Here is the NAS statement on the compatibility of science and religion. Sorry, but I'm not buying it. I understand the concerns behind it, and I'm not suggesting that it is realistic to change it. It could be a lot worse.
I just don't think it's true:
Compatibility of Science and Religion
Science is not the only way of knowing and understanding. But science is a way of knowing that differs from other ways in its dependence on empirical evidence and testable explanations. Because biological evolution accounts for events that are also central concerns of religion — including the origins of biological diversity and especially the origins of humans — evolution has been a contentious idea within society since it was first articulated by Charles Darwin and Alfred Russel Wallace in 1858.
Acceptance of the evidence for evolution can be compatible with religious faith. Today, many religious denominations accept that biological evolution has produced the diversity of living things over billions of years of Earth’s history. Many have issued statements observing that evolution and the tenets of their faiths are compatible. Scientists and theologians have written eloquently about their awe and wonder at the history of the universe and of life on this planet, explaining that they see no conflict between their faith in God and the evidence for evolution. Religious denominations that do not accept the occurrence of evolution tend to be those that believe in strictly literal interpretations of religious texts.
Science and religion are based on different aspects of human experience. In science, explanations must be based on evidence drawn from examining the natural world. Scientifically based observations or experiments that conflict with an explanation eventually must lead to modification or even abandonment of that explanation. Religious faith, in contrast, does not depend only on empirical evidence, is not necessarily modified in the face of conflicting evidence, and typically involves supernatural forces or entities. Because they are not a part of nature, supernatural entities cannot be investigated by science. In this sense, science and religion are separate and address aspects of human understanding in different ways. Attempts to pit science and religion against each other create controversy where none needs to exist.
I do agree that the methods that are distinctive to science are not the only ones used for obtaining knowledge and understanding. For example, I can gain knowledge by investigating historical records in a way that may be quite rigorous but not distinctively scientific. However, it's not true that science is unique in its reliance on empirical evidence. We seek knowledge based on empirical evidence all the time, without doing anything distinctively scientific. It's true, though, that our methods of investigation start to look more distinctively scientific when they use replicable experiments. Hypothetico-deductive reasoning is not unique to science, but science makes distinctive use of it, combined with replicable experiments to test hypotheses. But there are other features as well, such as the use of instruments that expand the capacities of the human senses, the use of mathematical models, and doubtless others.
Science is continuous with other kinds of inquiry that may use some of these techniques. It can also draw on knowledge gained from inquiries that don't especially utilise these distinctive techniques, such as ordinary sensory observation, testimony, historical records, our learned ability to interpret human behaviour and language, and so on. Thus, there's no clear boundary between the sciences and the humanities, such as history.
It is not so much that there is more than one way of knowing. Rather, there are different techniques for investigating different aspects or parts of reality. Not all aspects lend themselves to investigation through distinctively scientific techniques, and some lend themselves to investigation through other techniques (examining historical records, etc.). Still, we expect that knowledge and understanding obtained through different techniques will be consistent. Where lines of evidence obtained from different techniques show a convergence, we can be confident that we're getting at the truth.
Science has been especially useful for examining phenomena that are very small, very large, very distant in space or time, or indiscernible by our senses. By proceeding step by step, testing conjectures (by looking for other phenomena that conjectured entities, forces, etc., predict), using precise mathematics, adopting potentially consilient lines of inquiry, and so on, science is often able to reach robust, very well-evidenced conclusions about things that no one has ever seen with their unaugmented senses (such as the moons of Jupiter, long-extinct animals, and microscopic entities). In other cases, science's conclusions are more tentative, or problems are left with no plausibly-acceptable conclusions in sight. Because the evidence for scientific claims is typically circumstantial, and highly theory-laden, sometimes with very long chains of inference, those claims are always considered provisional, but some are so robust that they are unlikely ever to be overturned. E.g. it's inconceivable that the geocentric picture of our solar system will ever be restored. There is too much evidence from too many lines of inquiry. (And we should not forget that many ordinary, not-especially-scientific claims are also quite heavily theory-laden.)
Still, the picture of the world emerging from science is always incomplete.
On this view, there are methods of rational investigation that are not distinctively scientific, but there are no "other ways of knowing" that somehow compete with rational investigation itself. There is plenty of room for science (and, indeed, the humanities) to falsify various religious claims or render them implausible. For example, science has decisively falsified the religious claim that the Earth is 6,000 to 10,000 years old. Historical investigation has made it quite implausible that there was ever an Egyptian captivity of the Jews or a culture hero such as Moses. Some religionists may not give up such claims in the face of empirical evidence, but that doesn't mean that religion is fundamentally compatible with science; it means that some religious positions are (clearly) incompatible with it.
As for the inability of science to investigate the supernatural, this is either trivially (and unhelpfully) true or false. Unfortunately, the NAS statement doesn't nail down what is involved here beyond saying that religious faith typically involves "supernatural forces or entities". It is trivially and unhelpfully true that science cannot investigate such forces or entities if "supernatural" is defined to mean "that which science cannot investigate" (or in some other way that amounts to the same thing).
But it is false if it means that science is, in principle, unable to investigate claims about such paradigmatically "supernatural" things as ancestor spirits, water nymphs, fire demons, magic dragons, or astrological influences. If these things exist and behave in fairly regular ways - like lions, elephants, kangaroos, crocodiles, and the flow of water - then science can investigate them. Of course, if they did exist we might come to think of them as part of "nature", but that's just the point. There is no clear and meaningful line between "natural" and "supernatural", such that science cannot investigate beyond that line. It is simply that certain kinds of things, notably disembodied intelligences, don't actually seem to exist; in any event, hypotheses involving these things have had a lousy track record over centuries. It is usually good practice for scientists to avoid those kinds of hypotheses if they can (this is the grain of truth in "methodological naturalism").
Nonetheless, there is no reason, in principle, why science cannot investigate claims about, say, ancestor spirits as long as the spirits in question are alleged to behave in ways that are reasonably regular and affect things that can be detected by our senses (possibly via scientific instruments).
The kinds of hypotheses that are most difficult to test by distinctively scientific means - apart from those involving interpretation of texts, artworks, and so on; or alleging one-off, small-scale historical events - are not necessarily "supernatural" ones. They are any that: (a) are framed so vaguely that it is difficult to know what to look for out there in the world; or (b) are systematically protected by ad hoc modifications.
However, we don't have to falsify a conjecture together with all possible auxilliaries before we reject it; a conjecture that is preserved only by ad hoc modifications should also be rejected. Thus, if someone seeks to preserve the conjecture that the Earth is only 6,000 years old by postulating that an all-powerful being created the Earth in an "aged" form, we have good reason to reject that hypothesis. The auxilliary claim is too ad hoc.
In short, the relationship between science and religion/the supernatural is more complex than the NAS acknowledges, and it doesn't make religion as immune to scientific investigation as the NAS implies.
Once again, though, it doesn't follow that the statement should be changed at this late stage, or that it is not doing any social good. Nothing in this posts suggests that the policy implications of my discussion are clear. My best suggestion is that the current NAS statement be reviewed at some future time when the organisation is conducting a larger review of its communications strategy. At that stage, points such as I've made above can be taken into account.