ABSTRACT
Each of the behavior units in the system can be described
in terms of a set of "relevant variables." Just what is relevant
and what is not is a matter of judgment of the system-builder,
but we think of such things as states of war or peace, degrees
of hostility or friendliness, alliance or enmity, arms budgets,
geographic extent, friendly or hostile communications, and so
on. Having defined our variables, we can then proceed to postu-
late certain relationships between them sufficient to define a
path for all the variables through time. Thus we might suppose,
with Lewis Richardson, 1 that the rate of change of hostility
of one nation toward a second depends on the level of hostility
in the second and that the rate of change of hostility of the
second toward the first depends on the level of hostility of
the first. Then, if we start from given levels of hostility in
each nation, these equations are sufficient to spell out what hap-
pens to these levels in succeeding time periods. A system of this
kind may (or may not) have an equilibrium position at which
the variables of one period produce an identical set in the next
period, and the system exhibits no change through time.
