As Phil Johnson and my teachers pointed out, some people use
methodological naturalism as an over-arching (philosophical) principle
in an attempt to "explain away" God from all parts of creation,
history, and personal experience. Today, this spirit is pervasive in
most educational institutions. This attitude corresponds to what George
Murphy called "MN as a primary principle." (I should add that, while
this attitude is common in academia, I personally encountered it only
rarely in the physical sciences, especially amongst the younger
But as my teachers also pointed out, it is also possible to pursue
knowledge about the natural operations of creation, about human
psychology and society, and about history, in ways which are God-
It is possible to search for as-yet-unknown natural mechanisms in
cellular physiology, disease pathology, brain functioning, and, yes,
even biological history, in ways which are God-glorifying.
This corresponds to what George Murphy called "MN as a secondary
[limited, provisional] working hypothesis."
There is an important distinction between using MN as a primary
principle in all cases *in spite of* the evidence, and using MN as a
secondary working hypothesis for some particular problem *because of*
the preliminary evidence. I am disappointed that I don't see this
distinction made more often in writings about Naturalism. Acknowledging
this distinction won't weaken the arguments against the pervasive
MN-as-a-primary-principle. On the contrary, I think emphasizing this
distinction strengthens the case against pervasive Naturalism
If good knowledge about creation can be extrapolated by some to argue
for anti-theistic ideas about God and human beings, if a good technique
for learning about the natural mechanisms God created can be
extrapolated by some into a way of "explaining away" God from all parts
of creation, then Christian scholars face a difficult task. They must
separate "the good from the bad." They must carefully decide how to
present these distinctions -- to Christians and non-Christians alike --
in ways which minimize misunderstanding. And (here's the really
difficult one) they must decide what to do when, as Christians, they
disagree with each other. (One factor which makes disagreements so
disagreeable is the oh-so-human temptation to suspect that anyone who
disagrees with you must be doing so because of impure motives. More on
It's not difficult to see how pursuing new knowledge about natural
mechanisms in cellular physiology can be God-glorifying. But what
about those cases where the distinction between God-glorifying and God-
excluding pursuit of knowledge is less clear to some? Is such a
distinction still valid when discussing biological history, or miracles?
I believe the distinction is all the more important in these cases.
Paul Nelson wrote:
PN> Suppose a scientist doubted the standard theory of stellar evolution.
PN> Perhaps he decides that, for instance, the so-called "collapse problem"
PN> (Larson 1978) is unlikely to be solved by any natural mechanism.
PN> So he proposes that, on the grounds of the available evidence, stars
PN> are intelligently-designed, or created, objects.
PN> What do you suppose his chances are for publication at any of the
PN> major astrophysics or astronomy journals?
PN> I'd say his chances are nil. Not because of the evidence, however.
PN> In a curious fashion all empirical considerations will prove to
PN> be irrelevant. Rather, he will run afoul of MN. In the case
PN> of theories of *origins*, MN is all-important, and all-governing.
Yes, such a scientist might run afoul of the pervasive spirit of MN-as-
primary-principle. But it's worth dissecting this case a bit, because
more than one thing may be at work.
If this scientist pointed out real empirical problems with currently
accepted models, those calculations would certainly be published. But
what about the scientist's additional proposal of "design"?
There are many scientific problems for which natural mechanisms are
*believed* to be adequate but, because of the difficulty of the
calculations, this cannot be demonstrated within strong empirical
constraints. The data and the models are inconclusive. In these
cases, the question of whether natural mechanisms are adequate or
INadequate can be debated, but won't be resolved until further work is
done. If the scientist in your example had done the easier task of
showing that existing models do not match some data, but had not gone
so far as to demonstrate the more difficult claim that known natural
mechanisms are all INadequate, then claims about Design would probably
be rejected as premature. The scientist might believe that natural
mechanisms are inadequate, but until she can marshal enough empirical
arguments to make that case empirically, such claims won't make it
into the major research journals.
But suppose a scientist succeeds in the more difficult task of
marshalling strong empirical arguments that all known natural
mechanisms are inadequate. The scientist could argue for one or more
of the following: (1) A supernatural event occurred. (2) Super-human
technology brought about the event. (3) An unknown natural mechanism
is responsible for the event. (4) A very unlikely (natural) event
occurred. (5) A very unlikely (natural) event occurred, but there are
many different causally disconnected universes. Empirical science
can't really distinguish amongst those five. Now in this case, I agree
with you, major research journals will embrace the "unknown natural
mechanism" hypothesis FAR more happily than the argument for design or
supernatural intervention. This could be partly because the "unknown
natural mechanism" has had some success in the past. But I also think
that the pervasive belief in MN-as-primary-principle would, in many
cases, be a strong factor in accepting "unknown natural mechanism" and
rejecting the design hypothesis. And that, I agree, is not good. MN-
as-secondary-working-hypothesis would say that if the design hypothesis
can suggest additional empirical tests (and I believe that Design could
be made to do so), it *should* appear in the research journals. By
contrast, MN-as-primary-principle would exclude design no matter how
empirically useful it might become.
Phillip Johnson mentioned the study of miracles:
PJ> But granted the unique superiority of methodological
PJ> naturalism (MN) as a principle of scientific investigation (the premise of
PJ> my original comment), how can you be confident that you have such a
PJ> revelation? MN is not just applicable to evolution. It is also the
PJ> scientific way to investigate purported divine revelations and miracles.
If one could investigate a miracle that was accomplished through
natural mechanisms -- a "natural" event with special timing and
significance -- then MN-as-secondary-working-principle would conclude
that that is what happened. MN-as-primary principle, by
contrast, would conclude that there was no divine action at all.
If one could investigate a miracle which contradicted all known natural
mechanisms, MN-as-primary-principle would never give up the search for
a naturalistic explanation. MN-as-secondary-hypothesis, once it had
learned what it could, could be happily set aside.
Regard evolution, Phillip Johnson wrote:
PJ> My interest is in MN as held and applied in the culture of evolutionary
PJ> science. There any distinction between a "secondary working principle" and
PJ> a "fundamental presupposition about reality" collapses, because of the
PJ> determination to explain all of reality on naturalistic terms.
The key phrase is "...all of reality." MN-as-primary-principle seeks
to explain all of reality in naturalistic terms. MN-as-secondary-
working-hypothesis seeks only to explain the observed developments of
geological and biological history in ways consistent with the regular
operation of natural mechanisms over time. The two might agree about
particular empirical models and empirical predictions, but that is
much, much less than agreeing about "all of reality." Unlike MN-as-
primary principle, MN-as-secondary-hypothesis takes its cues from
particular theological and empirical data. MN-as-secondary-hypothesis
knows that its reason-to-exist in one area of study might not apply to
other areas of study, so it does not extend itself to other areas of
study without re-examination. MN-as-secondary-hypothesis acknowledges
that God can work effectively within natural mechanisms, through chance
and necessity, to produce desired outcomes. MN-as-secondary-hypothesis
acknowledges that there is much more to reality than the empirically
And if some members of the current culture of evolutionary science
can't distinguish between the two, then shouldn't we keep on pounding
the distinction until it sinks in? If some members of the church can't
make that distinction when considering scientific claims, then the church
needs more teachers like the teachers I had.
When pursuing new knowledge about the natural operations of creation,
when seeking for as-yet-unknown natural mechanisms, when attempting to
construct explanations for events in terms of known natural mechanisms
in an effort to learn more about those events and mechanism, there is a
very significant difference between MN-as-primary-principle and MN-as-
secondary-working-hypothesis. This is all the more true when studying
those "controversial areas."
But in these controversial areas, Christians can disagree. They can
disagree about where the theological data is pointing. They can
disagree about where the empirical data is pointing. Then what?
Phillip Johnson writes:
PJ> To put it in my words: Follow the evidence wherever it leads, even if it
PJ> leads to conclusions such as these: (1) the Darwinian mechanism has no
PJ> significant creative power; (2) the hypothesized Precambrian ancestors
PJ> never existed; (3) the documents hypothesis and the quest for the
PJ> historical (i.e. naturalistic) Jesus are mere naturalistic speculations;
PJ> and (4) the reality of irreducible complexity and the nature of genetic
PJ> information points to the necessity for attributing biological creation to
PJ> an intelligent cause.
and Phil goes on to say where he believes the empirical data is
PJ> The creative power of natural selection is not supported by evidence (other
PJ> than at the trivial peppered-moth level), and the whole Darwinian scenario
PJ> is contrary to the fossil and experimental evidence....
PJ> Explanations employing only natural causes may in other cases be true -- or
PJ> at least probable on the basis of an unbiased evaluation of the evidence.
PJ> Where that is the case, I have no objection to them. That is emphatically
PJ> not the case with Darwinian evolution;
But what if a Christian brother or sister disagrees about where the
data is leading? A Christian brother or sister may look at the same
data, plus the data of a dozen textbooks and the data of a few
hundred research journal articles and seminars, and conclude something
different: that gene duplication and modification, to name just one
mechanism, can produce novel physiological pathways; that there is
growing evidence of soft-bodied multicellular organisms and
phylogenetic divergence going back well before the Precambrian; that
irreducible complexity can self-organize by modification-and-feedback
mechanisms in systems with far simpler rules of operation than
biochemistry; that an unbiased evaluation of the evidence implies that
the fossil and experimental evidence is consistent with the hypothesis
of common ancestry and natural descent by modification. What then?
There's no simple answer about how to deal with a disagreement,
obviously. There's going to have to be some serious, nit-picky delving
into the evidence to make progress. One thing that won't produce
progress, however, is over-generalizations about the motives of one's
Yes, we're back to motives. Broadly stated judgments about other
people's motives *may* be true in some cases, though they are counter-
productive to dialog even then. More importantly, broadly stated
judgments about other people's motives are most certainly, in very many
Last year there was some discussion about what science can say, if
anything, about supernatural events. Here's one posting on that topic.
ABSTRACT: Science cannot prove that some past event was "supernatural;"
however, it can in principle establish that no known natural mechanisms
could account for that event. In that limited sense, science can address
When science investigates a puzzling event (either an "origins" event long
ago, or a more recent event such as an unexpected healing), science cannot
determine whether or not that event was supernatural. What *can* science
do? It can try to determine, to the best of its abilities,
--what the conditions were before the event,
--what the conditions were after the event, and
--what effect known natural mechanisms could have had during the event.
(I use the word "event," but it could also refer to a series of
events spread over time.)
As scientists study the initial conditions, final conditions, and
known natural mechanisms, they could reach three possible conclusions:
1) Sound empirical models predict that known natural mechanisms can
account for the event. (*1*) (Let's call these "natural events.")
2) We do not have sufficiently thorough empirical models, but
we believe that known natural mechanisms can account for the event,
and future improvements in empirical knowledge, elegant models, and
computing power will eventually allow us to prove this.
(For this letter, let's call these "non-empirical events.")
3) No known natural mechanisms could account for this event.
There are empirically sound reasons for ruling out all known
(For this article, let's call these "non-mechanistic events.")
There are two special categories which also deserve mention. I will
deal with them separately at the end.
SC1) Rare "natural events" whose timing and location have special
significance (especially religious significance) to a person or
group of people can be called "coincidental events."
SC2) Some objects and events indicate intelligent activity. (E.g.
stone tools found near hominid fossils.) The terms "natural"
and "non-mechanistic" become problematic, so we will call
these "intelligently designed events" and deal with them
(Please don't get hung up on the labels. I'm open to suggested changes.)
For any given event, there may be some disagreements in the scientific
community as to whether it is natural, non-empirical, or non-mechanistic;
however, it is possible for a great majority of scientists to agree. Most
would agree that the formation of the solar system falls into the first
category, "natural." Most would agree that galactic formation,
earthquakes, and zygotic development fall into the second category,
"non-empirical." A majority of scientists believe that abiogenesis and
macroevolution fall into the second category, but a minority
believe they belong in the third, "non-mechanistic." Most agree that the
Big Bang is in the third category.
What do scientists do with non-mechanistic events? Individual scientists
could reach (at least) five different conclusions:
A) A supernatural event occurred. (An event caused by intelligent beings
of an entirely different "reality" than our universe.)
B) Super-human technology brought about the event. (An event caused by
intelligent beings who are contained in and limited by our universe,
but with superior technology.)
C) An unknown natural mechanism is responsible for the event.
D) A very unlikely (natural) event occurred.
E) A very unlikely (natural) event occurred, but there are many
different causally disconnected universes.
Although these five are very different in principle, they play virtually
identical roles in *empirical* studies. Empirical science cannot
distinguish between these possibilities. Historical, philosophical, and
religious arguments are the decisive factors in each scientist's conclusion.
Note, however, that empirical science DOES play a vital role in
debating whether an event is "non-empirical" or "non-mechanistic."
Philosophical and religious arguments can also properly play some role in
this debate. This is the realm where scientific data, scientific
intuitions, and philosophical/religious expectations meet in the same
arena. For example, strongly Materialistic scientists will work hard to
push all events into the "natural" and "non-empirical" categories. This
effort might lead them to uncover new natural mechanisms sooner than
scientists who don't share their materialistic philosophy. Alternatively,
scientists with strong religious or philosophical reasons for believing
that certain events are supernatural can marshal scientific arguments to
show that those events are "non-mechanistic" rather than merely
"non-empirical." This effort might lead them to uncover flaws in proposed
naturalistic scenarios sooner than scientists who don't share their
religious beliefs. (*2*)
Ultimately, the development of new empirical models plays a decisive role
in determining whether a "non-empirical" event is "natural" (if empirical
models predict the event) or "non-mechanistic" (if empirical models rule
out known natural mechanisms). In the absence of such empirical models,
philosophical and religious arguments can play an important role in
persuasion, and to some extent, in formulating testable hypotheses.
Moreover, events which are deemed "natural" or "non-mechanistic" today
could change their status with the discovery of new natural mechanisms or
better empirical models.
Sometimes, "natural events" occur at a time and place and in such a way as
to have special significance (especially religious significance) to a
person or group of people; the argument can therefore be made that these
events have some non-mechanistic (e.g. supernatural) component. Empirical
science cannot resolve this argument; the most it can do is attempt
to determine the relative probability (infrequency) of the event, possibly
taking into account known "initial conditions." Additional historical,
philosophical, and religious arguments are needed to consider whether the
event was "purely natural," or had some non-mechanistic component. (e.g.
Was the event's timing and location predicted beforehand? How soundly
does the event fit within an established theological framework?)
Intelligently designed objects/events are an interesting special case.
For example, a paleontologist might determine that the breakage pattern on
the edge a particular stone could not have happened via "natural
mechanisms" (at least, not with any significant probability). If hominid
bones are found in the same area, however, the paleontologist might
reasonably conclude that the stone was crafted as a tool. The intelligent
activity of hominids becomes a special kind of "natural mechanism." A
similar argument is made in the search for extraterrestrial intelligence.
If a sufficiently complex repeating radio signal is discovered, the case
can be made that no natural mechanism could produce it except for the
"natural mechanism" of intelligent activity.
In the debate over biological origins, many people have pointed to the
analogous features between biological life and intelligently designed
objects or processes, thereby arguing that biological life was
intelligently design and perhaps, to some extent, assembled
"non-mechanistically." This is an example of a philosophical argument
used to persuade and, to some extent, as a starting point for marshaling
scientific arguments and to formulate testable hypotheses. (The extent to
which the argument is *convincing* is, obviously, a point of debate.)
One final point should be made. People can disagree about whether an
event is natural, non-empirical, or non-mechanistic. People can also
disagree over whether or not a particular event actually happened. Let
us use one particular issue. If everyone agreed that Jesus'
resurrection actually happened, then (almost) everyone would agree that
the event was non-mechanistic. (Science _qua_ science cannot determine
whether or not the disputed non-mechanistic event actually happened;
historical and philosophical arguments must be used in that decision.)
Since the event itself is widely questioned, the debate must often shift
to other events which are generally agreed upon (e.g. the written records
and the subsequent behavior of Jesus' followers); scientific intuition, as
well as historical and philosophical arguments, are then brought forward
to debate whether these agreed-upon events are non-mechanistic, or merely
(*1*) It is worth mentioning again that non-deistic theism asserts that
"natural events" are just as much dependent upon God's activity as
"non-mechanistic" events. In addition, even if natural mechanisms *could*
account for an event, that does not necessarily mean that is how the event
(*2*) Both of these biases could be pushed to the extreme, to the
detriment of science. One could imagine a scientific community so
obsessed with finding naturalistic explanations for non-empirical and
non-mechanistic events that it wastes vast resources on unproductive
pursuits which yield no secondary benefits. One could also imagine a
scientific community so complacent about supernatural explanations (or for
that matter, super-human or many-worlds explanations) that it makes
virtually no effort to search for new natural mechanisms for puzzling
events. Fortunately, the present-day scientific community does not seem
to fit either extreme. And of course, scientists from every philosophical
persuasion spend *most* of their time trying to push events from
the "non-empirical" category into the "natural" category.