The nature of problems in statistics is that you don't have a piece of complete information. You are dealing with samples and a limited number of outcomes. A hypothesis test allows you to state with some certainty and of course, you can never really be certain. So that is why there are statements such as we confirm this happened at the 5% significance level. We believe it, but inherent in any statement is the possibility of being wrong. In conclusion, the hypothesis tests allow you to have some belief system rather than saying all problems are uncertain and therefore unknowable.