Indirectly, through Lee Smolin, I finally get Peter Woit’s definition of falsifiable, which is actually nothing much more than the standard nonsensical nonsense that’s summarized by answers.com.

In Smolin on the Anthropic Principle, Woit says,

[Smolin] gives an eloquent explanation of the importance of falsifiability for a shared scientific enterprise.

I can assume, then, that Woit is satisfied with Smolin’s explanation of “falsifiable.”

In Smolin’s preprint, on page 3, Smolin writes,

According to Popper[1], a theory is falsifiable if one can derive from it unambiguous predictions for doable experiments such that, were contrary results seen, at least one premise of the theory would have been proven not to apply to nature.

This is definitely not phrased to impart knowledge, so it pays to keep reading. Smolin then clarifies by saying,

While the notion of falsifiability has been challenged and qualified by philosophers since Popper, such as Kuhn, Feyerabend and others1 , it remains the case that few philosophers of science, and few working scientists, would be able to take seriously a proposal for a fundamental theory of physics that had no possibility of being disproved by a doable experiment.

[First, I'd say that few working scientists have studied enough logic to know what a universal and existential quantifier are, which are at the core of falsifiable. And even Popper does nothing in his book to emphasize to the reader that falsifiable is meaningless once separated from the context of these quantifiers, which is what happens when people use a watered down version of falsifiable.]

I’ll rephrase Smolin’s last clause in the quote above: A proposal for a fundamental theory of physics should not be taken seriously if there is no possibility of disproving it by a doable experiment.

Immediately, to the uneducated and unsophisticated mind, such as my own, the question arises, “But what if your proposal is right, and it can’t be disproved?”

But the sophisticated and educated merely roll their eyes at such a question and know that it would be in vain to try and explain such lofty ideas as “falsifiable” to those who would ask such a question, although I remember Crazy Lubos basically making the same assertion, that something can’t be falsified if it’s true. However, although Crazy Lubos is educated, he’s rough around the corners, thus he’s unsophisticated, so we can ignore Crazy Lubos on matters involving falsifiable.

Now, I must admit that I’ve only read the first two pages of Popper’s book, but the fact is, falsifiable is meaningless if divorced from the context that science is 100% inductive by nature, and that falsifiable is 100% wrapped in the logic of the universal and existential quantifiers, the existential quantifier being the negation of the universal quantifier. And this context I’m talking about is laid down by Popper in the first 3 or 4 paragraphs of his book, The Logic of Scientific Discovery, although Popper doesn’t spell it out like I just did.

And on the surface, really smart, educated, and sophisticated falsifiable evangelists appear to be using the logic I just mentioned to justify why legitimate science must be falsifiable. For example,  Smolin tries to convince us that the negation of a statement, rather than confirmation, is what’s of ultimate importance. He says,

…confirmation of a prediction of theory does not show that the theory is true, but falsification of a prediction can show it is false.

So in other words, because science is 100% inductive by nature, the act of dropping a rock ten-thousand times and watching it fall to the ground ten-thousand times does not prove that the rock will fall to the ground the next time you drop it. After all, only God knows the future.

Well, okay, I gave a bad example, because really smart, sophisticated, and educated people don’t speak of such simple experiments when talking about falsifiable; they don’t use such simple examples from which we can derive propositions that can never be falsified.

I’ll end on an assertion. In experimental science, confirmation rules, because confirmation is what gets practical results. Where falsification of a prediction tells you that you’re wrong, confirmation tells you that you’re on the right track, and getting on the right track is what pays the bills.

No, that can’t be the end. If you understand that it doesn’t pay to get dogmatic, then you understand that falsifiable is much to do about nothing much.

And if you understand that confirming something N times is valuable information, as opposed to never being able to prove that “every time it will happen,” then you understand that falsifiable is much to do about nothing much.

Smolin says  that falsifiable is needed to keep science from grinding to a halt. Somehow, I don’t think Smolin’s needs are the same as the needs of an experimental scientist. I can imagine that the first time a scientist gets some results, his or her heart beats a little faster. And on the second or third time their prediction bears some fruit, they’re feeling good. And when they confirm that what they’re doing is the real thing, they’re not thinking about falsifiable at all.

Now, once they start trying to refine what they’re doing, then I suppose a concept similar to falsifiable comes into play. But the fact is, if a statement holds true N times under experiment, then something has been proved which can’t be unproved, something has been proved that can’t be falsified.

And once a statement has been modified, it’s not the same statement any more. So if a modified statement is falsified, then it’s not the same thing as falsifying the original statement. If you understand these kind of things, then you understand that falsifiable is much to do about nothing much.

Finally, here’s where the use of falsifiable can get absurd.

Falsifiable is supposed to be a safeguard against the situation where because an experiment has been successful N times, a dogmatic claim is made solely on that evidence. Because the N swans we’ve seen have all been white, we say dogmatically that all swans are white.

The absurdity is that certain propositions are  said to be falsifiable, thus qualifying them as science, when the claims have never been verified by experiment; that is, they’ve been verified zero times by experiment. Of course, this brings up another absurdity that’s needed for science by consensus, and that’s the equivalence of “in principle” with “actual experiment.”

And from the absurdity of this equivalence, we get some absurd logic, that if a proposition A is falsifiable in principle, that is, if in principle we have a hypothetical, doable experiment, then the negation of A, not A, must also be falsifiable, which means that both A and not A are both valid scientific propositions.

For if we’ve thought of an experiment that will confirm either A or not A, but we haven’t yet conducted the experiment, then for proposition A, the possibility exists that the experiment will show not A. And for proposition not A, the possibility exists that the experiment will show A. Therefore, both A and not A are falsifiable.

Rephrased, because the experiment hasn’t yet been performed, the possibility exists that proposition A will be confirmed and not A will be falsified, and the possibility exists that proposition A will be falsified and not A will be confirmed. Therefore, if proposition A has been given the status of science because of an in-principle experiment, then not A must also be given the status of science.

From Woit’s This Week’s Hype.

1. anon. Says:
   May 24th, 2007 at 6:15 am ‘I guess it’s all right that I don’t have time to comment on this, since no comment seems necessary…’ – Peter WoitGlad that you are not wasting time on fruitless arguments.Lubos is correct. Physics proceeds by asserting a theory is true and must be believed until it is disproved. Even when the theory is disproved, you must continue using it until a better theory comes along. It is very arrogant of certain people to assert that a person’s defense of extradimensional dogma isn’t physics. You must first prove string wrong, and provide the correct theory to go in its place.
2. Rick Says:
   May 24th, 2007 at 6:41 am anon, it should go without saying that what you’ve said is nonsense. Science doesn’t proceed via theories that -cant- be disproven nor bring much to the table.I’d say a theory can’t be taken seriously unless it can propose a way in which it could conceivably be disproven, even if it has some amazing utility and explains many things. Unfortunately string theory fails all 3 of these tests so it’s even more useless than a generally unfalsifiable theory.

Woit in Carroll and Johnson on Bloggingheads,

Sean gives a fairly standard defense of the landscape (”maybe it really is the way the world works”), with no discussion of the main problem with the landscape, that it shows no signs of making falsifiable predictions that would make it legitimate science.

From the comments:

• Eric Mayes Says:
  July 16th, 2007 at 9:45 am
  Peter,
  Bee made the connection with string theory. However, there are several falsifiable tests of string theory at LHC, namely supersymmetry and large extra dimensions. You may claim that a discovery of either of these does not prove string theory, however either of these would be very strong evidence in it’s favor.
• Peter Woit Says:
  July 16th, 2007 at 9:49 am
  Eric,You seem to have a non-standard definition of “falsifiable”. Supersymmetry and large extra dimensions are not falsifiable predictions. Their status is the usual one of string theory “tests”: if we see them, that’s great for string theory, if we don’t, no problem. Heads we win, tails doesn’t count…
• Eric Mayes Says:
  July 16th, 2007 at 9:57 am
  Peter,
  Sure, it’s falsifiable. All we need is a big enough collider. Now, it may be that LHC isn’t big enough, but this a technological liimitation not a fundamental one. However, it is clearly possible to find evidence in favor of string theory at LHC, even if LHC will not be able to rule it out.
• Peter Woit Says:
  July 16th, 2007 at 10:03 am
  Eric,“there are several falsifiable tests of string theory at LHC”“Sure, it’s falsifiable. All we need is a big enough collider. Now, it may be that LHC isn’t big enough”Funny the way these string theorists keep changing their story when you point out to them that they’re wrong…
• Eric Mayes Says:
  July 16th, 2007 at 10:18 am
  Peter,
  Do you deny that it’s possible, in principle, to find evidence in support of string theory at LHC?
• Peter Woit Says:
  July 16th, 2007 at 10:31 am
  Eric,Sure, although I don’t think it’s likely. But it’s really a tedious waste of time dealing with the comments that you keep posting here which make completely false claims (e.g. “there are several falsifiable tests of string theory at LHC”), which you never acknowledge are untrue, instead continually returning here to make the same false claims again and again.
• Eric Mayes Says:
  July 17th, 2007 at 4:52 pm
  Again, I would like to make the point that it is possible for string theory to be falsified by theoretical criteria alone, although this hasn’t happened. Peter’s main argument that it isn’t falsifiable is wrong for this reason alone. In addition, it is always possible that new, unexpected phenomena will be discovered that are not compatible with string theory.
• Santo D’Agostino Says:
  July 18th, 2007 at 9:09 am
  Yes, agreed, any theory can be falsified if it is logically inconsistent, and this does not require any experimental input. However, the point about falsifiability that has repeatedly been made by a number of people is that because no predictions have yet been made by string theory, it is impossible for any experimental result to falsify it. This is what I take the “not even wrong” label to mean.
• Eric Mayes Says:
  July 18th, 2007 at 9:50 am
  Santo,I’d like to emphasize that it may very well be possible to make definitive predictions in the near future and there very well may be experimental results in the future that may be able to confirm or falsify it. In my opinion, the claim that string theory is not falsifiable is itself ‘not even wrong’ because we really don’t know what new developments or discoveries my be found in the future. Saying we shouldn’t study string theory because we do know enough about it yet to make predicitions is rather dumb. How are we ever going to understand it well enough to make predictions if we don’t study it?
• anonymous Says:
  July 18th, 2007 at 11:36 am
  Eric,…You seem to have a problems with the present tense. Right now, at this very moment, string theory is ‘not even wrong’ because it makes no predictions of any value. Saying that it ‘may very well be possible in the near future’ to make predictions doesn’t obviate the fact that — right now — it does not.
• Eric Mayes Says:
  July 18th, 2007 at 11:58 am
  Anonymous,
  This is the nature of research in progress. I’m really getting tired of this abuse of Pauli’s phrase ‘not even wrong’. He certainly didn’t mean it in the strict, limited way that you and Peter use it. Essentially, he meant that a theory was ‘half-baked’, i.e. not consistent. The inability to test a theory due to current practical limitations is not a valid reason for dismissing a theory.
• Eric Mayes Says:
  July 18th, 2007 at 12:32 pm
  As an example, consider the theory that intelligent life exist elsewhere in the universe. This idea is almost certainly right, however I can’t think any way to disprove the idea. Is this idea ‘not even wrong’?
• Santo D’Agostino Says:
  July 18th, 2007 at 3:43 pm
  Scientific theories make predictions about the results of future experiments, then the experiments are carried out. If the experimental results match the predictions, within acceptable limits, then fine. If not, then the theory is discarded or modified. That is how the game is played.You have not responded to my requests for predictions of string theory, although you imply that there are some, just at energies that are too high to access now. Are there any? …
• Eric Mayes Says:
  July 18th, 2007 at 4:48 pm
  Ok, here are my predictions. Supersymmetry and the Higgs will be discovered at LHC with the Higgs mass between 116-121 GeV.