4.3. Adopting more reliable, more expensive reasoning strategies
К оглавлению1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1617 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84
85 86 87 88 89 90 91 92 93 94
One might adopt a reasoning strategy that brings more benefits but is
more costly than the old strategy. Such a change will sometimes (but not
always) lead to better reasoning (top right quadrant of Figure 3.5). Suppose
Test Taker is currently expending c resources on reasoning strategy
D; but he has the time and energy to employ c1 resources on these
problems (Figure 3.6). At this expenditure of resources, E is the most
reliable reasoning strategy available. Now should Test Taker quit D (at
cost c) in favor of E (at cost c1)? The answer is, of course, it all depends. If
he were to switch to E, he would increase his reliability on these particular
reasoning problems. But whether this change leads to better overall reasoning
all depends on whether the gain in reliability in this portion of the
test more than offsets the loss of reliability that results from spending
fewer resources on the other portion of the test. Although this may seem
odd to say, the most locally reliable reasoning strategy is not always the
best overall reasoning strategy. That’s because given resource limitations,
the optimization of global reliability often requires that local reliability not
be optimized.
All reasoning strategies have opportunity costs (i.e., what is forgone
by not devoting resources to the best available alternative). The devotion
of cognitive resources to one problem typically prevents or hinders us from spending time and energy on something else. Our point here is akin
to one made years ago by Simon (1982) about satisficing and about
bounded rationality in general: sometimes it is better to adopt reasoning
and decision procedures that are good and cheap rather than great and
expensive. We take Simon’s point to be that from the perspective of
prudential rationality, one ought not always use the ideal (in the sense of
‘‘maximally accurate’’) reasoning strategy. If given the choice between
reasoning in an ideal fashion about X or using the same energy to reason
very well (but less than ideally) about X and take your kid fishing, most
problems we face aren’t so significant that it would be worth it to miss out
on the fishing trip. But there is a stronger point to make. When a reasoner
has a choice between two tractable reasoning strategies (i.e., reasoning
strategies she can actually employ), sometimes the reasoner ought to adopt
the cheaper and less reliable strategy—even from a purely epistemic perspective.
This will occur when the opportunity cost comes in the form of a
forsaken epistemic benefit. In these cases, part of the cost of devoting those
resources to one problem is not having made any headway on some other
problem.
There are many examples of highly reliable reasoning strategies that
come with high costs. For example, any epistemological theory that
recommends Bayes’ Rule for updating belief is recommending a reasoning
strategy that is more reliable but also more expensive than ones a reasoner
is likely using. Another high cost prescription is that reasoners’ knowledge
should be closed under entailment (Cherniak 1986). Many have made the
point that strategies that are in practice impossible to implement cannot
enhance epistemic excellence. But given our discussion of opportunity
costs, there’s a brash lesson to draw: A reasoning strategy that is more
costly and reliable, but not so costly that it can’t be used, still does not
necessarily enhance epistemic excellence. By expending more resources on
a new and improved reasoning strategy, one inevitably takes resources that
could be used elsewhere. And if those extra resources could be better used
elsewhere, then one could be a better reasoner if one retained the less
reliable reasoning strategy and used the extra resources more effectively.
Ameliorative Psychology recommends a number of reasoning strategies
that would likely be more expensive to implement and execute than
the reasoning strategies most people currently employ. For example, deciding
to use frequency formats or a consider-the-opposite strategy (see
chapter 9) comes with nonnegligible start-up costs. And even if we ignore
start-up costs, they are likely to be more expensive to execute than most
reasoners’ default strategies. As a result, it is not inevitable that such
68 Epistemology and the Psychology of Human Judgment
reasoning strategies will make someone a better reasoner—even though
such strategies are more reliable than reasoners’ actual reasoning strategies
and can in practice be implemented and executed.
One might adopt a reasoning strategy that brings more benefits but is
more costly than the old strategy. Such a change will sometimes (but not
always) lead to better reasoning (top right quadrant of Figure 3.5). Suppose
Test Taker is currently expending c resources on reasoning strategy
D; but he has the time and energy to employ c1 resources on these
problems (Figure 3.6). At this expenditure of resources, E is the most
reliable reasoning strategy available. Now should Test Taker quit D (at
cost c) in favor of E (at cost c1)? The answer is, of course, it all depends. If
he were to switch to E, he would increase his reliability on these particular
reasoning problems. But whether this change leads to better overall reasoning
all depends on whether the gain in reliability in this portion of the
test more than offsets the loss of reliability that results from spending
fewer resources on the other portion of the test. Although this may seem
odd to say, the most locally reliable reasoning strategy is not always the
best overall reasoning strategy. That’s because given resource limitations,
the optimization of global reliability often requires that local reliability not
be optimized.
All reasoning strategies have opportunity costs (i.e., what is forgone
by not devoting resources to the best available alternative). The devotion
of cognitive resources to one problem typically prevents or hinders us from spending time and energy on something else. Our point here is akin
to one made years ago by Simon (1982) about satisficing and about
bounded rationality in general: sometimes it is better to adopt reasoning
and decision procedures that are good and cheap rather than great and
expensive. We take Simon’s point to be that from the perspective of
prudential rationality, one ought not always use the ideal (in the sense of
‘‘maximally accurate’’) reasoning strategy. If given the choice between
reasoning in an ideal fashion about X or using the same energy to reason
very well (but less than ideally) about X and take your kid fishing, most
problems we face aren’t so significant that it would be worth it to miss out
on the fishing trip. But there is a stronger point to make. When a reasoner
has a choice between two tractable reasoning strategies (i.e., reasoning
strategies she can actually employ), sometimes the reasoner ought to adopt
the cheaper and less reliable strategy—even from a purely epistemic perspective.
This will occur when the opportunity cost comes in the form of a
forsaken epistemic benefit. In these cases, part of the cost of devoting those
resources to one problem is not having made any headway on some other
problem.
There are many examples of highly reliable reasoning strategies that
come with high costs. For example, any epistemological theory that
recommends Bayes’ Rule for updating belief is recommending a reasoning
strategy that is more reliable but also more expensive than ones a reasoner
is likely using. Another high cost prescription is that reasoners’ knowledge
should be closed under entailment (Cherniak 1986). Many have made the
point that strategies that are in practice impossible to implement cannot
enhance epistemic excellence. But given our discussion of opportunity
costs, there’s a brash lesson to draw: A reasoning strategy that is more
costly and reliable, but not so costly that it can’t be used, still does not
necessarily enhance epistemic excellence. By expending more resources on
a new and improved reasoning strategy, one inevitably takes resources that
could be used elsewhere. And if those extra resources could be better used
elsewhere, then one could be a better reasoner if one retained the less
reliable reasoning strategy and used the extra resources more effectively.
Ameliorative Psychology recommends a number of reasoning strategies
that would likely be more expensive to implement and execute than
the reasoning strategies most people currently employ. For example, deciding
to use frequency formats or a consider-the-opposite strategy (see
chapter 9) comes with nonnegligible start-up costs. And even if we ignore
start-up costs, they are likely to be more expensive to execute than most
reasoners’ default strategies. As a result, it is not inevitable that such
68 Epistemology and the Psychology of Human Judgment
reasoning strategies will make someone a better reasoner—even though
such strategies are more reliable than reasoners’ actual reasoning strategies
and can in practice be implemented and executed.