Hume states that all reasoning concerning matters of fact comes from deriving a relationship between cause and effect (Hume, 296). However, the relationship between cause and effect is not discovered even by reason, but also by observation and experience (Hume, 297). Because they are connected only by observation, we can never know for sure whether causal relationships exist, and he argues that effects are distinct from their causes (Sepielli). He offers the example of two billiard balls, where one moves in a straight line towards the other. He thinks that there are hundreds of events that can result from the collision of billiard balls, and we cannot be sure which ones will occur (Hume, 298). Even though we examined the billiard balls and the table (Sepielli), and observed the same observation every time the balls moved towards each other, Hume believes that we can only use these observations to learn how the balls would influenced here, in the present, but we cannot use them to know what will happen in the future. In this case, Hume's problem of induction can be summarized as follows. If we truly believe that the future will be like the past, and we want to justify this belief, but we cannot justify it demonstratively or by repeatedly providing similar observations, then how should we justify it?