I treasure my lifetime subscription to Skeptical Inquirer magazine. I’ve probably learned more about evaluating scientific claims from reading it over the past several decades than from any other single source.
Recently, along with the typical torrent of terrific articles it unfortunately included a column “Whose Burden of Proof?”(1, SI, May-June 2013) with several potentially misleading ideas on Burden of Proof and Bayesian Analysis that are fundamental to critical thinking.
Here is my response letter published in the September/October 2013 issue. The author’s published response to this is included below.
(Note: The words in red (and between asterisks) were removed from the published letter.)
The article title I provided —
*“Big Difference Between Burden of Proof
for Skepticism and Denying”*
was changed to a different title and subject in the published letter to —
“Bayesian Analysis Prejudicial?”
Please allow me to suggest two clarifications to your article “Whose Burden of Proof?”
The article states that there is no difference between making a scientific claim and denying it — “to claim P is the same as to deny not P.”
While true, the article implies that denying and skepticism are equivalent when they are very different in substance and process.
Denying something is indeed a scientific claim and does have a burden of proof.
However, a skeptic requesting proof, evidence or definitions is not making a scientific claim and carries no burden of proof. Should we bear a burden of proof to ask a Dowser for evidence or to define his claims? I think not.
Fatal Bayesian Flaw
The second point is that a Bayesian philosophical framework has a serious flaw. Bayesian Inference is only one of many Philosophies of Science. It is distinguished by its partial dependence on prior judgments (“prior probability”) to alter conclusions about whether a hypothesis is valid or not.
By doing so Bayesian Inference essentially enshrines prejudice in an experiment’s result.
Imagine allowing a preconceived conclusion to affect the result of an experiment — the result is irreversibly tainted.
*Just like multiplying anything by zero [makes everything worthless] it doesn’t matter what the rest of Bayesian Inference involves in my view.*
Because Bayesian Inference allows use of a previous conclusion it loses any claim to purely merit-based decisions.
1. Whose Burden of Proof?, Massimo Pigliucci, Skeptical Inquirer, May/June 2013
2. “On Pseudo-Skepticism”, Marcello Truzzi, Zetetic Scholar, 12/13, pp3-4, 1987 :
“In science, the burden of proof falls upon the claimant; and the more extraordinary a claim, the heavier is the burden of proof demanded. The true skeptic takes an agnostic position, one that says the claim is not proved rather than disproved. He asserts that the claimant has not borne the burden of proof and that science must continue to build its cognitive map of reality without incorporating the extraordinary claim as a new “fact.”
Since the true skeptic does not assert a claim, he has no burden to prove anything. He just goes on using the established theories of “conventional science” as usual. But if a critic asserts that there is evidence for disproof, that he has a negative hypothesis—saying, for instance, that a seeming psi result was actually due to an artifact—he is making a claim and therefore also has to bear a burden of proof.”
3. Fun Article on Bayesian Analysis Flaws “Faster Than Light Neutrinos Succumb To Bayesian Method,” by Sascha Vongehr on Alpha Meme, Science 2.0
“If we allow belief to enter controversial priors, the scientific method will be rendered impotent. If scientists may do so, there is no reason why intelligent design (ID) should not insist on its own prior about that evolution is impossible.
If you are a real skeptic who is interested in science outreach and reestablishing the trust of the public, you do not force feed your beliefs via a perversion of the scientific method. You agree to unbiased priors and let the experimental data speak.”
The published response from the original author was . . . well, not at all clear to me. He wrote —
“When I pointed out the logical equivalency of p or ~p (sic) I did not mean to imply that denial and skepticism are themselves equivalent.
dd: That’s a good clarification to note, but a bit late as the article still leaves that point ambiguous – at the very least.
But skeptics cannot simply help themselves to the “you are making a positive claim” rhetorical move, since any positive claim can easily be turned into a negative one.”
dd: What ?!?
dd: While that doesn’t refute (or even address) the point that skeptics have no burden of proof, it appears to be an ironic “rhetorical move” by the author that confusingly misses the issue. Or is he contradicting his earlier clarification ?
He continues with —
“Concerning Bayesianism, it is certainly the case that it is one of several types of philosophical account of scientific epistemology, and it is also true that it has flaws (though I disagree with the modifier “deep”).
dd: He generally agrees with my point so far (though he distractingly disputes the term “deep” which I never used. I prefer the concept of “fatal” to describe Bayesian Inference’s flaws.)
“Still it is widely acknowledged to be both useful in both science and philosophy, and it is particularly helpful when thinking about burden of proof.
dd: He never explains how Bayesian Inference’s use helps understand Burden of Proof.
“Also, it is incorrect to say that ‘it enshrines prejudice in an experiment’s result.’ Rather, it helps make explicit and quantify the scientist’s judgment, which is inevitably subjective to begin with.”
dd: Again I find no rationale explaining why he disagrees. How is it incorrect? If he believes Bayesian Inference does not enshrine prejudice in an experiment’s result by its use of Prior Probability – shouldn’t we get some explanation of why not?
“In general, there is no such thing in science as ‘purely merit-based decisions,’ since all decisions are human judgments and therefore subjective. (Of course this doesn’t mean that such judgments are arbitrary or made regardless of evidence.)”
And that’s all he wrote. He never even addressed the “Prior probability” fatal flaw.
That contradiction without reasons is a logical error called “Appeal to the stone” meaning dismissing a statement without giving proof of its invalidity.
Since he didn’t rebut either of the two points, this means they (so far) stand unrefuted :
1. A skeptic requesting definitions, evidence or proof is not making a scientific claim and thus has no burden of proof.
2. Because Bayesian Inference allows (actually requires) conclusion altering by “prior probability” — Bayesian Inference essentially enshrines prejudice in an experiment’s result.
# # #