Hume’s Enquiry: II-V

In Section II, Hume focuses on what he calls ‘perceptions of the mind’.

These perceptions are divided into two kinds: Ideas and Impressions

Distinction between Ideas and Impressions

The characteristic by which we distinguish between Ideas and Impressions is their relative degree of force and vivacity

  • Less forcible and lively perceptions are thoughts or ideas.
  • More forcible and lively perceptions are impressions:

The other species want a name in our language and in most others, I suppose, because it was not requisite for any but philosophical purposes to rank them under a general term or appellation. Let us, therefore, use a little freedom and call them impressions, employing that word in a sense somewhat different from the usual.

By the term impression, then, I mean all our more lively perceptions when we hear, or see, or feel, or love, or hate, or desire, or will.

And impressions are distinguished from ideas, which are the less lively perceptions of which we are conscious when we reflect on any of those sensations or movements above mentioned. (A & W. p. 539, left column)

Relationship between Ideas and Impressions


Ideas are copies of impressions

Hume provides two arguments that are intended to establish this claim:

First argument:

Part I: analysis of ideas; analyze complex ideas, then find they resolves themselves into simple ideas; these simple ideas are copied from some prior feeling or sentiment. (God = infinitely intelligent, wise, and good being; reflect on our own mind and augment, without limit, the qualities of goodness and wisdom)

Part II: If you want to deny the UNIVERSALITY of the conclusion, produce a counter-example (an idea that is not derived from this source)—we will defend the view by producing the impression which corresponds to it.

Second argument:

Part I: A person who lacks one of the senses, won’t have any of the corresponding ideas (blind people don’t have ideas of colors, deaf people don’t have ideas of sounds).

Part II: Respond to possible counter-example—form an idea of x through comparison of impressions of things resembling x; have the idea of x, but it is not copied from an impression of x. Dismiss the counter-example and claim it does not force us to alter our GENERAL maxim.

With these claims on board, Hume then articulates a maxim or a principle for philosophical investigations like those he will carry out in the Enquiry.

When we entertain, therefore, any suspicion that a philosophical term is employed without any meaning or idea (as is but too frequent), we need but inquire, From what impression is that supposed idea derived? And if it is impossible to assign any, this will serve to confirm our suspicion. By bringing ideas into so clear a light, we may reasonably hope to remove all dispute which may arise concerning their nature and reality. (A & W, pp. 540-41)

The next thing Hume brings to our attention is the fact that ideas in the mind are connected in somewhat regular ways. They do not simply exist as individual ideas, but they also exists as members of chains of connected ideas that lead our thinking from one idea to another in predicatble and methodical ways. This association of ideas is something that Hume is interested in investigating: what are the particular ways in which ideas are associated with one another, and how do they get associated with one another in these particular ways?

Hume identifies three principles of association of ideas:

1. Resemblance (a picture leads to the idea of the original)
2. Contiguity (the idea of a part leads to the idea of other parts)
3. Cause or effect (the idea of bodily harm leads to the idea of pain)

It is this last kind of association that will be the main point of focus in Hume’s investigations into the human understanding, since it is the one that we make use of in reasoning to conclusions about the existence and characteristics of things and events that we do not immediately perceive.

General View of the Understanding

Humes general view of the understanding is outlined in the the two part section where he presents skeptical doubts concerning its use and a skeptical resolution of these doubts. The basic theses involved are:

  1. All reasoning about Matters of Fact is based on the relation of Cause and Effect
  2. Particular Knowledge of Cause and Effect is not arrived at through Reason but through Experience
  3. The fundamental basis for Inferences from Experience is not Reason, but Custom or Habit.

In other words, reasoning about what we do not directly experience is a process that is based on a general understanding of things that we have gained through experience. Experience teaches us about ‘how the world is set up’ and, once we have learned that sufficiently, we can use our understanding of the general order of things to ‘fill in the gaps’ left open by our immediate perception.  Experience provides the associations between ideas that we make use of when we follow out our chains of thinking and arrive at some conclusion about existing things.

If we think carefully about this mode of operation, however, we should realize that it rests on a rather significant presupposition; namely, the presupposition that the future will resemble the past, or, more generally, the presuppostion that what we do not experience is like what we do experience in ways that are sufficient for us to rely on the information provided by experience in drawing conclusions about things that are beyond our experience:

  • Under the presupposition that i) tomorrow will be relevantly similar to yesterday and all the days I have experienced prior to then, and ii) the sun has risen on each of these days: it is reasonable for me to conclude that the sun will rise tomorrow
  • Under the presupposition that i) the next swan I observe in nature will be relevantly similar to the other swans I have observed previously, and ii) the others I have observed have all been white: it is reasonable for me to conclude that the next swan I observe will be white

So, in addition to presupposing certain features of my previous experience (i.e., the sun always rising, swans always being white), I also have to presuppose a general principle about the similarities between known instances and (as yet) unknown instances:

This is a principle that I cannot possibly know through experience, since all my experience deals with known instances. I don’t have the access to unknown instances that would be necessary to establish that they are similar enough to known instances for me to make use of what I have learned from experience to reason correctly to conclusions about them.

This is a principle that I cannot possibly know a priori, since all a priori reasoning, according to Hume, concerns relations of ideas. The principle I am appealing to is a claim concerning a matter of fact (i.e., either the unknowns do, in fact, resemble the knowns or they do not, in fact, resemble the knowns).

I cannot directly perceive the truth of this general principle, and I cannot reason about it a priori, so the only way to decide the principle is true, and thus reliable, would be to reason about it on the basis of experience… but, what is the general principle we make use of every time we reason on the basis of experience? Oh, yeah, it is the very principle that is in question here…

This presents a problem for us, in Hume’s view, only if we insist that prior certainty concerning the truth of the principle is required for us to trust our use of it. That is, if we accept the Cartesian view that we should not trust our faculties unless we have independent verification of their reliability in leading us to truth, we might conclude that we cannot trust our reasoning about matters of fact at all. The more natural response, however, would be to conclude that we cannot trust our faculties completely, so we cannot trust the results of our reasoning about matters of fact completely, but that does not mean that all trust in them is misguided. We trust our understanding, not because we have independent verififcation of its reliability, but because it is all that we have to appeal to in circumstances like these.

Recognizing that our use of the understanding in reasoning about matters of fact is ultimately grounded in customory or habitual beliefs about the regularity of nature (that are themselves products of that regularity), rather than in rational insight concerning the necessary order of nature, is, for Hume, a way of humbling our pretensions and reminding ourselves that there are natural limits to the exercise of our capacities.

More Details about this General View

  • Reasoning about Relations of Ideas and Matters of Fact

Relations of Ideas = Cartesian Immutable Natures (logically necessary truths)


  • Claims in Geometry, Algebra, Arithmetic
  • Intuitively Certain (triangles have three angles and three sides–denial is a direct contradiction)
  • Demonstrably Certain (square of the hypotenuse is equal to the square of the two sides–follows from a chain of logical inferences that makes it such that denying the conclusion creates a contradiction with some step of the proof)
  • Discoverable by thought alone; i.e., without reference to any existing thing

Matters of Fact = Claims about Existing Things (logically contingent truths)

  • Claims in Astronomy, Physics, History
  • Not intuitively Certain (the sun will rise tomorrow–denying this claim may lead me to say something that is, in fact, false, but it doesn’t lead me to say something that is logically inconsistent. The truth or falsity of the claim depends on matters of fact concerning the sun and the earth and the future)

If we are going to treat it as being demonstrably certain, we can’t just start with things that are intuitively certain (e,g, spheres have these qualities, elliptical paths are described by this function, etc.) All of these mathematical truths hold independently of any existing things, so they can’t justify claims like ‘the earth spins on its axis while it traces an elliptical path around the sun’–these are all claims about matters of fact–not relations of ideas.
So, if we are going to have any kind of certainty concerning claims like this one, we have to discover another source for it.
We have sense experience and memory as sources of intuitive certainty concerning matters of fact. But, the claims of Astronomy, Physics, and History take us a long way away from our own sense experience and our own memories.
What is the evidence that assures of any matter of fact that we have not directly been witness to?

  • All reasoning about Matters of Fact is based on the relation of Cause and Effect

Intuitive knowledge about matters of fact is based on direct experience and memory.
All other knowledge about matters of fact is based on reasoning from matters of fact that we take to provide evidence for other matters of fact
A letter postmarked from Paris is taken as evidence that our friend is in Paris (though we are not there to see her, and to establish that on our own)
A watch on the beach is taken as evidence that some humans have been there before us (though we were not there to see them)
Hearing sounds that we can make sense of as linguistic expressions is taken as evidence for the presence of someone who is talking (though we don’t currently see that person)

Whenever we assent to the truth of claims about matters of fact that we do not directly experience, we do this on the basis of evidence following the formula:
y exists, and must have some cause x, which therefore also must exist and have the capacity to produce y.
(Sound familiar? Descartes? I exist with the idea of God, and must have some cause, who exists and is capable of producing me with this idea)
But, where do we get the knowledge of what particular things cause what particular other things that we have to have in order to plug in particular values for y and x here?
Why don’t I:
See a letter postmarked from Paris and take it as evidence that my dog is a terrier
See a watch on the beach and take it as evidence that watches grow in the sand around here
Hear sounds that I take to be linguistic expressions and conclude that tomorrow is Tuesday

  • Particular Knowledge of Cause and Effect is not arrived at through Reason but through Experience

Our knowledge of what particular things in nature cause other particular things in nature is a product of our experience with particular things in nature.
We can’t simply take some natural object, look at it carefully, and deduce a whole bunch of things about what effects it has.

We come to an understanding of how to connect particular causes and particular effects through repeated experience with things of similar kinds–the effect of the motion and impact of billiard balls is not something I can anticipate purely through reason–it is something that I know about through having played billiards and noting what happens when one ball impacts another.

I don’t know that people make watches just by knowing what a person is and deducing it from there. I have experience that leads me to understand what watches are and that and how they are made by people.

  • The fundamental basis for Inferences from Experience is not Reason, but Custom or Habit.

Inferences from experience rely on accepting a fundamental principle concerning the relation between past experience and future experience:
In my experience x’s have always caused y’s

Therefore, x’s will continue to cause y’s
(The future will resemble the past).

Okay, so let’s say this is actually what we assume whenever we are reasoning about matters of fact.

  • How do we justify this assumption?
  • Can we demonstrate that it is true?

Any attempt to demonstrate that it is true will have to beg the question (or make use of the very principle that it is supposed to prove).

  • Whether or not the future will, as a matter of fact resemble the past (or more generally, whether or not what we have not experienced is, as a matter of fact, relevantly similar to what we have experienced) is something that we cannot know through direct experience.
  • If we form any beliefs about such matters of fact, we will have to do so through a process of reasoning.
  • Processes of reasoning concerning matters of fact depend on knowledge of causes and effects
  • We gain knowledge of causes and effects through experience
  • Relying on experience for our reasoning requires us to accept that the future will resemble the past…

We cannot demonstrate that this principle the future will resemble the past (or unobserved objects/events are relevantly similar to observed objects/events) is true.

This means that one way of justifying it is not open to us.

Another way of justifying it might be open to us, however.

Justification for some belief does not always require some prior and complete assurance that the belief is true. That might be what we would like to have in our justifications. That might be what we mean by a wholly rational justification. But there are clearly ways of justifying beliefs that are not absolute in this sense (i.e., prior and complete).

Isn’t it also a certain form of justification to show the belief in question is one that we cannot function without? Assuming we do have some real interest in functioning (if we don’t the argument might not work for us, but who is going to be impressed by our picking this particular nit?), aren’t we to some degree justified in believing the future will resemble the past?