Call Us: US - +1 845 478 5244 | UK - +44 20 7193 7850 | AUS - +61 2 8005 4826


A generic problem is the difficulty of predicting the effects of efforts to change human behavior in the interest of environmental preservation, either directly or through the modification of technology. Actual effects often turn out to be different from what was desired. This point is illustrated by the consequences of some efforts to modify traffic patterns. Bypasses and beltways have been built to take traffic away from congested streets by providing alternative routes so that drivers do not have to go through the center of the city. Sometimes, however, these have increased traffic not just in the existing areas of congestion but in new areas as well. The provision of new roads has encouraged more people to drive into the city, and the appearance of large, empty roads has stimulated the development of housing to make use of them.

Highway safety provides another example of how an effort to reduce an undesirable effect of human behavior can itself have unanticipated consequences for human behavior. When antilock brakes were put into cars, the assumption that driving safety would be improved was sufficiently credible that insurance premiums were reduced for automobiles with this feature. Empirical evidence gathered by a taxi company in Munich indicated no significant decrease in accident frequency but a significant increase in driving speed (there is no speed limit on the autobahn) and in speed changes . Apparently, people drove faster because they believed their braking systems were safer and that as drivers they would be able to cope more effectively with emergencies. As a result of this study, insurance implications were reconsidered.

The general problem was discussed in terms of cybernetic theory almost 40 years ago . The problem involves whether to treat self-organizing systems as deterministic or stochastic. It is very difficult to predict how a complex system will evolve if it has many self-organizing characteristics; this complicates the task of the planning of interventions involving such systems. “Self-organizing” here does not imply a conscious or purposive response by the system to the inputs of the planners; it simply connotes a system for which the list of states that can be entered changes suddenly and “spontaneously” from time to time or a system in which the transition probabilities from one state to another are not constant. In general, very large, complex, and tightly coupled systems with many subsystems

show such self-organizing properties in response to disturbances, and such results may not be predictable, even stochastically