System dynamics and complex-adaptive systems are two distinct ways of analyzing complicated phenomena. These methods are normally employed by different researchers and considered alternative ways of representing processes. In reality, however, they merge: there is no reason to expect that in a real situation, an interesting or frightening tipping point in behavior will occur as the result only of 1) the interaction of mathematically expressible smooth curves (i.e., the growth rates, delays, feedback loops of system dynamics) or 2) unanticipated emergent behavior at one level (e.g., group) generated by behavior at another level (e.g., individual) in a system of semi-independent, co-evolving substructures displaying self-organization. Real tipping points, the stuff of history, are created by a very confusing combination of causal dynamics displaying features artificially separated by the methodological disciplines of system dynamics and complex-adaptive systems. Not only are the two hardly ever applied to world affairs, they are hardly ever even considered together in a theoretical context.
A simple tipping point results from the intersection of a presumably obvious dominant dynamic and a quite possibly unknown subordinate dynamic trying to push behavior in another direction, e.g., a steady off-shore breeze counteracted by the incoming tide or the growth rate of bacteria counteracted by medicine. Even a simple tipping point will take an observer by surprise if he is unaware of the point where dominance shifts, e.g., the tide becomes stronger than the wind or the death rate becomes faster than the growth rate. Conventional battlefield success in a civil war may be predictable for some period of time as a function of the balance of armaments and the flow of foreign aid only for prediction suddenly to fail as the result of the introduction of a new tactic. The new strategy represents learning, which does not necessarily occur linearly. Factional rivalry within a self-organized insurgency, i.e., an insurgency—such as the Syrian insurgency against Assad—lacking centralized control, may generate sudden leaps of strategic sophistication resulting from the ascendancy of a minor faction with a better plan, and a tipping point may result in the battlefield collapse of the adversary. In this example, one may be able to model mathematically the level of armaments and flow of foreign aid but will at best have only an abstract sense of the likely impact of the shift in strategy. If the shift in strategy provokes behavioral change by bystanders (e.g., initial battlefield successes of ISIS encouraging marginalized Iraqi Sunnis or frustrated global jihadis to join the battle), then the rate at which the tipping point is reached, i.e., the rate at which a new causal dynamic achieves dominance, only accelerates.
The analytical risk lies first in the great temptation to focus on such easily understood and easily graphed factors as weapons, funds, and manpower while ignoring potentially more important factors because their erratic impact is difficult to project forward in time mentally or to display visually. A second area of risk is the difficulty of foreseeing how the impact of other causal dynamics may change in response to the initial shift in dominance, e.g., by provoking a bandwagon effect (exemplified by secular Sunni Iraqis joining ISIS in July 2014) or, in contrast, by provoking a defensive alliance (exemplified by tacit U.S.-Iranian battlefield cooperation against ISIS in September 2014).
Understanding the causal dynamics is the key to dealing effectively with surprise. Unfortunately, the most important of these dynamics, and certainly the most difficult to perceive, are the ones that do not follow smooth paths. Exponential growth and oscillations caused by delays already challenge the cognitive capacity of humans, but the dynamics induced by complex-adaptive systems are far more difficult to understand. The combination explains much about challenge of designing effective foreign policy for no effective approach to assessing the impact of real-world political processes driven by this full mix of causal dynamics has yet been developed.