Reworked From Catallaxy:
Proxy data reconstructions should be considered valid if:
1. They have range-and-not-point implied-accuracy……..
2. If they have CONVERGENCE!!!!!
That is to say Convergence between alternate proxy types.
Alternatives that are independent of eachother.
So like today we have the satellite which corresponds pretty perfectly with the balloons.
The Satelite can then serve as proxy for the balloon measurements with a very narrow range… In fact a minimal, cosmetic shaded region there only for the purpose of maintaining decorum.
Then you have the ground readings. Then you have the tree-rings, the ice information the isotopes of this or that and so on.
You have the length of solar cycles and the number of sunspots and on and on depending on “what-it-is.”
Depending on “WHAT-IT-IS”…. or “THAT THANG”…..
Depending on WHAT-IT-IS-THAT-THANG you are trying to reconstruct by proxy.
As you trace back in time you start losing convergent information sources one after the other.
But you, by then will know, the range within which the various proxies can conflict-with-eachother in terms of their implied-estimate of what it is you are actually trying to estiMATE-by-proxy. (ie THAT THANG)
As one-proxy-after-another is LOST when you go back-in-time the range of the estimates for THAT THANG has to widen. To reflect the reduced CONVERGENCE.
This strikes me as precisely what Sami Solanki has done over at the Max Plank institute.
The first time I tried to follow his global air temperature estimates coupled with his solar-irradiance estimates…. going back in time many hundreds and thousands of years ….
((((((((so now you had two reconstructions compared to eachother…… ie… THAT THANG… compared to THAT OTHER THANG..))))))))
… As you went back his two bloody lines of pseudo-fucking-information that we were trying to compare got thicker.
And then got EXCEEDINGLY thick.
And I must say I was a bit frustrated and pissed about the whole thing.
I must confess to being immensely fucking-irritated by the thickness of the two lines.
I was just about to jump out of Samis’ computer and punch him in the shoulder….
But later I realised that this guys technique was in full accordance with my CONVERGENCE-epistemological-ideology.
So it makes me think that THIS IS THE WAY you look for the sound science.
(psssst: Are they doing it like Sami?)
You ask yourself a question:
(Are you doing it like Sami?)
And the question you ask yourself is THIS question:
(Are you swinging-it-like-Sami as opposed to bending-it-like real-climate)
ARE THEY DOING IT LIKE SAMI????
Because if you aint going at it like Sami then you might not be doing it right.
My best example of CONVERGENCE in economic science is combining Reismans inductive proof of the productivity-of-labour theory for the determination of real wages….
…. Paul Krugmans statistical study which proves that the two correlate, not only under free-enterprise conditions, but always and everywhere.
When you have this sort of convergence you have ‘rightful certitude’ or failing that at least ‘rightful near-certainty’ or being a bit of a wimp you can say that you have the ‘leading valid default-position.’
(((((((((And in this particular case that lead me to finally throw it in with the Union Power idea…..
.. Then that forced my hand in thinking that growth-deflation was a sort of moral necessity.
One social contract had died? Where do we go from here?
I went with growth-deflation.))))))))))
You shouldn’t look askance at PROXY-DATA as such.
You ought to hold these science-workers to account for NON-CONVERGENT-PROXY-DATA.
Proxy data is what we want alright and actually all we have.
But if it doesn’t have at least 3 lines of convergence, then its not really science.
Because if you have only two converging proxies and they disagree??????
WHERE’S THE TIE-BREAKER????
So I would say proxy-data is true science.
But only if you have three-proxy-CONVERGENCE.
What we are after is then:
Or more correctly:
And the upshot of that conclusion is that we ought to defund any studies that are deficient in this regard. Even in midstream.
I want to point out that my epistemological ideology is by no means an ivory tower thing.
We see here its immediate practical application.
It aint science unless it has 3-way (minimum) CONVERGENCE.
And there’s an opportunity to cut funds right there.
No ivory tower here.
Immediate implications to get better results and save money.
Now some smart cookie might realise that with my economics example we only have two lines of convergence.
And I sez we need three, two and then one to tie-break. Four would be better.
But 2 lines of convergence is pretty damn good for a humanities subject. Thats about as good as it gets in economics. And we can consider anyone that talks implicitly against the theory as “PLAYING-SILLY-BUGGERS” unless he can come up with some pretty good contrary information.