Falsification and falsifiability

[jay howard]

Interesting thread on the concept of falsifiability, but it seems to have totally misrepresented the concept. A few posters made the distinction between falsification and falsifiability, but by and large, this distinction went unnoticed.

Tried to post in that thread, but it was locked.

So, if anyone wants to take this issue up, please feel free. My understanding is that falsifiability differs from falsification in a critical way. Falsifiability, as the name implies, is the ability for a proposition to be false. If there is no way for a theory to be demonstrated to be false, then, Popper postulated, it can be guaranteed to be worthless as an explanation of phenomena, and thus, not scientific.

This seems to be a matter of testability. If a theory is not testable, then we know up front that the proposition at hand has no real predictive power. This is the distinction Popper was trying to make. Lakotos, Kuhn and others have attempted to show the shortcomings of this approach, and have, to a degree, succeeded. However, as a fundamental test of one's approach, falsifiability works quite well. It works, for instance to differentiate creationism and intelligent design from theories that can be tested and used to predict phenomena in the world. In this regard, falsifiability is quite useful.

This differs markedly from falsification, which appears to be what the other thread turned into a discussion of. Falsification is a product of falsifiability, but the reverse is not true.
.

 

[Moridin]

No, for a proposition to be falsifiable means that there exists, in principle, an empirical result of an experiment that would demonstrate its falsehood. Falsifiability is not the same as testability. For instance, tautological statements are testable, but not falsifiable.

No, intelligent design creationism can neither be tested or falsified, since they have not provided any explanatory model that can be subjected to testing. Do not confuse their negative attacks on evolutionary biology with their inconspicuously absent explanatory model: "How did the designer do it? What mechanisms? When?" and so on. This is not included in the ID flavor of creationism (or the earlier forms either).

The smallest demarcation criteria that should be advocated when it comes to creationism should be testability.

 

[jay howard]

"Falsifiability is not the same as testability."

True enough. Just trying to clarify that falsifiability is not the same as falsification. It seems to have been thoroughly confused in the other thread..

 

[mikelepore]

I will be grateful if someone would please criticize this thought of mine. It seems to me that some/sometimes and can-do hypotheses can be verified but not falsified, while all/none/always/never and can't-do hypotheses can be falsified but not verified. For example, the hypothesis "some ducks are brown" is equivalent to "it is possible to find a brown duck", this can be verified by the act of finding one. It cannot be falsified because we can never know if we haven't found one yet only due to probability. However, the hypothesis "all ducks are brown", which is equivalent to "it is impossible to find a non-brown duck", can be falsified but not verified. Find merely one white duck and the hypothesis is known to be false. It cannot be verified because the exception might be out there somewhere. (In saying this, I'm omitting the case of an exhaustive search of a population in a limited space, because I could certainly verify or falsify any statement I wish about all the ducks in my living room, etc.) Now I don't remember much about Popper -- everyone always talks about Popper dwelling on the subject of falsifiability -- but did Popper see the complementary roles of the verifiable and the falsifiable that I have opined about above, and, if not, why?

 

[Lord Ping]

I will be grateful if someone would please criticize this thought of mine. It seems to me that some/sometimes and can-do hypotheses can be verified but not falsified, while all/none/always/never and can't-do hypotheses can be falsified but not verified. For example, the hypothesis "some ducks are brown" is equivalent to "it is possible to find a brown duck", this can be verified by the act of finding one.

This sounds good. I suppose Popper's response would be that your hypothesis is not very interesting - real science searches for universal laws.

Incidentally "some ducks are brown" is not equivalent to "it is possible to find a brown duck." I think it is equivalent to "there is at least one x in the universe such that x is a brown duck." Anyway, I think you're right.

But Popper is all wrong. The sooner people stop reading him the better. The project to find a deductive "logic of scientific discovery" is utterly flawed.