Maurizio – Omnologos

Where no subject is left unturned

Perception of Risks, Social Networks and Globalization

with 2 comments

A radical rethink of the concept of risk is needed in light of globalization and the availability of social networking and communications tools


There was a time when big news would happen only to complete strangers on TV. Then came e-mail, and Usenet, and groups: and so when TWA800 exploded on July 17, 1996 I knew the brother of one of the dead. Time moved on, and on September 11, 2001 an acquaintance of mine was working in one of the World Financial Center’s buildings (he survived)

When the tsunami arrived at Phuket in Thailand on December 26, 2004, a very close friend of mine was there (he survived too). Finally during the attacks in London on July 7, 2005 one of the dead was Colin Morley, fellow member of Ecademy, the online social and business networking site I belong to since 2003.

Sounds ominous doesn’t it? But it should not be seen as a sign that there is an aura of bad luck around me.

Rather, as per Jeremy Waldron’s great insight in this week’s The New York Review of Books (“Is This Torture Necessary?“, Vol 54, N. 16 · October 25, 2007), the point is that security “is not just an individual good, enjoyed by each of us as a matter of [individual] statistical probability“.

Security must be rethought as a group’s, not just a person’s. Even if on 9/11 “99.999 percent of the US population […] were not killed“, the fact that 2,974 did was a hit on the sense of security of all that could imagine themselves being in the WTC, at the Pentagon or on United 93.

That’s why the fear of a major terrorist attack or any other large catastrophe appears superficially absurd, given each one of us’s infinitesimal probability to be involved. In reality, it’s not “how bad is the risk for me?” but “how bad is the risk for ‘my group’?”

This concept can be expanded further. The first ports of call for one’s feeling of belonging to a group are obviously family, friends and acquaintances/colleagues (in contemporary terms, one’s “active social network”, to the exclusion of the people one is not in touch any longer).

What may or does happen to one’s “active social network” will affect one’s sense of security.

This means in turn that the more people one knows and stay in touch with, the lower the sense of one’s own security. Actually, the cumulative chance of anybody in one’s active social network to be involved in a large catastrophe gets higher and higher as time goes by.

And in a globalized world where people travel around, and the possibility to get to know more people increases, things can only get worse. As something “bad” is bound to happen to somebody sometime, having a big enough active social network will guarantee a hefty supply of tragedies.

If we could be acquainted with every other human being, life would be very very hard to bear.

This is a rather unfortunate, unintended consequence of social networking tools, including basic e-mail.


Anyway, there are two silver linings to that.

First of all, just as tragedies will be in plentiful supply, so will causes of celebrations. We just have to work out a way to make those travel as fast as bad news already do.

Second, the more we know each other, the less we will be able to just harm each other. As all genocidal Dictators know, it is much harder to kill people when they are people, instead of faceless enemies.

Written by omnologos

2007/Oct/15 at 22:49:01

Posted in Risk, Social Networks

Tagged with

2 Responses

Subscribe to comments with RSS.

  1. Actually, if we stick to the traditional definition of risk in terms of volatility as it’s done in finance, enlarging the group by a factor of N should _reduce_ the risk by a factor sqrt(N).
    However, this assumes gaussian distribution of events. In cases of “fat tailed” distributions (e.g., paretian alpha-stable), the variance does not exist (the second moment diverges), and the volatility can be redefined in terms of e.g. interquartile distance; with alpha-stable distributions, this is related to to a parameter called gamma. In those cases, the volatility of the average of N independent variables is equal to the original volatility of each of them multiplied by a factor N^((1/alpha)-1), where alpha ranges between 0 and 2. The case alpha = 2 corresponds to the gaussian case; 1 to Cauchy/Lorenz (volatility left unchanged); and for values of alpha < 1 the volatility actually _grows_ with N. See e.g.:

    Click to access chap1.pdf


    2007/Oct/16 at 01:48:55

  2. […] parlo (in inglese) in un blog “Percezione del Rischio, Networks Sociali e Globalizzazione” dove cito Jeremy Waldron dalla The New York Review of Books (”Is This Torture Necessary?“, […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: