[Osx-nutters] Denial

[Stefano Mori]

On 2007-Aug-15, at 10:48, David Cake wrote:

> At 11:09 PM +0100 14/8/07, Stefano Mori wrote:
>> I doubt it because they say the debate is over.
>
> 	But if they said the debate wasn't over, you'd doubt it too
> (I mean, you'd be nuts to say otherwise).

I looked at this earlier. If they said, "we're actively pursuing  
several lines of research and we welcome knowledge from other  
disciplines, and here are the results so far", (words and evidence to  
that effect) then I would trust them much more.

It's like, if I go to a doctor who says, "I completely understand  
your condition, we'll sort you out no worries", that would make me be  
worried due to his apparent overconfidence.

If I go to a doctor who says, "it looks like you have this, these are  
the indicators and the counter-indicators, we have this treatment, we  
don't honestly know why it works, but it's all we've got, it's a bit  
tentative but so far 65% of people respond fairly well", then I would  
trust him much more. I trust him because he is being honest, I mean,  
I can Google the 65% figure to check.


> So, you've argued yourself
> into a catch 22 - there is nothing they can say that will make them
> think the scientific question is more or less settled.

You seem to think that we need to be told that it's settled, and I  
guess those scientists thought they needed to tell the public that it  
was settled. I don't need it settled. I am happy to proceed on a  
course of action even when the outcome is uncertain and the knowledge  
partial. Sometimes it's necessary to start in some direction, but as  
you start sailing you keep checking along the way that the chosen  
direction is still a good one. You are open to making course  
correction, even turning around 180 if necessary. But if they say  
it's settled, that means there's no more option for changing their  
minds. Nobody can have practically perfect knowledge of such a  
complex system, so there's plenty of scope for errors, some minor,  
and some major. A culture that believes it's settled is a culture of  
closed minded arrogance. If this were some little academic subject of  
no consequence then that wouldn't matter. But if it really is about  
planetary survival then there is no room for professional arrogance.  
As far as I am concerned, with that PR they shot themselves in the foot.


> 	I mean, sure, if they said 'oh, we are moderately confident
> of a particular outcome' you would probably take them at their word -
> but what if they are genuinely MORE than moderately confident?

Big complex system being forecast 100 years into the future.

Lets go back to your "predictive power of science because the sun  
rises" point.

According to Wikipedia the geocentrism model with the Earth at the  
centre of the universe goes back 5000 years.

So, the cosmologists 5000 years ago can predict that the sun will  
rise each day. We now know from experience that that prediction has  
proved to be correct one point eight million times.

Lets look at the modern climate models forecasting about +4C in 100  
years. That prediction has proven from experience to be correct  
exactly zero times.

They can't be more than moderately confident because they don't have  
the empirical experience that the prediction is correct even once.


>> Virtual certainly about forecasts of big complex systems only leads
>> me to suspicion.
>
> 	But somehow, the virtual certainty about facts like 'reducing
> smoking reduces lung cancer', despite the massive and poorly
> understood complexity of the human immune system and mechanisms of
> cancer, doesn't bother you.

The key word is forecast. As I said earlier, saying that 25% of  
smokers so far have gotten lung cancer is different to saying that in  
future, 25% of smokers will get lung cancer. This is a more complex  
issue than the sun rising, but you'll be interested to hear that the  
point still holds; once you start making forecasts it's a different  
game. We already know from experience that if a smoker starts later  
in life, or if they stop earlier in life, then this considerably  
alters the outcome, ie. they may not get it at all. Changing social  
trends may mean that in future only 10% of smokers get lung cancer,  
and changing work conditions may mean that the 20% of lung cancers in  
women who are non-smoking, may rise to 80% for all we know. For a  
good forecast we need to empirically examine the possibility that one  
day, other causes of lung cancer may be so great that reducing  
smoking doesn't actually reduce the incidence of lung cancer.  
Consider Chernobyl; you say to people, "look guys, just stop smoking  
and you'll be fine". To make your forecast scientific you need to  
first empirically quantify why these other factors are negligible,  
and why your method of quantifying these factors is empirically  
sound. In short, once we start making forecasts the facts are not so  
easy because the facts haven't happened yet.

Extend this to AGW and you have to start dealing with a far greater  
order of complexity. You have to even show why you believe that the  
actions you plan to take will succeed, because the damage of trying  
and failing may be worse than the damage of not doing anything at  
all, eg. Iraq. And it has to be empirical, otherwise you can't call  
it scientific.

Stefano