by what to be guided
The freaky thing about Mr Spock (and characters like him) is not that they come to their conclusions through logical inference, but that they do it so fast. As anyone who has actually worked on formal inference will happily tell you, going brute-force about any problem of a decent size--such as the kind they solve on Star Treck--usually takes too long to be helpful. The problem is bad enough that after two episodes, Captain Kirk would have simply given up on asking his Science Officer (ever again).
People deal with this problem in a variety of ways, as the study of experts has shown. People have good intuitions for picking models appropriate to a task (and experts have a larger bag of these than most other folk do). People have models worked out to the level of standard solutions, so that they can quickly slap together something that gets the job done. People let themselves be guided by heuristics such as "try the thing that most often worked in the past", or "try the thing that worked most recently". Through-out the thinking process, people monitor how they feel about the way the solution is developing, adjusting course if they become dissatisfied or feel they are not making progress.
Formally speaking, these pragmatic approaches are all guess-work. You cannot short-circuit the inference in this fashion, and solutions will only be correct up to the level of reasoning that has been done, and could be retracted in the next time slice due to running into a counter example or a contradictory fact. In order to really know whether an answer is correct or not, one would have to work the inference to exhaustion.
Enter then this interesting bit of research, which looks at the testerone and cortisol levels of London day traders. There are lots of thought-provoking stances in the piece, with some fun anecdotal evidence thrown in for good measure.
What irks me about the article though is the false dichotomy between rational and hormone level driven behavior. From my view of the problem, the hormone levels would be a reasoning heuristic that is always present--which presumably means that there are contexts where they are appropriately contributing to an evolutionarily helpful answer. (I understand and have worried about the difference between "logically correct" and "evolutionarily helpful", but for sake of argument will assume that in the majority of the cases, these two overlap considerably.) So, it is not that the reasoning is sometimes emotional and sometimes not; rather, that the emotive heuristic fails in particular circumstances to drive the inference toward success. But that follows, almost by definition, from the fact that the heuristics are an approximation of the overly expensive true computation.
So the authors of the study are seriously jumping the gun when starting with isolating the raised levels of specific hormones as the primary contributors to the failure of rationality. One has to also keep in mind that the domain of trading is one where the time span for decision-making is extremely short (i.e. short time to "first answer"). Such a situation requires one to increase the use of heuristics, just to cut to the chase as quickly as possible. Who knows, the hormone levels may be mere indicators of the increase in heuristic activity.
In order for the research to be convincing then, the researcher themselves would have to know what the rational answer was, free of all heuristics--asking experts merely swaps heuristics, effectively. I know that requesting they satisfy an impossible precondition seems unhelpful, but that observation does not alleviate my concern.
Finally, the analysis as presented seems to be insufficient, as the various stages of the reasoning process--model selection, model population, model combination, model use monitoring--are not adequately differentiated yet all employ heuristic techniques that could be expected to make use of hormone levels (whether successfully or not). This seems especially important in terms of the remedies that the researchers are interested in.
People deal with this problem in a variety of ways, as the study of experts has shown. People have good intuitions for picking models appropriate to a task (and experts have a larger bag of these than most other folk do). People have models worked out to the level of standard solutions, so that they can quickly slap together something that gets the job done. People let themselves be guided by heuristics such as "try the thing that most often worked in the past", or "try the thing that worked most recently". Through-out the thinking process, people monitor how they feel about the way the solution is developing, adjusting course if they become dissatisfied or feel they are not making progress.
Formally speaking, these pragmatic approaches are all guess-work. You cannot short-circuit the inference in this fashion, and solutions will only be correct up to the level of reasoning that has been done, and could be retracted in the next time slice due to running into a counter example or a contradictory fact. In order to really know whether an answer is correct or not, one would have to work the inference to exhaustion.
Enter then this interesting bit of research, which looks at the testerone and cortisol levels of London day traders. There are lots of thought-provoking stances in the piece, with some fun anecdotal evidence thrown in for good measure.
What irks me about the article though is the false dichotomy between rational and hormone level driven behavior. From my view of the problem, the hormone levels would be a reasoning heuristic that is always present--which presumably means that there are contexts where they are appropriately contributing to an evolutionarily helpful answer. (I understand and have worried about the difference between "logically correct" and "evolutionarily helpful", but for sake of argument will assume that in the majority of the cases, these two overlap considerably.) So, it is not that the reasoning is sometimes emotional and sometimes not; rather, that the emotive heuristic fails in particular circumstances to drive the inference toward success. But that follows, almost by definition, from the fact that the heuristics are an approximation of the overly expensive true computation.
So the authors of the study are seriously jumping the gun when starting with isolating the raised levels of specific hormones as the primary contributors to the failure of rationality. One has to also keep in mind that the domain of trading is one where the time span for decision-making is extremely short (i.e. short time to "first answer"). Such a situation requires one to increase the use of heuristics, just to cut to the chase as quickly as possible. Who knows, the hormone levels may be mere indicators of the increase in heuristic activity.
In order for the research to be convincing then, the researcher themselves would have to know what the rational answer was, free of all heuristics--asking experts merely swaps heuristics, effectively. I know that requesting they satisfy an impossible precondition seems unhelpful, but that observation does not alleviate my concern.
Finally, the analysis as presented seems to be insufficient, as the various stages of the reasoning process--model selection, model population, model combination, model use monitoring--are not adequately differentiated yet all employ heuristic techniques that could be expected to make use of hormone levels (whether successfully or not). This seems especially important in terms of the remedies that the researchers are interested in.