8 replies to this topic

[davidm]

[Split from here - HH.]

Hugo, I'd be interested in your view on how undetermination fits in with the current controversy (a controversy that most evolutionary biologists deny is valid) between the standard descent with modification view of evolutionary change and the intelligent design movement. A discussion of the philosophy of science as it pertains to evolutionary theory can be found at talkorigins.org, here.

From the above link:

Quote

One thing all three of these philosophers thought in opposition to Popper - there was no point that could be ruled off as the dividing line between 'rational' science and 'non-rational' non-science. Lakatos identified what he called the Duhem-Quine Thesis - nothing can be falsified if you want to make suitable adjustments elsewhere in your theoretical commitments. Get a result that upsets your favoured theory of gravitation? Then the instrument's in error, or something is interfering with the observations, or there's another process you didn't know about, or some other background theory is wrong. And the point of this is that all these moves are actually used - they are rational in the sense of good scientific practice. Positivism is irretrievably dead by this stage.

Edited to add: there's a whole blog devoted to the philosophyof biology, found here.

[Hugo Holbling]

The Talk Origins article was very disappointing, including the tired old reading of Feyerabend that has been refuted at length here and not taking account of the arguments against falsificationism beyond the simplistic notion that it is self-refuting (which i set out here). Still, at least it had the integrity to admit that the demarcation problem is unsolved and thus that this issue of whether ID is scientific or not depends on what "scientific" means, which is far from resolved. Telling us that "philosophers do conceptual tidying up, among other things, but scientists are the ones making all the sawdust in the workshop, and they need not be so tidy" dodges the point entirely, alas.

The relevance of underdetermination, in any case, is that there is more than a single way of explaining an available data set. The question, then, is whether ID does explain it and by what criteria we select theories. As i've been saying for some time, appeals to evidence alone are insufficient.

[davidm]

I didn't expect that the talkorigins article was very good, but what intrigued me was that it raised these issues in the first place. In reading various science books and also in perusing science blogs, I rarely if ever see the science writers or working scientists take into account these  philosophical issues. In fact, among biology blogs I've seen (for some reason I'm becoming more interested in biology), anybody who even seems to intimate any kind of residual sympathy, theoretical, philosophical or otherwise, with ID is hotly dismissed as a cretin. (Though perhaps I've been reading the Pharyngula blog   too much.)


Anyway, I raise this issue because I'm trying to see if there is some other modern counterpart  to the historical examples you raise in addition to, for instance, quantum theory. It seems to me that underdetermination, as I understand it, does come into play in trying to choose between ID and descent with modification via entirely natural means, though if I were asked which of the two corresponds with reality, I'd unhesitatingly say natural evolution, with no ID component. So now, I suppose, I expose myself as merely muddled.

[Hugo Holbling]

There's a degree to which many scientists seem to feel that the philosophy of science is pointless and irrelevant, but this very issue of ID (and/or creationism) show us why the demarcation problem is important still. I regularly hear or read people asking "is x a science?", whether the subject is ID, psychology or economics, without any clear idea of what separates science from non- or the much-maligned pseudo-science. If we have no answer to the demarcation problem or adopt an ignorant "i knows it when i sees it, i tells ya" approach, however, dismissals of ID (or anything else) as "nonscientific" are meaningless.

In general, anyway, underdetermination applies to all theories (given the caveats in my essay). I expect you noticed that the Talk Origins article didn't resolve any issues at all, so it should be no surprise that dialogue often seems to be at an impasse.

[davidm]

You know, Hugo, sometimes, you really make my head hurt.   Things that I had taken for granted seem much less obvious after reading your essays and posts.

I'm taking as a test case the conflict between the theories of evolution (broadly speaking, descent with modification) and the so-called theory of intelligent design, as a way of evaluating the philosophy of science issues that you have raised -- partly to deepen my understandinig of these issues, but also partly to decide if I really ought to be agnostic about deciding between unplanned evolution and ID.

Maybe you can  help me untangle this.

Most biologists would say there is no problem here, because the theory of ID is not a theory -- and hence (presumably) none of the problems that you have discussed, like undetermination or falsificationism, arise. On this view, ID is vacuous, because it predicts nothing and is really just a question-begging argument, assuming the existence of ID when it must be proved.

However, ID proponents would dispute this.  One example is Michael Behe. Behe would say that the theory of intelligent design predicts the existence of irreducibly complex structures -- and he cites, as an example, the flagellum. Behe argues that since the flagellum is irreducibly complex, it must have been designed. A failure to find irreducibly complex structures, Behe would presumably concede, would have falsified ID.

But, as you explained in another essay, falsificationism is a dubious project. So by this reasoning, even if we failed to find such structures, Behe would be justified in  tenaciously pursuing his theory (if it is a theory). The issue is made even more complex by the fact that biologists insist that we have failed to find such structures -- that Behe is simply wrong in citing the flagellum as irreducibly complex, and that moreover, evolutionary theory perfectly accounts for the way that seemingly IC structures in fact arose in a stepwise fashion, guided by little more than random chance shaped by natural selection.

Another ID prediction is due to Dembski, who argues that a structure in nature is designed if it displays specified complexity -- a complexity that could not have arisen by chance processes. Other scientists say they have shown that his entire premise is flawed, a flaw arising from Dembski's misunderstanding of statistics and probabilty. But the point, I think, is that Dembski can at least say he has produced a theory, and that even if others claimed to have falsified its predictions, isn't Dembski entitled to tenaciously pursue his project, and aren't we on shaky ground in claiming that it (or any theory) is ever really falsified? If these really are two competing theories, (evolution vs. ID) aren't they then underdetermined? And should we be agnostic in choosing between them? Or is the better strategy to choose the theory that  is more parsimonious, or more aesthetically appealing or makes more interesting predictions? And if that's the case, I would choose descent with modification and reject ID, but couldn't one argue that my choice is really nothing more than a matter of personal taste?

To go back to the theory of descent with modification: Biologists contend that one of the strongest selling points of evolutionary theory is precisely that it is so vulnerable to falsification. For example, so far as I can tell, it seems obvious that if mammal fossils started turning up in precambrian rocks, descent with modification is falsified. (Even with all the problems of falsification that you discussed, though it's possible I'm missing something.) That's because evolutionary theory relies on nested heirarchies of species -- common descent, with mammals much "higher" on the phylogenetic tree than precambrian species. In other words, it seems the theory of descent with modification would collapse by definition if such fossils were found. That is, there does not seem to be any conceivable ceteris paribus clause to fall back on here.

OK, now I'm going to take a breather. I hope my confusion had made sense.  

[FudoMyoo]

davidm]Another ID prediction is due to Dembski, who argues that a structure in nature is designed if it displays [i]specified complexity[/i said:

-- a complexity that could not have arisen by chance processes.

Interesting points David, I think they made sense. Another related issue that is unclear to me, is how do they define "complexity" or, as in the case of Behes argument; "irreducible complexity"?

[davidm]

[quote name='FudoMyoo][quote name='davidm] Another ID prediction is due to Dembski' date=' who argues that a structure in nature is designed if it displays [i']specified complexity[/i'] -- a complexity that could not have arisen by chance processes. [/quote]

Interesting points David, I think they made sense. Another related issue that is unclear to me, is how do they define "complexity" or, as in the case of Behes argument; "irreducible complexity"?[/quote]

What I'm aiming at is not to try to figure out whether ID or descent with modification is correct. It's to try to see how these competing theories (if ID is even a theory) should be viewed through the lens of the various philosophy of science issues, like undetermination and falsificationism, that Hugo has raised.

That said, to clarify the issues involved:

Dembski, a supporter of Behe, defines irreducibly complex as follows:

[quote]A functional system is irreducibly comple if it contains a multipart
subsystem (i.e., a set of two or more interrelated parts) that cannot be
simplified without destroying the system’s basic function. I refer to this
multipart subsystem as the system’s irreducible core.[/quote]

Anyone who wants to wade through 47 pages of Dembski can go [url=http://www.designinference.com/documents/2004.01.Irred_Compl_Revisited.pdf]here.[/url]

The idea is that there are such systems in living things, and the inference we are supposed to draw is that they could not have evolved by mutation and natural selection, for the reasons given above: the whole thing has to hang together as a piece from the start, or it will be completely useless for any purpose.

However, biologists do not support this idea (neither Behe nor Dembski is a biologist). They point out that evolution can commandeer different structures for different functions, and so over a long period of time, with many intermediate steps, a structure can evolve that may be irreducibly complex for the function that it currently serves, but would have served different functions in the past. An example, shown by the fossil record, can be found [url=http://www.talkorigins.org/faqs/comdesc/section1.html#morphological_intermediates_ex2]here,[/url] in a detailed discussion.


As far as specified complexity is concerned, a discussion of it (starting at the bottom of page 3) can be found [url=http://www.talkreason.org/articles/Stealth.pdf]here.[/url] As the author notes, Dembski claims to be able to prove, from modern information theory, that life and the universe cannot be the product of chance. The author shows why this is wrong.

But even if  Dembski is wrong on this particular point, it seems he would be entitled to invoke other hypotheses and assumptions, even if untestable, that would support his claim, and it would be very difficult to say why we should dismiss this approach. At least, this is in part what I get from the underdetermination thesis.

[Hugo Holbling]

That was a nice discussion of the many issues involved, David. I'm going to split it off into its own topic.

I think you may be right (by implication, at least) that the argument is taking place at more than one level. If we adopt a Lakatosian approach (as most scientists who discuss the philosophy of science seem to claim to - cf. his Methodology of Scientific Research Programmes) there is no reason why ID proponents should give up on it because it may seem to be a degenerating programme, because falsifications are claimed to exist or because its current arguments may be considered weak. Most theories have been in this position at one time or another, including what we often call our "best" ones. We can adopt methodological guidelines such as parsimony, beauty or an insistence on solely natural explanations, but these are due to whatever presuppositions we bring with us and not the matter at hand.

What i find interesting is that scientists readily concede that theirs is a fallibilist venture where certainty doesn't exist, in principle, and yet frequently describe select theories as "so well confirmed as to be fact". This may be one of the levels at which communication goes awry, since it's easy to understand it as dismissing any alternative as impossible (which, of course, it does not). If we include the necessary subtlety of "facts" as "confirmed to such a degree that it would be perverse to withhold provisional consent", then we can see the problem is probably terminological, at least in part. ("Perverse" is an unfortunate choice of word here, at any rate.) If we were to drop fact (after all, there are no facts in the traditional sense of the word, which is perhaps maintained by one side of this debate) and use highly-confirmed or something similar (even with an amplification like most highly-confirmed, say), there may be more understanding possible.

In any case, the difficulty in making any definitive statement regarding the ID controversy is that - as i said - it takes place at so many levels. The answer to the question "is ID scientific?" has implications for education, politics and religion at the very least. It seems to me that either there should be a change to educational philosophies as a whole, whereby we teach children (and adults) methods instead of "facts" (that is, how to come by (provisional) knowledge rather than collections of facts that end up being all but meaningless unless we have a way to make sense of them and apply any lessons gained to new situations where they don't give us ready answers), or we should make a change to the rhetoric and concepts we employ and render them more accurate. This would mean dispensing with problematic demands that only science be taught in the classroom and replacing them with more pragmatic requirements that, given limited time, only the most confirmed theories should be explained to students. Notice, however, that this in itself requires a change in educational philosophies since it would involve more emphasis on the fallibilistic character of science - something that no one on either side denies.

[davidm]

Hugo, thanks for this analysis. This pretty well sums up what I was driving at, and helps me clarify my thinking. I happen to think that the teaching of methods, and not just facts, is pretty important, and this very discussion underscores why.

Being the devil's   advocate that I often am, I sometimes find myself tempted to argue "don't give up" to the ID advocates, against the often brutally disparaging scientists who confront them.