Although the Charles Simonyi Professor of the Public Understanding of Science at Oxford University does not provide a single sentence (in the first four chapters of The God Delusion) defining or explaining what he means by a “scientific hypothesis” or by a “scientific question”, there are a few comments and phrases here and there that give a hint at what he had in mind. It is difficult, however, to assemble a complete definition or explanation out of the few crumbs that Dawkins leaves behind.

A scientific question is a question to which the idea of evidence is applicable.

Dawkins is opposed to PAP (Permanent Agnosticism in Principle) style agnosticism about the existence of God. So, his description of PAP, gives a clue as to why Dawkins thinks the question “Does God exist?” is a scientific question:

The PAP style of agnosticism is appropriate for questions that can never be answered, no matter how much evidence we gather, because the very idea of evidence is not applicable. The question exists on a different dimension, beyond the zones where evidence can reach. (TGD, p.47, emphasis added)

This suggests that one reason Dawkins views the God question as a scientific question is that he believes the idea of evidence is applicable to this question. In other words, evidence can be used to settle the question, or to provide us with a rational basis for preferring one possible answer over another.

Another comment by Dawkins about PAP, makes a similar point in different words:

PAP agnostics aver that we cannot say anything, one way or the other, on the question of whether or not God exists. (TGD, p.51, emphasis added)

The concept of relevance is implied here: we cannot say anything that is relevant to answering the question. If we can say something that is relevant to settling the question, then we can say something that provides a reason for accepting one possible answer over another possible answer. Providing reasons in support of a possible answer is closely related to the idea that evidence is applicable to answering the question.

This might work as a necessary condition of the concept “scientific question”, but it will not work as a sufficient condition. Evidence is also used to resolve historical and legal questions, such as: Was Jesus of Nazareth crucified by the Romans in the 1st Century? Did OJ Simpson murder Nicole Brown Simpson? Science might make a contribution towards answering some historical and legal questions, but historical and legal reasoning and methods must ultimately decide such issues, taking whatever relevant scientific information that is available into account.

Also, although “evidence” is not used to settle mathematical questions (mathematical proofs are required for that), it is used to support mathematical hypotheses. Swinburne gives the example of Goldbach’s conjecture that every even number is the sum of two prime numbers (Coherence of Theism, Rev. ed., p.45). It was rational to believe this to be true (apart from a mathematical proof) on the basis of the evidence that it held true for every one of the many even numbers that has been examined. Thus, evidence is applicable to legal questions, historical questions, and mathematical questions, as well as to scientific questions.

A scientific question is a question that is answerable in principle, if not in practice.

Again, consider how Dawkins describes PAP:

The PAP style of agnosticism is appropriate for questions that can never be answered… (TGD, p.47)

Dawkins quotes from Huxley, an advocate of PAP style agnosticism:

“… I… had a pretty strong conviction that the problem was insoluble”. (TGD, p.49)

Here is another comment from Dawkins about PAP:

The question [about whether God exists], for PAP agnostics, is in principle unanswerable… (TGD, p.51, emphasis added)

So, for Dawkins, a scientific question is one that is answerable in principle, if not in practice. Again, this will do as a necessary condition for the concept of a “scientific question” but it does not work as a sufficient condition. Any question that is worth thinking about should be answerable in principle, otherwise thinking about that question would be a waste of time. Historical questions, legal questions, and mathematical questions are answerable in principle too, not just scientific questions.

Science is not the only field in which questions are sometimes answered, and the unanswered questions in these other fields are usually unanswered because of practical obstacles (such as that the evidence currently available is insufficient to support a firm conclusion) rather than because the questions are in principle unanswerable.

Furthermore, it is not clear how we are to determine whether a question is in principle answerable or unanswerable. Until some clear test or criteria are spelled out for making this distinction, so that we can put questions into one bucket or the other, this necessary condition for being a “scientific question” is not very useful or helpful.