As I defined it earlier something is irrational when it results in self-defeating behavior. For example a set of desires could be considered irrational if they were desires for mutually exclusive outcomes, and a belief would be irrational if as a result of believing it we were led to think that we shouldn’t believe it. Being irrational doesn’t necessarily make a belief wrong, but generally we consider avoiding irrational beliefs and desires to be something we should do if we think that success (in terms of reaching outcomes that we want, irrelevant of the specifics) is something we should pursue if we pursue any goals at all.*
Another doctrine that we call often think of as rational is the belief that there exist regularities in the world, in the sense that the course of future events is determined by current events in a way that doesn’t change over time. Obviously if we accept that there are regularities then we can conclude that evidence is a good basis for judgment, and that causation really is a good description of the world (defeating Hume’s problem of induction). However, we can’t prove this belief in regularities though reason alone; it is not an a priori fact that such regularities must exist. But, as the names suggest, perhaps it is possible to demonstrate that the kind of “rationality” described above is also rational in the sense of avoiding irrationality, that the alternatives are self-defeating. If that were indeed the case we would have good reason to believe that regularities exist, and to reject Hume’s problem of induction, even though we don’t have a proof that believing otherwise is necessarily mistaken.
The first constraint that avoiding irrationality places on us is that if we have any goals at all we must believe that the world is predictable to some extent. If someone believed that the world was unpredictable (i.e. without any pattern at all) then they would have no hope of achieving any of their goals, since they would have no way of knowing which actions would bring them closer to their goals. Thus having any goals presupposes that the world is to some extent predictable.
Predictability alone, however, is not enough to establish that rationality forces us to accept the existence of the right kind of regularities. Even though a predictive theory must be in agreement with past events, there are still an infinite number of such predictive theories. For example, one predictive theory about gravity is that it pulls all objects downwards at 9.8 m/s squared, another is that gravity pulls all objects downwards at 9.8 m/s squared until 1 am on Jan 1, 2010, at which point it switches to 10 m/s squared, ect. Obviously if we are to be able to pursue our goals we must select one of these theories as the one we will work towards our goals under (even though there is no guarantee that it is correct); we can’t accept them all as equally possible theories since they contradict each other with respect to future events, the result of which would be to make us unable to predict future events at all, which as pointed out earlier is self-defeating.
Thus we must distinguish one of the possible theories as better than, or at least different from, the rest. Of these possible theories one is obviously distinguished, the unique theory which predicts that events will continue to unfold following the same patterns; there are an infinite number of theories that predict that the way things work will change, but only one that predicts that it will remain the same at all times. Just because we can distinguish this theory from the rest doesn’t necessarily make it the best candidate, however, so let us consider other possible ways to distinguish between the possible candidate theories. As it turns out there aren’t any other rational ways; there are, for example, and infinite number of theories about gravity that say it will change to 10 m/s squared at some point in the future, which differ only in the time they predict that it will change; what reason could we have to pick just one of them?**
Thus, since we must accept some theory if we are to attempt to fulfill our goals, we accept the theory that is uniform over time (which is a fancy way of saying that it predicts that things will keep operating by the same rules). And, if we accept that things follow predictable patterns which are the same at all times, then we can define causation in terms of those patterns (for example, A is the cause of B if the absence of A would “change the pattern” so that B didn’t come about). Similarly, accepting that the laws of nature are the uniform over time means that evidence can provide a good reason to believe something, since if B is usually caused by A then finding B means that A was probably the cause, reasoning which would be defeated if we believed that the causes of B may have been different in the past.
So, while we haven’t defeated the problem of induction and its associated worries we have definitely sidestepped it, showing that the principles of rationality show that we should consider it to be false, and that is definitely progress.
* It is this feature of rationality, that it has normative force even though rationality itself is simply a fact, which motivates some philosophers to make it the basis for morality, although that is not my goal here.
I should also mention that there is an additional principle which I call rational, which is to believe something only when provided with a reason, a principle I shall lean on when I point out that only some theories are naturally distinguished. This principle is rational because believing the opposite is irrational, since to believe things without reason would lead to contradictory beliefs, a self-defeating outcome.
** Of course some will point that the regularities do change over time, for example the Earth may gain or lose mass resulting in a change in the force of gravity. However, the reason we think that gravity would change is because there are other, more fundamental, regularities that are responsible for the Earth’s gravitational force, which predict that it may change in certain circumstances. This isn’t a counterexample to the points I am making here, then, simply an illustration that our theory about the world hangs together, such that some regularities may be predicted to change based on other regularities. You never get away from time independent regularities altogether though.