Induction is the practice of drawing general conclusions based on particular experiences. Although this method is essential to empiricism and the scientific method, there is always something inherently uncertain about it, because we may acquire new data that are different and that disprove our previous conclusions. Essentially, the principle of induction teaches us that we can predict the future based on what has happened in the past, which we cannot. Hume argues that in the absence of real knowledge of the nature of the connection between events, we cannot adequately justify inductive assumptions. Hume suggests two possible justifications and rejects them both. The first justification is functional: It is only logical that the future must resemble the past. Hume pointed out that we can just as easily imagine a world of chaos, so logic cannot guarantee our inductions. The second justification is that we can assume that something will continue to happen because it has always happened before.

To Hume, this kind of reasoning is circular and lacks a foundation in reason. Despite the efforts of John Stuart Mill and others, some might argue that the problem of induction has never been adequately resolved. Hume left the discussion with the opinion that we have an instinctual belief in induction, rooted in our own biological habits, that we cannot shake and yet cannot prove. Hume allows that we can still use induction, like causation, to function on a daily basis as long as we recognize the limitations of our knowledge.