make any apriori metaphysical or ontological claims about the world. In particular, it doesn’t assume a pre-existing natural world separate from the supernatural that gets identified on extra-scientific or philosophical grounds. Rather, science is in the business of providing transparent, unified, evidence-based and intersubjectively testable explanations of phenomena. In so doing, science identifies phenomena that end up playing well-confirmed roles in such explanations, and these end up constituting what we call nature. So science doesn’t presume naturalism; it isn’t, as the Kansas Board of Education says in its revisions, “driven by a naturalistic preconception.” Since science is metaphysically neutral it doesn’t need to be “balanced” by requiring the consideration of supernatural causes.

Importantly, explanations involving intentional, purposeful agents whose properties and characteristics cannot be specified are disallowed in science, but not because such agents are apriori supernatural. Rather, it’s simply that invoking an unspecified agent is an explanatory dodge, since particular intentions and actions can simply be posited to close a purported explanatory gap (e.g., to create the bacterial flagellum or human eye in one fell swoop). This is why when ID proponents claim that they aren’t necessarily invoking god, but perhaps a “natural” designer of some sort, they still aren’t doing science. There’s nothing wrong with invoking intelligent, intentional agents in explanations; after all, as such agents ourselves, we obviously play a huge role in explanations of culture and technology. But like everything else in an acceptable scientific theory, agents have to be sufficiently specified and thus at least potential candidates for empirical verification and explanation themselves. We human beings are pretty well specified by mainstream science, but ID notably fails to describe the designer.

Returning to the natural/supernatural distinction, we could ask: at what point were lightning bolts definitively shown to be natural phenomena, as opposed to the production of an angry god (Zeus, Jove, Jehovah, Yaweh)? Roughly, when science explained the various aspects of lightning (light, heat, duration, spatial distribution) in terms of other empirically verified, and thus natural, phenomena, none of which were agents. This rendered the agent hypothesis unnecessary since the explanatory gap had been closed. No longer was any intention needed to explain lightning bolts. This didn’t categorically disprove the existence of god, of course, but it helped define him as supernatural to the extent that in this domain, at any rate, he became explanatorily superfluous.[1] The same goes for evolution: no agency need be posited, or if it is, it must be well-specified on pain of explanatory inadequacy.

The upshot is that Kansas, although politically free (fortunately) to define science any way it likes, is powerless to change scientific practice to suit what might be a majority of citizens in favor of considering supernatural causation. Supernatural causation is ruled out in science because the supernatural is just that which can’t play a role in transparently explaining the world to ourselves, which is what science does. This is a function of scientific practice, not an apriori rejection of the agent hypothesis. Specify the agent (e.g., god), test it against the evidence, and if it turns out to play a well-qualified role in our best theories, then the agent hypothesis will be accepted by science. But of course at that point the agent will be included in the natural world – it will have been naturalized – not exactly what Kansas has in mind for god.

As is commonly noted, the nature of science isn’t decided by vote or fiat. It’s rather the historically refined practice of winnowing out the best – the most economical, predictive, unifying and (perhaps) aesthetically pleasing – explanations of the world. This practice doesn’t assume naturalism or any other contestable worldview. However, it rules out the consideration (at least beyond the first-pass consideration necessary to disqualify them) of entities, causes and hypotheses that have no empirically specifiable content. Such entities, causes and hypotheses, should they involve intentional agency, just are the supernatural, since anything that is adequately specified and evidenced gets included in the natural world. If Kansas insists on bringing supernaturalism into science, then science equally insists, by virtue of its methods, not any tendentious ontological commitment, on showing supernaturalism the door.

TWC, January 2006


1. The concept of the supernatural is of course traditionally tied to the idea of intentional agents (beings with purposes, intentions and plans), but the idea of the supernatural as a categorically separate realm emerged in parallel with the idea of physical nature as compositional and inanimate. Before then, the spirit world and physical world were fully intertwined, not distinct ontological categories. Minds and intentional agency, until very recently, have been thought outside nature in some respect precisely because they weren't susceptible to reductive explanation; they were (and still are to a great extent) considered non-decomposable essences and uncaused causers (see Joshua Greene and Jonathan Cohen, p. 1782 on this). Of course human agents are now virtually naturalized, but god resists naturalization by virtue of being insufficiently specified by his adherents, or if specified, empirically unevidenced.