Experiment Right Or Wrong

Allan Franklin. Experiment, Right or Wrong? CUP 1990

Experimental Results ‘6.1’ Outlines bayesian approach to philosophy of science; not much detail, includes cursory consideration of prior evidence problem, and ‘tacking’ problem

‘6.2’ Looks to provide a bayesian reading of experimental strategies to rule out error/decide validity of observation?not entirely convincing that the justification needs to be bayesian. The approach seems much more hypothetico-deductive rather than probabilistic (including the examples Franklin is using). ’ Hacking and (?) theory independence and intervention; regarding independent confirmation, Howson and Franklin have a bayesian account; a hypothesis receives more confirmation from two different experiments than from repetitions of the same experiment (Studies in History and Phil of Science 1984;15:51-62) cf Hartmann, centre for time ppt. ’ It is the details which are vital, and doing the work cf bayesian reasoning

‘6.4’ AF considers possible fallibilities within the experimental approach, topics under consideration include: selectivity (selecting the phenomenon out of the background effects, but what stops selecting out the wrong background); theory-ladenness of experiment and observation; experimental bandwagon ’ Each of these are accepted (as sources for fallibility) and countered as arguments denying the importance of experiment, key replies include: theory independence, reproducibility; experiment may precede theory

The roles of experiment ‘7’ A discussion of the relations between theory and experiment, and point that there are a number: at times the experiment takes on a life of its own; others theory precedes experiment, and others which theory covers experiment in some loose way ’ Franklin provides detailed example in physics after example; the approach similar to Weber in Experimental Biology, but harder for me to follow