I’m not sure the identify model is always the worst thing to do. There are contexts where it provides a useful heuristic. For example, consider someone who is somewhat rational and skeptical, but not very well-informed. Then someone mentions some alternative medicine with an unclear causal explanation which they know nothing about. They need to make (within their bounded rationality) whether or not this is worth investigating. What would the stereotypical skeptic do will most of the time give the correct answer. Or consider the mathematician who is trying to estimate an apparent double sum, and says “what do mathematicians try to do when given a double sum? Interchange the order of summation.” And surprisingly often, this works. The basic point is accurate, but one needs to be careful not to generalize here too much.
In the case of the mathematician, the identity model gives the same result as consequentialism would. Changing the order of summation has worked before, and there is a cheap experiment to see if it will work again.
Yes, of course. That’s the reason why the identity model exists in the first place. It is quick and easy to make a decision using it, and it often gives the correct answer, or nearly so. Nice and simple heuristics like this one are very useful in our day to day lives. However, to the extent that it differs from the consequences model, it is wrong. As such, we should do our best to minimize our reliance on it.
Or put it this way: The phrase “heuristics and biases” exists, and paints out a natural part of conceptspace. In the same way that we want to reduce biases in our thinking (to the extent that it is possible to do so, of course), we want to reduce the use of heuristics in our thinking. (Again, to the extent that it is possible to do so, of course.)
I’m not sure that’s accurate. Reducing use of heuristics shouldn’t by itself be a good. Understanding when to use heuristics and when not to, or when to use a different heuristic is a distinct goal than simply reducing our use of heuristics completely, especially given that we really do have limited cognitive resources.
I’m not sure the identify model is always the worst thing to do. There are contexts where it provides a useful heuristic. For example, consider someone who is somewhat rational and skeptical, but not very well-informed. Then someone mentions some alternative medicine with an unclear causal explanation which they know nothing about. They need to make (within their bounded rationality) whether or not this is worth investigating. What would the stereotypical skeptic do will most of the time give the correct answer. Or consider the mathematician who is trying to estimate an apparent double sum, and says “what do mathematicians try to do when given a double sum? Interchange the order of summation.” And surprisingly often, this works. The basic point is accurate, but one needs to be careful not to generalize here too much.
In the case of the mathematician, the identity model gives the same result as consequentialism would. Changing the order of summation has worked before, and there is a cheap experiment to see if it will work again.
Yes, but the identity model uses a lot less computational power.
Yes, of course. That’s the reason why the identity model exists in the first place. It is quick and easy to make a decision using it, and it often gives the correct answer, or nearly so. Nice and simple heuristics like this one are very useful in our day to day lives. However, to the extent that it differs from the consequences model, it is wrong. As such, we should do our best to minimize our reliance on it.
Or put it this way: The phrase “heuristics and biases” exists, and paints out a natural part of conceptspace. In the same way that we want to reduce biases in our thinking (to the extent that it is possible to do so, of course), we want to reduce the use of heuristics in our thinking. (Again, to the extent that it is possible to do so, of course.)
I’m not sure that’s accurate. Reducing use of heuristics shouldn’t by itself be a good. Understanding when to use heuristics and when not to, or when to use a different heuristic is a distinct goal than simply reducing our use of heuristics completely, especially given that we really do have limited cognitive resources.