« Rationality Quotes 17 | Main | Excluding the Supernatural »

September 11, 2008

Comments

Robin, I'm impressed by how you could slip in a betting markets reference in this post. Well done!

How would betting on what will be accepted centuries later work? Wouldn't effectively be betting on "This will be accepted in a n years AND I will somehow be able to collect in n years"? There is no reason to assume that "This will be accepted in a n years" and "I will somehow be able to collect in n years" are close to independent (especially for large n)", so it seems like the results could be badly bent out of shape. In particular theories directly supporting the extendibility of life will be taught out of proportion with their actually probability of being accepted (what the bets would be if we had some kind of "futurescope").

Or is there some trick to correct for this?

I'm sure you've covered this before, but how do you suggest creating a betting market where the payoff doesn't happen for centuries in a way that the odds will be meaningful?

Each one of your models has command and control written all over them...Why not teach the kids to think for themselves and then see what happens?

WTF, done much teaching? A perfect fact-free curriculum aimed at general intelligence is a lovely idea, but that's all it can ever be. If a teacher tells a kid that God made the world in seven days, a good percentage of them will still believe it when they're 30.

I saw this on TED last night and it seemed tangentially related to the main post and some of the comments. And, well, just interesting in it's own right:

http://www.ted.com/index.php/talks/jonathan_drori_on_what_we_think_we_know.html

I'm not sure academia's rejection of ID is mere snobbery. There is an inherent contradiction in what the ID people do. They use reason (which presupposes an orderly universe) to argue for a universe where God can snap his fingers and change the rules at any moment (such as the aging rate of certain rocks).

I plan on posting on long term betting markets someday. Not today though.

What I can't endorse is the paternalism of teaching an elite academic norm that intelligence design, ghosts, UFOs, etc. are too silly to even consider, in order to correct a public biased to think these options excessively likely.

I'm with you -- I think -- on the idea that confronting these public biases is the best way to correct them. If a curriculum offers an elective called "Examining The Case for Supernatural Phenomena," that's great. If taught properly, students will come away skeptical of these claims.

But if you're suggesting these concepts should be taught in a science class alongside science, you're missing the context. It's not that these concepts are too silly to consider, given unlimited time in a wide-open forum; it's that a 50-minute-per-day science class is not nearly enough time even to teach a decent amount of accepted science, let alone venture into subjects that are not science. (Not, as Nagel puts it, "dead science," but not science.) Every minute you spend on these matters is one less minute you get to impart scientific knowledge.

It's not paternal to exclude teaching German in French class. It's not paternal to exclude psychology from a math class.

In what way is it paternal to restrict science classes to theories for which there is credible empirical evidence -- along with, as ALL science classes do -- teaching the scientific method so that students can learn how to tell the difference between good science, bad science and "not science"?

Isn't the fact that children are forced to attend education, an inevitable distorter of the question here? It's obvious you ought to teach the truth. It's nowhere near so obvious you ought to force someone to hear the truth, even if you have their educational best interests at heart.

I've been thinking about this a bit lately, and I agree that Intelligent Design - the idea that life had an intelligent creator - is, or could be, a scientific hypothesis. There are definite predictions we could make from the assumption of ID; predictions that contrast with those of evolution.

But the most obvious predictions one would make from ID have been falsified. Life is full of imperfections and terrible design choices.

Nagel is almost correct. The only intellectually honest approach would be to explicitly teach atheism (which is one of the strongest conclusions of modern science) but this would be political suicide. So scientists instead use politically correct slogans such as "creationism is not science," etc. There is a certain amount of dishonesty involved but the outcome is more important. If we spent time refuting religious beliefs, UFOs, psi, alternative medicine, etc, we'd probably do more for the average student than teaching them basic science. But you couldn't do it.

I think there's a scientific case for the criticism that Darwinian evolution is not sufficient to explain all life we observe on earth. The problem is, critics generally do not offer a better theory (saying 'God did it' is obviously not scientific). But I don't think that we should hold off on criticisms of our best theory merely because there is not an alternative. After all, there are lots of things we do not understand. If it turns out Darwinism is not a sufficient explanation for the evolution of all life from a space cloud in 3 billion years, we will be like we were in 1800 and simply not know. Is that so inconceivable, that in 2008, we actually have no clue how something important works?

In an age where many serious people contemplate untestable ideas like 'many worlds', or that life is just a simulation run by aliens, thinking that something totally unknown (god, aliens) affected the evolution of life is plausible, though not provable at this time. After all, with punctuated equilibrium, you already have sky-hooks being introduced to explain gaps all the time. To say 'god did it' is pretty similar to saying 'a large event happened that we can never get a good record of'.

I don't see this affecting science adversely at all. When you read Watson's Double Helix, it is clear that evolution was irrelevant to the discovery of DNA. Jacques Monod's work follows from assuming we were designed, but constrained by laws of chemistry and physics. How many genetic differences are there between, man and his most recent ancestor? What is the overall state space that random mutations had to work upon? What is the probability, given assumptions about selection, and the number of deleterious to advantageous mutations, that a functioning person would have traversed this path? How do you correct for the anthropic bias? The list of interesting questions is long, and they are all scientific. The current thinking is, 'evolution is possible' and 'it's our only theory that isn't faith based'. But this is insufficient, because we know that a bunch of monkeys will type the complete works of Shakespeare, but we also know the odds of this are so low, a million monkeys typing every second from the beginning of the Universe would not have been enough time. Possible is not probable, and science is about probabilities, no possibilities. The fact that it's our best theory, does not make it more correct.

It is simply not true that evolution on the macro level is necessary for us to compete with the Chinese (as if it's zero sum), or build better drugs, or do molecular biology. Both sides view biology as something that at least looks designed, and uses a lot of common templates between the species (as evolution, or resource constrained designer would). Serious critics of Darwinism do not deny micro evolution, just the 'origin of species' part, or macro evolution.

What's taught now influences what is accepted centuries later. Thus feedback effects in prediction markets.

e.g. if most people now are betting on creationism (because they are creationists), then I figure it's likely creationism wins and is taught, thus significantly more likely than otherwise that it will be believed in a century (even if I thought if it weren't taught now it would perish quickly), thus I bet on creationism.

However, the claim that ID is bad science or dead science may depend ... on the assumption that divine intervention in the natural order is not a serious possibility. ...

No, that assumption is not necessary to rule out ID as science; all that is needed is a clear understanding of the modern scientific method, and ID fails. It looks like this:

Observation -->Testable Hypothesis -->Prediction -->Experiment -->Peer Review -->Reproduction of Results -->Parsimony -->Theory.

ID does not make it past step two. It proposes no testable hypotheses, no mechanisms, makes no testable predictions, has done no experiments, cannot get published in any scientific journals, and fails as the simplest explanation. It exists only as a criticism, religiously based (look at the Dover trial), of evolution. In essence ID says 'life looks complex, therefore goddidit.' The ID proponents might be better served if they went out and gathered some actual evidence, and performed some real science.

Teach ID and creationism, if there is any difference, in comparative religion where it belongs.

Ben, a "fact-free curriculum" is a straw man. A curriculum-free school is a better description of what can be done.
The Sudbury school model tries to do that. From what I've heard of this implementation (at least as of 12 or 13 years ago), they get good results when using a rule that the staff is only allowed to teach when and what the students ask them to teach.

@Peter McCluskey & Ben Jones

"Ben, a "fact-free curriculum" is a straw man. A curriculum-free school is a better description of what can be done."

I would say this issue can be approached two ways: Sudbury-esque or an Oxford way with a core curriculum in a Jacob Klein-type model. The success of each depends more on the personality and attitude of the student than the actual form, I might argue.

I would be thrilled to have a Supreme Court ruling that science and other factual subjects such as history, though not math, English and foreign languages, are by their nature contentious and as such are inappropriate subject matter for the obligatory and publicly financed education of a free people under an establishment clause. This would drive the middle class out of the public schools in droves, drive down the price of public school subsidized suburban real-estate, and focus the public schools on the basic curriculum that they have some chance of actually successfully teaching.

@ eric:

It is known as the Theory of Evolution(ToE), not 'Darwinism.' Your use of the latter term is very telling, even as you position yourself as a concerned critic of the scientific consensus. As I see it, your criticisms are 1) Punctuated Equilibrium is a sky-hook, ala Dennett; 2) Who cares whether an idea is testable?; 3) Micro evolution does not imply Macro evolution; 4) That thing with the monkeys and typewriters; 5) The ToE is really unnecessary, anyway.

My answers: 1) No, PE is just a really fast crane. Yes, the geologic record of life is incomplete, and gaps exist in the fossil record. Problem is, as soon as one gap is filled, two more magically appear, one to each side of the newly discovered transitional form. 2) But why go with something untestable, when we have a thoroughly tested Theory that has been vetted for 150 years? All that ID adds is unnecessary conjecture. 3) Micro evolution + Deep time = Macro evolution. There is no logical barrier between the two, unless you posit a young earth. 4) Yes, part of the ToE is random, as in the Sonnet Monkeys example (an argument from Creation Science, BTW), but part is not. Mutations happen randomly, but selection is a largely deterministic process. In short, bad analogy. 5) The ToE is the basis for a ton of medical research, especially in ERV's and oncology. We have seen evolution take place in the wild and in the lab, have made very useful predictions using the ToE, and have gained nothing from the ID people in the way of useful knowledge.

@ennui: Correct me if I'm wrong, but I think ID actually fails at the parsimony stage, not the testable hypothesis stage. Remember, ID proponents do in fact accept all of the biology that leads to the predictions we currently make, and thus, do have testable hypothesis.

That is, they can say, "I, as an intelligent design proponent, predict that this offspring will have genes from its parents."

What they do not have is a currently-testable hypothesis for the assumptions they want to add to the existing theory of how life came to be the way it is now. By adding an additional assumption that is not necessary to generate current successful predictions, the failure is at parsimony.

ennui: ID does not make it past step two. It proposes no testable hypotheses

If you thought there was an intelligent designer, wouldn't you predict that nature would run in a very harmonious way, with no pathogens, parasites, evolutionary arms races, cruelty, intra-genomic conflict, etc? Aren't these predictions ID makes - predictions which have however been falsified?

Re: Parsimony

It is true that the critics of ID have already proposed testable hypotheses, experiments, predictions, etc. But no ID proponent has muttered anything other than 'irreducible complexity.' They point at the immune system, blood clotting, the human eye, bacterial flagellum, et al, and simply say that a designer must have made them. Then other scientists go out and debunk the baseless claims.

I wouldn't predict harmony in nature as a result of design, unless I also assumed something harmonious about the designer. ID says nothing, nothing about who or what the designer is, or by what mechanisms it operates.

Serious critics of Darwinism do not deny micro evolution, just the 'origin of species' part, or macro evolution.

The "falling" part is the theory of gravity. The "infection" part is germ theory. The "origin of species" part IS the theory of evolution.

In other words, these critics keep the science that allows their dogma to be preserved and haphazardly throw out the other parts. This is not science, and thus, "serious critics" are not serious.

Denying "macroevolution" (as the term is incorrectly used) is denying the theory of evolution. The concept of species is post hoc. There is no algorithm to determine a species, and, thus, there is no defined barrier through which offspring passes from one to the other. Creating this barrier to justify your personal notion is unscientific.

There are many outstanding, interesting scientific questions regarding evolution, but none that threaten it without also overturning much of what we know about many other branches of science. If a "serious critic" of macroevolution were to present compelling evidence, it would be one of the most profound discoveries in history. That would be quite exciting, but I'm not holding my breath.

"Posterity futures - teach options to the degree betting markets say they'll be accepted centuries later. "

This it seems to me is asking for trouble. I am not sure how you intend these markets to work, but wouldn't it reduce to which party has greater leverage in the market? Take ID vs evolution - let's say a market was set up in 1900s to decide which one to teach. I have a suspicion that in this case we'd be teaching ID in school (or even genesis) for many centuries ...

"What I can't endorse is the paternalism of teaching an elite academic norm that intelligence design, ghosts, UFOs, etc. are too silly to even consider, in order to correct a public biased to think these options excessively likely".

I don't understand why you are grouping ID, ghosts, UFOs, etc. together. Could you explain why these are in the same sentence, other than the fact that the scientific (as opposed to an 'elite academic norm') is opposed to these 'options'? ID is a theory with clear claims and explanations. It is not clear to me what your 'theory' of 'ghosts' is?

If you thought there was an intelligent designer, wouldn't you predict that nature would run in a very harmonious way, with no pathogens, parasites, evolutionary arms races, cruelty, intra-genomic conflict, etc? Aren't these predictions ID makes - predictions which have however been falsified?

Movies and books are intelligently designed. Do books and movies have conflicts, cruelty, suffering, etc.?

I think this particular argument against intelligent design is a non-starter.

ennui: I don't see how TOE has been totally vetted. Darwin thought the ontology recapitulates phylogeny was his best data point, turned out wrong. Miller's experiments about the origin of life from primordial soup were exciting, but a dead end. Fossils of extinct organisms that are homologous to current ones really do not prove much, which is why untestable adjuncts, like PE, panspermia, and lateral gene transfer have been proposed. If science were about a popularity contest, I agree, the ToE passes, but that's not the metric.

Evolution is attractive to many, paradoxically, for the same reason religion is: because it 'explains' everything. Humans really want to know, they hate thinking there is something big out there we don't understand even in principle. But I think, like Stephen Wolfram's New Kind of Science, it creates a framework that is untestable. You would think that after playing with drosphilia for decades, we could create a new species, but speciation has very few actual observations (flowers), and these don't easily generalize to genera.

If you found a space ship on the dark side of the moon, and it looked like something from sci-fi movie, and from sattelite photos were certain it was not from the Chinese or Russians, you don't need to say where it came from to scientifically assert that the ToE is insufficient to explain this finding.

I'm an atheist. I also think variation plus random selection is implausible for creating all life as we know it on Earth. I don't think we need evolution, without which we stop doing science and start believing in anything. I'm a skeptic, that is all.

Movies and books are intelligently designed. Do books and movies have conflicts, cruelty, suffering, etc.?
Meaningless question. Asserting that something was designed by an intelligent being tells us nothing if we don't know what that intelligent being was trying to accomplish.

This is why programs like SETI do not look for "maximization" or "minimization" in the signals they receive, because the application of those concepts depends entirely on unknown standards. Instead, they look for improbability - signals that, given our understanding of astronomical physics, are unlikely to be generated by simple events.

The fact that postulating intelligence doesn't let us exclude any possibility of the design is part of what makes creationism not-science. It explains everything, which is logically equivalent to saying that it explains nothing.

You would think that after playing with drosphilia for decades, we could create a new species,
It's already happened with mosquitoes. Many plants, quite a few of which are cultivated. You just don't know anything about them.

You're not a skeptic. You're simply ignorant.

Allan's and ennui's thoughts should be combined.
If the "Intelligent Designer" is considered to be utterly inscrutable, then "Intelligent Design" makes no predictions.
If the Intelligent Designer has designed life with some particular purpose in mind, predictions would follow from the purpose. Personally, I can't think of any purpose which would be served by having so many parasites, particularly not the layers of parasites on parasites. But there are others more imaginative than I.

Nobody who isn't a particular type of religious person takes that stuff seriously, and no university worthy of the name studies it. That's enough to convince me that it's not science--it's religious doctrine disguised as science. Why is it "paternalistic" to treat this (along with the ghosts and UFOs) as unworthy of discussion in science courses?

Why do yourself supporting the "democratic idealism" option (teach whatever the public wants taught, including some creationism)? Would you support that approach to your own work also (college and graduate level economics education and research)?

should read "why do you SEE yourself supporting ..."

Meaningless question. Asserting that something was designed by an intelligent being tells us nothing if we don't know what that intelligent being was trying to accomplish.

You seem to think that I am arguing for intelligent design. I'm merely pointing out that one of the arguments against it posted here doesn't pass muster.

Isn't it a good thing if we point out flaws in arguments, regardless of whether they are for "our side" or not?

What I can't endorse is the paternalism of teaching an elite academic norm that intelligence design, ghosts, UFOs, etc. are too silly to even consider, in order to correct a public biased to think these options excessively likely.
A biology professor believes she is descended from an ape. The creationist believes he was created by a god.
Clearly, the biologist, with her high-falughtin' theories of her origin is the elitist.

An astronomer believes interstellar travel is far too resource intensive to justify myriad visits by aliens. A Ufologist believes in thousands of visits from aliens who are constantly revealing themselves to 'eye-witness' observers, but leaving nothing more than anecdotes.

Clearly, the astronomer, with his uppity assertion that maybe we're not that fucking important, is the elitist


I don't see how TOE has been totally vetted. Darwin thought the ontology recapitulates phylogeny was his best data point, turned out wrong.

But the essential insight, which explains why it looks like ontology recapitulates phylogeny --that ealier developmental pathways are more highly conserved-- is a confirmation of evolutionary predictions.

Miller's experiments about the origin of life from primordial soup were exciting, but a dead end.

Rumors of the death of abiogenesis research have been greatly exaggerated. It's a blossoming area of interdisciplinary research, and the Miller-Urey experiments are its foundation, not a dead end.

Fossils of extinct organisms that are homologous to current ones really do not prove much, which is why untestable adjuncts, like PE, panspermia, and lateral gene transfer have been proposed.

What a confusing jumble of half-truths and conflated concepts! While it is true that confirmation of a given prediction of a theory does not, in itself, prove the theory true, the homologies in the fossil record and the nested heirarchies they reveal are a fairly stunning confirmation of a central prediction of evolutionary theory. As for your "untestable adjuncts," PE is a hypothesis about the fossil record, specifically why it is that the record does not, in most cases, reveal complete, finely graded sequences of intermediate forms. It depends for its logic on what we already know about speciation and macroevolution. Panspermia is a red herring in the context, as there are no phenomena that call for it as an explanation. It's a "what if" kind of idea, talked about more at cocktail parties than in the lab. And lateral gene transfer is common in bacterial populations --it's a known phenomenon, not an "untestable adjunct," and I really cannot fathom what you're getting at by including it.

If science were about a popularity contest, I agree, the ToE passes, but that's not the metric.

Nice insinuation, but it misses the mark. The question is, why is evolutionary theory so popular among those who study it, and understand it the best?

You would think that after playing with drosphilia for decades, we could create a new species, but speciation has very few actual observations (flowers), and these don't easily generalize to genera.

Why would I think that? Speciation is well-understood, and there are ample observations. Look up Ring Species to get a sense of a spacial relationship between populations analagous to the temporal relationship between past and present populations. That our observations of speciation "don't easily generalize to genera" is word salad. You'll have to clarify what this means to you.

It might be worth noting here that if you take the Simulation Argument seriously, you believe there is a significant chance of some form of design. Given our history, "bored gamer" seems like a more likely designer than "omnibenevolent deity."

Of course, that passes the recursive buck to the origin of the simulation. Turtles, all the way down.

I have a page on the remaining potenially-viable intelligent design hypotheses:

http://originoflife.net/intelligent_design/

The page lists: Panspermia, Simulism, Optimisationverse and The adapted universe. I see no reason to cover any of these in schools.

They think, Anybody who is willing even to consider supernatural explanations is living in the past.

Has Thomas Nagel bothered to explain what constitutes supernatural explanation? I see the term "supernatural" as meaning nothing more than "incomprehensible", and thus "supernatural explanation" sounds to me like an oxymoron. Supernaturality doesn't really explain, it obscures.

A bit off-topic. What would be the best article for a layman to read about betting markets?

There are different versions of creationism. To the extent creationism makes wrong predictions a la theology before Darwin, it is bad science; to the extent creationism falsely tries to escape from its wrong predictions and make new post facto ones that match previous observations, it is dishonest science; to the extent creationism tries to avoid making predictions, it is not science at all, but is bad rationality.

So when you're considering an "intelligent design" theory, you first ask: "Are you making the same sort of predictions that theologians made before they knew the facts, are you making your theory fit the facts we now know, or are you avoiding making any testable predictions?"

In the first case, creationism is examined as a past hypothesis whose predictions were disproved, and therefore discarded, in a valid example of the scientific method in action.

In the second case, creationism is examined alongside N-Rays as an example of distortion of the scientific method: you must make your predictions before you see the facts, and discard your theory after it is falsified instead of trying to rescue it.

In the third case, creationism is a priori excluded from scientific consideration, but can be taught in philosophy classes as an exemplar of bad rationality, "belief in belief", etc.

Matthew: Movies and books are intelligently designed. Do books and movies have conflicts, cruelty, suffering, etc.?

No, they just have ink on paper.

And, uh, pixels on screens, or whatever movies have.

For the record, I was proposing that ID is an example of Eliezer's first type: many of the predictions an honest ID advocate would make (if he didn't know those predictions have already been falsified) have already been falsified.

But perhaps it's silly of me to think like this; after all, modern ID people are not actually making those predictions.

Eliezer, it seems clear to me that Nagel and the schools he discusses are talking about having classes discuss the general claim that someone designed something important in our evolutionary history. This general claim may be consistent with many specific claims, but the proposal isn't to allow or suggest any more specific claim.

ID *is* too silly to consider. Most versions of ID are too much like phlogiston. Thus they are too silly to be taught in a Science class.

Those that are not are very trivially refuted by evidence.

Eric, WRT fruit flies, you might want to take a look at this: http://www.talkorigins.org/faqs/faq-speciation.html

Talkorigins is a good site to explore. You might not believe this, but the arguments that you have chosen are familiar to all involved in these debates, and it tells us something about your sources. You should try to avoid the creationist propaganda.

Neanderthal Joke
A team of evolutionist/paleanotologist stumble upon an airplane parked in the bush. There is a campfire with pig bones scattered around..The "scientist" rush back to publish their Find in the Scientific Journals..."There was no evidence of humans at the Site, but there were homo porcinus sapien bones and technology equal to /better than the average homo sapien sapien"..Then PBS and the Discovery Channel produce documentaries based on articles in National Geograpic.. entitled "Darwin was Right, Pigs Can Fly."

Before deciding which theory should be learned at schools and how it is best to be done, shouldn’t we first consider the need to study any of these theories?

When children are taught to read, write and count, they acquire skills which will be really necessary to them in their adult life. Now why do we really want them to learn about the evolution or ID?

The real answer, I suspect, is that teaching evolution or ID in schools has a lot more to do with politics than with science. Each side wants to indoctrinate children with their own political values, so that when they grow up they learn to recognize the enemy – “godless liberals” or “brainless conservatives”.

Personally, though I have a firm opinion on which of the two is the correct theory, I would prefer either to leave them both out of the classroom or to allow both sides to present their arguments. Schools should not be the place for political indoctrination.

Robin: One potential difficulty with using betting markets is distinguishing between a theory that will undergo minor revisions and a theory that will be thrown out as wholly nonsense. Both might look the same to a futures market, or more similar than we'd like.

If people want to learn about these things, have them in "Crazy Speculation" class or "X Files" class or something; that'd be fine. I mean, if we're willing to go down that road, why not teach FlyingSpaghettism, or AliensAreKeepingUsInAPetriDishIsm, or ThisIsTheMatrixism?

The set of possible theories which don't currently have scientific evidence to back them
is unlimited. We could pick any random plot from a sci-fi novel and start "teaching" it. Are we going to teach them all? Pick some at random? No? Then which subset? If they're all equally unlikely, then how can we really decide on that?

I'm a fan of the "Wild and Unsupported but Fun To Think About" class idea.

"I'm an atheist. I also think variation plus random selection is implausible for creating all life as we know it on Earth."

This pretty much sums up my position. I am an atheist and so were my parents and siblings. I have never been to church except for weddings and funerals. I am also a software developer by profession and understand the power of a simple algorithm applied recursively. I have also done a graduate level Biotech course and looked at my share of gene sequences.

And yet I don't believe random mutation and death is an adequate explanation for the fact of life on Earth. It is part of it, but there is something else we're missing IMHO.

And yet I don't believe random mutation and death is an adequate explanation for the fact of life on Earth.

That's just a concise restatement of Natural Selection. It's an obvious straw man if you mean it to stand in for the entirety of modern evolutionary theory, which encompasses more mechanisms than variation and selection.

there is something else we're missing IMHO.

And that's it? You have nothing else to say on the matter? You can't articulate even one reason why you think this? Give me some detail about an observation in the biological world you do not think is adequately explained by the current state of the theory.

The comments to this entry are closed.

Less Wrong (sister site)

May 2009

Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30
31