Philosophy Pages

    Dictionary    Study Guide  Logic   F A Q s
  History Timeline Philosophers   Locke

Hume: Empiricist Naturalism

David Hume

Life and Works
. . Ideas
. . Belief
. . Cause & Effect
. . The Self
. . Skepticism
. . Morality
. . Religion
Internet Sources

Later in eighteenth century, Scottish philosopher David Hume sought to develop more fully the consequences of Locke's cautious empiricism by applying the scientific methods of observation to a study of human nature itself. We cannot rely on the common-sense pronouncements of popular superstition, which illustrate human conduct without offering any illumination, Hume held, nor can we achieve any genuine progress by means of abstract metaphysical speculation, which imposes a spurious clarity upon profound issues. The alternative is to reject all easy answers, employing the negative results of philosophical skepticism as a legitimate place to start.

Stated more positively, Hume's position is that since human beings do in fact live and function in the world, we should try to observe how they do so. The key principle to be applied to any investigation of our cognitive capacities is, then, an attempt to discover the causes of human belief. This attempt is neither the popular project of noticing and cataloging human beliefs nor the metaphysical effort to provide them with an infallible rational justification. According to Hume, the proper goal of philosophy is simply to explain why we believe what we do. His own attempt to achieve that goal was the focus of Book I of the Treatise of Human Nature and all of the first Enquiry.


Hume's analysis of human belief begins with a careful distinction among our mental contents: impressions are the direct, vivid, and forceful products of immediate experience; ideas are merely feeble copies of these original impressions. (Enquiry II) Thus, for example, the background color of the screen at which I am now looking is an impression, while my memory of the color of my mother's hair is merely an idea. Since every idea must be derived from an antecedent impression, Hume supposed, it always makes sense to inquire into the origins of our ideas by asking from which impressions they are derived.

To this beginning, add the fact that each of our ideas and impressions is entirely separable from every other, on Hume's view. The apparent connection of one idea to another is invariably the result of an association that we manufacture ourselves. (Enquiry III) We use our mental operations to link ideas to each other in one of three ways: resemblance, contiguity, or cause and effect. (This animal looks like that animal; this book is on that table; moving this switch turns off the light, for example.) Experience provides us with both the ideas themselves and our awareness of their association. All human beliefs (including those we regard as cases of knowledge) result from repeated applications of these simple associations.

Hume further distinguished between two sorts of belief. (Enquiry IV i) Relations of ideas are beliefs grounded wholly on associations formed within the mind; they are capable of demonstration because they have no external referent. Matters of fact are beliefs that claim to report the nature of existing things; they are always contingent. (This is Hume's version of the a priori / a posteriori distinction.) Mathematical and logical knowledge relies upon relations of ideas; it is uncontroversial but uninformative. The interesting but problematic propositions of natural science depend upon matters of fact. Abstract metaphysics mistakenly (and fruitlessly) tries to achieve the certainty of the former with the content of the latter.

Matters of Fact

Since genuine information rests upon our belief in matters of fact, Hume was particularly concerned to explain their origin. Such beliefs can reach beyond the content of present sense-impressions and memory, Hume held, only by appealing to presumed connections of cause and effect. But since each idea is distinct and separable from every other, there is no self-evident relation; these connections can only be derived from our experience of similar cases. So the crucial question in epistemology is to ask exactly how it is possible for us to learn from experience. (Enquiry IV ii)

Here, Hume supposed, the most obvious point is a negative one: causal reasoning can never be justified rationally. In order to learn, we must suppose that our past experiences bear some relevance to present and future cases. But although we do indeed believe that the future will be like the past, the truth of that belief is not self-evident. In fact, it is always possible for nature to change, so inferences from past to future are never rationally certain. Thus, on Hume's view, all beliefs in matters of fact are fundamentally non-rational. (Enquiry V i)

Consider Hume's favorite example: our belief that the sun will rise tomorrow. Clearly, this is a matter of fact; it rests on our conviction that each sunrise is an effect caused by the rotation of the earth. But our belief in that causal relation is based on past observations, and our confidence that it will continue tomorrow cannot be justified by reference to the past. So we have no rational basis for believing that the sun will rise tomorrow. Yet we do believe it!

Belief as a Habit

Skepticism quite properly forbids us to speculate beyond the content of our present experience and memory, yet we find it entirely natural to believe much more than that. Hume held that these unjustifiable beliefs can be explained by reference to custom or habit. That's how we learn from experience. When I observe the constant conjunction of events in my experience, I grow accustomed to associating them with each other. (Enquiry V ii) Although many past cases of sunrise do not guarantee the future of nature, my experience of them does get me used to the idea and produces in me an expectation that the sun will rise again tomorrow. I cannot prove that it will, but I feel that it must.

Remember that the association of ideas is a powerful natural process in which separate ideas come to be joined together in the mind. Of course they can be associated with each other by rational means, as they are in the relations of ideas that constitute mathematical knowledge. But even where this is possible, Hume argued, reason is a slow and inefficient guide, while the habits acquired by much repetition can produce a powerful conviction independently of reason. Although the truth of "9 × 12 = 108" can be established rationally in principle, most of us actually learned it by reciting our multiplication tables. In fact, what we call relative probability is, on Hume's view, nothing more than a measure of the strength of conviction produced in us by our experience of regularity.

Our beliefs in matters of fact, then, arise from sentiment or feeling rather than from reason. For Hume, imagination and belief differ only in the degree of conviction with which their objects are anticipated. Although this positive answer may seem disappointing, Hume maintained that custom or habit is the great guide of life and the foundation of all natural science.

Necessary Connection

According to Hume, our belief that events are causally related is a custom or habit acquired by experience: having observed the regularity with which events of particular sorts occur together, we form the association of ideas that produces the habit of expecting the effect whenever we experience the cause. But something is missing from this account: we also believe that the cause somehow produces the effect. Even if this belief is unjustifiable, Hume must offer some explanation for the fact that we do hold it. His technique was to search for the original impression from which our idea of the necessary connection between cause and effect is copied. (Enquiry VII)

The idea does not arise from our objective experience of the events themselves. All we observe is that events of the "cause" type occur nearby and shortly before events of the "effect" type, and that this recurs with a regularity that can be described as a "constant conjunction." Although this pattern of experience does encourage the formation of our habit of expecting the effect to follow the cause, it includes no impression of a necessary connection.

Nor do we acquire this impression (as Locke had supposed) from our own capacity for voluntary motion. Here the objective element of constant conjunction is rarely experienced, since the actions of our minds and bodies do not invariably submit to our voluntary control. And even if volition did always produce the intended movement, Hume argued, that would yield no notion of the connection between them. So there is no impression of causal power here, either.

Still, we do have the idea of a necessary connection, and it must come from somewhere. For a (non-justificatory) explanation, Hume refers us back to the formation of a custom or habit. Our (non-rational) expectation that the effect will follow the cause is accompanied by a strong feeling of conviction, and it is the impression of this feeling that is copied by our concept of a necessary connection between cause and effect. The force of causal necessity is just the strength of our sentiment in anticipating efficacious outcomes.

The Self

In a notorious passage of the Treatise, Hume offered a similar account of the belief in the reality of the self. Here there is the ordinary human supposition that lies behind our use of first-personal pronouns. Upon this relatively simple foundation, philosophers have erected the notion of an immaterial substance, a mind or soul that persists through time on its own. Hume's question is, "From what antecedent impression does the idea of the self arise?"

Hume pointed out that we do not have an impression of the self. No matter how closely I attend to my own experience, no matter how fully I notice the mental operations presently occurring "in my mind," I am never directly aware of "I." What I do experience is a succession of separate and individual ideas, associated with each other by relations of resemblance and causality. Although these relations may be extended through time by memory, there is no evidence of any substantial ground for their coherence. The persistent self and the immortal soul are philosophical fictions.

To suppose otherwise, Hume held, is to commit a category mistake: the self is just a bundle of perceptions, like the railroad cars in a train; to look for a self beyond the ideas would be like looking for a train beyond the cars. Our idea of a persistent self is simply a result of the human habit of attributing continued existence to any collection of associated parts. Like our idea of the necessary connection of cause with effect, belief in our own reality as substantial selves is natural, but unjustifiable.

External World

Another perfectly ordinary feature of human cognition is our belief in the reality of the external world. As I write this lesson, I readily suppose that my fingers are touching a keyboard, that the sun is shining outside and that the radio is playing a Clapton song. In Hume's skeptical philosophy, what is the status of these beliefs?

The primitive human belief, Hume noted, is that we actually see (and hear, etc.) the physical objects themselves. But modern philosophy and science have persuaded us that this is not literally true. According to representationalists, we are directly aware of ideas, which must in turn be causally produced in our minds by external objects. The problem is that on this view we can never know that there really are physical objects that produce our sensory ideas.

We cannot rely on causal reasoning to convince us that there are external objects, Hume argued, since (as we have just seen) such reasoning arises from our observation of a constant conjunction between causes and effects. But according to the representationalist philosophy, we have no direct experience of the presumed cause! If we know objects only by means of ideas, then we cannot use those ideas to establish a causal connection between the things and the objects they are supposed to represent.

In fact, Hume supposed, our belief in the reality of an external world is entirely non-rational. (Enquiry XII i) It cannot be supported either as a relation of ideas or even as a matter of fact. Although it is utterly unjustifiable, however, belief in the external world is natural and unavoidable. We are in the habit of supposing that our ideas have external referents, even though we can have no real evidence for doing so. Representationalism thusly implodes: the ideas, originally introduced as intermediaries between perceivers and things, end up absorbing both, rendering everything but themselves superfluous.

Mitigated Skepticism

Where does this leave us? Hume believed himself to be carrying out the empiricist program with rigorous consistency. Locke honestly proposed the possibility of deriving knowledge from experience, but did not carry it far enough. Bayle and Berkeley noticed further implications. Now Hume has shown that empiricism inevitably leads to an utter and total skepticism.

According to Hume, knowledge of pure mathematics is secure because it rests only on the relations of ideas, without presuming anything about the world. Experimental observations (conducted without any assumption of the existence of material objects) permit us to use our experience in forming useful habits. Any other epistemological effort, especially if it involves the pretense of achieving useful abstract knowledge, is meaningless and unreliable.

The most reasonable position, Hume held, is a "mitigated" skepticism that humbly accepts the limitations of human knowledge while pursuing the legitimate aims of math and science. (Enquiry XII 3) In our non-philosophical moments, of course, we will be thrown back upon the natural beliefs of everyday life, no matter how lacking in rational justification we know them to be.

. . . . .
History of Philosophy
Previous Next

Creative Commons License
The Philosophy Pages by Garth Kemerling are licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
Permissions beyond the scope of this license may be available at

©1997, 2011 Garth Kemerling.
Last modified 12 November 2011.
Questions, comments, and suggestions may be sent to:
the Contact Page.