- In my experience, all the New Yorkers I've met so far were rude.
- Therefore, in general, all New Yorkers are rude. (Or, the next New Yorker I will meet will be rude.)
(UN) What has happened in the past will continue to happen in the future. (Or, past experience is a sure guide for what will happen in cases that we haven't experienced.)But then what is the justification for (UN)? (UN) can be justified deductively or inductively. However, no deductive argument can prove (UN) true, since (UN) is a contingent, not a necessary, truth. And no inductive argument can support (UN), since such an argument would be circular, assuming the very thing that it aims to support.
Now, the problem of justifying inductive inferences is supposed to hold for all inductive inferences from the known or observed (i.e., that with which we have had experience) to the unknown or unobserved (i.e., that with which we have had no experience). But is there an epistemically relevant difference between inductive inferences from past to future and inductive inference from present to past?
In this NOVA Science Now episode, astronomers draw conclusions about the history of the solar system from meteorites they find on Earth in the present and evolutionary biologists draw conclusions about the evolutionary history of our species from lice they find in our hair in the present.
Should we have more rational confidence in these inductive inference from present evidence (e.g., meteorites and lice) to past (e.g., evolutionary history of the solar system and our species)?
Perhaps. Consider that the past is supposed to be closed, insofar as what has happened has already happened, whereas the future is supposed to be open, insofar as what will happen hasn't already happened. If so, then when we reason inductively from past to future, rather than from present to past, there is simply more that we can get wrong.
If this is correct, then does that mean that we are more likely to get our present-past inductive inferences right than our past-future inductive inferences simply because there is a lot more we can get wrong about the future? And if so, should we have more rational confidence in present-past inductive inference than in past-future inductive inferences?