In his “Of the Origin of Ideas,” from An Enquiry Concerning Human Understanding, Hume claims that thought amounts to no more than the Compounding, Transposing, Augmenting or Diminishing of sensory data (Enquiry, 1977, p. 11). As an example, he points out that while we may never have seen a gold mountain, the thought of one is no more than a combination of “gold” and “mountain”; likewise, a “virtuous horse” can be conceived because, “from our own feeling we can conceive virtue,” and a horse is an animal familiar to us (p. 11). Now the manner in which such objects as “gold,” mountains,” and “horses” are familiar to us poses a certain difficulty to this theory, but for the moment, let us just deal with his argument as it stands.
First, he begins by claiming that all thoughts are due to sensory impressions. This is illustrated with the golden mountain. When he hands us the virtuous horse, however, he has just made a move known as misdirection. If we ask which sensory impression gives rise to the idea of virtue, he tells us that it is a sensory impression of something “from our own feeling.” Calling it a “feeling” rather than an “idea” is disingenuous to the extreme. Virtue, perhaps even more than “horse,” is an idea, and as such is subject to the same rule he applies to all other feelings: which sense gives us the idea of virtue?
What he has done here is to introduce ideas of two quite different types (“virtue” and “horses”) as though they were the same. Having done so, providing that the reader accepts the bait-and-switch, he is then free to continue with his theory.
His next step is to challenge us to come up with a single idea “which is not derived from this source.” With our minds now drawn to concrete images, we find ourselves overwhelmed to discover that everything we can think of fits his criterion. “Smart cows,” “purple people eaters,” and “honest politicians” are all ideas we have never seen, but are derived in their parts from sensory impressions: providing, of course, that we have been preconditioned to accept “smart” and “honest” as objects which can be perceived.
Rather than attempting to meet this challenge all on our own (by mentioning such ideas as “attention,” “continuity,” “silence,” and so on), we will instead let the master meet his own challenge.
In “Against Substance,” from his Treatise of Human Nature, Hume adamantly claims that the idea of substance cannot be derived from sensory impression, and therefore that substance does not exist. In a tone of wonderfully controlled incredulity he says, “If it be convey’d to us by our senses, I ask, which of them; and after what manner? If it be by the palate, a taste; and so of the other senses” (Reality, 1994, p. 102). Now whether or not substance exists is at this point irrelevant: the real question we want to ask is, “If it is impossible to trace the idea of substance back to sensory impression, then doesn’t the idea of substance constitute the proof against his theory of the origin of ideas?” In “Against Substance” Hume asks the reader to point out how this idea of substance could originate from sense, with the obvious implication that it cannot — but in “The Origin of Ideas” he claims that the production of a single idea which is apparently not derived from sense would force him to “produce the impression or lively perception, which corresponds to it” if he were to maintain his doctrine.
But expecting this kind of consistency from Hume is to underestimate his powers of misdirection. In “Origins of Ideas” he insists that we root out incorrect philosophical “terms without any meaning” by asking “from what impression is that supposed idea derived?” Hence, what he is saying is that if we think we have an idea that is not derived from an impression, then the fact that it is not so derived thereby proves that it is not an idea.
This, of course, is the very core of Hume’s genius. First he issues the challenge to produce an idea which does not conform to his criterion, and then he claims that any idea so produced is necessarily excluded from being a true idea because it does not conform to his criterion.
Philosophically this is known as “begging the question” and should be outright condemned. But I am forced, as a fellow magician and deceiver of the public, am forced to admire such an audacious slight.
Earlier I mentioned that there were certain problems inherent even in the “sensory” objects of mountains and horses. To look at this more closely, we should examine what Hume means by the terms “sensory impression” and “perception,” and to distinguish carefully between them (something Hume does only occasionally, as befits his particular point).
Sensory impressions are the simple excitations of the sensory organs produced by a (presumably) exterior cause. Hardness, softness, red, cold, and the like are all sensory impressions. (Actually, as we shall see in a moment, even these are not strictly sense impressions, but simple perceptions derived from sense impressions.) Perceptions, on the other hand, are the coherent groupings of these impressions into perceived objects. We could say that cameras and tape recorders have sensory impressions, while humans and animals have perceptions.
To Hume, our perceptions represent something “added to” the sense impressions, hence a table, chair, or any other object is really a group of sense data that we have, somehow, put together by our imagination into a “thing.” This added value is not itself sense data, nor is it derived from sense data. According to Hume, it is the result of three principles, or laws of thought: Resemblance, Contiguity, and Cause and Effect. whether these are meant to replace, or co-exist with the previously mentioned mental principles of Compounding, Transposing, and Augmenting or Diminishing is not clear. When dealing with Humian rules, it is good practice to bear in mind that when he says “three rules” he’s just kidding. But whatever the actual number of these laws, it is by them that the mind is able to take sense impressions and create ideas.
Or is it?
These laws were introduced as a means to explain the association of ideas, not sense data, and sense data are never ideas. For instance, the sense data for a table might be: “white spot, white spot, hard, cool, white spot, brown spot” and so on. No matter how combined, however, these never form, in and of themselves, an actual table. Sense data does not form ideas, only perception does — but perception can’t occur, in Hume’s world at least, until the sense data have been pulled together somehow.
Here is where we have the problem with mountains and horses. According to Hume, a gold mountain is merely the combination of two ideas, gold and mountain, which are themselves derived from sense data. Sense data, though, is inadequate to form ideas — otherwise we would be inundated by opinions from our cameras and tape recorders.
Let’s go over this ground again to make sure it is clear. One of the problems in articulating anything about sense data is that while perceptions are quite familiar to us, we have never experienced a “sense datum.” In fact, sense impressions are known, or conjectured about, solely by inference. As soon as we begin to speak about the objects around us, we are speaking about perceptions (which is why the “white spot, hard, cool” impressions previously mentioned as forming a table are, themselves, already perceptions and not sense impressions).
Sensory data are abstract concepts of which we know nothing by direct experience, but must instead rely upon deduction and contemporary theories concerning the relationship of the nervous system to experience. Admittedly, someone might object to this by saying, “What? You’ve never seen a red dot?” But any shape or figure of any colour is already a perception, something upon which both Hume and psychologists would find themselves in agreement.
But this poses a problem: pointing to “perceived bits,” is not the same as pointing to “sensory bits,” since the perceived bits are themselves supposedly composed of sensory bits.
In other words, sensory impressions are abstractions that we have deduced — not experienced.
A “group” is not a sensory impression, as we can prove by Hume’s own method: “what colour is a group, what does it sound like, what does it feel like?” and so on. In the same fashion, “juxtaposition,” “distance,” “movement,” “continuity,” “discontinuity,” and a host of others are all excluded from being derived from sensory impressions.
Consider “movement.” It could be argued that surely movement is a sensory impression. But if we are rigorous in speaking only of sensory impressions, and not confusing them with perceptions, then it becomes obvious that this is not so. When I see sensory datum “A” in one location at one moment, and then see sensory datum “A” at another location the next moment, in order to conceive of it as “movement” I have to “identify” both as being the same object, realize that the “location” is different, “remember” where it was the moment before, “compare” the two sightings, and “intend” a perception from it.
Now Hume would possibly answer that identifying the two impressions as being the same is done by the matter of Resemblance. The first objection to this is that Hume proposed his “laws” (three? six? seven?), of which Resemblance is one, for the manipulation of ideas, and that impressions are not ideas — nor can they become so until they have first become perceptions.
The second objection is that even if we attempt to manipulate the two sensory impressions (in this case the two separate locations of the “A” which we see as “movement”) with the law of Resemblance, we must have an idea of what Resemblance it. It is not enough to merely call it a “law” or “principle” of the mind. According to Hume’s sensory theory, we compose an object, such as a table, from a great many sensory impressions of very dissimilar natures. Were they identical, perhaps some kind of law could identify them as such, but it would be a rather useless law since no two sensory impressions are likely that identical. Instead, these dissimilar impressions are grouped together under the basic similarity that they are all part of a table. In other words, the idea of “table,” which is supposed to be the end result of the process of Resemblance, is actually the root of it.
And this is the problem of “mountains” and “horses”: by Hume’s own criterion, these ideas have no meaning.
Having done away with the primacy of all objects, Hume has also effectively done away with cause and effect; but he is nothing if not thorough, and so devotes an entire section to making sure this myth is buried and not praised.
To this end he points out that Adam, even though “his rational faculties be supposed, at the very first, entirely perfect,” could have had no idea that submerging his head in water could quickly and efficiently put an end to his “rational faculties” — perfect or otherwise.
After noting first that we are back to the level of perceived objects, and have left the realm of mere sense impressions, let us quickly agree that what he says of Adam is quite true. The particular effect from any given cause cannot be determined until we have observed it — otherwise we would be born literally knowing everything. On the other hand, our “instinct” to reach out and touch things, even in our earliest hours after birth, testifies to our existent expectation that some kind of effect will result from the cause of a touch. the thing is, such “innate” knowledge is anathema to Hume, and since it’s his ball, we’ll play along.
Hume thus establishes, at least to his own satisfaction, that there is no inherent idea of cause and effect, that it is merely something we pick up from experience as we go along. The problem is, cause and effect are not sense impressions. True, a ball hitting another ball may be a sense impression (although actually it’s a perception, but let’s ignore that for the moment), and it’s true that the second ball moving off after having been hit is a sense impression (with the same caveat); however, and this is a big however, the “idea” of cause and effect is not a sense impression. It is an idea. But it is an idea without a sensory cause.
Which, according to Hume, is not possible.
On the other hand, it is undeniable that this concept of cause and effect is pretty strong. To get around this, he puts it forward as one of his (unknown) number of laws.
With the world, its relations and operations safely out of the way, Hume then allows them back again in a literally meaningless fashion through something called “Custom of Habit.” This Custom is what we gain simply because things have happened in a certain way previously, therefore we develop the habit of expecting them to happen in that way again.
But how could this Custom of Habit be brought about?
We start with nothing but a continual succession of events. “Suppose a person, though endowed with the strongest faculties of reason and reflection, to be brought on a sudden into this world; he would, indeed, immediately observe a continual succession of objects, and one event following another; but he would not be able to discover anything farther.” As we have already seen, these sensory data do not include the concepts of anything other than themselves — not even “succession” or “continual.” From here we — well, from here we have more sensory data, don’t we. We never learn about cause and effect because it is not itself a sensory impression, nor is it innate (unless it is — but he gives only one line to it being innate and an entire section to it being not innate). Since we never can get the idea of cause and effect, and since we cn never get the idea of groups or things or anything else, then we never have any ideas. We eventually die (or not so eventually since not only can we do nothing to keep ourselves alive, but we don’t even know that we should do so) and the “continual succession of events” stops.
Somehow in this process Hume manages to bring in Custom of Habit, but I’m afraid the mirrors are too well hidden on this trick for me to discern them.
With all the world and all aspects of the mind neatly disposed of, there is only one more task for Hume: to get rid of the “self.”
Since all we are is a bundle of perceptions (actually, a bundle of sensory impressions, but again, we’ll let that pass), then Hume declares there can be no “self” except in an illusory sense.
So let’s see where that leaves us.
If there is no self, then there is nothing to focus upon this or that group of sensory impressions and enact the three (or five, or two, or four) laws of thought upon them. Either that, or all the sensory data are equally processed, in which case the mental laws must be applied to all of them equally resulting in a grand total of nothing.
No world. No objects. No cause and effect. No reality. No self.
While it may appear that nothing is left, Hume has one more thing to discredit: the radical sceptic from whose “excessive skepticism [italics his] … no durable good can ever result,” and who, upon awakening from his dream, “will be the first to join in the laugh against himself, and to confess, that all his objections are mere amusement.”
And in this fashion, Hume ends his act with the Amazing Ouroborus trick in which he swallows his own tail.
Hume, David. An Enquiry Concerning Human Understanding. (Indianapolis, Indiana: Hackett Publishing Company, 1977).
Hume, David. A Treatise of Human Nature, from Reality, ed. Carl Levenson and Jonathan Westphal (Indianapolis, Indiana: Hackett Publishing Company, 1994).