An Enquiry Concerning Human Understanding and Other Writings: Sceptical doubts

January 24, 2025 | 1,526 words | 8min read

Paper Title: An Enquiry Concerning Human Understanding and Other Writings: Sceptical doubts concerning the operations of the understanding
Link to Paper: https://www.gutenberg.org/files/9662/9662-h/9662-h.htm
Date: 1748
Paper Type: Philosophy
Short Abstract:
The book, in general, is Hume’s attempt to explain human nature. The specific chapter, “Sceptical Doubts,” proposes that all objects of study fall into either “relations of ideas” or “matters of fact.” It presents a theory of knowledge.

Part 1: Types of Knowledge and the Law of Cause and Effect

Objects of human reasoning can be divided into two kinds: relations of ideas and matters of fact.

The statement two plus two is equal to four belongs to the first kind, as it expresses a relation between numbers. The famous formula a² + b² = c² is another example of this category.

Examples of this first category can be discovered without any observations—purely through thought. For instance, mathematics is a closed system that does not require empirical evidence.

On the other hand, matters of fact can never be disproven through pure thinking or deduction; instead, they require induction. This is because every fact and its opposite can be possible without implying a contradiction.
For example, the statements the sun will not rise tomorrow and the sun will rise tomorrow can both be true, depending on the circumstances of the world.

When we consider reasoning about matters of fact, it seems that they are founded on the relation of cause and effect. For example, if you ask a man why he believes his friend is in this country, he will provide some reason—such as having received a letter from him recently or having seen him yesterday.
Why does hearing such rational reasons and discourse assure us that the man’s friend is indeed in the country? Because this reasoning is based on the relation of cause and effect, which, in this case, is very direct and immediate.

Therefore, if we want to clarify the nature of evidence for knowledge, we need to investigate how we arrive at knowledge of the relation between cause and effect.

Hume claims:

We do not know the law of cause and effect a priori—by reasoning or deduction—but by induction. In other words, the law does not hold because we can prove it axiomatically; rather, it holds because we have experienced it to be true.

For example, no one believes they can discover—by reason alone and without experience—how gunpowder explodes or how to make fire with a flintlock. Such knowledge requires experience.

Humans often consider themselves intelligent enough to infer, purely through reason, that one billiard ball will transfer its impulse to another. However, this belief is merely arrogance, arising from habit and custom.

If you were presented with an unfamiliar object and asked to determine its effects and results without relying on past experience, how would you proceed? You might imagine various possible effects, but without observation, choosing one would be entirely arbitrary.

If we lift a stone or a piece of metal into the air without support, we know it will fall. But if we consider the situation a priori, is there anything inherent in the situation that indicates it must fall downwards rather than upwards, or produce some other effect?

All our reasoning a priori will never provide a foundation for predicting what will happen in such a case. Many other possible outcomes could be logically consistent, and selecting one without observation would be arbitrary. Thus, it is futile to pretend that we can determine any specific event or the general rule of cause and effect without the aid of experience and observation.

This limitation is also why philosophers cannot discover the ultimate causes of natural operations—such as the ultimate meaning of life. The inability of humans to resolve such questions is an unavoidable conclusion of philosophy. No theoretical framework can overcome this limitation.

Part 2: On the Validity of Experience

Our prior inquiry seems to answer the question: What is the nature of all our reasoning concerning matters of fact? The answer is that such reasoning is founded on the relation of cause and effect.
If we ask further: What is the nature of all our reasoning concerning that relation? we can answer in one word: experience.

However, we can take the inquiry even further and ask: What is the foundation of all conclusions drawn from experience?

Hume suggests that the best approach is to avoid such questions in the first place and to be modest about what we claim to know. Nevertheless, he provides some negative answers to the question.

There are many limitations to our senses. For example, sight allows us to observe the rough sequence of events when two billiard balls collide, but it does not reveal that the impulse is being transferred from one ball to the other.

If we see an object that has the consistency and color of bread, and we have previously eaten bread, we will not hesitate to eat it again, assuming it will provide us with nourishment.

However, there is no inherent connection between the visible properties of bread and its nourishing qualities—this connection cannot be inferred by reason alone without prior experience.
Our past experiences of eating bread provide us with direct and certain information about that specific object under those specific circumstances.

Why, then, should we assume that our past experience with bread will hold true in the future or for other loaves of bread?

Just because the bread I previously ate nourished me, does it necessarily follow that other bread must also nourish me at another time? The conclusion does not seem logically necessary.

It must be acknowledged that a mental inference is drawn:

I have found that such an object has always been accompanied by such an effect, and
I foresee that other objects, which appear similar, will be accompanied by similar effects.

These two statements are not the same, and one must infer the second from the first. Hume argues that it is not possible to logically infer the second from the first with certainty.

Reasoning can be divided into two types:

1. Demonstrative reasoning, which concerns relations of ideas.
2. Moral reasoning, which concerns matters of fact.

There cannot be a demonstrative argument for the validity of experience, as this would imply that denying it leads to a contradiction—yet the course of nature could change without contradiction.
On the other hand, if we rely on moral reasoning, we end up arguing for the validity of past experiences by using experience itself, which results in a circular argument.

We cannot determine that bread has nourishing qualities from its visible properties alone; experience is our only guide. When we encounter objects with similar appearances, we naturally expect similar effects—but this expectation requires justification.

If we assume that the future will always resemble the past, we rely on an assumption that experience itself cannot justify, since the hidden nature of objects could change without warning.

What, then, ensures that this assumption is valid? One might suggest skepticism as a solution, but philosophically, this answer is unsatisfying. Thus, Hume poses this question to the public for further contemplation.

Conclusion

There are two methods for gaining knowledge. The first is deductive reasoning, where we start with a system of axioms and apply rules of inference, such as modus ponens, to definitively prove things. The second is inductive reasoning, where we observe nature and the universe to draw conclusions. For example, if I observe that apples always fall downward when I drop them, I conclude that objects fall when released.

Inductive reasoning alone cannot be directly applied to the world, as it only provides us with models. We still need deductive reasoning to determine whether our models are accurate, appropriate for use, and applicable in a given context. For instance, in graph theory, we can create various graph models, such as the Erdős–Rényi graph. However, deciding whether it makes sense to use such a model in the real world—say, as a representation of social media networks—requires inference. In other words, we must observe social networks and determine whether they fit the random graph model.

Thus, all applicable human knowledge ultimately relies on deductive reasoning, which provides a solid foundation. This brings us to Hume’s challenge. David Hume argued that all inference in our world depends on the assumption of cause and effect. We believe that if we observe a phenomenon and formulate a natural law, that law will continue to hold in the future. For example, after repeatedly observing that objects fall when dropped, we formulated the law of gravity, assuming it will always hold.

But how do we justify this assumption? We claim that cause and effect hold true because we have inferred it from past observations: we have seen that causes lead to effects, and we believe that if we drop an apple now or a year from now, it will always fall.

Here lies the problem: what justifies the rule of inference itself? If our answer is “cause and effect,” then what justifies that? Once again, we rely on inference. In other words, we attempt to justify inference as a valid method of reasoning by using inference itself—resulting in circular reasoning.