We will soon grasp in a deep way how collective human behavior works, whether it' s action by small groups or nations. It will not be infallible, the complexity of such behavior makes exact prediction impossible.

You can expect the collective mind to take over from time to time, directly guiding the individual minds. Sometimes we act in hordes. Angry mobs and frightened crowds seem to qualify as examples of a collective mind in action, with nonlinguistic channels of communication usurping the individual capacity for rational behavior. Surely this tribal mind doesn't operate in normal, day–to–day behavior—or does it?

We have found that we can use measurements of social signaling to predict a wide range of important behavioral outcomes with high accuracy. Accurate predictions can be made, even for lengthy interactions, by observing only the initial few minutes of the interaction, even though the linguistic content of these "thin slices" of the behavior seems to have little predictive power.

We are examining some of the most important interactions a human being can have: finding a mate, getting a job, negotiating a salary, finding one's place in social network. These are activities for which we prepare intellectually and strategically, sometimes for decades, and yet the largely unconscious social signaling that occurs at the start of the interaction appears to be more predictive of its outcome than either the contextual facts or the linguistic structure.

So what is going on here? One might speculate that social signaling evolved as a method of establishing tribal hierarchy and cohesion. So here is what I suspect but cannot prove: A very large part of our behavior is determined by mainly unconscious social signaling, which sets the context, risk, and reward structure within which traditional cognitive processes proceed. In short, it may be useful to think of humans as having a collective, tribal mind in addition to personal ones.

— Alex Pentland