One of the problems that interests me is how best to learn the truth in controversial matters. There seem to be several approaches that different people use, but which I see as problematic. One is to simply go along with what your peers believe. This provides obvious social benefits but for people here who are interested in "overcoming bias" it requires some justification. One can in fact make a case that the majority view is often right based on the
Eliezer: Sorry, I did not mean to say that no individual could reliably distinguish good from bad science (at least in particular area). I meant to say that you can't tell which individuals to trust--even including yourself. (Linguistically speaking, I was using "reliably" to modify "individuals" in a grouped sense, not a distributive one, which is probably the less available interpretation of the two, and still imprecise.)
"On what? The Modesty Argument?"
Sorry, no. I mean how you compare to de Grey in your opposition to the consensus views of scientists in your field. Would Hal's argument that "on that basis I would predict that his project will not experience much success" apply equally to your research? I would say yes--the reason I support the SIAI nonetheless is that my limited expertise points me towards trusting my own judgment in the matter over that of the scientific consensus. Since I don't have as much expertise in biology, I must depend more on the scientific consensus in my evaluation of de Grey.
My post today about "To Win Press, Feign Surprise," and other posts I will make, show there are identifiable biases in academic consensus. So even better than accepting the academic consensus is to correct that consensus for these identifiable biases.
"To be able to identify when a framer is likely to be right, you have to be able to distinguish good science from bad, which is paramount to becoming an expert yourself. The scientific process can do this over time. Individuals can't reliably do this."
How does this follow? The social method of trusting individuals doesn't work, but that's because you can't tell which individuals to trust. I see no reason to assume that no individual human expert is capable of reliably distinguishing good science from bad, especially given that science is composed wholly of human beings, and does no thinking which occurs outside the confines of a human skull.
"I wonder how this compares to Eliezer's position."
On what? The Modesty Argument? Remember, I'm *against* Modesty.
I think the logical mistake (that positions or ideas that are shared by a consensus are necessarily superior to those held by a few) is made because known proven ideas are (almost by definition) are shared by a consensus, but that doesn't mean the converse is true.
When an idea has been proven by this I mean that scientists have accumulated deep understanding of how something works that is the result of feedback, for instance in physics lots of particle experiments meant that the model of how particles work was refined over time with new insights being added as the feedback dictated. The result is a model that is shared by almost everyone working in the field, because errors could easily be tested and proven wrong.
Where the idea is not proven (i.e. there has not been this gradual idea,test,verify idea,test cycle perhaps because it is impossible as with the anthropogenic global warming hypothesis) we should remain sceptical even if there is scientific consensus, for scientists are just as fallible (maybe even more so because of groupthink) than the rest of the population as evidenced by others above. Maybe they are fallible for different reasons (over-respect of other disciplines, peer review process stifling debate, need to get tenure etc), but they since they are human they have bias's. Even the greatest of them (Einstein) had strong convictions (god does not play dice) that turned out to be wrong.
How do we decide then whether and idea (if is not testable) is correct or not? In some cases Occams razor will work (for instance on religion) otherwise I take a sceptical approach - if it can't be tested it is unlikely to be an effect of any significance worth spending time evaluating.
Douglas, it's true, I was a bit contradictory there. When there is an ideological component to the controversy, and the layman studying it has an ideological position, I believe that they will usually find on studying the issues that they agree with the side supported by their ideology. At the same time, independent thinkers are in my experience more prone to extreme positions, which do not have a good historical record of coming true.
Matthew, the case of de Gray is a good example of the kind of challenge I am talking about. Via scholar.google.com I found an essay of his at:
in which he basically states that the scientific consensus is against him. Without getting into the question of whether his views are "pseudoscience" or not, on that basis I would predict that his project will not experience much success. It sounds like the gerentology community does not think that dramatic increases in mouse lifespan are possible with current and near future technology.
That's a pretty easy conclusion to reach. The question is whether reviewing his writings and those of his critics in more detail is worthwhile in terms of improving accuracy. Biology is extremely complicated and I am skeptical that a layman can learn enough to improve on the opinions of professional biologists. Add the fact that this is an emotional area where people have many latent hopes and fears and the prospects of a successful "think for yourself" operation look exceptionally dim.
Matthew: Well, in the extreme case, there's anti-gravity devices and alternate physics, and that sort of crackpot stuff. In the middle are people like Michael Chrichton (testifying to Congress against global warming). But then there are harder cases, like Aubrey de Gray. Here[1] are three papers claiming him of practicing pseudoscience.
Basically, if people don't appear to be reasonable, they probably aren't, and their support of a position isn't evidence for it. If they speak in vague abstractions and can't be pinned down in discussion or questioning, same thing. If they don't even bother to appeal to hard evidence, in areas where there is relevant evidence, same thing. There's more on Wikipedia under "Pseudoscience". This type of thing can be evaluated by people without a lot of scientific knowledge.
Similarly, claims that de Grey is doing pseudoscience are hard to evaluate, because he speaks concretely, cites particulars, references (and coauthors) papers, and doesn't immediately appear to be evasive in discussion (at least to my lay ears). So it takes people with much more expertise to evaluate his credibility. (And even then, it's hard for laypeople to distinguish infighting from legitimate complaint. Scientists don't have a reputation for being open-minded.)
"Now, I concede that in some controversies the "framers" are doing pseudo-science, and so to be able to find the truth you just have to be able to identify pseudo-science, which is usually much easier than becoming an expert."
I don't mean to understate the difficulties in discovering the academic consensus. Indeed it is often a challenge. I am looking for endorsement or criticism of this general approach. If it seems useful, ideally what I would like to see is greater efforts on the part of scientific and academic communities to improve their efforts to communicate established conclusions to the public.
Part of the difficulty in a third party doing it will be accusations of bias. Take the global warming controversy, where opponents have fought for many years against the claim that the scientific consensus favors human-caused global warming. The Intergovernmental Panel on Climate Change has issued a number of detailed reports supporting this position, but its governmental structure and the United Nations connection are targeted by conservatives who oppose its conclusions. The web site realclimate.org says it exists to promote the scientific consensus and oppose naysayers, but the other side found that its web address was registered by an environmental organization and claims that the site is biased. Skeptics also gained mileage from a widely circulated letter claiming opposition to global warming theories signed by "top scientists" and authored by a former president of the prestigious NAS. Global warming supporters have followed up with rebuttals and their own letters, but the net result is to muddle the issue.
On this issue I'd say the scientific community has done a better job of communicating in recent years, and it seems clear now that in fact proponents were right all along, and the scientific consensus has in fact been accurately reported by the IPCC and sites like RealClimate. Statements by other prestigious academic bodies, and articles in top science journals like Science and Nature, have been directed explicitly at this one goal, of communicating the consensus. This effort has been beneficial for society and it seems to be having an effect in improving understanding of the issue. It would be helpful if academic bodies would take on the burden of communicating the scientific consensus carefully and publicly as one of their main reasons for existence.
HF, you have two complaints about people who try to think for themselves. One is that they don't seem to add information: you can predict their conclusions. The other is that they reach "very unusual beliefs," where you seem to be complaining that they've only added noise. Could you elaborate on these two failure modes? I feel like your essay doesn't distinguish between them.
Hal Finney finds thinking for himself to be an unreliable way to find truth. One of the problems that interests me is how best to learn the truth in controversial matters. . . One is to simply go along with what your peers believe. . . Another is to t...
"controversies that occur primarily because certain belief systems are very efficient at filtering out facts that contradict their assumptions"
I think you could be right for controversies that meet that criterion. (I'd say many controversies that stay mostly within academia wouldn't.) But even still, to be confident that you're actually finding truth, you'd have to have a good, objective reason to believe that you're less biased and more rational that other participants in the debate. Even being disinterested might not be enough. How do you know when such a belief isn't self-serving?
On a different note, it strikes me that there are many controversies where the evidence is truly inconclusive, and the correct position is to sit on the fence.
"Actually, the scientific 'leaders' were generally 'fillers'"
Here again, there's the problem that most "framing" type people are wrong about most frames (not to diminish their importance too much). To be able to identify when a framer is likely to be right, you have to be able to distinguish good science from bad, which is paramount to becoming an expert yourself. The scientific process can do this over time. Individuals can't reliably do this.
Now, I concede that in some controversies the "framers" are doing pseudo-science, and so to be able to find the truth you just have to be able to identify pseudo-science, which is usually much easier than becoming an expert.
"Matthew, are you asserting that a layman living in those times could, through a reasonable amount of open-minded investigation, discover the truth more often than not?"
Yes, a layman back then could do so with some difficulty. And you can do it today much more easily in the information age today, on matters of scientific controversy such as population and gender distributions of IQ, the existence of parapsychological phenomena, and other controversies that occur primarily because certain belief systems are very efficient at filtering out facts that contradict their assumptions. But it has to begin with a sincere desire to ascertain the truth, rather than a sincere desire to have one's biases confirmed. . .
"The fact that the scientific consensus has been wrong, and that open-minded investigation by scientific leaders gradually shifted that consensus, doesn't mean that open-minded investigation is a good *individual* strategy for finding truth."
Actually, the scientific "leaders" were generally "fillers" (to reiterate a recent post here) and sided with the incorrect consensus in these instances, while the new "framers" forged ahead in the face of ridicule and dismissal from most quarters.
We should also look very carefully at the concept of "leadership" in science, and whether it is something we want to encourage or not. . .
pdf, yes, my question can be rephrased: how large a market is there for truth in controversy? If someone else provided the product would Hal buy and use it, or does he mainly only value academic consensus when he figures it out for himself? How about the rest of us?
Eliezer: Sorry, I did not mean to say that no individual could reliably distinguish good from bad science (at least in particular area). I meant to say that you can't tell which individuals to trust--even including yourself. (Linguistically speaking, I was using "reliably" to modify "individuals" in a grouped sense, not a distributive one, which is probably the less available interpretation of the two, and still imprecise.)
"On what? The Modesty Argument?"
Sorry, no. I mean how you compare to de Grey in your opposition to the consensus views of scientists in your field. Would Hal's argument that "on that basis I would predict that his project will not experience much success" apply equally to your research? I would say yes--the reason I support the SIAI nonetheless is that my limited expertise points me towards trusting my own judgment in the matter over that of the scientific consensus. Since I don't have as much expertise in biology, I must depend more on the scientific consensus in my evaluation of de Grey.
My post today about "To Win Press, Feign Surprise," and other posts I will make, show there are identifiable biases in academic consensus. So even better than accepting the academic consensus is to correct that consensus for these identifiable biases.
"To be able to identify when a framer is likely to be right, you have to be able to distinguish good science from bad, which is paramount to becoming an expert yourself. The scientific process can do this over time. Individuals can't reliably do this."
How does this follow? The social method of trusting individuals doesn't work, but that's because you can't tell which individuals to trust. I see no reason to assume that no individual human expert is capable of reliably distinguishing good science from bad, especially given that science is composed wholly of human beings, and does no thinking which occurs outside the confines of a human skull.
"I wonder how this compares to Eliezer's position."
On what? The Modesty Argument? Remember, I'm *against* Modesty.
The behavior or methods of argument are often key indicators of the quality of the argument.
Is precision and accuracy of measure recognized?
Is it the issue or the people (consensus) that is used to support an idea?
Is there an appropriate intellectual integrity (no common logical falsities, especially hidden ones)?
Judgmental or opinion? Hubris? Tone? - many more.
Hal, yes, a better solution would be for the academic communities to themselves summarize their consensus.
A bit off topic, but...
"in which he basically states that the scientific consensus is against him"
I wonder how this compares to Eliezer's position.
I think the logical mistake (that positions or ideas that are shared by a consensus are necessarily superior to those held by a few) is made because known proven ideas are (almost by definition) are shared by a consensus, but that doesn't mean the converse is true.
When an idea has been proven by this I mean that scientists have accumulated deep understanding of how something works that is the result of feedback, for instance in physics lots of particle experiments meant that the model of how particles work was refined over time with new insights being added as the feedback dictated. The result is a model that is shared by almost everyone working in the field, because errors could easily be tested and proven wrong.
Where the idea is not proven (i.e. there has not been this gradual idea,test,verify idea,test cycle perhaps because it is impossible as with the anthropogenic global warming hypothesis) we should remain sceptical even if there is scientific consensus, for scientists are just as fallible (maybe even more so because of groupthink) than the rest of the population as evidenced by others above. Maybe they are fallible for different reasons (over-respect of other disciplines, peer review process stifling debate, need to get tenure etc), but they since they are human they have bias's. Even the greatest of them (Einstein) had strong convictions (god does not play dice) that turned out to be wrong.
How do we decide then whether and idea (if is not testable) is correct or not? In some cases Occams razor will work (for instance on religion) otherwise I take a sceptical approach - if it can't be tested it is unlikely to be an effect of any significance worth spending time evaluating.
Douglas, it's true, I was a bit contradictory there. When there is an ideological component to the controversy, and the layman studying it has an ideological position, I believe that they will usually find on studying the issues that they agree with the side supported by their ideology. At the same time, independent thinkers are in my experience more prone to extreme positions, which do not have a good historical record of coming true.
Matthew, the case of de Gray is a good example of the kind of challenge I am talking about. Via scholar.google.com I found an essay of his at:
http://www.liebertonline.co...
in which he basically states that the scientific consensus is against him. Without getting into the question of whether his views are "pseudoscience" or not, on that basis I would predict that his project will not experience much success. It sounds like the gerentology community does not think that dramatic increases in mouse lifespan are possible with current and near future technology.
That's a pretty easy conclusion to reach. The question is whether reviewing his writings and those of his critics in more detail is worthwhile in terms of improving accuracy. Biology is extremely complicated and I am skeptical that a layman can learn enough to improve on the opinions of professional biologists. Add the fact that this is an emotional area where people have many latent hopes and fears and the prospects of a successful "think for yourself" operation look exceptionally dim.
Matthew: Well, in the extreme case, there's anti-gravity devices and alternate physics, and that sort of crackpot stuff. In the middle are people like Michael Chrichton (testifying to Congress against global warming). But then there are harder cases, like Aubrey de Gray. Here[1] are three papers claiming him of practicing pseudoscience.
http://www.technologyreview...
Basically, if people don't appear to be reasonable, they probably aren't, and their support of a position isn't evidence for it. If they speak in vague abstractions and can't be pinned down in discussion or questioning, same thing. If they don't even bother to appeal to hard evidence, in areas where there is relevant evidence, same thing. There's more on Wikipedia under "Pseudoscience". This type of thing can be evaluated by people without a lot of scientific knowledge.
Similarly, claims that de Grey is doing pseudoscience are hard to evaluate, because he speaks concretely, cites particulars, references (and coauthors) papers, and doesn't immediately appear to be evasive in discussion (at least to my lay ears). So it takes people with much more expertise to evaluate his credibility. (And even then, it's hard for laypeople to distinguish infighting from legitimate complaint. Scientists don't have a reputation for being open-minded.)
"Now, I concede that in some controversies the "framers" are doing pseudo-science, and so to be able to find the truth you just have to be able to identify pseudo-science, which is usually much easier than becoming an expert."
Can you elaborate on this?
I don't mean to understate the difficulties in discovering the academic consensus. Indeed it is often a challenge. I am looking for endorsement or criticism of this general approach. If it seems useful, ideally what I would like to see is greater efforts on the part of scientific and academic communities to improve their efforts to communicate established conclusions to the public.
Part of the difficulty in a third party doing it will be accusations of bias. Take the global warming controversy, where opponents have fought for many years against the claim that the scientific consensus favors human-caused global warming. The Intergovernmental Panel on Climate Change has issued a number of detailed reports supporting this position, but its governmental structure and the United Nations connection are targeted by conservatives who oppose its conclusions. The web site realclimate.org says it exists to promote the scientific consensus and oppose naysayers, but the other side found that its web address was registered by an environmental organization and claims that the site is biased. Skeptics also gained mileage from a widely circulated letter claiming opposition to global warming theories signed by "top scientists" and authored by a former president of the prestigious NAS. Global warming supporters have followed up with rebuttals and their own letters, but the net result is to muddle the issue.
On this issue I'd say the scientific community has done a better job of communicating in recent years, and it seems clear now that in fact proponents were right all along, and the scientific consensus has in fact been accurately reported by the IPCC and sites like RealClimate. Statements by other prestigious academic bodies, and articles in top science journals like Science and Nature, have been directed explicitly at this one goal, of communicating the consensus. This effort has been beneficial for society and it seems to be having an effect in improving understanding of the issue. It would be helpful if academic bodies would take on the burden of communicating the scientific consensus carefully and publicly as one of their main reasons for existence.
HF, you have two complaints about people who try to think for themselves. One is that they don't seem to add information: you can predict their conclusions. The other is that they reach "very unusual beliefs," where you seem to be complaining that they've only added noise. Could you elaborate on these two failure modes? I feel like your essay doesn't distinguish between them.
Apparent Truth
Hal Finney finds thinking for himself to be an unreliable way to find truth. One of the problems that interests me is how best to learn the truth in controversial matters. . . One is to simply go along with what your peers believe. . . Another is to t...
"controversies that occur primarily because certain belief systems are very efficient at filtering out facts that contradict their assumptions"
I think you could be right for controversies that meet that criterion. (I'd say many controversies that stay mostly within academia wouldn't.) But even still, to be confident that you're actually finding truth, you'd have to have a good, objective reason to believe that you're less biased and more rational that other participants in the debate. Even being disinterested might not be enough. How do you know when such a belief isn't self-serving?
On a different note, it strikes me that there are many controversies where the evidence is truly inconclusive, and the correct position is to sit on the fence.
"Actually, the scientific 'leaders' were generally 'fillers'"
Here again, there's the problem that most "framing" type people are wrong about most frames (not to diminish their importance too much). To be able to identify when a framer is likely to be right, you have to be able to distinguish good science from bad, which is paramount to becoming an expert yourself. The scientific process can do this over time. Individuals can't reliably do this.
Now, I concede that in some controversies the "framers" are doing pseudo-science, and so to be able to find the truth you just have to be able to identify pseudo-science, which is usually much easier than becoming an expert.
"Matthew, are you asserting that a layman living in those times could, through a reasonable amount of open-minded investigation, discover the truth more often than not?"
Yes, a layman back then could do so with some difficulty. And you can do it today much more easily in the information age today, on matters of scientific controversy such as population and gender distributions of IQ, the existence of parapsychological phenomena, and other controversies that occur primarily because certain belief systems are very efficient at filtering out facts that contradict their assumptions. But it has to begin with a sincere desire to ascertain the truth, rather than a sincere desire to have one's biases confirmed. . .
"The fact that the scientific consensus has been wrong, and that open-minded investigation by scientific leaders gradually shifted that consensus, doesn't mean that open-minded investigation is a good *individual* strategy for finding truth."
Actually, the scientific "leaders" were generally "fillers" (to reiterate a recent post here) and sided with the incorrect consensus in these instances, while the new "framers" forged ahead in the face of ridicule and dismissal from most quarters.
We should also look very carefully at the concept of "leadership" in science, and whether it is something we want to encourage or not. . .
pdf, yes, my question can be rephrased: how large a market is there for truth in controversy? If someone else provided the product would Hal buy and use it, or does he mainly only value academic consensus when he figures it out for himself? How about the rest of us?