Thursday, May 03, 2012

Falsification Does Not Solve the "Problem" of Induction: A Simple Example

The statement in the title is true in a number of ways, but here is a very simple argument.

Let us say we have decisively "falsified" a theory: rocks float on water, perhaps. So, per Popper, we now abandon that theory, and come up with a new bold conjecture.

Well, this abandonment depends entirely on an inductive inference! In fact, it depends on the principle of "conservative induction" being true: more of the same. Unless that is the case, why should the theory having been falsified today having anything whatsoever to do with whether we should employ it tomorrow?

Imagine that, instead, the principle of revolutionary induction is true: time for a change. In that case, the fact that our theory was falsified today should give us great hope for the theory tomorrow. And if that principle is true, then we should abandon all of our theories that have not been falsified right away!

So, falsification, far from solving the "problem" of induction by eliminating it, actually depends on conservative induction being justified. And this is not the only way in which falsification depends on induction, just the simplest that I have thought of.

These arguments are well known in the literature, and considered decisive by almost all professionals in the philosophy of science, which is why there are only a handful of Popperians left in the field. (I'd be surprised if, in the top 100 departments in the field, there are a total of more than five Popperians -- not five Popperian departments, but five individual Popperians.) And this is certainly not because Popper is a neglected figure or was out of the mainstream: no, Popper and Hempel were the two giant figures of the 50s and 60s. These ideas were carefully considered, and found to be wanting.


  1. I have exactly zero training in philosophy of science, so I admit that my ideas are probably going to look amateurish, but my own thinking is this. I used to be a falsificationist because it seemed a lot better than being a logical positivist. (Those were the only two alternatives I was aware of.) However, I soon realized that I had a very hard time coming up with a statement about empirical reality that could, in principle, be "falsified" or refuted.

    When we live in a world that's governed by many different laws, the conclusion that an experimental result is due to the *lack* of a particular law is always going to depend on the assumption that the result is not instead due to hidden laws which we haven't yet discovered, understood, and accounted for in our models. For example, if I release a rock on earth and it floats in the air instead of accelerating downward at 9.8 m/s^2, this would seem to be a straightforward refutation of Newton's Law of Universal Gravitation. But how do I know that there's not just some hidden force acting on the rock to make it behave that way, i.e., gravity is still exerting a downward pull on the object, but there's another, greater force which we don't yet understand that is *also* exerting force in the opposite direction?

    It seems to me that, whenever an anomaly appears to challenge a very well-accepted scientific theory, scientists are much more likely to hypothesize about such explanations rather than throw the original theory out and start from scratch. I'm not certain that that's such a bad thing, but I'm not sure where that leaves me, either.

    1. Mike, without training, you have hit upon a second way in which Popper's system depends on induction: a falsification never falsifies a single theory, but the whole complex of theories that are relevant to an experiment. How do we decide which to abandon? Well, we keep those that are better "corroborated," which is a way to avoid saying that we use inductive reasoning and keep those for which we have better evidence.

  2. "Falsification Does Not Solve the "Problem" of Induction: A Simple Example"

    Does anyone else see the irony of this title?

  3. It looks to me like you're giving us a counterexample, thereby proving that Popper's theory of science must be wrong.

    1. Well, if my post had been:

      "Falsification does not exist."


      "Falsification is of no importance."

      Then it would be ironic if I tried to prove it with a counter-example.

      But the idea of falsification is not Popper's. Nor is the idea that falsification is more important than passing tests -- I have found that in the 1600s. Popper's original idea is that falsification solves the problem of induction. Nice try, but it doesn't. And we can falsify the idea that it does!

  4. You have misread Popper. A corroborated theory is not chosen because of an inductive inference that theories that have been successful will continue to be successful, as David Miller has pointed out. Rather, the predictive power of theory is in the theory itself. Corroborated theories are chosen by which can be more easily falsified. Popper's philosophy does not rely on inductive inferences.

    1. I love Popper cultists: what a will to believe!


Distraction Deterrents in Small Contexts

"distracted from distraction by distraction" - T.S. Eliot I've been reading a little on how Facebook and other social netwo...