[Osx-nutters] Denial

[David Cake]

At 3:23 PM +0100 16/8/07, Stefano Mori wrote:
>Your belief that science is most likely to march forward inexorably 
>step by step, seldom taking a U-turn, is not useful in debating AGW.

	Why not? Its just the 'most likely' result. Science, of 
course, often changes its mind, just most often its

>The stock market marches mostly up, but that doesn't mean that your 
>own company's stock is going to do the same in the next 10 years.

	And, if you are looking for general investment advice, 'put 
money in stocks' is pretty good.


>  You wouldn't bet a
>thousand dollars on one company's stock just because the stock market 
>inexorably goes up in the long term, so why should we bet the whole 
>world's economy on one field of research?

	We have bet the economy on the environment already, Stefano. 
Your money is already there.
	You only get to decide whether you think science is a more 
reliable predicter of the behaviour of the environment than, say, the 
oil industry or the Republican party.

>"The debate is over... and here is our best guess".
>
>But that is not what the public hears. The public hears that the 
>scientists really do know what's going to happen. As I said earlier, 
>people keep screaming, "WE MUST ACT".

	Sure, public understanding of the scientific method and the 
mechanisms of science is lacking.
	And it would be nice to improve that. I'm doing my bit one 
mailing list at a time.
	But I don't think science itself is responsible for the poor 
quality of public political debate about scientific issues.

>But if it's just a guess, even a "best" guess, then there is no 
>"must" about it. It's a "probably we should act", "perhaps we should 
>do something about it",  "we guess we may be better doing something".
>
>Changes the nature of it, don't you think?

	Not much, no. You balance the likelihood of correctness with 
the consequences.

>  >	Which makes the point that the NUMBER of correct predictions
>>  is nonsense. Fitting all the observed facts is a much more useful
>>  measure of the value of a model.
>
>
>You prefer a complex model who's predictions haven't been tested even 
>once to a simplistic prediction that has worked millions of times?

	Hmmmm... do I think the heliocentric model of the universe is 
fairly correct? Nope, still no on that one.

>
>Are you naive? Sorry but I have to keep asking.
>
>Have you seen the cathedrals? Those were built using rules of thumb. 
>They built them that way because from experience they knew that it
>worked.

	No two were exactly the same. They extrapolated from their 
experience to build bigger things, with different designs. The way in 
which you think this is completely different from making predictions 
from theory? It isn't. There is no magic dividing line you can draw 
and put the bad scientists on the other side. Prediction is 
prediction, the difference is quantitive not qualititive.

>  >	The economy is a big complex system. No one has EVER tried,
>>  say, having the central banks tie interest rate changes to tarot
>>  readings. Maybe it would turn out be a good idea. I think we are more
>>  than moderately confident that its not.
>
>
>I think it's telling that you keep comparing AGW science to crazy 
>witch doctors as a way of making the scientists' predictions look good.

	After you've just been talking about how heliocentricism is 
good science? I think you have your rhetoric turned up a few settings 
too high, really - the bows have been getting very long indeed.


>  >	Well, sure. Because people might act on those theories, and
>>  change the outcome, for a start. AGW theory only makes predictions
>>  about the climate in 100 years IF conditions don't change, because we
>>  do nothing, and our use of CO2 doesn't change, and other relevant
>>  conditions don't change (I am sure AGW doesn't so 'and the
>>  temperature will rise by exactly 4 degrees, even if the sun goes
>>  nova'). I don't understand what criticism you are levelling here?
>
>The prediction can still be wrong without any of those things 
>changing. It's inherent in complex models that the small 
>insignificant errors can compound themselves in ways that make your
>prediction materially wrong.

	Well, actually, that is only true for certain numerical 
conditions, and certain values of 'complex'. I'm not saying its not a 
problem, I'm saying its a well understood one, of which no person 
with professional mathematics training is likely to be ignorant.


>  >	So, in order to make predictions about the death rate to road
>>  accidents in 2010, I first need to empirically quantify why I don't
>>  think we are likely to experience nuclear warfare or bird flu
>>  pandemic in that time? No, I just say 'if conditions do not change
>>  then' and move on.
>>	I think everyone actually grasps the idea that when science
>>  makes predictions, it isn't claiming the psychic power of
>>  precognition, and I'm not sure what else you complaining about here.
>
>It's actually a simple idea and yet nobody seems to even see it.

	Either that, or people understand it perfectly well and just 
don't understand why you are making a big deal over it. I think you 
should consider the possibility that the latter case is more true.
	Cheers
		David