Machines and Societies 
to predict and manipulate the behaviour of societies as we can 
that of machines? What new kinds of responsibility have now 
to be recognized? And who, in this case, are “‘we’’? 
SOCIETIES AS COGNITIVE INFORMATION-SYSTEMS 
There can be no question that an accurate information-flow 
model of a society could be a powerful tool in the hands ot 
anyone, in government or elsewhere, who wanted to predict or 
manipulate social processes to his ends. If the predictor 
himself were sufficiently isolated from that society, and could 
gain the information to keep his model up to date without 
significantly disturbing it, then his power might in principle 
seem unlimited. 
Is such a society then as helpless before its predictor as a 
piece of clockwork ? It is not, for a reason which has been hard 
to ignore at various points already. However mechanical may 
be the information-system embodied by a social structure, the 
significant fact is that its units are themselves cognitive in- 
formation systems. ‘This does not necessarily mean that the 
individual human organism, as a mechanism, is inexplicable in 
terms of an information-flow model. On the contrary, the 
logical relationship between physical, personal and even 
religious ways of talking about man requires no such “postulate 
of impotence”’ at the mechanical level. 
What it does mean is that the behavioural characteristics of 
the units of a human society are sensitive to information, 
including information about that society, if it comes their way. 
A notorious current illustration of such sensitivity is the way in 
which the publication of an opinion poll before an election 
can affect people’s voting behaviour, so strongly as to invalidate 
predictions based on it, even though the same poll, kept secret, 
would have given an accurate forecast. The very process of 
sampling public opinion can often change it significantly. The 
Same considerations expose what Popper® has called the 
“fallacy of historicism’’, namely the idea that human history 
is inevitably predictable by extrapolation from its past, so 
that all a wise man need do is to discern its direction and mount 
6* 163 
