Because I’m generally interested in such things, I’ve read a reasonable amount of philosophy of science. Most of it has made some sort of sense, but there always seemed to be something missing. But I think I can see where part of the problem is.
last week, John Wilkins posted his delight that physicists are taking up stamp collecting. At about the same time, Andrew Gelman posted on some comments by Irene Pepperberg saying that testing hypotheses is over-rated. This got me thinking.
Philosophers of science talk about it in terms of generating and testing theories: this is what Popper is all about, and the criticisms and extensions of his ideas (e.g. by Kuhn, Lakatos etc) are put into this context. My one anarchistic foray into philosophy of science was to argue against laws in ecology, and the nature of laws is still discussed in philosophy of science circles (they even have ceteris paribus laws, i.e. laws that apply except when they don’t). Again the context is theories and their properties.
But the point being made by both John and Andrew is that there is more to science than this. Andrew makes the point that his work is descriptive, e.g. showing how the voting changes with income in different US states. There is no testing of theory here. I often do something similar: the analysis of data is done to see what patterns are present, or to ask how a particular relationship looks, rather than to give a True/False answer to some question.
John argues that classification is the missing dimension to science, but I don’t think he goes far enough. Classification is part of a larger process: one of organising knowledge and observations. Raw field observations are pretty messy: there are a lot of factors affecting the real world, so it is difficult to see the patterns straight from the observations. One purpose of statistical analysis is then to describe the important features of the data, whilst taking into account all the confounding factors. Or, to put it another way, it is about organising our observations. Classification is the same thing, just that the observations are organised in a different way. So, I would criticise John not so much for being wrong, but rather for not going far enough.
This, then is part of what philosophy of science is missing: science is about learning about the real world, and this can be done by testing theories about how the world behaves, or simply by observing the world, and describing it. Both methods are surely valid, so if philosophers are to create a good working account of science, they need to pay attention to both aspects of science.
No doubt there are more aspects to science that they need to include, and to classify with obscure phraseology. I the mean time, I have to work out a strategy for blogging when there is a cat sat in front of the computer.