[There is a] long conceptual chain that starts from simple, ostensive definitions and rises to higher and still higher concepts, forming a hierarchical structure of knowledge so complex that no electronic computer could approach it. It is by means of such chains that man has to acquire and retain his knowledge of reality. [RM, 18]
To know the exact meaning of the concepts one is using, one must know their correct definitions, one must be able to retrace the specific (logical, not chronological) steps by which they were formed, and one must be able to demonstrate their connection to their base in perceptual reality. [IOTE, 50]
That some concepts are "wider" than others — that animal, for example, is wider than mammal, and mammal wider than deer — is something so trivial that hardly anyone has bothered making a fuss about it before Rand. But the way some Objectivists talk about the hierarchy of knowledge, you would think that only Rand noticed it, while everyone else is in denial that concepts have any such structure. "Knowledge is hierarchal," Rand's disciples keep insisting; to which the obvious retort is, "So what!" The problem with Rand's hierarchy of knowledge is not that it is wrong but that Objectivists exaggerate its importance.
Yet there is more to the hierarchy of knowledge than just its mere triviality. In the first place, technically speaking, it's not a hierarchy of knowledge, but a hierarchy of concepts. No concept, in and of itself, can tell us anything about reality. It's only when we assert something about that concept within the confines of a proposition that the adventure of knowledge begins. Even if we ignore this objection, and interpret the hierarchy of knowledge merely to mean a hierarchy of concepts, Objectivists are nonetheless guilty of drawing dubious inferences from the notion. Perhaps the most widespread dubious assumption has to do with the idea of what Objectivists call "reduction." Objectivists claim that high-level abstractions can be connected to reality by traveling "backward" through the hierarchy of concepts, thereby identifying the "logical sequence" by which a concept is related to perceptual data.
Here we find in the very bowels of the Objectivist epistemology a notion strongly reminiscent of Humean empiricism. In his Inquiry on Human Knowledge, Hume wrote:
If I ask why you believe any particular matter of fact ... , you must tell me some reason; and this reason will be some other fact, connected with it. But as you cannot proceed after this manner, in infinitum, you must at last terminate in some fact, which is present to your memory or sense; or must allow that your belief is entirely without foundation.
Hume is here interested in assertions about matters of fact. He is proposing a test for judging the accuracy or validity of such assertions. Objectivists are suggesting a test for judging the factual basis of concepts. As I have explained in earlier posts, concepts cannot be tested this way, since concepts, by themselves, don't say anything about facts: only propositions accomplish this task. But if we ignore this objection, it would seem that Objectivists are attempting something along the same lines as Hume. They seek a method of connecting their knowledge to reality. And since they are under the illusion that concepts (rather than propositions) constitute knowledge, they propose this reduction of high-level concepts "backwards" through the hierarchy of knowledge to the empirical knowledge at the bottom of the hierarchy.
There are two main problems with this. In the first place, unlike Hume, Objectivists are not really interested in fact checking their empirical claims. Objectivists may talk about traveling backward through the hierarchy, but that's not an intinerary Objectivists are likely to follow much in practice. Rand's brief foray into empiricism is mere lip service.
It's just as well that Randian empiricism is little more than an empty gesture, rather than a rooted habit of thought. For it turns out that Humean empiricism, however noble in its intention, is not entirely workable as a practical method. It was decisively refuted by Karl Popper in his article "Knowledge without Authority." In that article Popper notes that "most of our assertions are not based on observations [i.e., Hume's sense data, or Rand's "perceptual data"] but upon all kinds of other sources." And Popper goes on to show that the attempt to trace knowledge "backwards" to its empirical source is a fool's errand. If you went about questioning other people about the empirical sources of their knowledge, "you would in fact never arrive at all those observations ... the existence of which the empiricist believes." The chief problem with this sort of crude empiricism is that it assumes the existence of pure, unadultered observation which takes in nothing but the facts. Observation, however, does not work that way. As Popper notes:
Every [observer] must always make ample use, in his report, of his knowledge of persons, places, things, linguistic usages, social conventions, and so on. He cannot rely merely upon his eyes or ears, especially if his report is to be of use in justifying any assertion worth justifying. But this fact must always raise new questions as to the sources of those elements of his knowledge which are not immediately observational.
This is why the programme of tracing all knowledge to its ultimate source in observation is logically impossible to carry through: it leads to an infinite regress.
Although Rand probably never considered these objections to her empirical reductivism, some Objectivists might believe that, fortuitously, she escapes Popper's criticism by her differentiation between data and percepts:
A “perception” is a group of sensations automatically retained and integrated by the brain of a living organism, which gives it the ability to be aware, not of single stimuli, but of entities, of things. An animal is guided, not merely by immediate sensations, but by percepts. [VOS, 19]
The problem here is that Rand does not go far enough. She gratuiously assumes that sense data are mere undiscriminated stimuli, as if when Hume talked of checking one's facts via the evidence of the senses, he really meant checking them on the basis of mere stimuli. Hume meant no such thing. He had in mind something very similar to Rand's percepts, although he didn't devise any theory about how percepts arise out of stimuli. Nor are Rand's percepts quite the heuristic products that Popper has in mind. There may be some interpretion involved in Randian percepts, at least in terms of distinguishing objects, entitites, things, etc. But it clearly involves nothing as sophisticated as knowledge of persons, places, or things, let alone knowledge of linguistic usage or social conventions assumed by Popper. Rand's percepts are integrations; interpretation is too strong a word to describe them. After all, the Randian percept is identified with the "given" and the "self-evident." In short, Rand's perceptual knowledge is just the sort of pure observation that Hume had in mind when he talked about sense data. Therefore, Popper's criticism applies to Randian empiricism as well as the Humean variety. Tracing back one's concepts through the hierarchy to the perceptual data they are allegedly tied to is not an adequate test for claims about matters of fact. What might be an adequate test? Popper proposes the following:
[I]f we are doubtful about an assertion, then the normal procedure is to test it, rather than ask for its sources; and if we find independent corroboration, than we shall often accept the assertion without bothering at all about sources.