The Sun Isn't Guaranteed To Rise Tomorrow. Here's Why.

An exploration of David Hume's "Problem of Induction"

The sun isn’t guaranteed to rise tomorrow. In fact, nothing is. I’ll explain why in this post.

Today, I’ll be exploring David Hume’s problem of induction. By the end of this post, I’ll have explained why there is just as much rational justification for the claim that “a giant bowl of spaghetti will rise tomorrow” as there is for “the sun will rise tomorrow”. Before I get into the problem of induction, I’ll provide some context on who David Hume is.

---

---

David Hume

David Hume (1711-1776) was a Scottish philosopher, historian, and economist who contributed to a variety of intellectual realms. Hume’s ideas went on to have a profound effect on analytic philosophy, cognitive science, theology, and utilitarianism. Hume opposed rationalism, and believed that ‘passions’ control our behaviour rather than reason. As we’ll explore today, Hume also argued that inductive reasoning and our tendency to believe in causality cannot be rationally justified — this idea is known as Hume’s “problem of induction”.

Induction vs Deduction

When we interact with the world, we typically use three types of reasoning; inductive reasoning, deductive reasoning, and abductive reasoning. Today, I’ll be focusing solely on deduction and induction. This is because abduction is typically used to explain events that have already occurred, and therefore it has little relevance to this post. You’ll understand why in a second.

On a day-to-day basis, we employ inductive and deductive reasoning to interact with and make sense of the world around us. Inductive reasoning is when you take an observation, a set of observations, or other premises, and then use them to make a generalization or prediction about the world. Induction is typically used to show the likelihood of something occurring, and it makes probabilistic claims.

For example, if you observe 10,000 swans and they’re all white, you probably have a strong reason to believe that all swans are white. If you come to this conclusion, you’ve just used induction (taken a set of observations and used them to make a prediction about the future). Note that within this example, you’re not necessarily saying that all swans are white, but rather, that all swans are probably white. This is because induction cannot make deterministic claims, which are claims that guarantee truth or falsity and deal with absolute certainty. Instead, induction makes probabilistic claims, which are claims that are either “more likely” or “less likely” to happen.

Deduction, on the other hand, is when you use ideas, rules, or principles to come to a conclusion. Deductive reasoning produces conclusions with certainty and deals with deterministic claims. It is impossible for a deductive argument to be “probably true” - deductive claims aim for certainty. It is impossible for the premises of a deductive argument to be true and produce a false conclusion. For example, we can know that the example above (with the swans) is not a deductive claim because the conclusion is probably true, rather than being certainly, 100% true. Also, the premises are true (my observations of the 10,000 white swans), yet the conclusion is actually false (black swans, albeit rare, do exist). For hundreds of years, people believed that all swans were white, until a Dutch explorer discovered black swans on the Australian subcontinent in 1697.

Check out the following example of deductive reasoning;

All cats have ears. American Shorthairs are cats. Therefore American Shorthairs have ears.

Within my example above, I am dealing with ideas or principles which I then use to create, or deduce, a conclusion. Even if I was someone who had never seen an American Shorthair cat in my life, I would still be able to know that it has ears based on my premises, which are that all cats have ears, and American Shorthairs are cats. Again, deductive arguments deal with certainty (true premises will always produce a true conclusion), so the truth present within the premises is enough to guarantee the truth within the conclusion. This is why deductive reasoning deals with deterministic claims, rather than probabilistic ones. Unlike inductive claims, there is no room for doubt or something being “probable” within deductive claims. Deductive reasoning is often used in math and science, and is great for abstract and theoretical reasoning.

Due to the nature of deductive reasoning however, it is not typically used to make predictions about the future. This is because deduction deals with certainty, and results in conclusions that are either 100% true or totally false. Unlike induction, deductive claims are not “more likely” or “less likely”. While deductive reasoning may be used to make predictions within controlled environments, such as mathematical or theoretical contexts, when it comes to the real world, we primarily use inductive reasoning. You can’t make accurate deductive claims about the future because there are simply too many variables and unpredictabilities. This is why any and all claims that predict the future use inductive reasoning, because you can’t make claims with 100% certainty about the future (as deductive reasoning would). This is why your local meteorologist tells you that there’s a 90% chance of rain on Friday, rather than saying “it’s going to rain on Friday''. Any meteorologist who tells you there’s a 100% chance of rain on Friday has obviously never read Hume.

Hume’s Problem of Induction

Now that I’ve established and defined both induction and deduction, let’s dive into Hume’s problem.

Hume remarked that all predictions about the future (inductive claims) rely on the assumption that nature is uniform. In other words, when we make claims about tomorrow, we assume that aspects of tomorrow will be like today (we assume that nature is uniform). If induction relies on nature being uniform, we have to be justified in believing in the uniformity of nature, as it’s literally the foundation of any inductive argument.

Now, we could use either induction or deduction to explain the uniformity of nature. Right off the bat, we have to cross out deduction. We can’t use deduction to justify the uniformity of nature, because the claim that “nature is uniform” is a claim that goes beyond our observational and perceptual experience. It’s a claim that applies to tomorrow, the week after, and millions of years from now. Therefore, we can’t deduce that nature is uniform because it’s impossible to have premises, observations, rules, or principles that would support such a broad and all-encompassing claim with absolute certainty (remember, all deductive claims are deterministic, which means that they are either true or false). In other words, there are too many variables and too much unpredictability to use deduction within claims about the future.

Here’s the kicker; we can’t use induction to justify that nature is uniform either, because this would result in circularity (otherwise known as assuming your conclusion is true when building your argument). Circularity is sort of like if your friend were to say; “I always speak the truth because I’m an honest person”. It’s also the equivalent of believing in a crystal ball because it tells you that it only says true things.

Remember, induction is when you take observations and make a prediction with them. If you were to justify the uniformity of nature by using induction, you’d essentially be saying; “nature was uniform today, so it will be uniform tomorrow too”, and it makes you no different than your “honest” friend or your “truthful” crystal ball.

So, if we can’t justify that nature is uniform, we can’t justify induction. Now, you might be thinking: “big deal, why is this even important?”.

As I highlighted earlier, induction is our only way to make predictions about the real world. Deduction is great for making predictions within an abstract or theoretical framework, where variables and unpredictability are limited, but when it comes to predictions about the real world, induction is all that we’ve got. Induction is heavily employed within science, politics, and countless other parts of society. If induction isn’t rationally justified, and we can’t use deduction to make claims about the future, then technically speaking, all of our claims about the future aren’t justified! Everything, from your belief that the sun will rise tomorrow, to your weather guy telling you that it’s going to rain tomorrow, all are claims that aren’t rationally justified, yet we believe them anyway. Hume attributes our tendency to believe in these rationally unjustified claims to regularity. He argues that, since claims such as the sun rising every morning have happened so regularly, it’s basically impossible for us not to believe in them.

Anyone who claims that a giant bowl of spaghetti will rise tomorrow morning, replacing the sun, would quickly be labeled insane. But, if Hume is right, a bowl of spaghetti replacing the sun is no more rationally justified than the sun rising in the first place.

---