G. Gigerenzer claims that we often make use of 'fast and frugal heuristics' in coming to beliefs. Such fast and frugal heuristics in the case of the question, "Is this food safe?", might be,
1. Do I recognize this kind of food? If not, don't consider it safe.
2. Do I recognize it as being rotten or spoiled or contaminated? If so, don't consider it safe.
3. Have I had this kind of food before? If so, did I have an allergic reaction to it? If so, don't consider it safe.
4. Have I had this kind of food before? If so, is it from a clean establishment such as a supermarket? If so, consider it safe.
5. If I've not had this kind of food before, is it a food that I think is typical for people to be allergic to? If so, don't consider it safe.
6. If I've gotten this far without disqualification, are people telling me that it's safe? If so, consider it safe.
7. If not, don't consider it safe.
How in the world did I consider coming up with these heuristics instead of something totally different or opposite? Why isn't a heuristic, "Is the food the color red? If so, consider it safe," a heuristic cue that I would use?
Belief that a heuristic is adapted to the environment is rational when that heuristic helps the believer achieve beliefs which help them do it is that they want to do in the world. A person chooses the set of heuristics that optimizes for consistency and for usefulness towards achieving beliefs that correspond to increased success.
A belief that "the 7-step method [above] for considering food safe above is adapted to the environment" is rational if and only if it is formed from application of a set of heuristics such that belief that "the 7-step heuristic for considering food safe above is adapted to the environment" is positively correlated with behavior based on it that results in the satisfaction of the actual goals of the believer.
Here the method behind adopting this heuristic could be,
1. Has following this set of heuristics (e.g. the 7-step heuristic) always resulted in beliefs where behavior based on them resulted in satisfaction of my actual goals? If so, continue using them.
2. If behavior based on the beliefs formed from this set of heuristics sometimes results in failure, have I also met with a great number of successes such that there is still a positive correlation? If so, continue using them.
3. If the number of failures is considerable, try to come up with a better set of heuristics.
4. If the number of applications made is low, consider another set of heuristics if one presents itself to the mind and try and follow it if it seems promising.
The logical mind is now saying, what about this list of four? When is a belief that this set of four heuristics is adapted to the environment going to be rational? Here the method could be,
1. Has following this set of heuristics (e.g. this 4-step heuristic itself) always resulted in beliefs where behavior based on them resulted in satisfaction of my actual goals? If so, continue using them.
2. If behavior based on the beliefs formed from this set of heuristics sometimes results in failure, have I also met with a great number of successes such that there is still a positive correlation? If so, continue using them.
3. If the number of failures is considerable, try to come up with a better set of heuristics.
4. If the number of applications made is low, consider another set of heuristics if one presents itself to the mind and try and follow it if it seems promising.
This is not an exercise in
a priori reasoning or pontification about the self-evident. If the subject has applied this heuristic for coming to the belief that a heuristic is rational, and if the resulting beliefs (that a heuristic is rational) are positively correlated with behavior based on them that is successful in that environment, then this heuristic is rational. If the heuristic is applied to belief that the heuristic itself is rational, and if the heuristic itself is rational by its own guidelines, then the search can stop, and there is no infinite regress. Application again (and again and again and again) in increasing levels of abstraction would give the same result as the first case of application to itself. There is no question here of, 'but is the belief that the four-step heuristic reaches correct results
really true?' Such belief is rational and justified. And the belief that such belief is rational is rational, and so on ad infinitum.
Let me proceed to another example.
Suppose two people were playing Tic-Tac-Toe, and it was X's turn, as follows:
X * O
O X *
X O *
The belief under consideration is, 'is this move a good move?' One of the first heuristics could be, 'does this move connect three in a row?' Which reflects that the generalization that good moves include those that connect three in a row positively correlates with behavior that results in the satisfaction of the actual goals, which in this case is (presumably) to win the game.
Why is it that we are able to pick out food that is edible or to play a game like Tic-Tac-Toe? It is because we have evolved the ability to form intelligent beliefs about the world in order better to manipulate and survive in it. For this reason, nature provides the bootstraps of rationality in terms of methods that we cannot ignore and keep our sanity, such as that what you see is generally there, and that doing the same thing twice generally gets the same result. From these primitive methods spring up methods of increasing complexity, perhaps even increasing reliability, and very often increasingly esoteric application; they are accepted based on their consistency with methods already assimilated, which go back to the primitive methods supplied by human evolution.
The reason that this is accepted as the correct account of knowledge is that it corresponds to what we observe of human behavior with regard to knowledge. When determining whether something happens to be, the first question off the lips is, "How do we know?" This question skips past even "Is this part of what we know?" because the quality of the answer to the "How" question is presumed to answer the particular question of the state of this piece of information. This fact bodes ill for those who understand knowledge in terms of basic beliefs and bolsters those who see it in terms of methods applied.
And the way that the methods to be applied are themselves selected, that is the method of methods, is by optimizing for consistency with existing methods, and by seeding the procedure with a set of methods obtained in order to achieve rudimentary function in the world (a set of methods that help you find success as an organism in the beliefs formed). I use the word "seeding" because from that seed of mostly similar basic methodology spring several varied and sometimes incompatible more complicated methods, often because they apply to subjects that are far removed from the original domain to which knowledge applies, which is day-to-day human living.
This subject of epistemology is itself fairly abstract and removed from daily experience, which explains why it eludes a universally accepted solution. With slight irony, this can serve as weak confirmation for the theory that knowledge proceeds from methods learned for success in the realm of daily human experience.