I find myself nodding in agreement with the wisdom of the Kansas Board of Education. Not very much agreement, to be sure; the recent move to introduce official skepticism about evolution into its new public school science standards is just bad. Bad, bad Kansas.
But, amidst the bemoaning of this setback for Enlightenment values, we all had a little fun with the school board’s attempt to change the definition of science, as Risa has already pointed out. (See also John Rennie at the new Scientific American blog.) Seems that they have decided to open the door to explanations other than the purely natural — obviously, so that they can include religious (“supernatural”) explanations within a science curriculum.
But only after reading Dennis Overbye’s story in yesterday’s New York Times did I really understand what they had done. Here’s the new definition of “science”:
The changes in the official state definition are subtle and lawyerly, and involve mainly the removal of two words: “natural explanations.” But they are a red flag to scientists, who say the changes obliterate the distinction between the natural and the supernatural that goes back to Galileo and the foundations of science.
The old definition reads in part, “Science is the human activity of seeking natural explanations for what we observe in the world around us.” The new one calls science “a systematic method of continuing investigation that uses observation, hypothesis testing, measurement, experimentation, logical argument and theory building to lead to more adequate explanations of natural phenomena.”
At the risk of alienating all my friends — the school board is right. Science isn’t about finding “natural” explanations vs. “supernatural” ones; it’s about finding correct explanations, without any presupposition about what form they may take. The distinguishing feature of science isn’t in the explanations themselves, it’s in the process by which we find them. Namely, we toss out hypotheses, compare them to data, and look for the hypotheses which account for the largest number of phenomena in the simplest possible way. Simplicity here is in the sense of “algorithmic compressibility” — the number of bits, if you like, required to specify the mechanism that purports to do all this explaining.
What the Kansas school board has tried to do is to open the door for unbiased consideration of natural and supernatural explanations by a common standard — that of scientific investigation. This is just what I’ve been arguing for all along. Scientists have to get off this kick that science and religion are completely distinct magisteria that have nothing to do with each other. Quite the contrary; religion (at least in its common Western forms) goes around making claims about how the world works, and it’s perfectly appropriate to judge such claims by the same standards that we judge any other suggested hypotheses about nature.
The thing is, if we judge popular religious vs. naturalist explanations for how the universe works by a common scientific standard, naturalism wins. Without breaking a sweat, frankly; by the beginning of the second half, we have to send in the scrubs from the bench, at the risk of being accused of running up the score. Intelligent Design, to take one obvious example, is laughably bad as a scientific hypothesis. It explains practically nothing (since it refuses to say anything about the nature of the designer, so we have no clue what such a designer would ever choose to design), while introducing a fantastic amount of new complexity in the form of an entirely distinct metaphysical category (the designer). I have no problem saying that ID is a “scientific hypothesis”; it’s just such a bad one that no sensible scientist would give it a moment’s thought if it weren’t for the massive public-relations campaign behind it.
Science doesn’t home in on naturalistic explanations by assumption; it chooses them because those are the best ones. That doesn’t mean that we have to “teach the controversy” in high schools; the number of grossly inadequate scientific theories is far larger than we could ever address in such a context. But it’s about time that we admitted that science is perfectly capable of judging supernatural claims — and finding them sadly wanting.
I think this sort of thing risks descending into a definitional morass. That science is the process of looking for natural explanations is almost tautalogical. The process by which scientists arrive at scientific statements isn’t all that important as long as at the endpoint we do get scientific statements. It’s just that scientific statements can be a bit difficult to define. Somewhere in there, however, has to be a notion of falsifiability, and that is where things like ID and last tuesdayism are going to fall short.
I would claim that science is about “natural explanations” by definition. Take any supernatural phenomenon, show that it is real using scientific methods and it will no longer be considered supernatural. Before we understood electromagnetism, would an observer really have considered magnetism natural and dowsing supernatural?
Give scientists just a shred of real, hard evidence that God exists and they will soon be busy designing experiments to once and for all find out how and why He created the universe.
We already know the answer to that, it’s 42.
Yikes, this is an epistemic morass. The process of science cannot be divorced from the type of explanations it achieves. Because scientists are limited to tests conducted in the natural world, they are limited to natural explanations. Moreover, science famously achieves explanations that are “correct” only to a given degree of precision. It might be more accurate to say science seeks “useful” or “repeatable” explanations.
“Correctness” and algorithmic compressibility do not distinguish between natural and supernatural explanations. Indeed, these criterion would probably lead you to choose a supernatural explanation each time (“God did it” = one bit, “God did it this way” = correct by fiat).
George
George– No, it’s exactly the opposite. “God did it” explains precisely nothing, since you would have to simply list every single thing that God would do. That’s the least simple possible explanation. Which is why scientists prefer explanations that look for repeated patterns in nature, rather than arbitrary supernatural interventions.
Because scientists are limited to tests conducted in the natural world, they are limited to natural explanations.
Sigh. And this is a limitation how? Do non-scientists have the option of conducting tests in some other world? By definition, isn’t everything in the world natural? What the hell is an explanation that isn’t natural, anyway? All sensory data we have about the world around us is “natural.” Whatever structure is present in our brain that interprets these data is also “natural.” So any explanation that is not natural must be positing something unobservable. How is this reasonable?
Sean,
I think the point you make is an excellent one. Two small remarks though:
1) I think there is something to the claim that things are classed as supernatural precisely because they are non-scientific. For example walking on water is considered supernatural precisely because according to our well-tested and empirically sound scientific theories, it can’t happen. If that wasn’t the case there’d be no reason to class it as supernatural; there’s no a priori logical reason why walking on water should be impossible. It just so happens to be, as far as we understand things.
2) You say “The distinguishing feature of science isn’t in the explanations themselves, it’s in the process by which we find them.” I don’t really see why the process is important. In my opinion it doesn’t matter how the explanations are arrived at (whether from logical thought/supposed revelation from God, whatever); what matters is that they work, i.e. are well substantiated to a high level of rigor before being accepted.
I agree completely though that the qualification “natural” shouldn’t be part of the definition of science; at best it’s redundant and at worst it’s misleading.
I should really stop commenting here and get back to work, but:
“If that wasn’t the case there’d be no reason to class it as supernatural; there’s no a priori logical reason why walking on water should be impossible. It just so happens to be, as far as we understand things.”
No, no, no. Walking on water should be considered impossible because it doesn’t happen in the world we observe. There’s a reason it was considered miraculous long before science as such had really developed.
But then, there’s no “a priori logical reason” for anything. Stop me before I go on my tirade against “rationalism.”
Anonymous,
I’m sorry, I don’t understand why you think you’re disagreeing with me. I thought that was precisely my point; walking on water is classified as supernatural precisely because it is unscientific (i.e. not supported empirically) rather than because it is a priori impossible.
michaeld, you had written “because according to our well-tested and empirically sound scientific theories, it can’t happen.” I was arguing that it just doesn’t happen, independent of any theories. The empirical observation that it doesn’t happen long predates any scientific theory about why this is true. If you just meant that it is not supported empirically I don’t see why you mentioned theories. I think one should distinguish between “science” and “that which is supported empirically”, since science seeks a unifying explanation for what we observe. It is not simply the collection of all that is supported empirically.
And on that note I will refrain from further comments for a while.
Here is a discussion about walking on water:
http://star.tau.ac.il/QUIZ/96/A01.96.html
Anonymous, we probably have different notions of what a theory is but I don’t want to belabour the point, since it’s tangential.
Sean
Glad you said this: I had exactly the same reaction. To burden science with the idea of distinguishing the natural from the supernatural is really, really silly. It is a distraction from the real role of science, which is the evaluation of hypotheses and theories, and establishment of levels of confidence in the various theories we have. As you point out, the scientific method is perfectly capable of evaluating ID and relegating it to the list of abject failures whether they be, in any sense, naural or supernatural.
Count Iblis, nice link, and a nice way to show that scientists are just plain wierd for caring about stuff like this 🙂 The calculation does give you an appreciation for the power of a dolphin as it “walks” on water using its tail.
Myself, I’d use a trampofoil instead. They look increadibly cool the first time you see one:
http://www.trampofoil.com/
About the term “natural” :
I’m sorry but I don’t agree.
The “natural” term as in the natural sciences is very very relevant.
Science can’t be reduced only to the “scientific methodology” pioneered by Galileo. What is the object of Science in general ? It’s the study through the application of the scientific method to phenomena pertaining to our natural world.
Reducing science only to a method is really preposterous, since then the object of science wouldn’t be focused anymore on our natural world.
All this epistemological commentary (I was tempted to write “blather”) ignores the fact that the definition in the Kansas Standards occurred in a socio-political context that gives it a quite different meaning to the meaning that the definition in isolation might have. In context, the redefinition as generated by a predominantly Christian creationist state board means just that ‘The Christian God is now an acceptable explanatory construct in science, and school children may now be justifiably taught to disregard actual tested and corroborated scientific explanations of phenomena and substitute God as a sufficient explanation’.
RBH
I think a philosopher of science really needs to weigh in here, so that we don’t reinvent conceptual wheels. But here are two observations. First, in practice a supernatural process would distinguish itself from a natural one by the fact it did not fall into established patterns — either patterns directly observed or ones inferred from some broader theoretical framework. Following your argument that science is a search for patterns, it has a built-in bias against supernatural explanations, and therefore cannot be said to approach the natural vs. supernatural question agnostically.
Second, suppose I accept, for sake of argument, that Occam’s razor (aka algorithmic compressibility) a priori leads us to prefer natural explanations. Then, again, science has a built-in bias against supernatural explanations, with exactly the same effect as if I defined science at the outset to be restricted to natural explanations.
To put it differently, you seem to be using Occam’s razor as a metaphysical choice, which leaves it vulnerable to the fact that other metaphysical choices are possible. I prefer to think of Occam’s razor as an epistemic, or pragmatic, choice, which brings us back to my point that science is about finding *useful* explanations, as opposed to “correct” (whatever that means) ones.
George
Sean — the Kantian in me disagrees; I think you can separate religion and science adequately. I’m guessing not too many scientists or theists agree with me though. Scientists build and test models. There’s then some mapping function from the models to some unknowable truth. There’s the obvious and simplest mapping function which is the identity function — the models *are* the truth. But there’s no uniqueness proof, and religion is free to say there’s a different mapping function to a different truth.
I am not sure I follow. Never mind that there are several definitions and views on the meaning of ‘supernatural’ (see Wikipedia). But the general distinction seems to be that these things have never been observed in nature (gods, magic).
So, granted that these things are known by enormous amounts of observation to not exist, how can one claim to base scientific theories on them and what would be the purpose?
Take ID for example. The religious creator is supernatural so is out due to earlier observations. The panspermia creator is locally unprobable and globally doesn’t explain anything. So while ID is not entirely supernatural (only in its usual meaning) it is void of explanatory power because of what we already know.
Suppose science was characterized as a method of making assertions about how the natural world behaves that are subject to falsification by experiments which take place in the natural world. In such a characterization, we might assert that a charged particle will move in an electric field in a particular fashion because it was enchanted by the Invisible Elf. You couldn’t really test the elf hypothesis, but you could certainly perform an experiment to determine whether or not the particle moved in the way predicted. In that case, we would have the same descriptions of physical reality, but there would be this untestable hypothesis attached. Moreover, someone else could come along with a competing hypothesis – the electron moves the way it does because the Flying Spaghetti Monster so wills it – that makes the same predictions about how the electron moves. So, in terms of making falsifiable predictions about natural phenomena, the IF and FSM theories are indistinguishable and so, to most practicing scientists, irrelevant. However, I can easily see a bloody holy war arising between the passionate (and mostly non-scientist) adherents of IF and FSM.
I think RBH makes a decent point. To Ugo, of course it’s important that science is supposed to be about nature; it’s just I don’t think that science should rule out “supernatural explanations” (whatever that means) – at best such a restriction is redundant and at worst misleading.
It seems a lot of this argument is purely semantic. But I’ll just try and rephrase what I believe to be Sean’s original point (which I think is a very important point). To be concrete I’ll phrase it in terms of God rather than the supernatural (since the latter seems to be generating the semantics).
If the hypothesis of God led to a good solid, testable and empirically sound theory explaining the origin of species then it would be completely legitimate to class this as a respectable scientific theory (like evolution). Further, if the God hypothesis was useful in developing a quantative, testable theory of quantum gravity then that also would be great! The point is not that God is some sort of “no go” area that science isn’t allowed to go near. The point is that the God hypothesis simply doesn’t accomplish anything scientifically. If it did that’d be great, but because it doesn’t it is not science. The reason to reject ID and creationism (scientifically at least) is not because scientists have some sort of a priori prejudgice against the bible and religion but because ID and creationism have not led to one iota of scientific insight regarding the origin of species and appear to have no explanatory power whatsoever. Their only purpose is to support preconceived beliefs that people want to hold on to, not to help us understand the issues involved.
Ahem. Of course, if the religious creator only act is to create life, then it is not out due to earlier observations and Sean’s reasoning follows. I got momentarily fooled by the usual meaning of the creator, which one can not overlook as long as ID refuse to say anything about the creator.
Excellent post Sean. I often worry that many scientists run the risk of being hypocritical by dimissing supernatural claims offhand. We might win more hearts by giving ID and other such claims their rightful day in court. Which isn’t to say that we should teach these things in school. But we must be honest and at least admit to the world that ID is a legimate scientific hypothesis; and then soundly explain why it belongs to the dustbin of terribly bad hypotheses. Of course, most scientists don’t want to spend their time beating down every crackpot hypothesis that comes along. However, when that crackpot hypothesis gains such a following that it begins to affect the laws of our land and the education of our childern, I think it is our duty to take the time to do so.
To Jeff Olson :
Can you explain in what terms is the ID hypothesis a scientific hypothesis ?
Or more in general, what is a scientific hypothesis ? Because either we agree that “metaphysical” entities pertain to the language of the natural sciences and therefore we can elaborate scientific hypothesis about them or not.
What is then the scientific value of a hypothesis contaning metaphysical terms ?
I think that the view that science is simply hypothesis testing and that any and that where these hypotheses come from isn’t part of science is a little over-simplified. Read the description of intelligent design given at intelligent design network. They say that intelligent design is more a framework in which evidence of design is searched for. I believe that evolution is also a framework rather than a hypothesis. The concept of survival of the fittest is a tautology: that which survives is the fittest. It is useful if we can find simple ways of classifying what allows one to survive. In the same sense, Newton’s laws are useful if we find just a few forces out there.
So taking away the word “natural” in the type of explanations science allows is an appeal to use a new framework (or paradigm) such as ID. Again, its only valuable if we find just a few types of design principles. If not, its not useful.
So the appeal for ID is an appeal for a paradigm shift. And these are rather hard to argue against. You have to say things like, “ok, try using that for awhile and see how far you get.” I guess you judge the result by how simple the resulting picture is, which will often be to some extent subjective. To someone with God as a regular category of explanation in their mind, they will probably judge the result as more simple than someone who doesn’t invoke God.