APPLYING THESE METHODS TO FIND CAUSES

In stating the SCT and NCT and applying these tests to abstract patterns of conditions to eliminate candidates, our procedure was fairly mechanical. We cannot be so mechanical when we try to reach positive conclusions that certain conditions are necessary, sufficient, or both. Applying these rules to actual con- crete situations introduces a number of further complications, especially when using our tests to determine causes.

NORMALITY

First, it is important to keep in mind that, in our ordinary understanding of causal conditions, we usually take it for granted that the setting is normal. It is part of common knowledge that if you strike a match, then it will light.

Thus, we consider striking a match sufficient to make it light. But if someone has filled the room with carbon dioxide, then the match will not light, no matter how it is struck. Here one may be inclined to say that, after all, strik- ing a match is not sufficient to light it. We might try to be more careful and say that if a match is struck and the room is not filled with carbon dioxide, then it will light. But this new conditional overlooks other possibilities—for example, that the room has been filled with nitrogen, that the match has been fireproofed, that the wrong end of the match was struck, that the match has already been lit, and so forth. It now seems that the antecedent of our condi- tional will have to be endlessly long in order to specify a true or genuine suf- ficient condition. In fact, however, we usually feel quite happy with saying that if you strike a match, then it will light. We simply do not worry about the possibility that the room has been filled with carbon dioxide, the match has been fireproofed, and so on. Normally we think that things are normal, and give up this assumption only when some good reason appears for doing so.