There are a variety of schemes for assessing the causes of conflict: remote versus immediate, necessary versus sufficient, significant versus trivial, structural versus situational. Moreover, there is the question as to what one means by cause: an event (an occurrence) that can be dated, a condition, a mathematical relationship? There is finally the question of whether causes actually exist in social behavior, are statistical regularities, or are projections of our minds on that behavior.

My concern here is not to philosophically analyze the meaning of cause, but to cut through the various perspectives and interpretations and simply ask: "given my social field interpretation of conflict and the conditions and forces presumably operating, what do I mean by, for example, x causes or is a condition of war?"¹

A number of aspects of the international field bear on an answer to this.

• Possible causes and conditions are interrelated and relative within a gestalt comprising the process of conflict (the conflict helix).

• We are dealing with states as parties to conflict, where "state behavior" comprises an outcome of a complex process of domestic and foreign decision- making, organizations, and processes (so far this process has been the focus of this book).

• Decision makers are basically teleological: mainly deciding in order to satisfy interests rather than because of specific causes (thus the emphasis on process and balancing).

• The field may be in a state of local equilibrium or disruption, thus influencing the salience of possible causes or conditions.

Moreover, in defining causes and conditions I am concerned with their systematic, empirical evidence. With this in mind, I will demand of a "cause" or "condition" that it:

• have a logical (or statistical) relationship to conflict,²
• have a theoretical explanation as to why it should affect conflict,
• be empirically and scientifically testable.

Now, denote some cause as x, and the conflict behavior it affects as y. Then, logically, a cause in my terms can have one of three meanings. First, the cause can be sufficient for y: if x then y. Whenever x occurs, y follows. But y can also be an effect of other causes, so that if x is sufficient for y, we cannot say that not x implies not y, or that y implies x. If heavy drinking is a sufficient cause of automobile accidents, this does not mean that sober driving implies no accidents. For accidents also are caused by excessive speed, falling asleep, mechanical breakdowns.

Second, the cause x can be necessary for y: if y then x. If y occurred, then x must have also occurred. Or, y would not occur without x being present. Thus, gravity is a necessary cause for one to fall on an icy sidewalk. But gravity is not sufficient. Even though gravity is present you may not slip and fall.

Third, a cause may be necessary and sufficient: y if and only if x. That is, y will occur only when x occurs. For example, the launching of American ICBMs involves a complex set of simultaneous procedures to be carried out by more than one person. These procedures have been designed so that together they are necessary and sufficient for a launching: a launching would not occur without following these procedures, either by accident or will.

There is a special type of sufficient cause that I will call a trigger. This is a class of causes x that operates within a system of conditions highly favorable for producing y, such that any member of the class x could have produced y and would not have been a cause without these conditions. The trigger is often called the proximate or immediate cause. Any kind of spark (class x) may trigger an explosion once sufficient explosive gas has accumulated; any additional car (class x) on a bridge may trigger its collapse, given it is already overloaded. Any overt attack (class x) on a regime, such as an assassination, may trigger a revolution or war, if the political conditions are ripe.

Specific triggers are therefore random causes. They are not of intrinsic interest except as a class of events. What is of interest are the causes and conditions that must cooccur for triggers to be successful and the specific, sufficient causes that make certain that some random trigger will spring y.

Then there are conditions z of y. Conditions, in my sense, retard or enhance the operations of a cause. Setting a match to kindling is a sufficient cause of fire. But, if the (condition of the) wood is wet, the sufficient cause will be hampered, possibly defeated. If the (condition of the) wood is desert dry, the wood could burst into vigorous flame at the touch of a match. World opinion similarly could inhibit or aggravate violence depending on whether it was strongly opposed or favorable.

In theoretically explaining conflict behavior, then, the causal model I will use treats causes and conditions as operating within a field and through a medium of meanings, values, and norms. The model is illustrated by Figure 16A.1.

With this understanding of causes and conditions of Conflict Behavior in mind, then, Table 16A.1 organizes the theoretical causes and conditions of Conflict Behavior. The proposition relating them to conflict and discussion of their meanings are given in Appendix 16B; the evidence is given in Appendix 16C.

In reading Table 16A.1 keep in mind that Conflict Behavior includes violence or war. Therefore, what is a necessary cause for Conflict Behavior would also be so for war. What causes or conditions are specified in addition for violence or war are therefore those that theoretically discriminate them from other Conflict Behavior. Thus, given that the necessary causes have occurred, then nonviolent conflict escalates to military action and war if neither party is libertarian, the status quo has been disrupted, and each is confident of success. This escalation will be made, more or less, intense depending on the presence of the tabulated aggravating and inhibiting conditions. Especially, escalation to war will be probable if the parties have power parity and it is a widescale class conflict.