July 20, 2010

Four months ago, I claimed that the right to vote ought to be balanced by the duty to learn. Joe Keohane's essay on "how facts backfire" appears to put a damper on my suggestion (*). If "facts don't necessarily have the power to change our minds", then compelling citizens to face them may not eliminate the influence of money and defeat the forces of pronaocracy.

The import of Joe Keohane's analysis however goes beyond political science. It is nothing less than an enquiry about the nature of truth, the necessary foundation of our Information Age. Before drawing practical lessons from its findings, we should perhaps understand them to begin with.

If there is a flaw in the essay, it is that it avoids defining what a fact truly is. As the author contrasts "facts" with "beliefs" and "things [that] are objectively, provably false", this flaw is minor. What Joe Keohane calls a fact is a statement backed by science, as opposed to one backed by either popular opinion or faith in an authority, which he calls beliefs.

With this clarification in mind, it is easy to share the author's rhetorical puzzlement. Meeting with someone who tells you the earth is flat or insists President Obama is not a natural born citizen is as frustrating as it is exhilarating to prove a mathematical theorem by oneself and know it for a fact. So "what's going on?"

Joe Keohane provides two explanations. One is the sheer amount of bad data, "endless rumors, misinformation, and questionable variations on the truth", only too happy to feed our voracious appetite for any and all data. Stating however "part of the answer lies in the way our brains are wired", the essay focuses on mental processes. "It's absolutely threatening to admit you're wrong" is how "political scientist Brendan Nyhan" summarizes the scientific study he led on the subject at the University of Michigan. Willfully ignoring facts proves to be an effective, if not efficient, way to cope.

I find these two explanations convincing but, as Joe Keohane implies, other causes may also play a role. If Man is flawed, Society sure is no help.

First comes the deep love Americans have for their court system. Adversarial truth finding can make for great drama, witness the countless novels, plays and movies revolving around a trial (1). The issue is that few react by identifying with the jury and most with the lawyers. As the latter battle in court to shed the most favorable light on their clients, "rather than facts driving beliefs, [their role] dictate[s] the facts [they] chose to accept"..

Take Marissa Mayer's defense of Google against accusations of manipulating search results (**). She is quite right in pointing out that transparency invites abuse. She is right again when she says that mandating "neutral search" rules on every recommender would undercut the originality of their recommendations. But she skillfully omits the one fact which singles out the Google search engine as a public utility, i.e. its current dominance.

The second factor is the rise of statistics as today's most influential tool. Scientific facts mean nothing outside of their context. What for instance makes it provably true, no matter the subject, that if, out of 10 people, two persons agree, everyone will agree (2)? In statistics however context includes a complex modeling of uncertainty and too often conclusions are presented as if the hypotheses which define the model were facts.

Recall how unkind real facts have been to the efficient market hypothesis and the assumption of Gaussian distributions, both widely used in pre-crisis financial market modeling. All scientists understand model dependencies, quite a few unfortunately forget their vocation and become advocates.

Justin Gillis's report about a climate research scandal is proof to this disturbing trend (***). One report "find[s] [the accused scientists'] rigor and honesty as scientists are not in doubt", but adds they "[failed] to display the proper degree of openness". Another agency found that "the climate change panel had tended to emphasize the negative effects of global warming while playing down positive ones". Truly, this is how lawyers behave.

Scientists of course have every right to harbor and defend their own beliefs. But by mixing their personal advocacy and their role as scientists, they blur the border between truth backed by science and truth backed by authority. They encourage the confusion between facts and beliefs.

Third factor, some go even further and willfully turn science itself into a new religion. Auguste Comte's latter day disciples (3) are entitled to their beliefs. They may be provoked by proponents of other religions whose narrow-minded interpretations are hostile to science. But they should refrain from sacrificing science under the pretense of protecting it. The resulting noise rivals the Babel tower effect.

A national bookstore chain puts religious books in its science section. The Wellesley magazine prints an online survey "to test [the] understanding of basic science" with the following question, "which best describes your ideas about intelligent life beyond Earth?", and answers such as "it probably exists - philosophical reasons" and "it probably doesn't exist - religious reasons" and "aliens regularly visit Earth and abduct people" (****).

Left with the choice whether astronomy professor Kim McLeod is in fact a sleeping alien checking her own cover or a scientist with an unhealthy penchant for privacy violations, I am afraid, very afraid. How much the answer to her question by a future Hillary Clinton (4) forty years from now?

If both Man and Society conspire against facts, can citizens pass informed judgments on electoral office candidates? The solution I advocate requires each eligible voter to pass a test to confirm his or her having achieved the required level of learning on what each and every candidate stands for. To ensure the freedom of the process, the questions would be picked by each candidate concerned. To enforce its objectivity, the answers would be vetted by an independent "jury of professional test manufacturers".

Suppose then candidate X claims President Obama is not a natural born citizen. X is free to include a test question as to "whether President Obama is a natural born citizen" but X will not be allowed to insert "No" as the answer by which the citizen will be found having understood X's message. The only correct answer will be "Candidate X falsely believes President Obama is not a natural born citizen".

Utopian! Yet weighing answers on a scale of real facts, genuine beliefs and false beliefs would "increase the "reputational costs" of peddling bad info", as Brendan Nyhan recommends. Candidates will sue. Yet isn't it better their complaints get to court before rather than after the election, as commonly done today? Besides these court dramas would educate the public like no other media if properly focused on the jury rather than the lawyers. Impossible? What is the US Supreme Court but a jury of nine professional experts who love to humble the lawyers appearing before it.

Joe Keohane's essay reinforced my own beliefs, it is a fact, I am afraid. But perhaps you now share my defense of democracy, and hence privacy.

Philippe Coueignoux

• (*) ....... How Facts Backfire, by Joe Keohane (Boston Globe) - July 11, 2010
• (**) ..... Do not neutralise the web's endless search, by Marissa Mayer (Financial Times) - July 15, 2010
• (***) ... Scientists Cleared of Rigging Climate Research, by Justin Gillis (New York Times) - July 8, 2010
• (****) . Questions to test your science knowledge, by Kim McLeod (Wellesley Magazine) - Summer 2010
• (1) To Kill a Mockingbird, Inherit the Wind and The Caine Mutiny will do for a sample.
• (2) using base 2, as computer scientists will have already guessed.
• (3) for more information, see Auguste Comte in the wikipedia
• (4) for more information about this Wellesley 1969 class member, see Hillary Clinton in the wikipedia
  ...... as for the usual drivel on online interactions being anonymous and confidential,
  ...... how then can college send their own students to face charges of online piracy?