4.1. Resource reallocation

The first way to improve one’s reasoning is not depicted on Figure 3.5. It is

possible for one to become a better reasoner without adopting any new

first-order reasoning strategies, without changing our strategies for reasoning

about the world (i.e., our first-order strategies). We can, instead,

change our (second-order) strategies for allocating resources to our firstorder

strategies. For example, suppose Test Taker devotes a lot of time

to the quantitative section of the aptitude test, thereby leaving insufficient

time to do well on the analytical section. If Test Taker’s cognitive

resources are allocated in such a way that the MERs of all his reasoning

strategies are not all equal, then he could improve his reasoning (i.e.,

generate a greater number of right answers) by transferring resources from

reasoning strategies with lower MERs to reasoning strategies with higher

MERs. He might reason better (i.e., get more correct answers) by simply

spending less time on the quantitative section and more time on the analytical section of the test. (For a similar discussion, see Goldman 1999, 87–94.)

The Test Taker example does not do justice to the potential practical

importance of the insight that one can become a better reasoner simply by

reallocating resources. That’s because our toy Test Taker model presupposes

that the only way to improve our reasoning is by reallocating

resources so as to get more truths. But excellent reasoners reason reliably

about significant matters. So we can reason better by reallocating resources

in such a way that we end up with more significant truths—even if the

total number of truths we end up with is not much changed.

Some people might object that they don’t need to be reminded to

spend resources on problems that matter. Still, the advice we offer here is

not idle. Whether a reasoning problem is significant is an empirical

question, and there is now considerable empirical evidence that identifies

some of the factors that most affect people’s happiness. The burgeoning

research area of ‘‘affective forecasting’’ has amply demonstrated the radically

mistaken character of people’s predictions concerning what will

make them (and others) happy. For example, once a person is a decile or

so above the poverty level, money contributes little to happiness (Diener

and Oishi 2000). Yet, people who pursue money in the belief that it will

increase their happiness express greater frustration than their peers (Myers

2000). Money is only the tip of the illusion. People believe that winning

the lottery, getting tenure, and moving to a sunnier climate will make

them happier. They believe they (and others) will be made substantially

less happy in the long run by a paraplegia-inducing spinal cord injury and

getting denied tenure, and in the short run by having a painful colonoscopy

of longer rather than shorter duration. The scientific evidence shows

that they are wrong on all counts. Our attachments to these personal forecasts

can be extremely costly and personally damaging (Kahneman 2000).

These results are counterintuitive, but fortunately, science (unlike some

branches of philosophy) isn’t about respecting our intuitions.

The very happiest of people—those ranking in the top 10% of global

satisfaction with their lives—have a number of things in common, but

none of these is especially specific, occult, or exotic. They do not exercise

more than others, experience more ‘‘good’’ events, nor are they wealthier

or more religious. Their common ‘‘secret’’ is disarmingly simple: They tend

to be more social, with stronger social and romantic relationships than the

less happy groups (Diener and Seligman 2002). When monitored over electronic

pagers, people report the most enjoyment not when eating chocolate

Extracting Epistemic Lessons from Ameliorative Psychology 65

mousse as they recline in a chaise lounge on a Caribbean beach (or any of

its slothful, self-indulgent correlates), but when they are unselfconsciously

absorbed in a mindful challenge (Csikszentmihalyi 1999).

Given its attention to significance, our epistemological theory has the

resources to guarantee that our normative recommendations will be informed

by the lessons of hedonic psychology. Good reasoners adopt

reasoning strategies that operate reliably on significant problems, including

problems that are important to our well-being. We have already

discussed how all manner of false belief and personal conceit will intervene

to tempt defection from excellent reasoning strategies. But our best psychological

theories can predict these siren songs, and our theory can

recommend a resolute course of action in reply. As a result, our theory will

be able to recommend that people allocate cognitive resources to reasoning

strategies that tackle problems that are most likely to promote their

well-being and away from reasoning strategies that tackle problems that

are likely to undermine their well-being. Our theory has the wherewithal

to make such recommendations even if the prescribed change of focus

does not lead to a greater number of truths.

The first way to improve one’s reasoning is not depicted on Figure 3.5. It is

possible for one to become a better reasoner without adopting any new

first-order reasoning strategies, without changing our strategies for reasoning

about the world (i.e., our first-order strategies). We can, instead,

change our (second-order) strategies for allocating resources to our firstorder

strategies. For example, suppose Test Taker devotes a lot of time

to the quantitative section of the aptitude test, thereby leaving insufficient

time to do well on the analytical section. If Test Taker’s cognitive

resources are allocated in such a way that the MERs of all his reasoning

strategies are not all equal, then he could improve his reasoning (i.e.,

generate a greater number of right answers) by transferring resources from

reasoning strategies with lower MERs to reasoning strategies with higher

MERs. He might reason better (i.e., get more correct answers) by simply

spending less time on the quantitative section and more time on the analytical section of the test. (For a similar discussion, see Goldman 1999, 87–94.)

The Test Taker example does not do justice to the potential practical

importance of the insight that one can become a better reasoner simply by

reallocating resources. That’s because our toy Test Taker model presupposes

that the only way to improve our reasoning is by reallocating

resources so as to get more truths. But excellent reasoners reason reliably

about significant matters. So we can reason better by reallocating resources

in such a way that we end up with more significant truths—even if the

total number of truths we end up with is not much changed.

Some people might object that they don’t need to be reminded to

spend resources on problems that matter. Still, the advice we offer here is

not idle. Whether a reasoning problem is significant is an empirical

question, and there is now considerable empirical evidence that identifies

some of the factors that most affect people’s happiness. The burgeoning

research area of ‘‘affective forecasting’’ has amply demonstrated the radically

mistaken character of people’s predictions concerning what will

make them (and others) happy. For example, once a person is a decile or

so above the poverty level, money contributes little to happiness (Diener

and Oishi 2000). Yet, people who pursue money in the belief that it will

increase their happiness express greater frustration than their peers (Myers

2000). Money is only the tip of the illusion. People believe that winning

the lottery, getting tenure, and moving to a sunnier climate will make

them happier. They believe they (and others) will be made substantially

less happy in the long run by a paraplegia-inducing spinal cord injury and

getting denied tenure, and in the short run by having a painful colonoscopy

of longer rather than shorter duration. The scientific evidence shows

that they are wrong on all counts. Our attachments to these personal forecasts

can be extremely costly and personally damaging (Kahneman 2000).

These results are counterintuitive, but fortunately, science (unlike some

branches of philosophy) isn’t about respecting our intuitions.

The very happiest of people—those ranking in the top 10% of global

satisfaction with their lives—have a number of things in common, but

none of these is especially specific, occult, or exotic. They do not exercise

more than others, experience more ‘‘good’’ events, nor are they wealthier

or more religious. Their common ‘‘secret’’ is disarmingly simple: They tend

to be more social, with stronger social and romantic relationships than the

less happy groups (Diener and Seligman 2002). When monitored over electronic

pagers, people report the most enjoyment not when eating chocolate

Extracting Epistemic Lessons from Ameliorative Psychology 65

mousse as they recline in a chaise lounge on a Caribbean beach (or any of

its slothful, self-indulgent correlates), but when they are unselfconsciously

absorbed in a mindful challenge (Csikszentmihalyi 1999).

Given its attention to significance, our epistemological theory has the

resources to guarantee that our normative recommendations will be informed

by the lessons of hedonic psychology. Good reasoners adopt

reasoning strategies that operate reliably on significant problems, including

problems that are important to our well-being. We have already

discussed how all manner of false belief and personal conceit will intervene

to tempt defection from excellent reasoning strategies. But our best psychological

theories can predict these siren songs, and our theory can

recommend a resolute course of action in reply. As a result, our theory will

be able to recommend that people allocate cognitive resources to reasoning

strategies that tackle problems that are most likely to promote their

well-being and away from reasoning strategies that tackle problems that

are likely to undermine their well-being. Our theory has the wherewithal

to make such recommendations even if the prescribed change of focus

does not lead to a greater number of truths.