October 16, 2007

John Kay mounted his horse last week in defense of scientific truth and lead a mighty charge against that ungainly dragon, "scientific consensus" (*). His motives are pure, his analysis impeccable. Indeed truth comes under the distinct species of science, authority and popularity. Combining all three, "scientific consensus" is a devilish chimera. And yet, with all due respect, presumption tainted John Kay's argument when he declared "if [one] converted all members of the Royal Society to creationism, that [...] would not affect my belief in evolution".

Look for the key in a book by Karl Sabbagh (1). Its title promises a look at the Riemann hypothesis (2). It turns out to paint how professional mathematicians go about their work. Among them is Louis de Branges (3), who wrote a proof of this difficult problem and duly put it on the Internet for all to read (4).

Publication of a scientific proof is no proof in itself though. It must be either verified or found erroneous. But what if the popular consensus among the scientific community is against risking the necessary resources? As a site (5) puts it in regards to this and similar claims: "If you are a university mathematics lecturer who teaches analytic number theory, you might want to consider setting your students the task of deconstructing the more serious of these [proposed proofs]. They may otherwise never be given any serious attention, which would be a shame". Out of the mainstream, truth must stay hidden. Even if someone ventures to prove you wrong, chances are nobody will bother endorse your critique either (6).

I should know. Recently Google and Facebook have created some controversy with their plans, whether made public or still under wraps, to enter personalized advertising. To unprejudiced observers, it appears there is a trade-off between the future of advertising and the right to privacy. The truth is, no such trade-off is required as ePrio has shown and patented. Society though is not well equipped to shoulder the verification of my claims.

Indeed privacy has proven to be a poor prospect for venture capital, whose goal is to make money. Can ePrio succeed where ePrivacy Group (7) and Zero-Knowledge (8) have failed? Meanwhile the media establishment is there to record the voice of those who have one, not to give one to those who have none. To balance Eric Schmidt's pleas for Google, which it published prominently (**) on its Comment page, the Financial Times printed a letter to the editor from Epic's president, Marc Rotenberg (***), rather than from ePrio's. But lest they compromise their independence by creating favorable conditions for commercial gains, Epic and other non for profit organizations are paid not to find a solution. However relentlessly, they can only complain about attacks on privacy and thus indirectly reinforce the notion that a trade-off is inevitable.

Scientific truth cannot be verified against the consensus of the time. By proclaiming heliocentrism as a fact, Galileo stepped out of the mainstream of astronomy in his day and suffered the consequences. How can John Kay claim himself so ready to lose his reputation, his FT column, his income? I wish him never to be put to a test less hypothetical than his, which assumes a sudden and complete reversal between mainstream and fringe science.

Some reversals happen though. May truth about privacy find its public sooner than later.

Philippe Coueignoux

• (*) ............ Science is the pursuit of the truth, not consensus, by John Kay (Financial Times) - October 10, 2007
• (**) .......... Global privacy standards are needed, by Eric Schmidt (Financial Times) - September 19, 2007
• (***) ........ Letters To The Editor: Google's proposals on internet privacy do not go far enough, by Marc Rotenberg (Financial Times) - September 24, 2007
• (1) The Riemann Hypothesis, by Karl Sabbagh, Farrar, Straus and Giroux, New York - 2002
• (2) see the Riemann hypothesis in the wikipedia
• (3) see Prof Louis de Branges' home page
• (4) see Prof Louis de Branges' papers
• (5) quoted from a list of proposed proofs of the Riemann Hypothesis
• (6) see this unendorsed refutation of Louis de Branges' proof
• (7) see Eprivacy Group in the wikipedia
• (8) see Zero-Knowledge in the wikipedia