Aversive responses, such as pain, fever, vomiting and panic, were shaped by natural selection because they gave selective advantages in the face of various dangers. Optimal decisions about when to use our growing pharmacological powers to block these responses will require signal-detection models of how defenses are regulated.
Nesse's First Law
An optimal mechanism to regulate an all-or-none defensive response such as vomiting or panic will express the response whenever CD< ∑(pH x CH w/o defense) –∑(pH x CH w/defense). That is, expressing a defense is worth it whenever the cost of the defense (CD) is less than the estimated reduction in harm, based the probability (pH) and cost of various harmful outcomes (CH) with and without the expression of the defense. This means that optimal systems that regulate inexpensive defenses against large somewhat unpredictable potential harms will express many false alarms and that blocking these unnecessary responses can (and does) greatly relieve human suffering. Blocking responses yields a net benefit, however, only if we can anticipate when a normal response is likely to be essential to prevent catastrophe.
Nesse's Second Law
An optimal mechanism to regulate a continuously expressed defense, such as fever or pain, will increase the defensive response up to the point where the sum of CH and CD is minimized. At this point the marginal increase in the cost of the defense becomes greater than the marginal decrease in harm. This helps to explain why so many defenses, such as those involved in inflammation and the immune responses, so often seem excessive.
Many will recognize this analysis as a less grand and somewhat more practical variation on Pascal’s Wager. So far, however, few in the pharmaceutical industry seem to recognize the importance of routinely assessing the effects of new drugs on normal defensive responses.