This essay is written in response to those who have questioned me about ‘naturalism’, what it means to me, and why I am a naturalist. It is the product of several informal exchanges.
First off, I would note that ‘naturalist’ is a title that I adopt with some hesitancy. It is amazing how much one can do to damage the position of an opponent by simply labeling it as an -ism.
Naturalism may be defined as the belief that things in reality are consistent, coherent, and predictable in nature. “Consistent” here means that they act in accordance with certain fundamental regularities. “Coherent” here means that they can be understood with theories. “Predictable” here means that one can predict with some probability that they will do a particular thing. Closely associated with naturalism is uniformitarianism, which says that the same natural laws govern the universe in different places and different times.
Now let me elaborate some more, just on what naturalism means to me. After this explanation, I will focus on objections and arguments.
Concepts do not exist in a vacuum. By ostensive definitions, naming things that confront us in reality, we may form basic concepts. We do not determine the natures of a thing in reality, but rather we discover its nature – what it is, what it does.
For example, to understand what a ‘tree’ is, we observe several examples of trees and abstract or draw out the similarities common to all known trees. The implicit assumption here is that, when we cross another thing fitting our concept of ‘tree’, it will be the same sort of thing as the known examples. This is basically induction, starting from several known examples and extrapolating to the next one.
The way in which we discover the nature of something is by induction from experience. We frame a hypothesis about it, and we find that it is confirmed by experience in several cases. As the consistent verification increases without bound, so does the degree of certainty with which we may say that we discovered a bit of the nature of something–what it is, what it does–and this principle may be (at least tentatively) called a part of ‘the laws of nature’ (descriptive of the nature of things, not prescriptive).
Logical possibility is to hypothetical fictions what nomological possibility is to reality. In fiction, we get to define the nature of things. All things are (logically) possible, but not all things are (logically) compossible, as Leibniz said. A logical impossibility is an internal self-contradiction in a hypothetical fiction. But in reality we don’t get to say what things are and how they work; reality confronts us with the nature of things. A nomological impossibility is something that contradicts the way things work in reality, that is, something that is contrary to the natural order of things.
If we can know the nature of anything at all, if we can formulate general principles that describe the way things work, by consequence things contrary to our understanding of nature must be regarded as impossible or at least highly improbable. We might as well not have stated a principle at all, if we are to regard violations of this principle as being just as likely as fulfillments.
Of course, we do not have infallible knowledge of the nature of things without recourse to experience, so we cannot claim absolute certainty for our formulations of the laws of nature. Our theories should undergo revision if faced with adverse evidence interpreted with general principles of nature that are borne out by experience to a greater extent. But empty suggestions without evidence that a general principle with consistent confirmation just might fail in a particular instance needn’t be regarded seriously.
The main question here is: Can we reasonably extrapolate from consistent confirmation of a principle to its application in an unspecified case–with what certainty and with what exceptions (if any)?
My answer to this question is quite simple. The degree of certainty with which an inductive conclusion is accepted correlates with the extent to which the theory is consistently confirmed by empirical means. An inductive argument from experience is defeated only by a stronger inductive argument from experience. This reduces to the epistemological truism that one rationally ought to take all available evidence into account.
These basic principles may be called naturalistic inquiry.
Question: Could you elaborate on what kind of evidence would make you accept an event contrary to a supposed law of nature?
Basically, the same process of reasoning that led me to accept something as a law of nature would lead me to revise it, perhaps replacing it with a theory that is more complete and consistent with experience.
As I said, our theories should undergo revision if faced with adverse evidence interpreted with general principles of nature that are borne out by experience to a greater extent.
All evidence is interpreted with general principles about the nature of things. A set of fingerprints on an object matching a person’s own can verify that the person held the object only if we use the general principle that people have unique fingerprints. Radiometric decay of carbon can estimate the age of an object only if we use the general principle that C-14, on average, has a half-life of about 6000 years. One must have both the specific fact and the general principle to draw a conclusion, because a bare factoid does not tell us how to make a deduction.
Now, the most reliable class of evidence must be our own sober experience from our own eyes and ears. All other evidence ultimately comes to us through these means. So if someone sees something happen, he or she has a reason to believe the experience. However, there may be a misunderstanding in interpreting the experience (as magicians are happy to show us). And if you have one trippy experience in contrast to a vast coherent and consistent body of sober experience, you may rightfully regard the former as misperception. But you must not cling to a theory that directly contradicts everything you see, for it is by experience that theories are established and by experience that they are taken away.
The next most reliable class of evidence is the repeatable experiment. You cannot experience everything for yourself, and the experience of other people is objectively worth just as much as your own, so not all facts can or should be directly observed by you. But since we do not want to base our theories on tainted data, high standards of testimony must be set for use in the hard sciences. Typically the procedure of an experiment along with the results are published, and thus others in the scientific community can check them, and they can even make a name for themselves by discrediting others. Since most scientific laws are based on this kind of intersubjective experience, they can be overturned by the same evidence.
But since there are sciences such as paleontology in which the data points cannot be repeatedly verified, and since the scientific method has only been around for the last few hundred years, not all data must be intersubjectively verified by scientists. The documented eyewitness testimony of consistent impartial observers is the next most reliable class of evidence. Notice the key words: Documented, eyewitness, consistent, impartial observers. The “documented” part means that the person left a record that everyone knows to be authentic. The “eyewitness” part may be assumed, unless there is evidence to the contrary, if the person claimed to be an eyewitness and was actually in a position to see the event. “Consistent” here merely means that the different observers do not directly contradict each other or themselves on this data point. The “impartial” part may be difficult to determine at times, but it becomes quite relevant if there is only a single observer to the event. If there is an established motive to lie or self-delude, we should want independent confirmation from other observers, preferably with different motivations.
It is on the basis of these three classes of empirical evidence that we may form laws of nature: (1) personal experience, (2) repeatable experiments, and (3) documented eyewitness testimony of consistent, impartial observers. The same kind of evidence can establish an event contrary to a supposed law of nature. That is, if the principle behind these classes of evidence is stronger than the theory itself. A single data point from class three may be overlooked in favor of an elegant theory with overwhelming support from classes one and two. However, in most cases, all pieces of empirical data from these three classes should be taken into account in forming a theory.
This is not to say that we cannot use lower standards of evidence in a field such as, for example, ancient history. It is just that we must be cautious in using a lower quality of evidence in making a data base for scientific theories.
Note that the present relics of the past cannot be interpreted as historical evidence at all, unless we presume that the same fundamental regularities obtained then as still obtain today. Ancient history is largely based on the hard sciences, such as the reliability of carbon dating for example.
And low quality evidence (such as hearsay) should be disregarded if contradicted by theories of the hard sciences with overwhelming confirmation. Because the event is already highly improbable given the veracity of a consistently confirmed empirical theory based on eyewitnesses, a mere rumor or legend is just about as good as no evidence at all.
Some people sum up the above by simply saying, “extraordinary claims require extraordinary evidence.”
Question: Even if I accept your methodology of ‘naturalistic inquiry’ described above, how does this prove that miracles are impossible?
First we should try to define ‘miracle’, and then try to determine in what sense they are ‘possible’.
What have people traditionally identified as miracles? Virgin births, resurrections, exorcisms, sudden healings, prophecy, etc. In one sense (contranatural), a miracle is contrary to the natural order of things.
Again, if we can know the nature of anything at all, if we can formulate general principles that describe the way things work, by consequence things contrary to our understanding of nature must be regarded as impossible or at least highly improbable. We might as well not have stated a principle at all, if we are to regard violations of this principle as being just as probable as fulfillments.
Of course, what is considered miraculous can come to be accepted as fact if there is “extraordinary evidence” as defined above. But if there is little more than hearsay, it can rightfully be rejected because, without evidence, there is no reason to entertain the idea that a general principle with consistent confirmation just might fail in a particular instance.
If we consider highly improbable events to be impossible, then we may well call miracles ‘impossible’. If we define ‘possible’ in terms of logical possibility, anything is possible, but then logical possibility isn’t very interesting.
Question: Isn’t the ‘supernatural’ more a matter of entities above Nature than events contrary to natural order?
Even on this view, a miracle is the intervention of this supernatural being, in a way that would be considered contrary to natural order. So there are all the same problems in proving miracles in this sense.
Now, one might want to say that entities above Nature are part of the way things work in reality. The claim that natural laws have built-in accommodations for supernatural beings is just a strange claim, one to be resolved based on whether observations support it. In this sense, this is an empirical question to be resolved within naturalistic inquiry. Even if entities existed that were not part of the physical space-time universe, that would disprove materialism or physicalism, not naturalism.
Physicalism would be empirically defeated if an immaterial being were established as a necessary theoretical entity. It would not be enough to show that all known explanations are insufficient. This would not help establish a supernatural explanation; if all known explanations fail, that shows we don’t know how it happened, not that it involved a being above Nature. In addition to showing other theories false, a supernatural explanation must be provided, which explains the characteristics and mechanism of this theoretical immaterial entity, before being tentatively accepted as the most probable and plausible theory. To the best of my knowledge, this has not been done.
Question: Isn’t the evidence for [the Resurrection, the Holy Qur’an, reincarnation, etc.] quite extraordinary?
The specific topic of whether there is extraordinary evidence for this or that extraordinary event will not be detailed here. This essay is intended to provide a general framework for discussing miracles. If you want to work within this framework to prove a miracle, I wish you the best of luck. But my own research into Christianity, Islam, and Hinduism shows that it’s hard enough to provide any evidence for these claims, let alone extraordinary evidence.
Question: What about supernatural context? Doesn’t that raise the background probability of some miracles?
There are two ways to interpret ‘supernatural context’. In one interpretation, a motive has to be shown for a supernatural being to intervene. In another interpretation, a motive only has to be imagined. On both views, miracles without a supernatural context are rejected as improbable, but miracles with one have a fair chance.
Here are a few criticisms of the first interpretation:
First, before a motive can be established, it must be established that there is a supernatural being that can intervene in Nature. Otherwise, the belief in an omnipotent Creator is just an unnecessary assumption.
Second, if this supernatural being is the omnipotent Creator of the universe, it is not clear why He would need to intervene in His world to achieve His ends, as if the original act of Creation were insufficient.
Third, it is difficult to see how the will of a supernatural being could be discerned unless it revealed itself to us. Even if it can be argued in principle that something is an accurate revelation of a supernatural being’s dealings with man, in fact such arguments fail. There are established scientific and historical errors and contradictions in the many scriptures of the world, and there is no reason to assume that the writers were more accurate when it came to discerning the will of beings above Nature.
Fourth, in practice, this methodology entails an acceptance of a particular religious tradition (the one assumed to have the revelation of a supernatural being’s will) and rejection of all contrary claims as ‘out of context’. Of course, if one assumes the inerrancy of the Bible, one will accept the biblical miracles and reject all arguments to the contrary on the grounds that they contradict the Bible. This is not an argument.
Here are a couple more criticisms that apply to the second interpretation:
Fifth, a supernatural thing needn’t be an intelligent agent that has motives for its doings. Yet this methodology seems to give special treatment to supernatural entities with imagined reasons, which is just another of its unnecessary and inelegant assumptions.
Sixth, a supernatural context can be imagined for any alleged miracle. Some say they disbelieve miracles that are ‘giggly’, but there is no reason to assume that a supernatural being (such as Loki) isn’t a prankster. Some say they disbelieve miracles that are ‘evil’, but there is no reason to assume that a supernatural being (such as a demon) wants all the best for humanity. Again, this methodology in practice implies just the acceptance of a particular religious tradition without reason.
Here are a few more criticisms of both views:
Seventh, things in reality do not seem to have the built-in caveat “unless a supernatural being wills otherwise.” That is, the nature of things can be described in a way that cuts gods out of the equation. Such a description is internally consistent and veridical with experience. One can make a valid inductive argument that, of all the known causes of things, nothing above Nature has been demonstrated to be one. Although not complete, we are always edging closer to an accurate description of Nature, one that doesn’t include gods.
Finally, it is not clear that the methodological principles of a ‘supernatural context’ are necessary to a complete and consistent theory of knowledge (epistemology). My answer to the question of demonstrating events contrary to supposed laws of nature assumes little more than the validity of induction from experience. This response includes numerous exceptions and caveats, akin to Ptolemiac epicircles.
Question: Since theories are always being replaced, isn’t it arrogant to assume that today’s theories are any better? The ancients also thought a lot of things were impossible that technology has made possible.
First, today’s theories are closer to the truth for the very reason that they are always open to revision and correction. If we never changed, we would be stuck with notions such as that the function of the brain is to cool the blood or that the sun and other planets revolve around Earth. Looking for evidence that conflicts with the current theories is the essence of the scientific method. Today’s theories have withstood many man-years of revision and correction by scientists, which is what makes them credible.
Second, since the advent of the scientific method with men such as Newton, our theories about physical laws have not been so much replaced as revised. Einstein’s general relativity is not a radical departure from Newton, but it is more complete and accurate with regards to high-speed particles. Students are still taught Newtonian physics because it is useful in engineering and approximately accurate as far as it goes. Likewise, if a quantum theory of gravity emerges, it will apply to distances below the Planck scale but will not really disprove Einstein.
Third, when it comes to miracles in religious traditions, this is just irrelevant. These alleged events are not reported as happening with technological contrivances that are unknown to scientists today. Indeed, the ancients were right in thinking that men did not have the ability to fly, because they actually couldn’t with the means available then. I do not deny that there may be technological ‘miracles’ in the future, because they wouldn’t be miracles at all, and besides the question is whether such events happened in the past. This is like saying that a baby might have been able to lift a ton because a muscleman can.
Question: If God is possible, aren’t miracles possible?
Who cares whether miracles are ‘possible’ in some undefined sense? Alleged miracles are highly improbable, and thus are to be disbelieved. We may call this high improbability an ‘impossibility’.
But there are various uses of the word ‘possible’. Here are three:
1. Logically possible. If a statement is not self-contradictory, it is logically possible that it’s true.
2. Nomologically possible. If a statement is not contrary to the natural order of things, it is nomologically possible that it’s true.
3. Epistemologically possible. If the opposite of a statement is not known, it is epistemologically possible that it’s true.
Now, in the first case, I have not argued that a miracle is self-contradictory, as I can imagine a miracle such as a virgin birth (although not a square circle). This is the kind of possibility that doesn’t matter; in this sense, anything’s possible.
In the second case, most agree that miracle claims are nomologically impossible. This is an important sense of possibility, particularly for engineers; if it is nomologically impossible for a structure to stand, no matter how good it might look on paper, it won’t work.
In the third case, a statement is known to be false only if it can be shown to be significantly improbable. The question above doesn’t penetrate the arguments against the probability of miracles at all–which, if successful, make miracles epistemologically impossible.
What about the epistemological possibility of God? To draw an analogy, we may not know that one John P. Dooberfunkle didn’t see Elvis alive (yet), but we do know that Elvis is not alive. Likewise, we may know that miracles don’t happen, and then only on reflection realize that a miracle-working God doesn’t exist (since this is implied). And, of course, we may reach this conclusion from independent means as well.
Question: Aren’t you assuming determinism?
A causal determinist can be a naturalist. But a naturalist needn’t be a determinist, and I am not one.
On one interpretation of quantam mechanics, the laws are ultimately probabilistic in nature. This allows for events that, while not absolutely impossible, are nonetheless highly improbable given these statistical laws. The philosopher of science Mary Hesse explains:
“There is no question that most events regarded as significantly ‘miraculous’ in religious contexts would, if they violate Newtonian laws, also be excessively improbable on well-established quantum laws, and therefore would be regarded as violations of these also. Thus, if we consider only the currently accepted theories of physics, the credibility of such miracles is no greater than in Newtonian theory.”
Christian apologist William Lane Craig agrees: “It would be crazy, for example, for a person accused of murder, who was known to have been alone in the room with the victim at the time the murder occurred, to offer as his defense the claim that another man quantum tunneled into the room spontaneously, shot the victim dead, and then, before he could be apprehended, spontaneously quantum tunneled back out again.” (Reasonable Faith, p. 142)
It should be noted that some laws of nature, such as general relativity, are not stated in probabilistic terms. But in light of quantum mechanics, an event may be regarded as nomologically impossible if it is highly improbable given the statistical laws of nature. We can still make reasonably accurate predictions with statistical laws.
Question: How do you justify your methodology? Isn’t naturalism just a presupposition? Why should I be a naturalist?
Our knowledge of reality indeed depends on the reliability of sense experience and the validity of induction. If you do not grant some reliability to our senses, you are left with solipsism. But solipsism is psychologically impossible to believe, and in fact rejected by those who mean to accept it (with the curious sight of a man telling other men that they do not exist). If you do not further grant some validity to induction, you could believe no more than what you directly observe. Induction is necessary to organize experience into general concepts and principles of nature. Based on such principles, we can infer the veracity of events that are not personally experienced. If nothing else, induction is a practical necessity; we could not get up in the morning and expect the sun to rise, and we could not expect the food we buy at the supermarket to be relatively safe, without the principle of induction. So the reliability of experience and the validity of induction are clearly necessary to a complete and consistent theory of knowledge (epistemology). Without them, we wouldn’t even be having this conversation right now.
The basics of naturalistic inquiry can be deduced simply from the reliability of experience and the validity of induction. (If this is not clear, read the original explanation and the first question again.)
So whether you know it or not, you already adopt a naturalistic epistemology or methodology of sorts. When one becomes self-conscious of these principles of rational inquiry, they manifest themselves as ‘naturalism’.
Let me offer a couple examples of professed supernaturalists who implicitly assume a form of naturalism.
My chemistry teacher remarked one Friday: “Your theoretical must always be more than the actual [yield], unless God steps in; otherwise…” Note that Mrs. Walker never takes the god-factor into account when doing reactions.
Craig argues that “thermodynamics shows the universe began to exist” (Reasonable Faith, p. 116). Along with the fact that there is usable energy, Craig assumes the First and Second Laws of Thermodynamics, which state that energy is neither created nor destroyed and that entropy (unusable energy) in a closed system increases. Craig assumes that these laws hold in different times and places and that no beings outside Nature have intervened to add usable energy. Yet it is difficult to see how Craig explains his acceptance of such naturalistic principles (even saying on p. 114 that this is an “atheistic view”).
It is no wonder that even religious people usually apply naturalistic inquiry to most areas of life. The alternative to naturalism is profound ignorance of empirical reality. Supernaturalism offers no grounds for criticism, as supernatualists profess belief in a Walt Disney wonderland where pumpkins can turn into coaches, oranges into spaceships, and women into pillars of salt. What could be unreasonable to such a person?
Of course, since they have to function in the world, supernaturalists typically do not extend the uncritical eye to claims outside of their religion. So they will reject claims about Elvis being alive, for example. But this only shows the fundamental inconsistency of their presuppositions.
Note also that, through naturalistic inquiry, a remarkable amount of consensus has been reached among scientists of many different backgrounds in just the last couple centuries. There is one physics, one biology, one scientific community; and despite many differences of opinion, there is large agreement as to the demonstrable facts among scientists. This consensus wasn’t forced by persecuting heretics, but rather by encouraging new ideas to be tested for validity. This pays witness to the objectivity of the scientific method. On the other hand, claimed means to knowledge such as mysticism fail miserably in this regard. There are thousands of religious sects in the world, and more cropping up every day, and there is even a large amount of disagreement within any one religion. This is difficult to explain if mysticism is a reliable means of acquiring knowledge.
So to answer the question: Naturalistic inquiry is based simply upon the reliability of sense experience and the validity of induction. These naturalistic principles are part of any complete and consistent theory of knowledge. Even if one has religious belief in the supernatural, methodological naturalism is assumed, out of practical necessity if nothing else.