Let me propose a dichotomy between two kinds of possible definitions of rationality
(A restatment of Cameron Taylor’s idea):
1) A positive quality of a step in the decision making process.
2) The efficiency of a decision making blackbox. Its expected utility.
I think that accurate definitions of the first type are difficult to give and in scientific practice difficult to use, for two reasons:
-Because they presupose a model of the thinking process of the agent.
-Because they (usually?) presupose a way to judge the quality of intermediate states of the decision making process of the agent.
Evolutionnary biologists, for example, use the concept of fitness which clearly has a definition of the second kind.
While trying to define rationality, it is very tempting to describe how to be rational. Since being rational is so difficult, resisting this temptation could be beneficial. Is there a definition of the first kind which doesn’t say too much about how to be rational?
Let me propose a dichotomy between two kinds of possible definitions of rationality (A restatment of Cameron Taylor’s idea):
1) A positive quality of a step in the decision making process.
2) The efficiency of a decision making blackbox. Its expected utility.
I think that accurate definitions of the first type are difficult to give and in scientific practice difficult to use, for two reasons:
-Because they presupose a model of the thinking process of the agent.
-Because they (usually?) presupose a way to judge the quality of intermediate states of the decision making process of the agent.
Evolutionnary biologists, for example, use the concept of fitness which clearly has a definition of the second kind.
While trying to define rationality, it is very tempting to describe how to be rational. Since being rational is so difficult, resisting this temptation could be beneficial. Is there a definition of the first kind which doesn’t say too much about how to be rational?