Hume concludes bluntly (p. 28) that "All inferences from experience ... are effects of custom, not of reasoning." Thus we are justified in expecting the sun to rise tomorrow, though only by habit or custom -- it is not a proposition supported by reason. This seems a very strange claim.

Has he simply defined reason too narrowly, to include only deductive inference? We are tempted to say that it is an inductive inference, which we believe to the degree of certainty that the evidence (our past experience, reports of others) warrants -- and that the weight of probability is very heavy in this case. Given our individual and collective experience and known history (including geologic history), the sun will rise tomorrow to a near certainty, we want to say, and inductive reason strongly warrants our belief that it will.

But Hume might reply that we have only made the proposition seem highly probable by artificially excluding myriad other possibilities of which we have no knowledge. We don't know what we don't know, so we cannot accurately factor in all possibilities to calculate the genuine liklihood of its occurring. Thus the probability calculation we used to justify the belief is not meaningful data after all.