Posted on 04/06/2003 10:12:13 AM PDT by Hank Kerchief
Lately we have seen the notion of falsifiability represented as a fallacy. This is itself, a fallacy.
The concept of falsifiability is a greatly misunderstood but legitimate part of the scientific method (a rigorous application of reason to evidence). Consider this statement made as an objection to falsifiability, "Falsifiability can be a valuable intellectual tool: it can help you to disprove ideas which are incorrect. But it does not enable you to prove ideas which are correct." In fact, that is exactly what "falsifiability" does do, and without it, no scientific hypothesis can be proven.
In science, a proposed hypothesis is not considered valid if there is no experiment that can be performed that would, if the hypothesis is incorrect, fail. If such an experiment can be performed, and it "fails to fail," it is proof (or at least very good evidence) the hypothesis is correct.
No doubt the prejudice against this very useful objective method lies in the name, "falsifiability." It does not mean the scientist must attempt to prove a hypothesis false, but the very opposite. "Falsifiability," is the method by which a hypothesis may be proven true. It also does not mean that a hypothesis must be assumed correct until it is falsified.
The idea of falsifiability protects the field of science from being obliged to entertain as, "possible," any wild hypothesis on no other basis than it cannot be disproved. If a hypothesis is correct, there will always be a test or experiment that it would fail, if it is incorrect, which when performed proves the hypothesis correct by not failing (or incorrect by failing).
If no test can be devised for testing a hypothesis, it means the hypothesis has no consequence, that nothing happens or doesn't happen because of it and nothing depends on it being right. If this were not true, whatever depended on the hypothesis could be tested. There is absolutely no reason to entertain a notion that has neither purpose or consequence.
"But why not perform experiments to verify rather than falsify?" In fact, all experiments performed to test a hypothesis are attempts to verify it. If such a test could "pass" even if the hypothesis were incorrect, passing the test would prove nothing. Passing a test is only, "proof," if passing is only possible when the hypothesis is true, which means the test must fail (the hypothesis will be falsified) when the hypothesis is untrue. A test which cannot falsify a hypothesis, if it is incorrect, cannot prove it, if it is correct.
To say a hypothesis is not falsifiable means that it cannot be proved (or disproved), and, therefore, is unacceptable as a scientific theory.
It is very unfortunate that this concept is misunderstood by many who are otherwise quite rational and objective. The principle not only applies to science, but almost all complex or abstract concepts. The attempt to verify any conjecture by means of a method that cannot discriminate between those conjectures which are true and those which are false can never discover the truth. Only a method which distinctly demonstrates a conjecture is false, if it is, can verify those conjectures that are true.
The concept of falsifiability sweeps away mountains of irrational rubbish masquerading as science, philosophy, ideology, and religion. One question that must be asked about any doubtful proposition or conjecture is, "how can this be disproved if it is false?" If there is no way to test if the proposition is false, there are no rational grounds whatsoever for assuming the proposition to be true.
(Excerpt) Read more at hpamerica.com ...
We'll get no further than does the UN in our debates until we all agree to certain definitions. Even Clinton knows that.
Couldn't disagree with you there.
Feyerabend! Can you imagine what Socrates and he would have said while walking among the shades!
There is some evidence that could be used by so-called "social scientists," but they usually don't want to be bothered. For example:
1. Whenever a socialist and a free-market society exist side-by-side, people almost always flee to the free-market society. Examples: East & West Berlin, Red China & Hong Kong, Cuba & Florida. The refugees never go in the other direction.But such experiments are rare, and zealots for controlled economies can always argue that their system hasn't really been given a good trial. There are other experiments too. The war on poverty has failed, the war on drugs has failed. But the advocates of these programs won't ever admit it. So whatcha gonna do? If "social scientists" don't want to behave in a truly scientific manner, and accept the failures of their experiments, one must conclude that they aren't engaged in science at all.2. Whenever people have a choice of living communally or with private property, they very rarely choose communal living. Examples: Israel's kibutz system, occasional communal experiments in the US.
3. Controlled economies never out-produce private economies, when all factors are equal. Examples are rare, but the best is the private plots that Stalin permitted to exist on his communal farms, which were about ten times as productive as the nearby land.
4. Cutting marginal tax rates always increases tax revenues. At least that's the lesson of Kennedy's tax rate cuts, and Reagan's.
Math is fundamentally different from science. Math is about constructs of the mind, such as "numbers" and "primes" and "equalateral triangles" and their relationships. It may or may not apply to physical reality, but there need not be such a reality for the concepts to have strict logical consistancy.
Science, OTOH, is about nature and the way things work. Light moves at a certain speed, bends through water and gravitational fields, and can dislodge electrons from certain metals. When you mix certain chemicals, they change color or explode. Electric current causes compass needles to move. Etc. Scientific "facts" are less strongly known than mathematical facts - there could be some level of external gravitational field, or velocity, at which what is believed to be true turns out to be measurably not so. But the fact that 2+2 = 2x2 = 2^2 does not depend on gravity, the speed of light, or anything else.
there are an infinite number of primes is untestable. This truth is easy to prove but impossible to test.
I disagree: the proof is not only simple, but it certainly generates what must be an infinite list of primes. The proof is, take all the primes you know of, multiply them together, and add 1. The new number can't possibly be wholly divided by any of its divisors - it therefore must be prime or be a composite number containing prime factors not in the original list. No matter how many primes you multiply, there is always at least one more. Hence, the number of primes must be, quite simply, infinite.
But Im not a mathematician and perhaps a rigorous proof is equivalent to a test. I differentiate the two myself.
A proof is different from a test: A proof is a demonstration of certainty, while a test is merely a demonstration of likelihood. To my knowledge there are no proofs in science: It might yet be shown that everything presently understood is wrong within some domain, that we haven't explored yet and might not even be aware exists.
Try going back and studying the fallacies link you provided and then identifying the fallacies in your post and you will be able to state whatever position you have in a manner that I can understand it. Otherwise, we might as well be speaking similar but different languages. Languages just similar enough to cause us to believe we are speaking the same language without realizing we are speaking some type of gibberish to each other.
Excellent observations, PatrickHenry, all of 'em.
The many communal experiments in the United States -- the Owens Colony, Oneida Colony, Brook Farm, Fruitlands, et al., were all miserable failures. They were "home-grown" social experiments that didn't work here just as they haven't worked anywhere else. We are talking about 150+ years of "experiments." How much longer do we have to wait for folks to just reasonably concede that the fundamental premise of all such experiments -- that one may legitimately deal with human problems from the standpoint of group rather than individual -- is fatally flawed, ill-founded, from the get-go? Why do we need to give the folks who want to keep running this same experiment over and over again, "more time?" Isn't it time just to move on?
I've recently read that enthusiasm even for the kibbutz system is declining....
Thank you so much for writing, PH.
Forgive the poorly worded sentence.
There are scientific truths that are not "falsifiable" in theory.
In other words, simply because a belief or theory is not subject to repeatable and observational "operational" scientific testing, this does not mean that belief or theory is false.
I don't think it was ever wildly popular. And I've read that most of the young people raised in such places leave them and don't return. I think the system survives mainly because in some of their political parties it's kind of an advantage to have it in your background, like being raised in a log cabin or something. The only thing that is acceptable about the system is that it's voluntary. That's what makes it such a neat experiment. About 97% of the population rejects it.
I suspect that Wittgenstein was a rather lonely man. But that's probably true of anyone who has only one pair of glasses with which to view the world, be it philosophy or natural science or religion.
What he as said is basically that if we can't speak of something in an understandable manner we should say nothing at all. The Tractatus just lays down a set of rules for doing this within the confines of the world as it can be understood by natural science. There are many things in human experience that cannot be discussed by science or other epistemology.
If, in principle, a belief can't be tested, it isn't a scientific belief. It may, as you say, be a true belief. But how would one know?
I assume by "testing" you mean (as I do) at least repeatability and observation.
Any scientific singularity would not be subject to such testing.
There are several in number theory, such as Euclid's second theorem that there are an infinite number of prime numbers. Provable, but untestable.
Why not? In other words, how would you define a scientific belief?
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.