Sunday, July 20, 2008


The Philosophy of Confusion

Wes Elsberry at The Austringer has a post up, entitled "Intelligent Design: Philosophical Bogosity," about Bradley Monton, a philosopher at the University of Colorado at Bolder and a self-described atheist, who is humping a manuscript of a book about Intelligent Design that argues that ID is science. Given that the title is tentatively "An Atheist Defends Intelligent Design," the Discovery Institute will no doubt be more than helpful with contacts to such publishers as the Foundation for Thought and Ethics, one of William Dembski's favorites.

Be that as it may, Wes investigates Monton's blog for some idea of what we are in store for and finds Monton criticizing Ken Miller's book, Only a Theory for calling ID a "science stopper." Monton endorses a response proposed by Alvin Plantinga:

[W]hile theistic scientists could choose to stop investigating the world, and be satisfied with the answer "God did it", they need not. What theistic scientists can do is investigate the questions: "what did God do?" "What structure did God choose to give the world?" As long as scientists are willing to investigate those questions, then science can go on in pretty much the standard way. Allowing supernatural hypotheses won't really change anything.
Wes correctly responds:

The essential point is conceded by Plantinga and Monton in this summary: the supernatural explanation fails to explain, and explanation must await someone willing to seek a naturalistic secondary cause that will itself actually explain the phenomena of interest. The mere possibility that someone working in a theistic science could choose to do so does not validate "theistic science" as something good and to be desired.
Go read Wes' post for his further cogent arguments as to the superfluous nature of ID vis a vis science. What I want to discuss is the category error that Monton and the IDers make but you'll have to hang in while I set the stage.

In another post, Monton criticizes Miller claim, as Menton puts it:

... that the intelligent design movement doesn't just want to "win the battle against Darwin"; the intelligent design movement wants to "win the greater war against science itself" (p. 183).
Menton states that "as far as I can tell" the only evidence that Miller gives for that "strong claim" is a passage from William Dembski:

The implications of intelligent design are radical in the true sense of this much overused word. The question posed by intelligent design is not how we should do science and theology in light of the triumph of Enlightenment rationalism and scientific naturalism. The question is rather how we should do science and theology in light of the impending collapse of Enlightenment rationalism and scientific naturalism. These ideologies are on their way out … because they are bankrupt. (p. 190)
This itself raises questions of Menton's agenda since, on p. 183, in the very paragraph before the one Menton quotes in connection with Miller's claim, Miller cited to the Wedge Document, which makes all the same arguments Dembski does in the cited passage. But to continue just on the issue of ID's status as science, Menton gives two "interpretations" of Dembski, one Menton calls "anti-science" and one "pro-science." Here is the "pro-science" interpretation:

On the pro-science way of reading the passage, one would hold that naturalism is a key part of Enlightenment rationalism, and there is a style of science where one takes an assumption of naturalism to be part of the methodology of science. One would hold that intelligent design is opposed to the naturalism in Enlightenment rationalism, and naturalistic science, but one would not hold that intelligent design is opposed to science itself.

It is pretty clear to me, judging from everything I've read by Dembski, that he intends the latter, pro-science, reading. ... ([I]n my opinion, at least) it becomes clear that Dembski is pro-science; he's just not pro-naturalism, and hence he's not pro-naturalism-as-a-scientific-methodology. Now, Miller apparently thinks that if one drops methodological naturalism, then science will stop, because one can simply appeal to God as an explanation of any scientific phenomenon.
So, let's follow along: It's okay to posit a supernatural cause for natural phenomenon because a scientist can freely choose to ignore his own, science stopping, "hypothesis" about God and continue seeking naturalistic causes "pretty much the standard way." Then what does it mean to be "pro-science" if one rejects "naturalism-as-a-scientific-methodology"? What, pray tell, is non-naturalistic science anyway, if supernaturalist explanations are admittedly "science stoppers"? If science only advances through ignoring the supernatural, how can rejecting methodological naturalism be a part of science?

It's clear enough that the only way you can hold these two contradictory ideas together is by smearing science together with philosophy and/or theology and denying it has its own status separate and apart from them. It is a simple category error and a failure of proper definition. It borders, at least, on the post-modernist maneuver of claiming that science is only a "social construct" no different than the IDers' theological musings.

You can see this impulse to trample the borders between science and philosophy, in an example given by Miller, at Dembski's blog, Uncommon Descent. Dembski cites to a "fire rainbow," a phenomenon he, himself, gives a naturalistic explanation for:

Clouds have to be cirrus, at least four miles in the air, with just the right amount of ice crystals; and the sun has to hit the clouds at 58 degrees.
But Dembski goes on to say "It's the gratuitousness of such [beauty] that leads me to rebel against materialism." There is no conceivable way that the "gratuitousness" of beauty can be measured scientifically, nor can the conclusion that materialism is wrong on that grounds be tested by scientific means. Dembski is, of course, free to make whatever philosophical conclusions he wants but if he wants to do science, he must restrict himself to ice crystals and the refraction of light in naturalistic interaction that can be tested by scientific methods.

The main objection to ID is not that it lacks "arguments," as Menton seems to think. After all, it is an ancient concept dating back as far as Empedocles (c. 490 – 430 BCE), at least, and has been kicked around by philosophers ever since. The objection is that the arguments are not scientific or amenable to scientific testing, which means that calling it science only confuses. No one seriously objects to teaching about ID in history or religion or philosophy classes, even classes on the history and philosophy of science. The objection is against teaching that ID is science, instead of philosophy or theology.

But get to know the name of Bradley Monton. He's going to be the next star of the Intelligent Design movement ... whether he wants to be or not.

My time is finite, so I'm liking the idea of more people piling on for examination of Monton's claims. By his weblog, it appears to be a target-rich environment. Monton is quick off the mark to dismiss criticisms of IDC as "fallacious", but it looks to me like he's either ignored swaths of criticism (my stuff with John Wilkins and Jeff Shallit, and "Why Intelligent Design Fails", for a few of instances), or strains at gnats and swallows camels, as in the item about "science-stoppers".

Wesley R. Elsberry
Yeah, there are a couple of more of his posts I've got my eye on. Maybe Wilkins would like a crack at him, too.
As you know, John, I'm basically sympathetic to your point of view. But I have a concern about the lines you'd like to draw here.

The concern is that it's proven to be exceedingly difficult to formulate a non-arbitrary demarcation criteria for scientific theories. The various attempts I'm familiar with all seem to promise either too much or too little. And, more importantly, it seems important to recognize that in fact the nature of science itself has evolved as scientific theories evolve. What counts as successful science -- i.e. the criteria of good science -- evolve along with the body of theories.

Among other things, this means that it was not entirely inappropriate for Steven Fuller to have pointed out in the Dover case that admitting ID as science would involve changing the definition of science. Science is not some a-historical entity.

Unfortunately, I'm not quite sure where this leaves us. But it does tell me that it can be exceedingly hard to draw any hard-and-fast distinctions between scientific theories and philosophical -- or even theological -- reflections on those theories.

On the other hand, we might be able to get some mileage out of the distinction between arguments and explanations, and maintain that just having a good argument for some claim doesn't amount to an explanation of the phenomena at issue. We might be able to hold that line without being committed to anything carved in stone about the criteria for assessing the relative success of explanations.
Calr Sachs wrote

Unfortunately, I'm not quite sure where this leaves us. But it does tell me that it can be exceedingly hard to draw any hard-and-fast distinctions between scientific theories and philosophical -- or even theological -- reflections on those theories.

That is undoubtedly the case, but as someone once said, "I may not know exactly where day ends and night begins, but I sure as Hell can tell the difference between them."

ID in the style of Dembski and Johnson and Behe and Fuller (and, apparently, Monton) is well into the "night" category. Over a cluster of criteria -- e.g., testability, research fecundity, explanatory power -- each admittedly fuzzy when taken alone but jointly illuminating when taken together, ID has given us nothing.
Yes, demarcation criteria are difficult if not impossible to craft, because science is a sorites heap. But a structual necessity of science, as it is understood today, is testability. Unless any hypothesis can be tested by the rest of the scientific community, it is not science.

There can be issues of what constitutes "testability." String theory has been rejected by many scientists because they not only see it as presently untestable but unable to be tested under any conceivable circumstances. But that's why I asked what a non-naturalistic science would look like, especially since both Plantinga and Monton concede that "goddidit" is a science stopper.

Sure, if we someday invent Wilkin's "divinoscope" that can observe God (and we could get around the problem of how to test that the divinoscope is really showing us God), then maybe goddidit can become part of science. But in the meantime, admitting the supernatural as a "scientific" explanation prevents ruling out natural causes as well as supernatural. How could we even rule out Plato's four elemental particles in the form of the "perfect solids," if any result can be explained away as "God's whim," the way Johnson explains the peacock's tail? We only think there are electrons, protrons, neutrons, etc. because God's whim makes the world look that way when it really isn't. On the other hand, if we accept the "appearance" of the world when it comes to elemental particles, why not with regard to the rest? What demarcation will the IDers draw?

Sure, we could change the definition of science to anything we want and include anything we want, up to and including broadway choreography. The question is how changing the definition improves or degrades the objectivity of our knowledge of the natural world. In a point I want to blog on further, Monton drags out the old "if God acts in the world but you rule that out a priori, science will be wrong." Yeah? So what? If we cannot test that scientifically, who'll know? If you want to test it by revelation or the personal experience of God, who needs science? You've already got religion. If you want to test it by argument from "first principles," instead of by testing naturalistic hypotheses by naturalistic means, who needs science? You've already got philosophy.

I think it is science-envy in both cases -- theists and relativists wanting the aura of science without the standards.

The only difference I can see between an "explanation" and an "argument" is that one can be tested by human (i.e. empiric, natualistic) means, and the other can't be.
What theistic scientists can do is investigate the questions: "what did God do?" "What structure did God choose to give the world?" As long as scientists are willing to investigate those questions, then science can go on in pretty much the standard way.

That is, the standard way of a few centuries past, when people could satisfy themselves describing what they saw, without constructing ideas about how such things came about or why they function as they do.
As I understand it, the "demarkation problem" is to give criteria which draw a line between science and non-science, with all of science on one side, and all of non-science on the other.

We don't need a solution to that problem to determine that some things are not science. We can draw a line which includes all of science, as well as lots of other things, on one side, and some things on the other.

We could, for example, draw a line which includes science, history, law, music, and baseball, but excludes "intelligent design". It might not be very interesting to philosophers of science, but it would serve our purposes.

Tom S.
Thank you very much, John and RBH, for your helpful comments -- esp. the thought that science is a sorites heap, and the thought that even if individual criteria are fuzzy, then can be illuminating when taken together.

But I still want to know more about what we mean when we talk about "testability." For -- if I may play devil's advocate -- surely in philosophy and in theology there is a sort of "testability" according to criteria that are partially constitutive of those activities. In philosophy we want conceptual clarity and sound argument, we want to avoid confusions and contradictions wherever possible, and we want to arrive at genuine insights about the nature of knowledge, of reality, of human action, etc. (Or to arrive at insights into the limits of our knowledge of such things.)

So, how are these desiderata not forms of "testing" of philosophical positions and systems?

One might restrict science to "empirical testability," but I worry that this move opens up whole new problems about what counts as 'empirical' -- since the boundary-line between observational claims and theoretical claims shifts along with the progress of science itself. (Atoms, for example, used to be purely theoretical entities, but now are observational ones, thanks to improvements in technology.)
So, how are these desiderata not forms of "testing" of philosophical positions and systems?

They may well be forms of "testing" by philosophical standards. What's more, conceptual clarity and sound argument, the avoidance of confusions and contradictions, and insights about the nature of knowledge and reality, and human interaction with both, are important to the scientific enterprise (without being part of it), just as they are important to law, government and many other human enterprises (without being part of them).

Yes, I think the testing in science is of a specific type. It is empiric in nature and it is both science's primary tool and its sine qua non. It consist of experiment and observation that can be and is shared with and repeated by people who do not need to share any commitment to "higher" philosophies or religions in order to make the tests. The only commitment they need share is to make the experiments/observations and share the results. Science achieves this goal by reducing its subject to the least common denominator: the physical universe as it works through "laws" which are invariant in operation (or invariant in the way that they vary, which is the same thing).

In short, science is the attempt to find out what invariant laws govern the physical universe and how the universe actually operates under those laws. It is a much narrower enterprise than either philosophy or religion and that is (IMHO) its strength. Postulating a being that can consciously violate those laws makes the enterprise impossible since, in the absence of invariant knowledge of that being's acts and intentions, it renders the universe a place where there are no invariant laws that can be agreed on. At the very least, in order to do science, the practitioner must agree that the apparently invariant laws will be treated, within the discipline of science (hence methodological naturalism), as if they are, in fact, invariant.

Naturally, no one need accept this assumption, minimalist though it may be, but if they don't, they cannot be said to be doing "science." One measure of how successful science has been and, therefore, one measure of the soundness of science's underlying methodological assumption, is the number of people who want to claim its mantle without committing to its standards.

Of course there can be gray areas in theorizing – I mentioned one: string theory, which I think is a fully realized mathematical "theory" that may never make it to being even a "hypothesis" in physics. Furthermore, as Popper and his opponents showed, there can be tricky problems as to what constitutes "testing." But the absolute bottom requirement is that any theory must be capable of being subjected to empiric testing of a sort that confirm (more than merely "not contradict") or deny the theory.

The more honest IDers will concede that God cannot be subjected to such empiric tests. Certainly the scientific community as a whole (the only judge in such matters) does not accept that God is a testable hypothesis. If the IDers squirm and say that the "Designer" could be a space alien or some such (see Ron Bailey's purple space squid), who are theoretically empirically testable, then their refusal to even attempt to propose or conduct such tests demonstrates their lack of serious scientific intent. In either event, ID fails the criteria of science.
Bradley Monton wrote a cordial response to my post.

I wrote a pointed response to his response.

Wesley R. Elsberry
Yeah, I saw it this morning but you beat me to the punch. I'll still probably have more to say later today.
"Our" position -- i.e. the anti-design theory position -- often seems to assume, I think, that supernatural claims are immune to empirical testing. Firstly, I'm not entirely sure that's right. Secondly, I think we would benefit from an argument to that effect, and not just stating it as something obvious.

Consider this example, from recent work by Eliot Sober: "there exists a supernatural agent who is solely responsible for the natural order, who is fully omnipotent, and who would have done everything in its power to make everything purple."

It seems clear to me that this is a claim about supernatural causation that is empirically testable, insofar as observation shows it to be false. So claims about the supernatural are not, just for that reason, outside the scope of empirical testability. We might need a better argument than the "if we admit the supernatural, everything goes" line we've been using so far.
The reason that such a test (i.e. "God" as I define him does not exist unless the universe is only 6,000 years old) works is because you are not testing the existence of some supernatural being, you are testing my definition of "God," which I have conveniently cast in empiric terms. Try it another way: "there exists a supernatural agent who is solely responsible for the natural order, who is fully omnipotent, and who would have done everything in its power to make exactly the number and type of purple things we happen to see in the universe." Is that God testable? That is, in fact, what ID proposes and, indeed, what most mainstream religions propose.

Sure, if you set a criteria for "God" that is empirically testable and universe does not match that criteria, you can say with some empiric warrant that such a God does not exist ... short of denials of empiric evidence itself, a la Omphalos, which denials are not themselves empirically testable.

But trying to reverse that and to take the empiric world on its own terms and attempt to "detect" the existence or non-existence of some supernatural being based on the empiric world's nature must fail, unless, of course, you sneak empiric criteria back in through some back door. Monton tries this in another post where he suggests a "test" for God by doing a large statistical test of mutations and, if a preponderance were beneficial, that would provide some evidence for the existence of God. Quite apart from the fact that mutations are only "beneficial" within certain environments (see Abbie Smith's recent post at ERV re the error of creationist "research"), meaning there is a problem of definition of "beneficial", such a procedure would only work as a "test" if we can rule out any presently unknown naturalistic cause for the result and there's no way to do that. Nor would the finding that most mutations are not beneficial be a "test" showing the non-existence of God, without assumptions being about the intent of God and its actions or inaction upon empiric nature.

Someone might find such a "test" a convincing philosophical argument for the existence of God, while others might say "why would a God put beneficial mutations in most creatures when most creatures fail to have their genes make it beyond a generation or two?" But in either case, it would not be a scientific theory that is empirically testable.
John, this is extremely helpful to me, and I appreciate your willingness to help me out. (Philosophy of science is not my strength -- I mostly do history of modern and pragmatism.)


(1) There exists a supernatural being, solely responsible for the natural order, who wanted (or would have wanted) to make sure that everything is purple.

(2) There exists a supernatural being, solely responsible for the natural order, who wanted (or would have wanted) to make sure that there as exactly as many and type of purple things as there are.

(1) is testable, and can be found to be false; is (2) not even testable?

I think it isn't testable. It's not testable because a hypothesis must specify some set of conditions which are not merely "read off" from already established observations. That is, a hypothesis needs to --

(a) specify some set of observable conditions;

(b) which conditions would not obtain if the hypothesis were not true;

(c) and where the observable conditions are not part of pre-existing store of knowledge.

The strength of (1) is that it specifies an observable condition ("everything being purple") which is true if and only if the hypothesis (1) is true.

The weakness of (2) is that the observable conditions specified by the hypothesis are true regardless of the truth of the hypothesis -- there are exactly as many and types of purple things are there are, regardless of whether or not there are any supernatural entities.

How does that sound? What else am I missing?

In other words, I fully agree that one cannot infer from any facts about the natural world to the existence (or nonexistence) of any supernatural beings. Hume recognized this quite clearly, and so too did Kant. As Kant put, the argument from design to the existence of God only works by presupposing that we already know what God is.

(As a minor aside, it's because of Kant's devastating criticism of the argument from design that that argument never developed much of a following in Germany or France. But I find it very interesting that Hume's criticism did nothing to affect the popularity of the argument from design in England -- so far as I know, Paley completely ignores Hume. I think that Kant's criticism of the design argument is much more general than Hume's criticism but could be wrong about that.)
Yep, that's about right, I think.

It's not testable because a hypothesis must specify some set of conditions which are not merely "read off" from already established observations.

This is the concept of the "critical observation" which (if my increasingly decrepit memory serves) Popper started to develop after the weaknesses of his "falsifiability criteria" were exposed ... or maybe it was Quine.

Of course, within a limited field, such as science, where you are also willing to tolerate tentative answers, you can do a more limited test, say between two or among a few competing theories that you think exhaust the likely possible explanations. Then you need not have an "if and only if" situation, as long as one theory explains the set of observations better than any competitor.

I also seem to remember that Paley was said to have written Natural Theology almost entirely as a rebuttal to Hume but that as a pious divine, who didn't want to promote the great heathen, and given the style of the day, which did not necessarily give references to people or arguments that the author expected his/her audience to recognize, he managed to do it without mentioning Hume once. If I find the reference, I'll get it to you one way or another.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?

. . . . .


How to Support Science Education