Duhem's non-falsifiability thesis

Started by Geremia, July 30, 2019, 10:03:47 AM

Previous topic - Next topic

Geremia

Some (Sungenis in ?) think that a single crucial experiment (experimentum crucis) can falsify a physical theory; however, all a single experiment can do is show that at least one of the physical theory's assumptions is wrong, but it cannot say which.

The French Catholic physicist, philosopher, and historian of medieval physics Pierre Duhem formulates this non-falsifiability thesis as follows (Aim & Structure of Physical Theory p. 185), when describing an experiment that contradicts theory:
Quote from: Pierre Duhemthe only thing the experiment teaches us is that, among all the propositions used to predict the phenomenon and to verify that it has not been produced, there is at least one error; but where the error lies is just what the experiment does not tell us

[yt]hM83cBa9wDM[/yt]
(The development of the doctrine of the Trinity is mentioned @11:54.)

gsas

I thought that this is the reason, why theoreticians and experimenters always try to isolate their subject to just one probe at a time. 

Daniel

#2
I'm not sure I get the point. If at least one of the theory's prime assumptions is false, then the whole theory falls apart. And if an experiment proves that a valid argument's conclusion is false, then it follows that at least one of that argument's assumptions must be false. It doesn't really matter which of the assumptions is false though. (Well, it does matter if you're trying to fix the theory. But if you're merely trying to falsify it, it doesn't matter.)

Geremia

Quote from: Daniel on August 26, 2019, 07:04:58 PMI'm not sure I get the point. If at least one of the theory's prime assumptions is false, then the whole theory falls apart. And if an experiment proves that a valid argument's conclusion is false, then it follows that at least one of that argument's assumptions must be false. It doesn't really matter which of the assumptions is false though.
Yes, that's exactly what the thesis says; one cannot isolate which assumption has been falsified. Physics theories must be viewed holistically.

In a letter to Fr. Garrigou-Lagrange, O.P., Pierre Duhem writes:
Quote(1) We shall never have the right to affirm categorically of any one of the principles of the mechanical and physical theory, that it is true. (2) We are not allowed to affirm of any one of the principles on which the mechanical and physical theory rests, that it is false, so long as there has been no discovery of phenomena that disagree with the consequences of the deduction of which this principle constitutes one of the premises."

What I have just said applies particularly to the principle of inertia. The physicist has not the right no say it is certainly true; but still less has he the right to say it is false, since we have so far met with no phenomenon (if we leave out of consideration the circumstances in which the free will of man intervenes) that compels us to construe a physical theory from which this principle would be excluded.

Geremia

Quote from: gsas on August 26, 2019, 09:30:15 AMI thought that this is the reason, why theoreticians and experimenters always try to isolate their subject to just one probe at a time.
That's not really possible, though. Quantum theory, in which an observer influences the outcome of the experiment, makes this very apparent.

Geremia

Quote from: Daniel on August 26, 2019, 07:04:58 PMit does matter if you're trying to fix the theory
Probably the more controversial aspect of the thesis is that theories can always be fixed. (It's not clear Duhem thought this, but Quine certainly did).

Kreuzritter

Quote from: Daniel on August 26, 2019, 07:04:58 PM
I'm not sure I get the point. If at least one of the theory's prime assumptions is false, then the whole theory falls apart. And if an experiment proves that a valid argument's conclusion is false, then it follows that at least one of that argument's assumptions must be false. It doesn't really matter which of the assumptions is false though. (Well, it does matter if you're trying to fix the theory. But if you're merely trying to falsify it, it doesn't matter.)

As it is and in the strict sense, yes, but the theory can be modified to "preserve" it, and this can be done even by changing implicit assumptions through the introduction of a new concept, as was done with "dark matter". In that case, it's pretty much a denial that the conditions of the test experiment were properly met: "We didn't get the predicted result because we hadn't accounted for this unexpected stuff."