Anti-Semitic Auto-complete?

By Symi Rom-Rymer

Journalist Stéphane Foucart posed an interesting question in a recent article in the French daily paper, Le Monde: what can Google tell us about our prejudices?  Intrigued by an earlier piece in Télérama, a weekly French magazine, that pointed out that the word ‘Jew’ often appeared in the Google search drop-down menu when someone typed in the name of almost any French top media executive or public leader, Foucart undertook his own unscientific study; producing the same outcome.

Based on this experiment, he took the results as a sign that the canard that Jews run the media or exert undue influence on French politics still hold sway among the general population. He furthermore argues that this is uniquely a French problem since the word ‘Jew’ did not  come up in the American or Spanish versions of the search engine, given similar inputs. Therefore, he concluded that Google had laid bare the French people’s latent anti-Semitism.

But what does this experiment really reveal?  Much of the answer to that question relies on the nature of Google searches.  Search results are derived from an algorithm designed to avoid a small group of people, intentionally manipulating the system to associate one word with another.  Common search words that pop-up in the search window cannot, according to a Google’s spokesperson, be created by a small group of people wishing to influence larger searches.  For example, a neo-Nazi group in Paris could not force the Google search engine to automatically link the word ‘Jew’ when someone types in ‘get out of France.’  Instead, Google judges the popularity of searches based on those carried out by each unique IP (Internet Protocol) address.  In order for a search term to grow in popularity, then, it must be typed in on a significant number of distinct computers each time for it to be recognized by Google.  Thus, the reason ‘Jew’ comes up so often in tandem with public figures is because a significant number of French people are entering those terms into the Google search engine on their own computers.

But Foucart ends his inquiry too soon.  If one actually clicks on the search term in question—let’s say ‘Nicolas Sarkozy’ + ‘Jew’—the picture become more complicated.  Contrary to what Foucart might expect, the top three links that come up are not, in fact, to extremist organizations.  Instead, they are links to Sarkozy’s Wikipedia page, a pro-Jewish personal blog praising the French president’s ascension to power, and an article in Haaretz.  Similarly, ‘Bill Clinton’ + ‘Jew’ (another automatic paring) brings up articles about his daughter’s marriage to Marc Mezvinksy. Because the most popular sites that appear in a Google search are those that appear at the top of the page, it seems that more French-speakers are more interested in reading Sarkozy’s Wikipedia page or about Chelsea Clinton’s wedding than anti-Semitic conspiracy theories. If ‘Jew’ is a common search term in the Google French engine, but the most popular pages are innocuous or even laudatory, as in the Sarkozy example, it is questionable that can we confidently conclude that the ‘Jew’ search, however discomforting it may be, is the most trustworthy indication of wide-spread anti-Jewish feeling in France.

Without detailed study, it is impossible to know why the French conduct the searches that they do or what they are hoping to find.  Perhaps they are obsessed by their leaders’ supposed Judaism.  Or perhaps they turn to the internet, just as Americans do, to confirm or deny rumors they might hear from the media or others about certain public figures.  Reports of increasing anti-Semitism in France, like anti-Semitism anywhere is always troubling, but sometimes the supporting facts are more ambiguous or are simply not borne out by deeper examination.  Sometimes, a cigar is just a cigar.