4 ms·

I see what you mean but I think you are overlooking how easy it is to muddy the waters. You come up with a good evidence of evolution. Great: but it will take t

by nanairo 16y ago

I see what you mean but I think you are overlooking how easy it is to muddy the waters. You come up with a good evidence of evolution. Great: but it will take the anti-evolution just one guy that makes a somehow plausible alternative scenario, or finds one whole in your argument and suddenly people are back to square one. Despite we are _not_ back on square one. Between a theory that almost works (like evolution) and one that is pure conjecture, the fact that there's a whole in the former all it shows is that we need to improve the theory. It is the same thing in climate modelling. It takes a nobody to say something which sound somehow plausible for the scientist to need to prove themselves again, often taking several years and a lot of effort. Most scientist, including at the IPCC, do describe hypothetical scenarios with all their doubts, including error bars etc... They also state all their assumptions, and their full method. Then one guy says: "Hey, it's been crazy cold in England this winter... so much for global warming", or "between 2001-2002 the global temperature went down" and suddenly everyone feel they are back to "we don't know". But why? The data from the IPCC is the result of many different models, with many different assumptions. The opposition has provided no better model: all they need to do to keep the status quo is point out some small unknown or flaw. That's WRONG! And btw, climate science is possibly one of the hardest disciplines around: you are making a prediction in the future of a chaotic system. But the anti-science groups can find objections just as easily to pretty much any bit of science they dislike.

zeteo 16y ago
It's true that there's a certain amount of fuzziness, but also there's a gradient of trust between 100% and zero. Predictions that consistently come true start to dwarf alternative explanations after a while. Darwin said that the great apes are our closest living relatives, and 150 years later research finds the genetic code similarities in the 90+% range. That's pretty strong. On the other hand, it is indeed possible that the field of study is a chaotic system, and accurate predictions impossible. But then I will argue that's not a branch of natural science, but rather of history. A good historian can tell you in small details the reasons why Lincoln got elected in 1860, but is as clueless as the next guy as to the next presidential election. Would you trust him if he said Republicans are always best? Similarly, there's a lot of great work done in historical climatology, with pollen analysis, ice cores etc. and we can find out a great many things about the past. Computer models can be built to fit this history perfectly (as I'm sure that even I can code some curve that spits out the order of Democrat and Republican presidents, or even last year's Dow Jones). But unless we see these models also making verifiable and accurate predictions (without 50% margins of error! what if chimp genome turned out to be 49% identical to ours?), it might well be that the system is inherently chaotic and muddy. Then it might simply not lend itself to scientific analysis, at least not with our present methods.
- nanairo 16y ago
  Indeed I agree that there exist a gradient of trust. But it is this that seems to be so difficult for the non-scientific literate to understand. In part I think it's a problem that society seem to use "shown scientifically" for "it's true". If there is one thing that people should learn at school is statistics and the scientific method: a LOT more useful than trigonometry nowadays! However my point is that it's not enough to have consistent predictions to show you are right. The examples of alternative medicine or even evolution are a good case. In the former we can do new tests, and we have: a lot of money has been spent trying and trying again different studies on homeopathy and friends. However the alternative practitioners always come up with an excuse why the scientist is wrong, or his result not meaningful. Their answers are not scientific in any way, but they can muddy the water enough that the audience feels they can't know either way. As for climatology, I agree with you that it's hard to call those predictions. Partially I think the very clear statements of the scientists get lost and simplified through different layers of repetitions so that when you read it in the mainstream newspaper you get a very partial version. There is also a problem that part of IPCC was _not_ strictly science. The IPCC has 3 panels from what I remember: the scientific stuff, the effect these will have on society, and ways to prevent them. The former gives proper scientific assessment but the later two less so. One of the problem of the IPCC is that many of its complains came from mistakes in the latter two groups, but they cast a shadow of doubt to the first one (the scientific one) despite the three groups have little connection to each other. (the stuff about the Himalaya melting, for example was something added by the second group, not the first one). I am not sure how easy it is to predict climate. We do better than random guesses, so I guess there's some help there, but it's not great so it shouldn't be reported as such. (of course if people knew what error bars and standard deviation were it wouldn't be so difficult). One problem though with chaotic systems is how difficult is to "fix" them. Maybe you can increase a variable (say pollution) up to 1000 (in an arbitrary scale) without any effect. Then if you take it that extra 10 points the whole system goes bad. People seem to think that in that case we can simply fix it by lowering the variable again, but that's not how it works. If your system moved from one attractor to another (in state space) it may be extremely hard to move back. This is why I think our best bet by far is to keep things as they are: better to err on the side of caution. (of course those are just my opinions: scientist should make predictions, but then it's society that decide what to do with them ^_^).

by nanairo 16y ago

zeteo 16y ago

It's true that there's a certain amount of fuzziness, but also there's a gradient of trust between 100% and zero. Predictions that consistently come true start to dwarf alternative explanations after a while. Darwin said that the great apes are our closest living relatives, and 150 years later research finds the genetic code similarities in the 90+% range. That's pretty strong. On the other hand, it is indeed possible that the field of study is a chaotic system, and accurate predictions impossible. But then I will argue that's not a branch of natural science, but rather of history. A good historian can tell you in small details the reasons why Lincoln got elected in 1860, but is as clueless as the next guy as to the next presidential election. Would you trust him if he said Republicans are always best? Similarly, there's a lot of great work done in historical climatology, with pollen analysis, ice cores etc. and we can find out a great many things about the past. Computer models can be built to fit this history perfectly (as I'm sure that even I can code some curve that spits out the order of Democrat and Republican presidents, or even last year's Dow Jones). But unless we see these models also making verifiable and accurate predictions (without 50% margins of error! what if chimp genome turned out to be 49% identical to ours?), it might well be that the system is inherently chaotic and muddy. Then it might simply not lend itself to scientific analysis, at least not with our present methods.

nanairo 16y ago
Indeed I agree that there exist a gradient of trust. But it is this that seems to be so difficult for the non-scientific literate to understand. In part I think it's a problem that society seem to use "shown scientifically" for "it's true". If there is one thing that people should learn at school is statistics and the scientific method: a LOT more useful than trigonometry nowadays! However my point is that it's not enough to have consistent predictions to show you are right. The examples of alternative medicine or even evolution are a good case. In the former we can do new tests, and we have: a lot of money has been spent trying and trying again different studies on homeopathy and friends. However the alternative practitioners always come up with an excuse why the scientist is wrong, or his result not meaningful. Their answers are not scientific in any way, but they can muddy the water enough that the audience feels they can't know either way. As for climatology, I agree with you that it's hard to call those predictions. Partially I think the very clear statements of the scientists get lost and simplified through different layers of repetitions so that when you read it in the mainstream newspaper you get a very partial version. There is also a problem that part of IPCC was _not_ strictly science. The IPCC has 3 panels from what I remember: the scientific stuff, the effect these will have on society, and ways to prevent them. The former gives proper scientific assessment but the later two less so. One of the problem of the IPCC is that many of its complains came from mistakes in the latter two groups, but they cast a shadow of doubt to the first one (the scientific one) despite the three groups have little connection to each other. (the stuff about the Himalaya melting, for example was something added by the second group, not the first one). I am not sure how easy it is to predict climate. We do better than random guesses, so I guess there's some help there, but it's not great so it shouldn't be reported as such. (of course if people knew what error bars and standard deviation were it wouldn't be so difficult). One problem though with chaotic systems is how difficult is to "fix" them. Maybe you can increase a variable (say pollution) up to 1000 (in an arbitrary scale) without any effect. Then if you take it that extra 10 points the whole system goes bad. People seem to think that in that case we can simply fix it by lowering the variable again, but that's not how it works. If your system moved from one attractor to another (in state space) it may be extremely hard to move back. This is why I think our best bet by far is to keep things as they are: better to err on the side of caution. (of course those are just my opinions: scientist should make predictions, but then it's society that decide what to do with them ^_^).