Hume draws a distinction between impressions and thoughts or ideas (for the sake of consistency, we will refer only to "ideas" from here on). Impressions are lively and vivid perceptions, while ideas are drawn from memory or the imagination and are thus less lively and vivid. Impressions comprehend, according to Hume, "all our more lively perceptions, when we hear, or see, or feel, or love, or hate, or desire, or will." Thus, both the color red and the feeling of anger are considered impressions. Ideas are what arise when we reflect upon our impressions, so the memory of seeing the color red or a thought about anger are considered ideas.
While we might consider the human mind an unlimited organ, able to conceive of imaginary creatures and far-off lands with great facility, Hume points out that our imagination in fact consists merely of a complex of ideas. For instance, if we imagine a gold mountain, we are compounding our idea of gold with our idea of a mountain. Hume provides two arguments to support this claim. First, he suggests that all complex ideas are compounded out of simple ideas, which are in turn derived from simple impressions. For instance, our idea of God as supremely good and intelligent comes from taking our simple ideas of human goodness and intelligence and augmenting them without limit. Second, he points out that our imagination is limited to those ideas of which we have had impressions. Thus, a blind man is unable to imagine colors, a deaf man to imagine sounds, or a mild-mannered man to imagine cruelty.
Hume admits one objection to his distinction. He points out that I can imagine certain colors without ever having perceived them. For instance, if I have seen several shades of blue, I might be able to imagine some other shade of blue that falls between them. Though he has no answer to this objection, he remarks that the counter-example is so singular that is does not upset his general maxim.
This distinction between impressions and ideas is valuable for clearing up our philosophical vocabulary. While ideas are faint, obscure, and easily confounded with other ideas, impressions are vivid and clearly defined, and we are not likely to fall into error with respect to them. Thus, when we find ourselves discussing a philosophical term that we suspect may not refer to any idea, we may simply ask from what impression its supposed idea might be derived. Since all ideas are derived from impressions, a term that is not connected to any impression is meaningless.
In a footnote, Hume notes that his distinction between impressions and ideas clears up some confusions found in Locke's rejection of innate ideas. Hume complains that Locke fails to clarify what he means either by "innate" or "idea." In Hume's vocabulary, we could assert that impressions are innate and ideas are not.
In section III, Hume discusses the connections that exist between ideas, asserting that all ideas are linked to other ideas. Hume lays out three principles by which ideas might be associated: resemblance (where a picture of a tree might make us think of the tree), contiguity in time or place (where mention of one apartment might lead us to discuss others), and cause and effect (where the thought of a wound makes us think of the pain that follows from it). Hume admits that he has no reason for laying out only these three principles except that he cannot think of any others that would be needed. For instance, association by means of contrast or contrariety can be seen as a combination of resemblance and causation.
Here we begin to see Hume's empirical method at work. By looking inward and observing his own mental processes, Hume brings to bear three important distinctions. The first, and most important, is the distinction between ideas and impressions. This distinction is original to Hume and solves a number of difficulties encountered by Locke. A proper discussion of Hume's footnote would take us too far afield, but we should remark that Hume's criticism of Locke is exact and powerful. The distinction between impressions and ideas might seem quite obvious and of no great importance, but Hume is quite clever to identify the full importance of this distinction. An empirical philosophy asserts that all knowledge comes from experience. For Hume, this would suggest that all knowledge comes from impressions, and so ideas are set up as secondary to impressions.
The second distinction, between complex and simple impressions or ideas, helps draw out further the power of the first distinction. A simple impression might be seeing the color red, while a complex impression might be seeing the totality of what I see right now. A simple idea might be the memory of being angry while a complex idea might be the idea of a unicorn (composed of the idea of a horse and the idea of a horn). Complex ideas and impressions are compounded out of the simple ones.
With these first two distinctions, Hume is creating a hierarchy of mental phenomena. Since the complex is compounded out of the simple and ideas are derived from impressions, everything in our mind is based ultimately upon simple impressions. A complex idea is compounded out of several simple ideas, which are in turn derived from several corresponding simple impressions. Hume thus suggests that a term can only be meaningful if it can be connected with an idea that we can associate with some simple impressions. Hume, we should note, is silently implying that every term must be connected with some idea. In the eighteenth century the philosophy of language had not yet flourished, and it was not clear how difficult it might be to determine precisely how words, ideas, and reality link up. Hume's suggestion that all terms can be analyzed into simple impressions anticipates Russell, who argues that we can analyze all terms into simple demonstratives like "this" or "that." Hume's suggestion comprehends a picture of language according to which the words we use are a complex and opaque expression of a simpler underlying language which proper analysis can bring out.
The third distinction is the three laws of association. If the previous two distinctions give us a geography of the mind, describing its different faculties, this distinction gives us a dynamics of the mind, explaining its movement. According to Hume, any given thought is somehow related to adjacent thoughts just as any given movement in the physical world is somehow related to adjacent moving bodies. His three laws of association, then, might be seen as equivalent to Newton's three laws of motion. With them, Hume hopes to have described fully the dynamics of the mind.
There are a number of objections we might want to raise to Hume's distinctions and the way they are introduced, but we will touch on only a few briefly. First, we might ask how strictly we can distinguish between impressions. Hume argues that ideas can be vague, but that impressions are exact and that the boundaries between them are clearly defined. Is the boundary between the impression of a 57" stick and a 58" stick that clearly defined? There is some level of vagueness in our impressions that Hume does not acknowledge. We could also point out that while we are experienced in distinguishing colors, we are not so good with some other sensations. For instance, we often have trouble distinguishing between tastes.
Second, we might object to Hume's implicit philosophy of language. It seems closely linked to the idea that simple impressions are clearly defined and infallible. It is far from clear, however, why it should be desirable or possible to reduce all our language to simple impressions. What, we might ask, is the simple impression from which is derived the word "sake," for example?
Third, we might ask Hume to be clearer in his distinctions. For instance, are dream images impressions or ideas? Most likely they are ideas, since they consist of a mixture of imagination and memory. However, dreams are (arguably) phenomenally indistinguishable from waking experience: we cannot prove that we are dreaming from within a dream. Thus, all our impressions from within a dream are as real to us as we dream them as waking impressions are to us when we experience them.