Nice example, Kathy. To emphasize more on this distinction between "rational" and "rationalization", many repeatable experiments have been done on decision making among inmates of psychiatric hospitals. In many cases involving token economies, so called irrational people who give off the wall reasons for their decisions, actually display decisions about trading that are very rational. I believe Gary Becker is one of the authors.
The reason why I do not find this definition "operational" is because the evidence being collected is more appropriate for a study of a theory about how an individual can understand and articulate his inner thought process than about how he actually makes decisions. The appropriate evidence about the latter is the decisions made. A knowledge engineer may need to inquire about how a conclusion was reached by the expert. This may result in a better structural model with several "hidden" nodes. But that inquiry must be carefully guided to avoid all the problems associated with anchoring and the like. And, ultimately, the inquiry is only valuable to the extent it is consistent with decisions actually observed. Bob - -----Original Message----- From: Kathryn Blackmond Laskey [mailto:[EMAIL PROTECTED] Sent: Wednesday, August 20, 2003 9:53 PM To: [EMAIL PROTECTED] Subject: Re: [UAI] Allais' paradox >On the contrary, if rational behaviour is defined, following Aristotle, >as behaviour for which an actor is able to provide reasons when >questioned, then I should think this definition is quite operational. Yes, you are right, that is an operational definition. But have you had the experience of asking someone the reasons for his/her behavior, and receiving a response that was blatantly inconsistent and self-contradictory? For example, when I was a teenager, I worked in a nursing home. One of the patients was convinced that the staff was hiding her mother upstairs. She kept demanding that we let her see her mother. She had what she considered good reasons for her beliefs, and could provide them when questioned. In fact, the staff found questioning her an amusing sport. Her reasons for her opinions and behavior made for endless lunchtime stories. They changed from day to day and were wildly at variance with observable facts, which she blithely denied. This patient would satisfy your definition of rational. Yet most people would consider her irrational. Would you consider her rational because she could provide reasons for her beliefs and behavior? If so, is there any meaningful content to your theory of rationality? Kathy
