Hacker News new | ask | show | jobs
Linguistic Harbingers of Betrayal (vene.ro)
102 points by gcv 3937 days ago
5 comments

I feel the planning is the most decisive symptom. A person who doesn't see a future won't project themselves into one.
The excess of positive language prior to betrayal point definitely matches my personal experience.
The reduction in politeness is also interesting. As someone who really enjoys formality, I like to think that forms are a kind of social guardrail to reduce the likelihood of misbehaviour (which would of course include betrayal), but that's probably far too pat.
Formality won't reduce betrayal but instead prescribe behavior that lets one fully mask intention of betrayal. When expected behavior norms aren't fully prescribed, or when they're impaired (as when one is drunk), it's harder to maintain politeness. Downton Abbey-style formality, on the other hand, lets one maintain the facade of perfect society while stabbing someone's back.
> The reduction in politeness is also interesting.

It's very counter-intuitive. I'd like to hear a rationale for this strategy.

My guess is dehumanization. The betrayer has to shove down their empathy, and that blocks them from thinking of actions that flow from caring about another's feelings - except instrumentally. So direct feeling manipulation, "positive sentiment", goes up.
This may be measuring relative power or favorability of position for the two players. Weaker players are more polite to their stronger allies (and less likely to betray), strong players are less polite and more likely to betray, for example.
They don't have an app for this yet? There's an obvious startup here.
I was thinking the same thing. Diplomacy is one of the best multiplayer games ever designed. This should be ideally suited for writing as a mobile app. I'd love to know if there any out there worth checking out.

There's this at the iPhone App Store: "The Game of Diplomacy by Chris Hughes" https://appsto.re/us/iCIzt.i

>There's an obvious startup here.

Maybe, but the related startup idea that came to mind for me might seem a bit odd; similar linguistic patters might exist in statements made by startup founders and they might be indicative of success or failure. The "excessive positivity" and "unnatural politeness" seen prior to betrayal may also be present prior to startup failure.

The real problem with attempting to prove or disprove this conjecture seems to be getting the needed data. It might be possible from public statements from founders, but I think the more significant gems would only be found private correspondence between the founders and investors (or possibly customers). I sincerely doubt YC partners or VC's like Marc Andreessen (@pmarca) and Ben Horowitz are going to let "just anyone" do linguistic analysis against all of their private personal/business email content.

Not sure exactly what you meant, but an app that does sentiment analysis on your text conversations on your phone would be pretty cool.

It wouldn't just detect potential betrayal; it could show all kinds of cool (but useless) stuff.

that is not purple
"Our men at the Wernicke and Broca Station Posts suggest that the good Ambassador Moustache is not as trustworthy as he seems..."