What Is the Problem of Induction? Can It Be Solved?

What is the problem of induction? Can it be solved?

In this essay I will begin by outlining the differences between deduction and induction, and I will examine the problem of induction. I will then go on to assess the various responses to the problem and seek to refute them. I will finish by concluding that the problem of induction is insoluble as the responses fail to deductively justify induction.

In order to fully understand the nature of induction and the problem associated with it, we must first examine the distinction between induction and deduction. Deductive arguments are arguments where the premises must entail the conclusion, meaning that it is logically inconceivable for the premises to be true and the conclusion to be false. For example:

Premise 1: All bachelors are unmarried

Premise 2: John is a bachelor

Conclusion: John is unmarried

As the example shows, the premises of deductive arguments make its conclusion necessarily true, indeed if john is a bachelor we know he is unmarried through “the mere operation of thought.”^[1] However Inductive arguments differ from deductive arguments, in that the premises of inductive arguments do not always entail the conclusion, as the nature of the reasoning leads to a certain epistemic uncertainty regarding the conclusion. Indeed inductive arguments extrapolate and generalize knowledge from past experiences to posit strong arguments, which only support the truth of the conclusion. For example:

Premise 1: All sun cream I have ever used prevents me being sun burnt.

Premise 2: This cream is sun cream

Conclusion: This cream will prevent me being sun burnt

Inductive arguments appear to be one of the most common forms of reasoning, and it is vitally important in both every day life and in scientific method. Indeed inductive arguments allow us to make predictions based on previous experiences, such as ‘the sun will rise tomorrow because it has done so every day in recorded history.’ Induction largely relies on ‘more of the same inferences,’ in other words because such a thing has happened so often in the past this gives us a good reason to believe it will happen again in the future.

However David Hume in his work ‘An Enquiry Concerning Human Understanding,’ challenges the reliability of induction by arguing that we cannot assume the uniformity of nature. Hume argued induction was problematic, because all of our inductions rely on the uniformity of nature, in that they assume “the future will be conformable to the past.”^[2] However this was an issue for Hume, because we don’t seem to have any justification for believing in the uniformity of nature, other than past experiences of its success. Hume’s argument can be broken down like this:

Step 1: All arguments are either deductive or inductive

Step 2: There is no valid argument for the reliability of induction

Step 3: An inductive argument for induction would be circular.

Indeed Hume is arguing that in order to justify our belief in the uniformity of nature, we would need to provide a deductive argument for the reliability of induction. However in step 2, Hume notes that no valid deductive argument for induction actually exists. This is because in order to construct a deductive argument for induction that does not beg the question, the argument would have to limit itself to premises that do not require induction, such as past observations. However we can’t construct a deductive argument out of past observations, because they are merely past experiences and only guarantee the success of induction in the past. Indeed induction failing in the future “implies no contradiction,”^[3] as the argument is based on past observations of induction being successful, which “imposes no logical restrictions”^[4] on the reliability of induction in the future. In addition step 3 argues, “We can never use experience to prove the inductive principle,”^[5] because this would be a circular argument, as induction is “a principle itself derived from probable reasoning”^[6] and therefore obviously fallacious.

For Hume therefore there is nothing that justifies our belief in the uniformity of nature, and thus the future reliability of induction. As a result Hume argues that predictions based on past observations are just as reliable as predictions that are not based on previous observation. For example if a person were to get a piece of wood and declare that if they put it onto a fire that it would not burn, that prediction is just as reliable as saying the wood would burn. Indeed induction would argue that of course the wood would burn, because every time we have put wood on fire in the past it has burnt. This is because induction is relying on the uniformity of nature, and because it has worked well in the past it will obviously continue to do so. However a person could respond by arguing that because nothing guarantees the uniformity of nature, and because fire has successfully burnt wood in the past so many times there is a chance it will be unsuccessful this time. Hume’s point here is that appealing to something because it has consistently been unsuccessful and thus likely to change is just as bad as arguing that because of past successes there is no reason to believe it will change. The problem of induction therefore is that nothing justifies our belief in the uniformity of nature, which in turn undermines the notion that we can make predictions based on previous observations.

However a number of responses have been put forward against Hume’s problem of induction, one of which comes from Karl Popper who has argued that induction has no place in science at all. For Popper “induction cannot be logically justified,”^[7] but this was ultimately irrelevant in regard to science, because science only ever relies on deduction. Popper argued that when someone asks ‘how can we justify our theories given that they rely on induction,’ that that person is asking the wrong question. Indeed according to Popper science doesn’t try to seek justification for its theories, as theories are not ‘confirmed.’ Rather science is a deductive process that aims to formulate hypotheses that can be falsified via experimentation. Popper argued that “every good scientific theory is a prohibition: it forbids certain things to happen,”^[8] in order to rigorously criticize them and search for errors within the theory. For example: