11/11/2014
On Hume’s “Problem of Induction”

David Hume presented the “Problem of Induction” in the 18th century. Several philosophers including Hans Reichenbach, Peter F. Strawson and Laurence BonJour have attempted to solve the problem in different ways. In this paper, I will lay out the problem of induction from Hume and the counter-arguments from the three philosophers above. Lastly, I will close with my personal argument against Hume’s problem. Before explaining Hume’s problem of induction, we must first explain induction itself. Induction is simply creating a general, repetitive claim based on past experiences. For example, if every time I threw a ball into the air and it came back down, I could make the inductive inference that all balls thrown into the air come back down. This has always been a popular way of inference, even providing the basis for science itself. However, Hume raises the question, “What actually gives us justification for believing past patterns entail future repetition of the patterns?” Hume claims we have only two types of reasoning available, demonstrative (relation of ideas, a priori) and experimental (based on experiences, a posteriori) reasoning (Hume p. 1). Demonstrative reasoning allows one to make conclusions off of reason alone, for example “all bachelors are unmarried” is demonstrative. Experimental reasoning depends on experiences and interactions with the outside world. For example, if I see fresh deer tracks in the snow, I know they are fresh deer tracks because I have seen deer make the fresh tracks in the past, and only deer are capable of making them. Therefore, there are deer in the area. This is an example of inductive reasoning. Hume believes this cannot work because it cannot ever lead to a contradiction, despite a contradiction being entirely possible in the real world (p. 3). Hume’s doubt of induction presents a massive problem to the current conception of knowledge that humans have. Nearly everything we consider to be a factual pattern is being called into question and doubted. However, Hume believes that we have no reliable justification for thinking these patterns will continue. There is no way to immediately know if the future predictions are true until that certain time in the future arrives, even then it would be presupposing the reasoning it used in the first place. Induction not only fails to produce contradictions, but it also begs the question in claiming the cause is also the effect. Hume also argues if we were to create a system to reliably test induction, that system itself would also require another system to verify it and an infinite cycle of verification systems would begin, leaving us in a worse problem (Hume, p. 16). Hume concludes that induction cannot possibly be a reliable way of obtaining knowledge and should not be trusted.

Hans Reichenbach agrees with Hume that we cannot demonstrate an inductive inference before it occurs and that demonstrating one in the future presupposes itself. However, Reichenbach believes that induction is our best pragmatic approach in attempting to learn the truth about the world (Reichenbach pp. 19-20). Although it is not a 100% reliable system, induction is still a good habit to use to analyze the world. Reichenbach argues that any other method that we could possibly use to reason with could not be as good as induction. He gives the example of a sick patient who is close to death. The patient’s doctor offers a surgical procedure that may or may not save the patients life, however, without the surgery the patient will certainly die. Reichenbach likens induction to the patient opting for the surgery. Although it may not be 100% reliable, it is the best chance we have for reasoning. Therefore, Hume is correct about induction’s flaws, but induction should still be considered reliable because it is the most sensible and realistic approach.

Peter F. Strawson disagrees with Hume on the basis that seeking justification