Hume on Causation

Sections 4-7 of the Enquiry present Hume’s treatment of the rationality of beliefs concerning unobserved or future entities or events, based merely on our current or remembered experience. Hume comes to rather famous (or infamous) skeptical conclusions concerning the rationality of such judgments, and thus of whether we can have genuine knowledge of unobserved or future events. His argument became the basis for what has come to be know as the problem of induction. Well look at various parts of this argument below.

Two Kinds of Knowledge

Section 4 of the Enquiry introduces what is now known as ‘Hume’s Fork’. This is the division of all reasoning into two categories—viz. relations of ideas and matters of fact. Hume will argue that causal knowledge as traditionally conceived relies on a notion of necessity that does not fit into either of these two categories. Hence, such “knowledge” is merely specious. Let’s look at the two prongs of “Hume’s Fork.”

Relations of Ideas

Hume describes the knowledge we have by means of perceiving relations between ideas as “discoverable by the mere operation of thought”, and as not dependent on the actual existence of what is thought.

Though there never were a circle or triangle in nature, the truths, demonstrated by Euclid, would for ever retain their certainty and evidence. (EHU §4.1)

Knowledge of relations between ideas is, according to Hume, certain, necessary (in that its denial entails a contradiction), universal, and justified independently of any experience. Hume thinks that our knowledge of mathematics and logic falls into this category. Importantly, according to Hume such knowledge gives us no insight into how things actually are in the empirical world. That is the cost of our being able to have such knowledge purely by thinking.

Using Leibniz’s terminology, Hume is arguing that knowledge of relations between ideas is analytic and a priori. Moreover, Hume is giving a reductive account of the a priori in terms of relations between ideas. His explanation of how it is that we can know the truth of some proposition a priori (i.e. independently of experience), is that the truth depends on a relation between the ideas constitutive of the proposition. For example, the proposition “all red things are red” is one whose truth can be known without having to go look at the world. We can know it to be true just in virtue of the meanings of the terms composing the proposition. Similarly, the proposition “all bachelors are unmarried” is knowable a priori just by knowing the meanings of “bachelor” and “unmarried.” We can know that this proposition is true just by virtue of possessing the relevant ideas. A bachelor just is an unmarried man. So the predicate of the proposition just states a feature which is by definition part of the subject of the proposition. It is in this sense that there is a class of knowledge for Hume that is “discoverable by the mere operation of thought.”

However, Hume and Leibniz radically disagree concerning the extent of analytic truth and our a priori knowledge. Recall that Leibniz construes all truth as analytic, in the sense that all truth depends on the predicate’s being ‘contained’ in the subject. So, for any true judgment, there is in principle an analysis of the subject concept in which it can be shown that the predicate of the judgment is part of (or ‘contained in’) the subject’s complete concept. Hume, in contrast, denies that all truths can be explained in terms of containment. The reason we have a priori knowledge of relation of ideas, for Hume, is because we (i) have privileged access to those ideas, in the sense of stipulating the meanings of the ideas whose relations we are considering (such as “bachelor”), and (ii) the knowledge we have of relations between ideas is thus parasitic on our knowledge of the meanings or content of the ideas, and not anything else (such as the way the world is). Hence one can know that all triangles have three sides is true and necessarily so, just in virtue of knowing what ‘triangle’ means.

Matters of Fact

In contrast to knowledge of relations between ideas, Hume argues that our knowledge of matters of fact is contingent (its denial is not a contradiction), local, uncertain, and dependent for its justification on empirical inquiry. Knowledge of matters of fact includes such propositions as those concerning the number of houses on one’s street, the names of one’s parents, the identity of the current President of the United States, and the atomic weight of gold. Knowledge of all of these propositions requires that we consult how things actually are in the world. In none of these cases can we earn the knowledge merely by thinking.

In the terminology of Leibniz and (after Hume) Kant, knowledge of matters of fact is synthetic and a posteriori. It is justified only by current or past experience and cannot be attained purely by any process of conceptual analysis. Hume is thus taking up a diametrically opposed position to that of Leibniz, who argued that knowledge of empirical matters of fact can be obtained via conceptual analysis of the complete concepts of the substances constituting the actual world. For Leibniz, it may be the case that we cannot complete such an analysis, but one is in principle available (to God).

Hume’s fork results in the position that there is no substantive knowledge of the empirical world that we can have via analysis of our ideas (i.e. our concepts) alone. All empirical knowledge including, e.g. knowledge in the physical sciences, is knowledge of matters of fact, and is thus contingent, local, and uncertain. Hume attempts to drive this conclusion home by arguing that the process of reasoning by means of cause/effect relationships has no ultimate rational basis, and thus cannot give us knowledge of how things are beyond our current or remembered experience. We’ll turn to an examination of that argument now.

Hume on Causal Reasoning

Hume’s argument concerning casual reasoning proceeds in two stages. He first gives a negative argument, arguing that knowledge of causal relationships is not something we can establish via knowledge of the relations between our ideas. Causal knowledge is not knowledge of relations between ideas. Then Hume argues that the causal relationships discovered through experience tell us nothing about how things will be in the future, only how they have been in the past. There is no knowledge of matters of fact concerning necessity. Hence, there is no knowledge of necessary connections between what we have experienced and what we might experience in the future. Thus there is no rational (i.e. demonstrative or logical) basis for our beliefs concerning anything we have not experienced, including future events and events outside our current ambit.

The positive phase of Hume’s argument then consists in articulating why we do have the causal beliefs that we do (recall that the relation of cause-effect is one of the three central associative forms of thought for Hume). The only basis, he thinks, for our beliefs that are the result of causal thinking is that they are formed by ‘custom’ or habit. It is, according to Hume, part of our nature as human beings to form strong beliefs concerning the relationships between events that we experience when we perceive there to be regular connections between the occurrence of those events. Causal relationships that we believe to exist in the world are thus nothing more than habits of association, and thus features of our psychology. Let’s look at Hume’s arguments more closely.

The Negative Argument

Hume first argues that all causal reasoning concerns matters of fact and could never be acquired by any a priori process, which he equates with knowledge of relations between ideas, as for example is gained in demonstrative reasoning (EHU §4.6). He gives several examples. If we imagine a fully informed, perfect reasoner (‘Adam’), that reasoner could still never determine that a transparent and fluid substance (water) could drown, or that, by means of reason alone, any particular entity actually exists. Part of the reason for this conclusion is Hume’s principle that there are no necessary connections between distinct existences.

The mind can never possibly find the effect in the supposed cause, by the most accurate scrutiny and examination. For the effect is totally different from the cause, and consequently can never be discovered in it. (EHU §4.9)

every effect is a distinct event from its cause. It could not, therefore, be discovered in the cause, and the first invention or conception of it, à priori, must be entirely arbitrary. And even after it is suggested, the conjunction of it with the cause must appear equally arbitrary; since there are always many other effects, which, to reason, must seem fully as consistent and natural. In vain, therefore, should we pretend to determine any single event, or infer any cause or effect, without the assistance of observation and experience. (EHU §4.11)

Hume’s point here is that there is no logical contradiction entailed by saying that a cause in one case has one effect (e.g. ingesting caffeine and being more alert) and in a different but relevantly similar case a different effect (e.g. the caffeine causing drowsiness). Part of Hume’s explanation of this is that causes and their effects are distinctly existing things (events), and thus there is nothing about the existence of the one that logically requires the existence of the other (though there may be reasons to think that it physically requires the existence of the other). This is in contrast to the kinds of connections had by our ideas. The idea of a three sided plane figure does entail, according to Hume, the truth that this plane figure has interior angles which equal 180˚.

Once Hume has established that causal relationships cannot be known via a priori demonstration, independently of our knowledge, via experience, of matters of fact, he goes on to examine whether our causal reasoning concerning such matters of fact has any rational basis. Hume is straightforward about his aim.

I shall content myself, in this section, with an easy task, and shall pretend only to give a negative answer to the question here proposed. I say then, that, even after we have experience of the operations of cause and effect, our conclusions from that experience are not founded on reasoning, or any process of the understanding. This answer we must endeavour, both to explain and to defend. (EHU §4.15)

To see Hume’s argument let’s consider the following bit of causal reasoning.

  1. In the past, when I have drunk coffee I have felt more alert
  2. ∴ When I drink this coffee (and presuming that circumstances are relevantly similar), I will feel more alert – i.e. the coffee causes alertness

What basis, Hume asks, do we have for the transition from the first belief to the second? More specifically, with what right or justification do we hold the second belief, based on the first? Hume grants that we often form beliefs like (2) on the basis of beliefs like (1), and thus that there is absolutely nothing unusual about our doing so. He simply wishes to know what the reasoning, or rational process is, by which we come to (2) given (1). Hume’s claim is that we have no justification or rational basis for holding (2) given (1). Why not?

Hume first points out we do not intuit the truth of (2) given (1). That is, it is not self-evident or otherwise incontrovertibly obvious that (2) holds on the basis of (1). We can see this, for example, by asking whether there is any logical contradiction in denying (2) given (1). Further, Hume argues that there is no process of rational demonstration that could lead us to to (2) given (1). Again, we can see this by means of the fact that there is no logical contradiction generated in denying (2) while accepting (1).

it implies no contradiction, that the course of nature may change, and that an object, seemingly like those which we have experienced, may be attended with different or contrary effects. May I not clearly and distinctly conceive, that a body, falling from the clouds, and which, in all other respects, resembles snow, has yet the taste of salt or feeling of fire? Is there any more intelligible proposition than to affirm, that all the trees will flourish in December and January, and decay in May and June? Now whatever is intelligible, and can be distinctly conceived, implies no contradiction, and can never be proved false by any demonstrative argument or abstract reasoning à priori. (EHU §4.18)

Hume concludes that since the reasoning process here cannot be intuitive or demonstrative, then all that remains is for it to be based on probability. That is, the inference from (1) to (2) is a form of inductive inference, where the truth of (1) is taken to increase the likelihood that (2) is true. But, Hume asks, what reason have to think that the truth of (1) does make the truth of (2) more likely? The only basis we have, Hume thinks, is the assumption that the future will be like the past. But reason have we to accept this assumption? It cannot be one accepted on the basis of intuition or demonstration, for reasons given above. And it cannot be accepted by appeals to probability, for that is exactly what we are trying to prove, and hence the argument would be circular. So, Hume concludes, there is no rational basis for accepting conclusions such as (2) on the basis of (1).

We have said, that all arguments concerning existence are founded on the relation of cause and effect; that our knowledge of that relation is derived entirely from experience; and that all our experimental conclusions proceed upon the supposition, that the future will be conformable to the past. To endeavour, therefore, the proof of this last supposition by probable arguments, or arguments regarding existence, must be evidently going in a circle, and taking that for granted, which is the very point in question. In reality, all arguments from experience are founded on the similarity, which we discover among natural objects, and by which we are induced to expect effects similar to those, which we have found to follow from such objects. (EHU §4.19-20)

It is impossible, therefore, that any arguments from experience can prove this resemblance of the past to the future; since all these arguments are founded on the supposition of that resemblance. Let the course of things be allowed hitherto ever so regular; that alone, without some new argument or inference, proves not, that, for the future, it will continue so…it is not reasoning which engages us to suppose the past resembling the future, and to expect similar effects from causes, which are, to appearance, similar. This is the proposition which I intended to enforce in the present section. (EHU §4.21-23)

It is important to note the generality of Hume’s argument. It doesn’t just hold of the specific example I gave concerning coffee given above, but rather for any kind of causal inference based on observed effects to any inference concerning how things must be, or will be, or are in places not yet observed. In effect, it condemns the entirety of scientific reasoning to the elucidation of perceived regularities, rather than to the realization of any sort of deeper metaphysical truths concerning the natures of perceived existing things.

The Positive Argument

If Hume’s negative argument is successful, we have no rational basis for the causal inferences we routinely make, both in the course of everyday experience and in scientific theory construction. This doesn’t mean, Hume thinks, that we shouldn’t make such inferences (though in doing so we shouldn’t construe them as rational), nor could we stop making them even if we wanted to! Hume asks us, in §5, to consider a person endowed “with the strongest faculties of reason and reflection” and the judgments this person would make.

Suppose again, that he has acquired more experience, and has lived so long in the world as to have observed similar objects or events to be constantly conjoined together; what is the consequence of this experience? He immediately infers the existence of one object from the appearance of the other. Yet he has not, by all his experience, acquired any idea or knowledge of the secret power, by which the one object produces the other; nor is it, by any process of reasoning, he is engaged to draw this inference. But still he finds himself determined to draw it: And though he should be convinced, that his understanding has no part in the operation, he would nevertheless continue in the same course of thinking. There is some other principle, which determines him to form such a conclusion. (EHU §5.4)

Hume’s answer is that this “other principle” is ‘custom or habit’. That is, the causal inferences we make are not grounded in the actual perception of any real connections in nature. Instead, we make such inferences because of our psychological makeup. Part of human nature is to be apt to make such inferences, regardless of whether there is anything in reality corresponding to them.

For wherever the repetition of any particular act or operation produces a propensity to renew the same act or operation, without being impelled by any reasoning or process of the understanding; we always say, that this propensity is the effect of Custom. By employing that word, we pretend not to have given the ultimate reason of such a propensity. We only point out a principle of human nature, which is universally acknowledged, and which is well known by its effects. Perhaps, we can push our enquiries no farther, or pretend to give the cause of this cause; but must rest contented with it as the ultimate principle, which we can assign, of all our conclusions from experience. (EHU §5.5)

Custom, not reason, explains why we make such inferences, and there may be nothing which explains why this custom or habit exists. It may just be a brute fact about human nature. We can’t say for sure either way. Hume thinks that this conclusion is not particularly catastrophic.

it is certain we here advance a very intelligible proposition at least, if not a true one, when we assert, that, after the constant conjunction of two objects, heat and flame, for instance, weight and solidity, we are determined by custom alone to expect the one from the appearance of the other. This hypothesis seems even the only one, which explains the difficulty, why we draw, from a thousand instances, an inference, which we are not able to draw from one instance, that is, in no respect, different from them. Reason is incapable of any such variation. The conclusions, which it draws from considering one circle, are the same which it would form upon surveying all the circles in the universe. But no man, having seen only one body move after being impelled by another, could infer, that every other body will move after a like impulse. All inferences from experience, therefore, are effects of custom, not of reasoning. (EHU §5.5)

Thus there is no more to the relation of cause and effect than the experience of constant conjunction. The best we can do is base our causal inferences on the experience of such constant conjunction, with the awareness that these inferences are not rational, and thus not ‘justified’ in any epistemic sense. They are simply the natural human response to experience of ‘constant conjunction’ or empirical regularities of any kind. These responses are practically useful, and are the means by which human being successfully get around in the world, but they are not rational, in any deep sense of that word.

Custom, then, is the great guide of human life. It is that principle alone, which renders our experience useful to us, and makes us expect, for the future, a similar train of events with those which have appeared in the past. Without the influence of custom, we should be entirely ignorant of every matter of fact, beyond what is immediately present to the memory and senses. We should never know how to adjust means to ends, or to employ our natural powers in the production of any effect. There would be an end at once of all action, as well as of the chief part of speculation. (EHU §5.6)

Necessary Connections

One argument, that might be offered to counter Hume’s skeptical musings regarding our causal reasoning, is that such reasoning really is justified by a mind-independent relation in the world. The causal relation, it might be thought, is a necessary relation between a cause and its effect. If A causes B, then given A, B must necessarily occur. For a great many thinkers in the seventeenth and eighteenth centuries, this was the definition of causation.

Hume counters by questioning whether we have any such concept of necessary connection. In accordance with his broader philosophical psychology, he asks whether there is any impression, whether from within or without, from which we might trace the idea of necessary connection.

When we look about us towards external objects, and consider the operation of causes, we are never able, in a single instance, to discover any power or necessary connexion; any quality, which binds the effect to the cause, and renders the one an infallible consequence of the other. We only find, that the one does actually, in fact, follow the other. The impulse of one billiard-ball is attended with motion in the second. This is the whole that appears to the outward senses. (EHU §7.6)

Hume argue here, in keeping with what he said previously, that we perceive only (constant) conjunction not connection. For example, we perceive one event (the impact of two billiard balls) followed by another (the motion of the struck ball). We haven no impression that the second event necessarily connects with or must follow from the first.

Not perceiving any impression of connection, necessary or otherwise, in our experience of external objects, Hume asks whether we have any internal impression, via reflection on the mind’s activities (here recall Locke and Leibniz), that might furnish us with the idea of necessary connection.

Since, therefore, external objects as they appear to the senses, give us no idea of power or necessary connexion, by their operation in particular instances, let us see, whether this idea be derived from reflection on the operations of our own minds, and be copied from any internal impression. (EHU §7.9)

Hume considers our volitions, that is, the acts of will that determine our bodies to move. But again, he finds only conjunction and not connection. There is nothing in our experience of our own volitions which makes any clearer how or why our bodies move in response. There is, however, an important internal impression that is in part responsible for our idea of causation, though not in the way required to establish the idea of necessary connection.

after a repetition of similar instances, the mind is carried by habit, upon the appearance of one event, to expect its usual attendant, and to believe, that it will exist. This connexion, therefore, which we feel in the mind, this customary transition of the imagination from one object to its usual attendant, is the sentiment or impression, from which we form the idea of power or necessary connexion. Nothing farther is in the case. (EHU §7.28)

So, given the existence of external regularities (or regularities in occurrence of external sensations), there is an important internal source for our idea of a cause, but it is a mere feeling of expectation, not a rational inference, or impression of genuine necessary connection.

Hume then suggests several alternative ways of defining ‘cause’, that don’t result in any misconstrual of what we actually experience (EHU §7.29). Causation should be understood as meaning,

Suitably to this experience, therefore, we may define a cause to be [i] an object, followed by another, and where all the objects, similar to the first, are followed by objects similar to the second. Or in other words, [ii] where, if the first object had not been, the second never had existed.

Or alternatively,

We may, therefore, suitably to this experience, form another definition of cause; and call it, [iii] an object followed by another, and whose appearance always conveys the thought to that other.

It is not clear that all three of these definitions are synonymous, so it isn’t clear which Hume thinks is the proper definition. But Hume thinks that any of these definitions is preferable to the definition of causation in terms of necessary connection. He has argued that the notion of necessary connection has no cognitive content, since it cannot be traced to any impression. This problem does not hold, he thinks, of any of his three proposed definitions. Hence, Hume’s psychological method gives, he thinks, clear rewards in metaphysics and epistemology. We have learned what causal ‘reasoning’ is (i.e. it isn’t reasoning at all), and we have a better understanding of what “causes” are and why traditional metaphysical theories have been mistaken in their claims concerning the nature of causation, causal relations, and our knowledge thereof.


  1. [@allison2008; @bennett2001a; @bennett2001b; @garrett1997; @garrett2014a; @loeb2006; @morris2013; @norton2009; @ott2009; @owen1999; @read2000; @strawson1989; @winkler1991] ↩︎

Icon by Nun from The Noun Project. Website built with Org-mode, Hugo, and Netlify.