Books on Amazon
|Method||Fraassen I 77
Realism/Science/Methodology/Method/Boyd: only realism can explain the scientific activity of the experimental setup (method, experiment). This is needed for the legitimation of intertheoretical considerations. This is to explain the role played by accepted theories in experimental setup.
1. Principle: (according to Boyd anti-realistic) if two theories have precisely the same deductive observation consequences, then every experimental evidence for or against the one is simultaneously one for or against the other.
BoydVs: this is simply wrong as it is stated there, and it cannot be improved either.
Empirical equivalence/FraassenVsBoyd: I have a completely different definition of empirical equivalence than he has.
2. Principle: (according to Boyd accepted by all philosophers): Suppose a scientific principle contributes to the reliability of a method in the following minimal sense: its application contributes to the likelihood that the observational consequences of accepted theories will be true. Then it is the task of epistemology to explain the reliability of this principle.
Fraassen: I also believe that we should agree with that. It is itself a principle about principles.
Boyd/Fraassen: they have a special example in mind:
(P) a theory must be tested under conditions which are representative of those in which it is most likely to fail in the light of accompanying information if it can fail at all.
Fraassen: this is harmless as it is stated there.
Problem: "Accompanying information": I assume that he "understands" here "knowledge" as "light", i.e. as knowledge about the underlying causal mechanisms that are based on previously accepted theories.
Boyd: e.g. Suppose,
M: chemical mechanism
C: Bacterial type
L: Theory, which, together with accompanying information, assumes that the population of bacteria develops as a function of their initial population, the dosage of A and the time.
Experiment: Question: what must be taken into account when constructing the experiment?
1. E.g. a substance similar to A is known, but it does not dissolve the cell walls, but interacts with a resulting cell wall after mitosis. Then we must test the implication of the theory L which is to be prooved, which does not work in this alternative way.
Then the sample should be viewed in such a short time that the typical cell has not yet split, but it is long enough that a large part of the population is destroyed by A (if there is such an interval).
2. E.g. one knows that the bacteria in question are susceptible to a mutation that mutates the cell walls. This leads to the possibility that theory L will fail if the time is long enough and the dosage of A is low enough to allow selective survival of resistant cells. Therefore, another experiment is required here.
In this way accepted theories lead to a modification of experiments.
Knowledge/Fraassen: we must understand knowledge here as "implied by a previously accepted theory".
The Philosophy of Science Cambridge 1991
B. van Fraassen
The Scientific Image Oxford 1980