That book makes sense, to a degree. Language does not tell us how to think, but it can constrain what we say. I find that our words reflect what we think, particularly when it comes to our prejudices.
Common indicators are language that is inclusive or exclusive (we/us/I Cs them/they), notions of gender (eg. Feminism indicates a prejudice toward one gender).
I have found that when someone has a biased mindset (particularly an exclusive one) and you point out their mindset through their use of language, most moderate people start to think differently and as such, their language changes.
However, I find that emotive people and political correctness break this pattern. This parallels the "faith" gambit - logic never enters the equation.
Facebook already has a good mechanism for validating account owners which involves showing users random unlabeled photos of their friends and asking them to guess who's who.
I guess an excuse like "you got hacked" helps you get much more important information from your users.
No one is talking about how all of the sudden, a lot of accounts supposedly got "hacked".
How on earth could this "solution" to trick its users possibly scale to more than 1 billion active accounts?
Don't get me wrong, I absolutely believe that Facebook would like to link government IDs to each and every account if they could, but trying something like this would be the end of Facebook.
