On Professor Gray’s Explanation of the Scientific Method
Written by Dr Jim Petch
Vincent Gray’s recent contribution on ‘The Scientific Method’ prompts me to respond with an extended set of points on Karl Popper’s ideas. My first article for PSI addressing Popper’s work is found here.
It is refreshing to see reference to the work of Peter Medawar who seems to be neglected by today’s commentators on science including climate science. He has much to say of sound common sense and in my view climate scientists and their fellow travelers (on all sides) would do well to try to take from his writings something of his spirit and humanity.
It is good also to see emphasis on the lessons Karl Popper teaches us but again on Principia Scientific, important parts of his ideas are poorly presented and I feel it is important to keep making points of argument which at first sight may appear pedantic but which, in fact, are pivotal to a proper understanding of Popper’s philosophy and to a proper appreciation of how the empirical science should be conducted.
Vincent presents a set of diagrams that illustrate what he calls the ‘methods approved by Popper”. In fact Popper, to my knowledge, made no such approval and was most clear that for him there was no scientific method. He did prescribe the method of conjecture and refutation as the way to eliminate error.
Strangely the diagrams Vincent presents retain the inductivist error of going from observation to hypothesis and not the other way around as it should be with the deductivist Critical Rationalist approach of Popper.
The systems Vincent presents may, as he says, use a mixture of induction and deduction but if so Popper would surely have rejected them since he found no place at all for inductivist reasoning. Again as far as I know, the only diagram Popper gives us is his schema for conjecture and refutation in the sequence
P1 – TT – EE –P2 and so on and on.
This embodies an altogether different idea than the diagrams Vincent provides. Its focus is on the repeating cycle of problem situations, which arise from the testing of ideas against observation.
As regards Vincent’s comments on Climate Science, while I fully agree with his remarks about repeatability, I disagree that in this field no observation can be repeated’. If this were true then no observation anywhere at any time could be repeated, which is of course silly. Observation are intrinsically repeatable and in reality, within reasonable bounds, observations are repeated. The fact that there are differences of instruments etc etc does not negate the intrinsic repeatability of meteorological measurements (you cannot, of course, measure climate, but that’s a different story!!).
And it is not clear why a supposed lack of repeatability has anything to do with identification of trends. Two different issues with no real connection are conflated here.
There is also a logical non sequitur in that whether or not scientific conclusions based on observation alone are unreliable or not has nothing to do either with repeatability or with forming premature conclusions. That view, which I agree with, rests on quite different points of argument.
And again at this point Vincent slips on the induction banana skin. There is, if we accept Popper’s ideas, no criterion of validation, only a criterion of falsification. The point is that ideas should be tested critically, not validated, and those ideas retained which have withstood severe testing. A small point, you cannot test for falsifiability, only for falseness in ideas. Falsifiability is not a matter of testing but of logic.
And, as an aside, Planck’s idea are not validated in the inductivist sense but have withstood severe testing (for the time being).
And the question for us is not ‘can Climate change theory be validated?’ (that asks for inductive reasoning). But should be ‘how do we design tests that will show climate change theory to be wrong if it is?’ A quite different matter methodologically and in terms of the practicality of experimental design.
Again, whether or not ideas/models are tested against past or future data is beside the point. The issues are their testability and the severity of the tests we subject them to.
Dr Jim Petch was until late 2007 the head of the Distributed Learning Centre at the University of Manchester, when he retired. His focus of study is towards Critical Rationalism.