Pascal's Wager and Open Access

Sally Morris (ALPSP) wrote in the SPARC Open Access Forum:

SM:There's a difference between 'refute' (= produce evidence to disprove) and 'rebut' (= argue against). Stevan's letter does the latter, not the former; there is no evidence whatever that self-archiving will not damage journals or those who produce them.

(Umm, first, that's Berners-Lee et al's letter, not Stevan's letter...;>)

Second, there is no evidence to refute Creationism either: Just no evidence for it, and all existing evidence against it (in both cases). So one can only rebut, not refute, in both cases. (I might add that the ALPSP reponse is likewise merely rebutting RCUK, not refuting, but the difference is that the ALPSP adduce no valid evidence in support of their rebuttal, and that is precisely what our rebuttal points out: along with the positive evidence in support of the RCUK policy and contrary to the ALPSP rebuttal).

I suggest that Sally look into the logic of hypothesis-testing and empirical inference. One does not, in the real empirical world, say "I conjecture, and you cannot refute": Refutation (disproof) is only possible in mathematics -- by proving that something is logically impossible, self-contradictory. For anything else that is not logically impossible, we seek not refutation but supporting or contrary evidence. For the proposition "Self-archiving will ruin journals" (or even that it will reduce subscriptions) there is no supporting evidence to date, and all evidence to date is to the contrary: that self-archiving is neither ruining journals nor even reducing their subscriptions.

Sally would do well to look at "Pascal's Wager" (as I have urged her to do before):

Pascal thought that it was more rational to behave as-if the Creed (that there is an afterlife, with eternal damnation for nonbelievers) were true, because the costs of behaving as if the Creed were false if it was in fact true (eternal damnation) were so much greater than the costs of behaving as if the Creed were true even if it was in fact false (leading a slightly more constrained but finite life).

What Pascal missed was that the force of this unassailable logic came from one unquestioned but questionable premise: The (arbitrary) threat of eternal damnation, merely on the Prophets' say-so. It was the direness of the purported consequences that made the logic look unassailable. (Any rival Prophet could have raised the Wager by promising even more dire consequences [e.g., one's soul splitting into an infinity of sub-souls, all suffering one another's anguish for an eternity of cardinality Aleph-1, where each instant lasts an eternity] if one fails to behave according to that Creed, and so on.)

What this shows is that one does not make a point by just positing the dire consequences that would ensue if one does not take the point.

I, for my part, am not prophecying ruin for research if researchers fail to self-archive. I am merely demonstrating exactly what research and researchers are actually losing, daily, monthly, yearly, as long as they don't:

Sally should give up the Doomsday business too...

Stevan Harnad