Threshold Models of Collective Behavior Granovetter 1978

This paper is thick with the vocabulary of game theory. he is influenced by the stuff, and talking to an audience preoccupied with it, but he is proposing that 'thresholds' in model of behavior can effectively model contagion and other things modeled in the same way as contagion. Judging by the citation of the 2005 paper, this must have been very influential.

I think he end sup using non-dimensional values. I wonder if it made the work easier.

Since this is before easy access to computers, I think he is making the mean field approximation, though i haven't verifired that. What I mean by that, is 'well mixed' 'no space'.

Much of the work at this time, according to a quote on p 1433, is reacting agaisnt previous work that assumes irrationality is central to riot behavior and other group behaviour. They are showing, though non rigorously, that it can be rational, or that individually rational behavior can lead to 'irrational' outcomes.

Roger Brown is a big deal here.

Compared to pay-off matrices of game theory, threshold analysis allows more complicated systems, but has to settle for statistical conclusions. it is a gross, or slightly scaled out measure. Is this paper about using statistical measures on game theoretic models?

I've never seen 'sissy and faggot' in a scientific paper before. guess I shouldn't take it out of context. p 1435.

Great changes in behavioor must not be confused with corresponding changes in threshold. 

he gives the examples of a lynchmob, saying that though we would expect momentary irrationality, this is not necessary to lead to a lynchmob.