ABSTRACT
There is a general human need for resources that can help with decision-making under uncertainty, and there is a specific human need for resources adapted to scant numerical information. Chapter 4 responds to these needs by recommending a plausibilistic form of decision theory. The strength of this form of decision theory is its adaptability to decisions that must be made without the standard decision-theoretic inputs of sharp probabilities and utilities. The volume does not argue that standard forms of decision theory should not be used; it actually argues the opposite: standard forms of decision theory should be used—when they can be. When they cannot be used, however, it is often possible to rely on a form of decision theory that takes comparative inputs of plausibility and desirability and yields comparative outputs of plausibilistic expectation. This comparative form of decision theory has standard decision theory as a special case. Therefore, when beliefs and desires can be reliably expressed in terms of probabilities and utilities, standard decision theory can still be applied. But when beliefs and desires cannot be quantified yet can be compared in terms of ‘less than’, ‘equal to’, or ‘greater than’, comparative decision theory can be used as a guide to decisions under uncertainty.
