Philosopher Thomas Nagel, writing in the journal Philosophy and Public Affairs, criticizes the exclusion of Intelligent Design from science classes on the grounds that evolutionary science too rests on an assumption: the naturalistic assumption. He argues that both evolution based on natural selection and ID have untestable assumptions.
Frankly, I think that Nagel is wrong partly because he doesn't understand the people pushing ID and partly because he doesn't understand science. With respect to the first, he seems to give the IDers like Michael Behe credit as honest brokers pushing a position of principled skepticism rather than backdoor Creationism. So be it. I think that he is being dangerously naive to do that, but I don't like to start an intellectual argument by impugning the motives of the opposition.
What I will argue is that Nagel is wrong to characterize ID as science because the naturalistic assumption is intrinsic to the definition of science.
Nagel argues that because evolution operates with the a priori assumption of a naturalistic explanation -- allowing no supernatural intervention -- it is not being scientific:
The claim that ID is not a scientific theory implies that even if there were scientific evidence against evolutionary theory, which was originally introduced as an alternative to design, that would not constitute any scientific evidence for ID. We might have to give up evolutionary theory, but then we would be constrained by the canons or definition of science to look for a different scientific, i.e., nonpurposive, explanation of the development of life, because science prohibits us from even considering ID as a possible alternative explanation, one whose eligibility would otherwise be enhanced by the rejection of the leading scientific explanation, namely evolutionary theory.
What would it take to justify the claim that there are propositions such that the discovery of evidence against them can qualify as science, but evidence in favor of them cannot? Someone who accepts this view would probably extend it to propositions about ghosts or extrasensory perception. Research showing that effects that some benighted souls have attributed to ghosts or mental telepathy can be explained in a perfectly naturalistic way would count as science, but any argument that the evidence does not support those explanations, and that significant experimental or observational data are better explained by ghosts or by ESP, would not count as science, and could therefore be ruled out of consideration. On this view it would not even be a false scientific claim.
The idea is that any naturalistic or nonspiritual explanation of a phenomenon can be either confirmed or disconfirmed by empirical evidence together with causal and probabilistic reasoning. No empirical evidence against such a nonspiritual alternative, however, nor any other kind of empirical evidence, could provide a reason for believing the spiritual hypothesis. Belief in something like that is necessarily the result of a different cognitive process, having nothing to do with the scientific evaluation of empirical evidence (rank superstition or blind faith, to give it its true name). I submit that this way of drawing the boundaries around science depends not on a definition but on the unspoken assumption that all such propositions are obviously false -- there are no ghosts, there is no ESP, and there is no god -- so that to invoke such things to explain any observed phenomenon, even one for which no other explanation is available, reveals a disposition to take seriously a possibility that a rational person would not consider. Without this assumption the exclusion of ID from consideration cannot be defended.
In order to think that the refutation or very low probability of all the available alternatives provides some evidence for an explanation E of some observation, one has to believe that E is at least possible. So if one thinks that the existence of ghosts is not a possibility, no spooky manifestations, however elaborate and otherwise inexplicable, will be taken as evidence, however weak, that a ghost is behind them. The real issue over the scientific status of ID is over what determines the antecedent belief in the possibility that a nonphysical being should intervene in the natural order. Opponents of the scientific status of ID are moved by the fact that those who believe this is possible, and who therefore can regard certain empirical observations as evidence for its actuality, usually believe in the possibility as a result of faith or ecclesiastical authority, rather than evidence. This nonscientific element, which is a necessary condition of their interpretation of the empirical evidence, is thought to undermine the scientific status of the whole position.
Unfortunately it also seems to undermine the scientific status of the rejection of ID. Those who would not take any amount of evidence against evolutionary theory as evidence for ID, like those who would not take evidence against naturalistic explanations of spooky manifestations as evidence for the presence of a ghost, seem to be assuming that ID is not a possibility. What is the status of that assumption? Is it scientifically grounded? It may not be a matter of faith or ecclesiastical authority, but it does seem to be a basic, ungrounded assumption about how the world works, essentially a kind of naturalism. If it operates as an empirically ungrounded boundary on the range of possibilities that can be considered in seeking explanations of what we can observe, why does that not undermine the scientific status of the theories that depend on it, just as much as a somewhat different assumption about the antecedent possibilities? (Emphasis mine.)
Nagel is arguing as if the presence of the assumption of naturalism is inherently anti-scientific. He presents science as a method for dealing with evidence without any biases about the nature of that evidence.
The problem with this point of view is that it is a caricature of the scientific method. Yes, science includes a method of dealing with evidence, but it also includes the naturalistic assumption about that evidence. Science divides evidence into what is admissible and what is not. There is a reason that before the term "scientist" was coined the term commonly used was "natural philosopher." This is because science deals with the class of phenomena that physically exist in nature. It has nothing to say about supernatural phenomena; it's principles do not apply there.
I think that Nagel is confusing science with philosophy. Philosophy -- and more importantly logic -- is a method of evaluating evidence that does not privilege one type over another. Science, on the other hand, does privilege evidence. We only deal with natural evidence. Thus, the statement that science should admit ID to avoid making an unscientific assumption is a non sequitur. It is like arguing that algebra should admit other maths that violate the transitive property because algebra assumes the transitive property. I agree that there are other maths. I agree that those maths may have meaning or usefulness. I just reject the notion that the place for them is in algebra.
Nagel also seems to be frustrated with the normative nature of the naturalistic assumption. This is just silly. The selection of particular books to be read in English class is a normative assumption. It is based on what the teachers feel is most useful. More importantly, you would never ask an English teacher to include books in French because the "decision to include only books in English is a normative one." There is a place for French books: French class.
Further, if you want to argue for the usefulness of assumptions, you would be hard pressed to find one more useful than the naturalistic assumption. Let's set aside the fact that all the wonders of the modern world are predicated on an assumption of physicality. (No one thinks prayer will move the subway.) Let's make a wager. Give me one person who believes in spirit healing and one person who believes in science, and let's see who cures AIDS first. My money is on the scientist. Naturalism is the most experimentally verified assumption in the history of humanity.
I am perfectly willing to admit to Nagel (or anyone else) that science has assumptions about the world. I am intellectually honest enough to admit that naturalism is an assumption -- albeit a well-tested one. What I suggest is that this assumption is so intrinsic to science that to abandon it would corrupt and destroy the entire enterprise. Nagel seems to believe that if we just let the IDers in but qualify their entrance with the statement that, "there really is little evidence for their beliefs," everyone will be satisfied and life can go on. I counter that to do that we would have to sacrifice science on an altar of superstition. You cannot run an experiment assuming that there is a non-zero probability that an angel changed the results. If there is a non-zero probability that angels are interested in experimental findings, then all the experiments I have ever done are rendered uninterpretable.
Let me get some things straight, though. I don't think it is wise for scientists to push the argument of how science negates religion. I think that is a bad political choice for reasons that I have discussed before. Further, in great contrast to many of my liberal friends, I don't give a rat's ass what the IDers choose to teach their children. They are more than welcome to teach whatever perverse twaddle they choose to dignify with the term "science." In doing so they will likely render their children incompetent for anything other than an evangelical world, but so be it. I will even go so far as to say that I do not like the idea of "public" education because it puts me in the position of needing an opinion about how other people raise their children. I would much prefer a system of school choice.
However, this is the situation we are in, and we had better make the best of it. So long as we have "public" education, scientists and educators will continue to argue that ID is by definition non-science and therefore has not place in science class. Science by definition includes the naturalistic assumption. You could -- if it were legal -- teach ID in any other course you like, but not in science class. This is the essence of Stephen Jay Gould's idea of Nonoverlapping Magisteria -- a notion with which I wholeheartedly agree:
The lack of conflict between science and religion arises from a lack of overlap between their respective domains of professional expertise -- science in the empirical constitution of the universe, and religion in the search for proper ethical values and the spiritual meaning of our lives.
ID, because it requires supernatural causes, simply does not fall into the realm of science.
Science has no problem with supernatural causes. Otherwise it would be impossible to do scientific research on e.g. whether prayer helps people get better. The James Randi Foundation even offers a "one-million-dollar prize to anyone who can show, under proper observing conditions, evidence of any paranormal, supernatural, or occult power or event." All that is needed is that the hypothesized supernatural cause has an empirically observal natural effect. But then, supernatural causes that have no effects in the material world are completely superfluous.
The way I always see this discussion of "teach the Controversy" as fairly simple. I say appease the IDers, that is, if I was running a biology class, I'd start the very first lecture discussing ID. It takes 2 minutes at most. Say something along the lines of "Okay, not everyone believes evolution is random, and they feel that life was designed. This class will focus on evolution principles as there is nothing to study in ID. If ID is correct, then all there is to know about ID is that there was a creator who designed life. Now lets get on to real biology" I think that's a good description of ID, and it takes no time to describe, as there isn't "evidence" just faith.
The idea of "a naturalistic assumption", "metaphysical naturalism", is originally apologetic I believe, and frankly I find it both unsupported and unnecessary. Predictive theories are naturally [sic!] "naturalistic" by way of observable mechanisms, while unpredictive theories like "designers" can be rejected from the get go.
There could very well be posited a powerful but natural "designer" of, say, observable spacetime bubbles and their constituents and development. But it isn't predictive, "a theory that explains everything explains nothing". Standard cosmology is.
I don't think it is wise for scientists to push the argument of how science negates religion.
Similarly negation described as "philosophical naturalism" is AFAIU apologetics and besides the point. Science makes religions fact claims unnecessary, likely even wrong as far as so called supernatural 'mechanisms' goes, but it shouldn't be blamed on scientific methods and their realized "assumptions" (such as predictivity).
D'oh! "Methodological naturalism".
Will someone point out to me where the theory of evolution. or the evidence for it, relies on the assumption that there are only natural causes? Except for some bogus ones from creationists, I have never seen an explanation of evolution, or argument for it, that embeds that assumption.
For that matter, will those who think that assumption is necessary to science explain to me the procedure science should use to distinguish a natural cause from a supernatural cause?
Is Nagel honestly arguing that science is not scientific for privileging "causality" over "acausality"? It's not any old "non-physical being" that ID proposes - because most beings like "ghosts" could simply be subsumed into a natural explanation once they have laws that determine their behavior.
No, the argument here is about non-causal explanations - explanations by positing a being that is not constrained by natural law. If he's saying that, well it appears poor Nagel has reached that point where one should consider retiring before one besmirches one's early good work, n'est-pas?
Martijn: All that is needed is that the hypothesized supernatural cause has an empirically observal natural effect.
No, it needs one more step. It must have both observable effects in the natural world, and it must follow a law of causality. You can only study prayer if you not only assume that prayer "causes" healing, but also that prayer has that effect in some consistent manner - whether 1% or 99% of the time. You can't have a situation where the "theory" is that Job's God will sometimes punish you for praying in a completely arbitrary manner.
A pedantic distinction, but I feel an important one - you have to have both observations of a certain class, and a theory of a certain class.
Careful, frog. Modern physics rejects classical notions of causality. Quantum mechanics is about as ghostly a theory of physics as one can imagine.
The interesting question to those who think that science relies on "methodological naturalism" is whether we should eject QM from the realm of science? And if not, why not? What, exactly, is the procedure for dividing proposed explanations into those that are "natural" and those that are "unnatural"? There needs to be a procedure for this, if it is supposed to be part of a method. Larsson should note that anthropologists of necessity appeal to explanations that are teleological and not fully predictive, since their subject -- people -- are teleological and not fully predictable. Does that mean we should remove anthropology from the sciences, and insist that its practitioners hob-nob with artists and poets? (I half suspect they do that anyway, just for the better parties.)
But the naturalistic assumption is what makes it "natural science." If people want other assumptions, supernatural ones, there is already a classroom for that. It's called theology. If they can demand their theology in the science classroom, I propose scientists start making counter-demands. Put physics into English Lit class, pre-calculus in American History, make Earth Science a mandatory section in Phys Ed. (This last could be easier than one thinks - when I was a kid we had a "drug education" section in PhysEd.)
If churches were doing their job, they wouldn't have to try to get the government to do it for them.
zy, "But the naturalistic assumption is what makes it 'natural science.'"
Well, that's the story. But just because it is the story, doesn't mean that it is the case. In all my study of science, and of scientific arguments, not once have I ever seen that assumption used. All I ask, of those those who think there is something behind this story, is two things. First, a procedure to distinguish natural causes from other kinds of cause. And second, examples of where this assumption is used.
I lost most of my respect for Thomas Nagel when I realized that what may be his most famous paper "What is it like to be a bat?" is nothing but an extended "argument by failure of imagination." Martijn's comment is in line with the position of philosopher John Searle, that science has the conceptual power to take any phenomenon, even ID or ESP or whatever, and transform it into a natural phenomenon via its method of observation, hypothesization and testing. That is, "naturalism" isn't an assumption, it's a result. Nagel is just as guilty of incoherent reasoning as the other ID supporters, although his academic credentials may be far more distinguished.
I will add to Martijn's point that this whole discussion about whether science allows for supernatural explanations presumes without warrant that there is anything to explain. It is not as if we had a sphere of water hovering mysteriously in the air in violation of every known law. None of the IDer's arguments about the inadequacy of evolutionary mechanisms hold water. Their biology is bad, their statistics are worse. They are in the exact same boat in this respect as people like John Edwards. I don't need to reject paranormal explanation for his abilities, because those abilities have not been demonstrated through rigorous science to exist in the first place.
Science Avenger, that would be John Edward. A lot of people have lost respect recently for John Edwards, but that is an entirely different matter!
I think Jake should subtitle this post, "What is it like to be a batty ID apologist?"
It is like arguing that algebra should admit other maths that violate the transitive property because algebra assumes the transitive property. I agree that there are other maths. I agree that those maths may have meaning or usefulness. I just reject the notion that the place for them is in algebra.
A bad analogy as non-transitive algebras are also called algebra!
I'd like to address a side statement you make:
" I will even go so far as to say that I do not like the idea of "public" education because it puts me in the position of needing an opinion about how other people raise their children. "
I love the idea of public education, because the beliefs and knowledge of the people in my society dramatically affect my well being. If enough people believe vaccines cause autism, then politicians - who by the nature of our political system MUST pander - will weaken public health measures, making it far more likely that I or a loved one will be effected by some third world disease like measles or polio. If enough people don't understand the constitution, then politicians can get away with infringing upon, say, my fourth amendment granted right to safety from unwarranted search (oh wait...)...
If education were totally privatized, then it would become completely balkanised (like the internet, which is more a network of echo chambers than a "system of tubes")... and how long would it be before we completely succumbed to one sect or another of religious wackos, who believed, with absolute 100% certainty that what they believe is correct, and who are willing to kill and be killed for their beliefs?
We all have an incredible vested interest in insuring that our neighbors don't make their kids as crazy as they are. (ok, growing up in west texas made me paranoid. But having survived my first 16 years by not stating any of my views - in light of having heard repeatedly how people with my beliefs regarding, well, everything, should be denied the right to vote / expelled / imprisoned / shot... I desperately want these peoples' kids to not be like them!)
Relative to quantum mechanics, a few quotations are in order.
Richard Feynman: If you think you understand quantum mechanics, you don't understand quantum mechanics.
Steven Weinberg: Quantum mechanics is a totally preposterous theory which, unfortunately, appears to be correct.
Lawrence Krauss: Nobody understands quantum mechanics.
Having quoted the above savants, however, it is quite reasonable to assume that quantum mechanics is a methodological naturalistic theory. The reason is very simple. Unlike supernatural explanations, it is falsifiable.
1. Quantum mechanics makes predictions of the energy levels of the hydrogen atom which are verified by experimental observation. If the experimental observations differed from the quantum mechanical predictions, the theory would be falsified and would either have to be modified or discarded.
2. Quantum electrodynamics makes a prediction as to the anomalous magnetic moment of the electron which has been verified by experimental observation to 10 significant digits! As Richard Feynman pointed out, this is equivalent to measuring the distance between the Empire State Building in New York and City Hall in Los Angeles to the nearest foot. Again, if experiment and observation differed, this would falsify quantum mechanics and would require modification or discarding of the theory.
I would point out that there is nothing new here. Back in the 18th century when Newton presented his law of gravity, the notion of action at a distance was just as mysterious at that time as quantum mechanics is today. Well might Newton and Laplace have said (AFAIK they didn't) that nobody understands gravity.
What?! Methodological naturalism just means falsifiability?
That fails in both directions. Moses, rather famously, did some experiments with the Pharaoh. That story may be myth, but the god it paints is falsifiable. Does that make Yahweh natural?
Conversely, physicists occasionally come up with theories that are either untestable in theory, or untestable as of yet. Those theories are correctly criticized on such grounds. But I'd be leery of saying that those are therefore theories of the supernatural.
Maybe "causality" is too strong a word. I'm reaching here for a close link between observation and theory, sufficient to eliminate bad theory and bad observation. This linkage must include something temporal, and must be heavily constrained. It requires a theory that is self-consistent and powerful - in the end, one must be working towards a mathematical description of reality.
I don't see how ID can possibly become a mathematical theory of reality.
Let's take the most non-physical physical theory (that I know of), Max Tegmarks idea about physical reality being isometric to reality. What makes that scientific, while "supernatural" descriptions are not? Well, there are clear, mathematical constraints that can be verified or negated: if we develop the mathematics sufficiently, it should be possible to describe all possible mathematical universes that contain self-aware entities, and find out whether we a) are not in one of them, b) are in a marginal one, or c) are sitting smack dab in the center of an island of possible universes. If a), Max is wrong and we know that science can not fully describe reality - science falls silent at some point. If b), Max may be wrong, but science should still speak about everything. If c), Max is most like right.
This should be determinable solely by matching observation and theory. On the other hand, how can we possible have an ID theory without having the intelligence within the theory? Otherwise, we get into an interminable game of "God did X to test our faith", where X is whatever we please. It's not a constrained theory, amenable to mathematicization (in whatever time frame), because the essential player is undefined.
In short, it's crap. On the other hand, if ID was explicitly positing that green men from Venus directed the genesis of life - we could actually test that! We could hope to know someday, to a high degree of certainty, that that did or did not happen, since we could know a) whether they existed, b) what technology they had, c) whether they existed at the right time, d) whether their specific technology left behind evidence, evidence tightly constrained by the limits of that technology. Aka, just like what we would do if we were determining whether a stone age axe was really an axe.
If Nagel can not distinguish between theories that are mathematical, and at least create a well-defined envelope of consistency (ala QM and such), and Just-So stories (ala Mark Twain), well, he has little business philosophizing about science.
Conversely, physicists occasionally come up with theories that are either untestable in theory, or untestable as of yet. Those theories are correctly criticized on such grounds. But I'd be leery of saying that those are therefore theories of the supernatural.
This is a misuse of the term theory in science. What Mr. Russell is describing is termed a hypothesis in science. Indeed, physicists (and other scientists) ofter propose hypotheses. However, in order for a hypothesis to become a theory, it must provide predictions that are testable. In fact, the issue if testability is intimately connected with the issue of falsifiability. As I previously stated, quantum mechanics qualifies as a testable scientific theory of methodological naturalism, despite its many apparent absurdities, because it provides testable hypotheses, a couple of examples which I provided (there are hundreds if not thousands of such predictions).
An example of a hypothesis which is not yet a scientific theory is string theory, which is, in actuality, a theory of mathematics which may or may not have application to physics. There is nothing new or startling here. As examples consider Riemannian geometry which is a theory of mathematics which was shown by Einstein to be applicable to cosmology and group theory which was first shown by Wigner to be applicable to nuclear physics. Riemannian geometry and group theory are now considered scientific theories because they provided testable hypotheses which have been confirmed.
SLC: This is a misuse of the term theory in science. What Mr. Russell is describing is termed a hypothesis in science. Indeed, physicists (and other scientists) ofter propose hypotheses. However, in order for a hypothesis to become a theory, it must provide predictions that are testable.
I disagree that "hypotheses become theories". They are two different classes of ideas. String theory is a theory, not a hypothesis - a hypothesis is a slice of a theory that you test in the lab. I can hypothesize, on the basis of theory, that rats will die if I deny them food; that will not, ever become a theory but will remain a correct hypothesis, as part of a theory of alimentation.
The testable sets of predictions are the hypothesis (or hypotheses, depending on how you structure your proposal). Theory is independent of it, but must provide hypotheses (which are testable by definition) in order to be a productive and accepted theory.
As far as I know, string theory has not yet produced any hypotheses - which is why it is still a borderline case, and not a generally accepted theory. But at least it is potentially productive of hypotheses - which distinguishes it from ID, which has not, and can not, produce any useful hypotheses without specifying a designer that is constrained by the "theory" (mathematically constrained).
frog, I agree that ID is complete crap. But the criticism that it continually relocates its claimed causal mechanism into ever smaller gaps has nothing to do with some alleged distinction between causes that are "natural" or not. That precisely is the point I've been pressing. The notion that ID falls outside the realm of science because it talks about things called gods just gives its advocates a legitimate beef, that their thinking is being rejected purely on the issue of domain propriety. The simple and stronger truth is that it is crap because of the intellectual shoddiness of its arguments. Which would be shoddy regardless of whether its subject were the gods or ancient reptiles. "Methodological naturalism" continues to seem to me an unnecessary smokescreen.
SLC, Reimannian geometry and group theory both are mathematical structures that are incorporated into a variety of physical theories.
Russell: The notion that ID falls outside the realm of science because it talks about things called gods just gives its advocates a legitimate beef, that their thinking is being rejected purely on the issue of domain propriety. The simple and stronger truth is that it is crap because of the intellectual shoddiness of its arguments.
The latter point is right; but there's more to it. It's not just that they argue that the explanatory agent is a "god" -- but that it's an undefined god. If they told us it was Zeus, and he lived on Mt. Olympos, then we can go and check whether mail was being delivered there.
That's not what they do - they give us a being outside of laws, who can do anything, anywhen, anywhere. We can't pin him down - not just in gaps, but even outside the gaps. "God" could, conceivably under their system, design when he wants, and evolve when he wants. The argument is even shoddier than the details - it's shoddy in essence.
Not a question of propriety - it's a question of what kinds of arguments can actually be answered via science. Give me an actually well-defined designer, and we just reach the level of simple shoddiness - but without that, they're not even wrong.
frog writes, "It's not just that they argue that the explanatory agent is a 'god' -- but that it's an undefined god."
Yes. And their argument would be equally bogus were it an undefined particle or undefined field. See, we're in vigorous agreement about the various flaws in creationist/ID thought. Notions that are so nebulous that no one can tell what they mean are not even wrong. Yes, absolutely. Claims that are continually pushed beyond the realm of empirical investigation become no better than fantasy. Of course.
This criticism is all valid.
But none of it relies on distinguishing some causes as "natural" and other causes as "supernatural." "Ill-defined" and "continually redefined to be impossible to test" are not synonymous with "supernatural." The fact that so many modern believers in the gods practice such rhetorical trickery doesn't mean that science is somehow excluding the supernatural realm by definition. It means that believers have resorted to completely bogus reasoning to maintain their faith.
Another problem with supernatural explanations is that, as Ken Miller and Neil Tyson explain, they are science stoppers. The best example of this is Issac Newton who invoked intervention by god to preserve the stability of the solar system after he became convinced that the interaction of the planets with each other would eventually cause the system to become unstable. Thus he said that every once in a while, god had to give each of the planets a nudge to preserve stability. However, some 100 years later, the French mathematician, Laplace used the technique of perturbation theory to calculate the interactions between the planets and proved the system to be stable. When questioned by Napoleon as to what role god might play, Laplace replied that he had no need of that hypothesis. The lesson here is that Newtons' invocation of the supernatural stopped his investigation. Laplace, rejecting the supernatural explanation, was not stopped and was able to supply a methodological naturalist solution. Attached is a link to a presentation by Dr. Tyson in which this issue with Newton is discussed, starting about 9 minutes into the video.
A problem with the criticism in the post is that Nagel is a naturalist himself; he's what's known as a nonreductive naturalist, but he's very definitely a naturalist. He is not against the naturalistic assumption, with which he explicitly agrees (as long as it is not assumed to require that every form of naturalistic explanation is reducible to one form). What he is criticizing in the passage in question is a particular argument against ID -- which he thinks treats science as a presuppositionless endeavor, despite the fact that the argument being criticized itself assumes that science presupposes a kind of naturalism in order to get its conclusion. I don't think this refutation has quite as general a scope as Nagel seems to think, given that, despite being common among unsophisticated opponents of ID, it is hardly a major part of the genuinely serious arguments for keeping ID out of the classroom; but the criticism in the post misses Nagel's argument entirely. It also ignores the fact that Nagel goes on to suggest that evolutionists are better off opposing ID not on the grounds that it is "not science" but on the grounds that it is bad science; although he still thinks this runs into difficulty with the Establishment Clause.
The post also completely ignores the fact that Nagel ends by agreeing that ID is "just the latest manifestation of the fundamentalist threat" and pointing out that it's not this but the particular approach that he's criticizing. This is all quite explicit; it's unclear to me why some of the above commenters have been so quick to mouth off about an argument they haven't even bothered to read. (For those who can't access the InterWiley link, Nagel has the paper up on his own website).