2

Following Do we apply scientific skepticism at Skeptics.SE? I'm left with a curiosity, but I'd like to learn more. Can you help me out?

According to some we are not using all tools allowed in scientific skepticism, and specifically we're "missing" pure logic arguments, unpublished theoretical proofs and unpublished empirical research.

Restricting this to non-empirical tools, it seems to me that these tools are not really usable in skepticism in general, so here's my doubt.

A factual claim might contain poor logic, however pointing that out does not disprove the claim.

For example: "Men are taller than women on average, therefore any man is taller than any woman" is not disproven by pointing out the fallacy. Either all men are taller than all women or not.

Another example is debunking any claim of causation due to correlation by pointing out that correlation is not causation (correlation is necessary for causation so certainly it does not disprove it).

It's furthermore obvious that a factual claim which contains solid logic is not necessarily true as its factual premises might be wrong.

For example: "All introvert people are lonely. All programmers are introverted, therefore programmers are lonely." This claim contains a valid syllogism but showing that doesn't prove that the claim is valid: it does not show that all introverts are lonely, nor that all programmers are introverts, let alone its conclusion.

Equally a claim might be based on a solid theory, but showing that in an answer that does not prove the claim is correct.

For example: "Do gravitons exist?" can't be answered by simply citing general relativity. While they are predicted by it, we haven't seen any yet and it's perfectly possible that they do not exist, and that we have to change the theory, after all.

The vice-versa is also true:

If someone claims that a medicine works based on a paper containing mathematical mistakes, showing that it does, does not "prove" that the medicine is ineffective, or that it is effective, nor it "proves" that we don't know whether it works or not.

The above paragraphs are examples of why stating that non-empirical logical/theoretical answers are enough to confirm/debunk a claim seems ...alien to me.

I'd like to know what valid examples people think of, when they state that these tools are allowed in scientific skepticism

Basically, I'd like to know what we are missing.

Can you think of examples of any on-topic, factual claims that can be settled by pure theory or pure (formal or mathematical) logic?

  • 2
    I'm pedantically mentioning "Is Fermat's Last Theorem True?" as an on-topic (!?) factual claim, so that I may dismiss it as a pedantic special-case, where the claim is entirely mathematical, not real-world. – Oddthinking Jun 9 '15 at 5:48
  • 1
    @Oddthinking would the actual proof to the theorem be a valid skeptical answer? Simon Singh, certainly a skeptic, wrote a whole book on it, and it wasn't the proof itself. OTOH, math theorems can be proven by logic, but I'm not sure the questions are on topic. I guess they could be rephrased as "has the fermat theorem been proven?" – Sklivvz Jun 9 '15 at 8:19
  • 2
  • 1
    @ChrisW Strange comment, yours: in fact, I am asking about questions (but you link to an answer) and furthermore the answer is not of the kind I'm speaking of, as it has 7 references. Maybe you are linking to the wrong post or haven't read my question carefully? – Sklivvz Jun 9 '15 at 13:55
  • 1
    IMO that is a theoretical answer, i.e. a claim settle by theory: it says why, in theory, a conspiracy is implausible. Also the references are merely cosmetic, added post facto to satisfy your requirement. Of course you can say, "yes, but that's not pure theory". If you want to insist that pure theory is never sufficient that's OK too: even an example like this one might need evidence that all humans are mortal and than all Greeks are human. – ChrisW Jun 9 '15 at 14:32
  • 1
    Say someone wanted to argue that atmospheric CO2 rises not due to fossil fuel use, but due to degassing from the oceans in response to an increase in global temperatures and that CO2 doesn't cause temperatures to rise. They demonstrate that CO2 doesn't cause temperatures to rise by showing a graph of increasing CO2 and the "hiatus" in GMSTs. Pointing out the internal contradiction there ought to be enough to show the hypothesis is faulty. – Dikran Marsupial Jun 9 '15 at 15:24
  • 1
    @DikranMarsupial that would show their demonstration is incorrect, which doesn't imply the claim is incorrect. Compare with this "I claim that aliens visited earth because they can travel faster than light" being answered with "Nothing can travel faster than light". – Sklivvz Jun 9 '15 at 16:41
  • 1
    @ChrisW no, I merely note that you are just giving examples of what is allowed. I specifically asked for answers not allowed here but allowed in "scientific skepticism". It seems self evident that answers already allowed here are not good examples ;-) – Sklivvz Jun 9 '15 at 16:44
  • 1
    @DikranMarsupial "a lack of correlation between temperature and CO2 pretty much refutes the claim", indeed, but that's a factual statement that is best settled by experiment, or are you arguing that that could be settled otherwise? – Sklivvz Jun 9 '15 at 16:48
  • 1
    @DikranMarsupial We already allow to reference something in the answer when its validity is not in question. Why should we not? – Sklivvz Jun 9 '15 at 16:52
  • 1
    @ChrisW it's not an edge case at all: it's a theoretical answer which is referenced according to our rules. It's missing some reference for some of its claims but there's a lot of non-cosmetic stuff in there (i.e. stuff which is much more than definitions but directly addresses the points in the answer). – Sklivvz Jun 9 '15 at 16:54
  • 1
    @DikranMarsupial If a reference is used in a question this doesn't necessarily prevent an answer to use it otherwise. For example if I ask a question about a news article reporting on some experiment, the answerer may refer and use the article itself to answer my question. Clearly, if it's something as trivial as "you didn't read it properly" we are going to close the question as non-notable. But there are cases in which more meaningful questions and answers were given. – Sklivvz Jun 9 '15 at 16:58
  • 2
    I think that if it weren't referenced according to our rules then it would still be theoretical, and still an answer (albeit no longer "according to our rules"). Apparently it's not the kind of answer you're looking for here, though, so ... – ChrisW Jun 9 '15 at 17:00
  • 2
    I tried to answer your original question, and my answer was converted to a comment. I cannot improve on that answer nor explain it further, so I think that's the end of that; please forgive me if I don't try to address a new/follow-on question i.e. "What answers are prevented by requiring references?" – ChrisW Jun 9 '15 at 17:09
  • 2
    That seems pretty illogical to me. To agree that claims can be answered by purely logical means in some cases, but that they are not valid skeptical tools seems like a contradiction to me. Certainly in science, scientists would be highly skeptical of claims with internal contradictions/inconsistencies! – Dikran Marsupial Jun 9 '15 at 17:14
1

What claims can be debunked or confirmed by pure logic or theory?

Ones where you can point out an internal inconsistency in the logical basis for the claim.

  • True. So claims like "this is a true demonstration!" can be debunked by "no it isn't, here's the error". Factual claims on the other hand can be true even if the "demonstration" preceding them is inconsistent. – Sklivvz Jun 9 '15 at 17:15
  • I gave an example in the comments where the experimental data given as part of the claim are enough to show that the claim is false, hence that point has already been addressed. – Dikran Marsupial Jun 9 '15 at 17:32
  • 1
    @DikranMarsupial: Yes, I've used the original reference to disprove the claim. In those cases, I have a reference - it just happens to be the one cited in the question. – Oddthinking Jul 1 '15 at 18:01

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .