Last modified on Sunday, October 11, 1998, by Malcolm R. Forster

Background: Creationists are happy with either of two claims: 1) Creationism is not a science, and neither is evolutionary theory (which should not be taught in schools), or 2) evolutionary theory is a science but so is creationism (which should be taught in schools along side evolutionary theory). They would be happier with 1) but will settle for 2). In recent years, anti-creationists have concentrated on blocking arguments to the second conclusion.

1. Science looks for patterns in nature, order, and natural regularities (laws).
2. Science is explanatory, and the use of natural regularities is necessary for this purpose.
3. Science makes predictions, and the use of natural regularities is necessary for this purpose.
4. Science looks for testability. This has two aspects. (a) Confirmation, or positive support for the theory. (b) a theory must be open to possible falsification. E.g., if a planet were discovered going in squares, then the laws would have been shown to be incorrect.
5. Science is tentative. A scientist must be prepared to reject his theory.
6. Science should strive for simplicity and unification.
7. Scientists should be intellectually honest.

Criticisms and Comments:

1. It is not clear to me why natural regularity is necessary for explanation. Nor is it clear to me that explanation is necessary for science, although if science is often explanatory.
2. In 4, Ruse should not claim that theories are open to logical falsification. He need only insist that the theories models are open to falsification. Theories are and should be rejected, but that is covered 5.
3. There is no mention of the distinction between prediction and accommodation, which I think is important.
4. The emphasis on natural laws plays an important role in Ruse's argument because he quotes two different creationists, who emphasize that the process used by God in the act of creation are now not operating in the natural universe. However, I do not see why science should care about the difference between natural and unnatural laws. Show me an unnatural law and I will make good scientific predictions and explanations.

1. Creationism does make testable empirical assertions; e.g., that the earth is 6,000 to 20,000 year old, and most of the features of the earth's surface are diluvial (arising for the great flood). They are testable, and they have failed those tests (c.f. astrology.)
2. Creationists have changed their views.
3. To say that science is a matter of natural law is rather fuzzy. Darwinism was accepted by many scientists well before the laws of genetics were known. Continental drift was accepted before the mechanism was understood. Smoking is accepted as a cause of cancer, even though the mechanism is still not fully understood.
4. Ruse’s standards of testability and revisability are exceedingly weak. They could be satisfied by the declaration "I will abandon creationism if we find a living specimen between man and the apes." Exceedingly unlikely, but strong enough for falsifiability.
5. The key issue, according to Laudan, is whether evolutionary theory is better supported by the evidence than creationism. Debating the scientific status of the theories is a red herring.

Argument 1: Scientific theories make predictions (= observational consequences deduced from the theory). If the predictions prove to be false, then the theory is false. That is, scientific theories are falsifiable (the Popperian demarcation criterion). Creationist theory is not falsifiable. Therefore creationist theory is not science.

Objection: Predictions are not deduced from scientific theories alone. Therefore, if the prediction prove false, the theory is not falsified. Thus, a key premise in the argument is false.

Argument 2: Scientific theories together with auxiliary assumptions form models. Predictions are deducible from models. Scientific models are falsifiable. Creationist models are not falsifiable. Therefore creationism is not a science.

Objection: (Kitcher 1982, "Believing Where We Cannot Prove" (reading 72), p. 66) Imagine that we want to expose some self-styled spiritual leader as a fraud. We point out that the teacher's central doctrine "Quietness is wholeness in the center of stillness" is unfalsifiable. Call this doctrine D. But when it is coupled with other statements, it produces observational consequences. For instance, let O be any observational statement. Then D combined with "If D then O" has the observational consequence O. So, there are models of the "theory" (M = {D, if D then O}) that are falsifiable. In other words, this criterion lets in too much, and it would be easy for creationists to use it to argue that creationism is a science.

Argument 3: (Kitcher's) In genuine science, auxiliary hypotheses are independently tested, scientific theories tend to be unified, and scientific theories suggest new lines of investigation and new models (fecundity). Creationism does not display this cluster of features. Therefore, creationism is not a science.

Objection: While I agree that these features mark a difference between genuine science and creationism, it is not clear why these features of science are relevant to the purpose of science as a source of knowledge. This is not really an objection, but it does suggest that the argument is incomplete.

Argument 4: (Mine) Science succeeds at more than accommodating experience. Its models succeed at anticipating facts not used in the construction of the models. While creationism can easily accommodate the facts, it is not very successful in anticipating new facts in the required sense (there are some small exceptions to this, so it is a matter of degree).

Note 1: This gets around the objection to argument 2 because a model of the form {D, if D then O} will not predict a new observation O'. One can merely add a new auxiliary assumption "if D then O'" in order to accommodate the new fact.

Note 2: The independent testability of auxiliary assumptions (in argument 3) arises when the same auxiliary assumptions enter into other models. E.g., the auxiliary assumption that Madagascar separated from Africa at the certain point in geological history is also used in geological models, and gains independent support from them. This is also an example of unification (though not of theory unification).

Note 3: Theory unification aids in the anticipation of novel facts because it allows for the extension of a model constructed in one domain of explanation to another.

Note 4: My criterion is very close to Lakatos’s criterion (a research programme is scientific if and only if it is theoretically progressive) since ‘theoretically progressive’ comes down to ‘making novel predictions’.

Clarification: It is important that anticipation, or prediction, of novel facts does not have to be prediction of the future. It is unfair to count only a prediction of the future course of evolution as the only thing that will count as prediction in evolutionary theory. It should also be allowed that predictions do not have to be strictly deductive consequences. It is enough that they fit the model well.

Remark 1: Argument 4 captures, I think, all the virtues of argument 3, except that it is clearer that prediction is central to the purpose of science in its pursuit of knowledge.

Remark 2: In order to successfully apply Argument 4, we need to argue that Darwin’s theory of evolution has made successful predictions. Here I would appeal to the example of protein homologies as a prime example.