One of the most common arguments raised by the rationalists against religious faiths is that many claims made by the religious tend to be unfalsifiable. Many times we hear or read arguments in the line of:

Under what conditions would you accept that there is no God / that there was no miracle here / that no supernatural beings showed their presence in this case? The problem of your thinking is that it is fail-safe. You have presupposed the answer to this problem and are willing to do whatever it takes to make this answer sound plausible.

I would like to ask a similar question here, though in the other direction. Under what conditions would a naturalist accept the incorrectness of their position? What conditions would be necessary to accept that the supernatural exists AND has influenced or does influence the natural?

In their advocacy of their worldview, atheists sometimes say:

The view that atheists believe that there is no God is incorrect. Atheism is the null hypothesis, not a claim. Atheists only claim that there is insufficient evidence to accept that there is God.

Thus it would seem that rationalists do not presuppose the lack of the supernatural.

However, quoting the Wikipedia article about Historicity and origin of the resurrection of Jesus:

Post-Enlightenment historians work with methodological naturalism, and therefore reject miracles as objective historical facts.

Furthermore, from another Wikipedia article, Naturalism (philosophy)§Methodological naturalism:

Steven Schafersman states that methodological naturalism is "the adoption or assumption of philosophical naturalism within the scientific method with or without fully accepting or believing it ... science is not metaphysical and does not depend on the ultimate truth of any metaphysics for its success, but methodological naturalism must be adopted as a strategy or working hypothesis for science to succeed. We may therefore be agnostic about the ultimate truth of naturalism, but must nevertheless adopt it and investigate nature as if nature is all that there is."


Schafersman asserts that "while science as a process only requires methodological naturalism, I think that the assumption of methodological naturalism by scientists and others logically and morally entails ontological naturalism", and "I maintain that the practice or adoption of methodological naturalism entails a logical and moral belief in ontological naturalism, so they are not logically decoupled."

This is a significantly different position. Here we presuppose the lack of supernatural influences on the natural and do everything it takes to interpret the reality in such a way that we may maintain the correctness of naturalism. In this view naturalism seems just as unfalsifiable as many religious claims. This brings the suspicion of naturalists' detachment from reality: regardless of what the truth may be and regardless of what is happening around them, they will keep believing in naturalism.

Superficially, it would seem to me that the correct, rational, open-minded position is to not presuppose the supernatural, but do not apriorically reject them either. Thus, whenever there is evidence for something strange happening, we should first seek natural explanations; but if we fail to explain the situation naturally, we admit the hypothesis that this even is/was supernatural, until natural explanations become available.

This view of mine has been challenged. For example:

I am concerned about the false dichotomy being offered by the question: The only options given are: either the child's recovery is completely understood by modern science, with many known examples OR the recovery was entirely due to a single appeal to God by his mother.

(source: Oddthinking's comment under my question Did a hypothermic, clinically dead boy revive in response to prayer defying “every expert, every case history, and every scientific prediction”?)

Furthermore, this view seems dangerously close to the (in)famous God of the gaps principle, rejected not only by rationalists and scientists, but by many theologians as well. Finally, it can be argued that the view I presented has not, historically, been useful in advancing knowledge.

But the only alternative I can see is to unfalsifiably, apriorically reject the supernatural in a most close-minded manner.

I must be missing something here.

  • "we presuppose the lack of supernatural influences on the natural" Exact; the central "metaphysical" postulate of scientific naturalism is that only "natural" exists and thus whatever "interacts" with natural facts is natural. Commented Feb 15, 2020 at 20:34
  • 1
    The key methodological approach to understanding natural facts is their "reproducibility"; we cannot reproduce miracles... Commented Feb 15, 2020 at 20:36
  • 2
    Does this answer your question? Is any aspect of the supernatural testable? What level of proof is possible for the supernatural?
    – Conifold
    Commented Feb 15, 2020 at 23:38
  • If one of the answers below is enough for you, please accept it and we can "close" the post. Commented Feb 17, 2020 at 13:52
  • I don't think the prior question fully answers this question.
    – Dcleve
    Commented Feb 18, 2020 at 1:27

5 Answers 5


There are many shades of Godless belief and atheists debate their own "theologies" every bit as much as religious believers do. Just because one atheist makes a claim it does not mean that all the others will accept it. And there is always the example of Buddhism as a Godless religion to be accommodated.

But really, as Dr. Susan Blackmore has argued (in relation to paranormal phenomena), descriptions such as supernatural, paranormal, Godless are negative ones. They are not defined by a testable hypothesis about what the cause of certain phenomena might be, they are defined only by the suggestion that certain causes can be ruled out. Refuting such a proposition is next to impossible.

She makes the case that the only way forward is to devise testable hypotheses about specific causes and then to test them experimentally. That is what science is. From this viewpoint, something supernatural affecting reality becomes a contradiction in terms.

In other words, to say that the supernatural can influence reality is to move outside of any scientific viewpoint.

  • Thank you for your answer. However, may I ask 2 clarification questions? (a) "they are defined only by the suggestion that certain causes can be ruled out. Refuting such a proposition is next to impossible" Why? "The only way forward is to devise testable hypotheses about specific causes and then to test them experimentally." I don't understand - you now propose to falsificate the hypothesis about a specific cause, and falsificating it means demonstrating that this is not the cause ie this cause has been ruled out. But previously you said this could not be demonstrated!
    – gaazkam
    Commented Feb 16, 2020 at 9:13
  • To refute a proposition that X is not a cause of Y, we have to demonstrate that X is indeed a cause of Y. Scientists demonstrate causal relationships routinely. If the hypothesis about a causal relationship is falsifiable - and if the phenomenon is natural, then it often will - then, having failed to refute it, we accept its correctness. We thus have refuted the negative proposition that the phenomenon was not natural by providing a plausible natural explanation. On the other hand, failing to produce such an explanation = failing to falsify the supernatural hypothesis.
    – gaazkam
    Commented Feb 16, 2020 at 9:18
  • In any case, could you please elaborate how can the proposition that X is not a natural phenomenon be self-refuting? Even if it is unfalsifiable, how can it be a "contradiction in terms"?
    – gaazkam
    Commented Feb 16, 2020 at 9:20
  • Demonstrating that a certain cause (say clairvoyance) can never result in the phenomenon observed requires that every single observation be unarguably attributed to some other cause. Since most observations remain moot as to their ultimate origin, proving that any particular cause never applies is next to impossible. The psi hypothesis requires that all but one possible cause be ruled out, which is harder still. Even if genuine cases were to exist, no sceptic would ever accept that all possible mundane causes had been ruled out. Commented Feb 16, 2020 at 16:54
  • I am not aware that self-refuting propositions are implied or entailed by anything I wrote, I am sorry if I somehow gave that impression. Next to impossible is not the same as logically impossible. Commented Feb 16, 2020 at 16:58

Three points that I hope go together to suggest an answer:

1) Can you define supernatural in a way that does not 'eat itself'? Are we not allowed to define everything as natural that we observe in nature? If God exists, he is in some sense part of nature, no? What makes something supernatural?

Given something that cannot be defined, I can definitely say I would never accept it as a cause. "I would know it when I saw it", or "My culture told me this exists" aren't definitions.

2) Supernaturalists are participating in the process of science in bad faith. If the 'supernatural' is real, whatever definition you can give for it, then eventually exploration would naturally discover it. It would not need to be injected by positing a category on the basis of cultural tradition. So why don't the religious just wait out the game, and let exploration of the universe figure out what God is? Because this is not about God, it is about social control.

I can watch history evolve and see that every time someone defines anything supernatural, they end up making false assertions. So I can as a matter of observation refuse to accept that entire category of explanations until all others are exhausted.

Throughout history, science adopts provisional positions about what constitutes evidence, based on what has failed previously all the time. Nobody is ever going to accept stellar influences as predictive of events, it was tried too earnestly and failed too often. We do not still have a lobby within the scientific community driven by astrologers.

But we do have large religious lobbies. For some reason, this particular provisional position, clearly supported by observation, offends people. As you point out, you don't even have to accept this observation as a philosophical fact or natural law in order to find its guidance productive.

3) I can also see that when a large enough group of people insist on something supernatural at the same time, they end up killing one another over it. So I can ethically choose not to further their violence.

To me, atheism is clearly lying about not being a religion. It is a religion, because it really arises by creating a causal excuse for one's moral positions. This is the way religions arise, and what keeps them alive. The moral system in atheism involves opposition to pointless authoritarianism and the waste of lives in thinly-veiled tribal rivalries that pretend to be about 'supernatural facts'.

Given that this is a religious position, I cannot expect others to accept it, and I will not proselytize for it. (Though obviously, given that I am writing this, I will explain myself, when asked.) I do consider folks who do so to be hypocrites. If your religious position is really a bias that covers an ethical assertion about freedom of thought, then forcing it to the front in all public decisions, above all other positions, and proposing it as a new public norm, the way folks like Dawkins do, is a grand hypocrisy.

  • Thank you for your answer. There is still something I don't understand... "So why don't the religious just wait out the game, and let exploration of the universe figure out what God is?" - However, is it not the case that by the definition God cannot be explored? On the other hand, God may have revealed something to us. If so, then simply rejecting this message may be stupid. This is, at least, what I understand religious traditions say. And while it may be debatable whether God has revealed anything, we still do not seem to be able to explore God.
    – gaazkam
    Commented Feb 16, 2020 at 9:24
  • If this is what you think religious traditions are telling us, then explain traditional European theology as a phenomenon. Catholicism counts as a religious tradition that has extensively endeavored to explore God. You may define God in a way Thomas Aquinas would not, but that is the whole problem, isn't it? Commented Feb 17, 2020 at 2:01
  • I am not saying Christians need to stop believing anything until this is all worked out. But they need to allow other people to do their business. If the Christians are right, those other people will be forced to come around. Why push the agenda, if you have any real faith? So this is not real faith, it is a drive to control the process. Commented Feb 17, 2020 at 2:09
  • Earlier science surely did not agree that this was the message. When Christianity and Islam dominated science (for at least 11 centuries, from the fall of Rome through the Reformation), they had their opportunity to make their point, they injected God everywhere, (even if it was the God of Aristotelian physics, they conflated it with their own). They constituted physical theories based on explorations of God. This did not improve their results -- they ended up with Alchemy. So we have turned away from this premise. It could come back. But not through insistence -- through relavance. Commented Feb 17, 2020 at 2:11

Let me start by noting that the concept of a 'null hypothesis' is strictly delimited. A null hypothesis is an analytical reference to the status quo: a statement that some particular effect under consideration is not occurring. There is no meaning to the concept except as a negation of a specific proactive claim. A null has to nullify: it is always a statement of what we expect to see if what we want to see does not happen.

With that in mind, it should be clear that neither the assertion that God exists nor the assertion that God does not exist is a null hypothesis. Both statements are general theoretical claims about the world, claims from which we might derive proactive hypotheses. Those derived hypotheses will each invoke their own null if and as they are operationalized to be tested. We do not use nulls to evaluate general theoretical claims against each other. Instead, we test hypotheses derived from each general claim, and evaluate which general claim best fits the observable world as we test it.

All general theoretical claims have the same intellectual standing until they are confronted by observable phenomena. Trying to dismiss one or another ab initio by puttering around with the concept of nulls is a category error; these are theories, not hypotheses, and different rules apply.

The problem with the term 'supernatural' as it's generally used in these discussions is that both sides rest on a form of absolutism that makes discussion pragmatically impossible. Supernatural is mainly used in the Kantian sense of transcendent — i.e., "not being realizable in experience" — but both sides project that inability as a universal. Atheists and anti-theists will claim that only those things we can 'objectively' experience exist, and anything we objectively experience is ipso facto not supernatural; theists reject the claim that objective experience defines existence, and allow subjective and indirect forms of experience, or even mere inference, to demonstrate the existence of supernatural essences.

If we set aside that absolutism, what becomes clear is that we can never know empirically whether there is anything we cannot know empirically. The whole question of the supernatural — which the modern polemic has reduced to a question of 'experiencing the inexperiencable' — is what Wittgenstein dismissed as vacuous philosophy: a mistake in language that we treat as though it were meaningful and relevant. The politics of the situation is understandable, because there is an ongoing struggle between religious and secular worldviews for social and political power. But if we set aside that turf war, there isn't much philosophical value to asking the question.


Under what conditions would we have to accept that the supernatural has influenced the reality?

From a systemic point of view, any interaction between an object (in this case, the supernatural) and a subject (reality, in this case), produces a change on both of them (even if almost insignificant, it is a change). Otherwise, there is no interaction. Interaction literally means mutual action and reaction between two systems. In consequence:

Condition 0. The concept of the supernatural has reached an objective definition and its existence can be proven.

Condition 1. It can be verified that the supernatural is able to interact with objects of reality.

From a strictly logical perspective, if condition 1 is fulfilled, condition 0 would be fulfilled (e.g. Descartes follows such path: if he's able to interact with himself by thinking[condition 1], then, he exist [condition 0]). Condition 0 is added because in contrast with Descartes' concept of I, the term the supernatural, and the definition of the existence of such a transcendental entity, are quite arbitrary.

Remember that all objects of reality exist only in each one's minds (as Hume, Berkeley suggest). The scent of a rose does not exist without a person able to interpret it as such. Existence of the supernatural is a troubling element for the concept of existence.


Your question is a good one, and there is an answer -- that pursued by methodological naturalism.

One of the problems you have encountered in this question is the multiple equivocations that have infected the usage of naturalism and supernatural.

For naturalism, there has been significant effort spent to clarify different meanings, and among philosophers, these different meanings are generally given different terms. Each of the below are sometimes called "naturalism", but it is better to use the more explicit terms below:

  • Materialism -- there is only matter in the universe
  • Physicalism -- there may be other things than matter (ideas and experiences) in the universe, but they are all entirely dependent on matter, and matter is causally closed
  • Metaphysical naturalism -- Physicalism is true, and the only way to gain knowledge is the scienctific method, and reasoning (this position is often referred to as scientism)
  • Methodological naturalism -- the scientific method is a refinement of informal empiricism, and formal reasoning is a refinement of informal reasoning -- and these two tools are the most reliable method of investigating our world. What that world consists of is open, based on those investigations.

I have also seen naturalism used as a rejection of agent-effects -- which bizarrely would mean rejection of psychology or the possibility of conscious agency (and therefore the study of consciousness) from science.

In general, Wikipedia is a poor reference for philosophical questions. For a better discussion of naturalism, see the SEP https://plato.stanford.edu/entries/naturalism/ which absolutely distinguishes metaphysical from methodological naturalism.

Supernatural suffers from the same equivocation, but has had much less effort spent to disentangle its very different meanings. The term can be used to describe:

  1. all non-material claims (including consciousness, ideas, will, math, logic, or platonic forms)
  2. all claims relative to Gods or spirits
  3. or it can be an assertion that significant parts of nature are not resolvable by the methods of methodological naturalism. (Note, not all supernaturalists are non-physicalists -- the "Mysterian" movement is mostly physicalists who reject absolute confidence in methodologic naturalism https://www.edge.org/response-detail/27017)

The way the equivocation of "supernatural" plays into this problem, is that often the advocates of metaphysical naturalism equivocate a "supernatural" view about Gods, spirits, math, ideas, with a "supernatural" claim that none of these subjects are investigatable in any way. And then any "uninvestigatable/unfalsifiable" beliefs are then dismissed outright as meaningless.

This reasoning process is fallacious, depending on an equivocation fallacy, plus a version of the Logical Positivists "Verification Principle" despite the VP having been shown to be self-refuting (the VP itself canto be verified, and is therefore meaningless per its own terms). It is also easily refuted, by showing that all sorts of God and spiritual claims are testable and refutable (see all examples of proofs of God, citations of contradictions in the Bible and Koran, the Problem of Evil, and other non-optimization problems, etc).

So a possible answer is -- embrace METHODOLOGICAL naturalism, and yes, physicalism can be held to with religious dogmatism.

  • Thank you for your answer. I must say I'm surprised by you claim that showing that all sorts of God and spiritual claims are testable and refutable (see all examples of proofs of God, citations of contradictions in the Bible and Koran, the Problem of Evil. I recall having been told that Kant already proved that one cannot prove nor disprove the existence of God? As such 'proofs' like Anselm's proof or Aquinas' Five Ways are invalid if viewed as proofs and likewise the problem of evil fails to disprove anything.
    – gaazkam
    Commented Feb 18, 2020 at 11:07
  • However, this does not have to mean that we can say nothing about the existence of God. While for example Aquinas' Five Ways are not proofs, they may be valid as a heuristic approach. Is this reasoning in line of what you said above?
    – gaazkam
    Commented Feb 18, 2020 at 11:13
  • Yes, the existence of flawed proofs, and the ability to critique them is in itself a demonstration of the ability o apply reasoning to the question, refuting the claim that it is beyond reason. And yes the Problems of Evil, of Flawed Revelations, and of Design Imperfections are not proofs, they are empirical test cases -- and one can always kluge a theory, deny the evidence, or otherwise try to evade a conclusion in empiricism. Once again, they serve to disprove the claim that empiricism cannot apply,
    – Dcleve
    Commented Feb 18, 2020 at 11:36
  • Also, I followed to your question on Skeptics. The comment you quoted was correct -- the argument claiming the recovery was a miracle assumed falsely that medicine is a deterministic and predictable process, not a spectrum of outcomes with a probability curve to it. A single surprising recovery, within the spectrum of medical outcomes, while under medical care, is not evidence for the efficacy of prayer. Prayer efficacy has to be shown statistically, vs a control.
    – Dcleve
    Commented Feb 18, 2020 at 12:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .