Machines and Societies 
abstract an information-flow model of the complex nexus 
by which prices, production, employment, investment and the 
like are interconnected in a particular society such as our own, 
we may expect benefits both to our understanding of present 
ills and to our planning of future improvements?. 
It is important to stress, however, that the safest function of 
such models is to suggest questions worth pursuing, rather than 
answers. Any manageable model is bound to over-simplify; 
so the process of using it should be envisaged as one of progres- 
sive refinement of the model, through the occasional dis- 
covery of unexpected answers to questions based on its present 
form. 
Thus to an information-system theorist confronted with the 
instability of the old boom-slump cycle, our stock recommenda- 
tions (a) to (d) (page 156) automatically suggest questions like 
the following: (a) Is any coupling or regulation tighter than it 
need be? (4) Can any stages in the chain of regulation be 
by-passed ? (c) (i) Are any of the major elements known to be 
hypersensitive to change, or excessively sluggish? That is to 
say, is this instability likely to be due to too much or too little 
ability to respond rapidly to rapid changes? (It could, of course, 
have several components of either kind). (i) If sluggishness is 
present, can we suitably introduce some sensitivity to the rate 
of change (or the acceleration) of the quantities which are fluc- 
tuating? (The precise amount and form, of course, would need 
to be estimated or adjusted in detail). Conversely, if “antici- 
patory”’ responses are causing instability, can we introduce 
some inertial factor (not necessarily at the same point) to 
compensate for it? (d) Ifa further reduction of overall sensitivity 
would at present lead to unacceptably poor regulation of the 
economy against the external factors that can disturb it, can 
we make it tolerable by introducing any sensitive devices to 
“feed forward”? rough adjustments to our regulators, so as to 
compensate automatically for changes in these external 
factors ? 
Now it must be said that some—though not, I think, all—of 
these questions have been raised already by economists, without 
157 
