Critics have often ignored Rand's epistemology as well. The fact is, it's Rand's ethics and politics that stirs up the animals on both sides, pro and con. Those critics that have tried to analyze the Objectivist epistemology have either gotten lost in the thickets or have become consumed by purely technical issues that most people don't care about. For me, Rand's epistemology could be reduced to two salient points: a denial (or at least mis-characterization) of the unconscious, intuitive phases of human thought; and the insistence that every word has an "objectively correct" definition. Those are the most important, or at least the most relevant, points of Rand's epistemology. By importance I mean: they are the most fundamental to what Rand was trying to accomplish in her overall philosophy. Admittedly, this is not obvious at first glance, so some explanation is in order.
Critical to Objectivism --- in some sense the reason why its progenitor regarded it as an "objective" system of philosophy --- is that it holds that every aspect of human experience (or at least everything that can pass for "knowledge") is reducible to "objective" conscious deliberative reasoning. Rand was kind of anal about this. She was deathly afraid of anything that smacked of the "arbitrary." She wanted the right to claim that everything she believed, including her moral and political preferences, constituted a "valid" form of knowledge which had been reached entirely through consciously directed "reason." Knowledge attained by unconscious thinking processes (such as "intuition" and "judgment") she regarded as "invalid" --- a horror show of the "arbitrary."
This is one of those views that sounds plausible but is in point of fact wrong. Now when I was writing Ayn Rand Contra Human Nature, I could not find an easily explicable scientific source that proved the validity of "intuition" (although many such books would appear after ARCHN was published). So for the most part I let the issue drop, opting instead attack Rand's epistemology from a Popperian point of view. Popper is a wonderful philosopher, very rational and clear-headed, and a pleasure to read. He is far better philosopher than Rand. But I choose to attack Rand using Popper's critical rationalism not because I was a died-in-the-wool Popperian (I was no such thing), but because I believed the clarity of his thought would make for a more accessible book. My deeper sympathies were with the philosophy of Michael Polanyi, as presented in his book Personal Knowledge. The trouble with Polanyi is that his book is one of the most difficult I have ever read, and I did not think I would ever be able to use its insights to refute Rand's deeply flawed ideals of "objectivity" in way that the normal intelligent reader could understand. This meant the subject of intuition verses "reason" and faux-objectivity would not be covered with any depth in Ayn Rand Contra Human Nature.
Popper might have been useful in refuting Rand's insistence on the importance of defining one's terms, but I have found that partisans of that view are indifferent to the arguments against definition. Any emphasis on the importance of definitions raises in insoluble problem. Words must be defined in terms of other words. If I am asked, for example, to define what philosophy is, and I claim "it's the study of wisdom," must I not also define the words used in that definition? There's the rub, because of course we are now in the presence of an infinite regress. The insistence on defining all one's terms is a false ideal of knowledge, as is the belief that all words have a "corrrect" definition.
As I delved more deeper into these issues, I was drawn into the philosophy of American critical realism, particularly as it was developed by George Santayana and Arthur Lovejoy. These forms of critical realism attempted to establish the representationalism of Locke and Kant on a surer footing. Now Rand despised representationalism because she associated it with Kant (whom she never read nor understood). Within a realist ontology there is only two alternatives: either a "direct" realism or an "indirect" realism (i.e., representationalism). Rand was attracted to direct realism --- it seemed to her more objective --- but she seems to have realized, in a rare moment of insight, that a direct realism wouldn't work with conceptual knowledge because concepts cannot be conceived as literal copies of existents. There exists two opposing strains of thought in the Objectivist epistemology. Underlying Rand's so-called "solution" to the problem of universals is the implicit assumption that the highest ideal of knowledge is literalism. This is something that Rand held unconsciously, without fully realizing it; we only know that she was tacit literalist because her solution to the problem of universals is literalist all the way through. She believed she had solved this problem, not by presenting scientifically accumulated evidence, nor by abstruse logical argument, but rather by merely asserting how concepts are connected to reality (i.e., "a concept is a mental integration of two or more units possessing the same distinguishing characteristic(s), with their particular measurements omitted"). This is a literalist "solution" to the problem. Rand is tacitly explaining what it is exactly and literally that concepts refer to in reality.
The other strain in Rand's epistemology derives from what she called "unit-economy." "In any given moment, concepts enable man to hold in the focus of his conscious awareness much more than his purely perceptual capacity would permit," she wrote. That is essentially a principle of representionalism. The fact that a concept can represent far more than can be perceived in a given instant indicates the essentially symbolic (and therefore representational) nature of concepts. Concepts are items of description: they are not literal copies of their referents. Hence Rand's solution to the problem of universals contradicts her principle of unit-economy. At its core, Rand's epistemology is a confused mish-mash of literalism and representationalism, of direct and indirect realism.
This became apparent to me as soon as I grasped Santayana's notion of essence. Santayana posits that the mind perceives the world through a veil of what are essentially Platonic ideas --- what he calls "essences." Our conception of the world cannot be literal or direct because the world is out of scale with our minds (as Rand herself acknowledges). Santayana's essences, which are essentially his version of "concepts," are the building blocks of knowledge, but they are not knowledge when regarded in of themselves. The realm of essence includes all possible meanings, which are infinite in scope (they include all numbers, which are infinite). However, not all these meanings represent something in reality. Does every number, the entire infinity of them, represent something in the external world? What about unicorns or "honest politician"? By making knowledge the principle unit of knowledge, Rand created all kinds of artificial problems. She has to regard "imaginary" concepts, such as jaberwocky or 20 gazillion, as "invalid" because they're not connected to anything in reality. How much sense does that make? But it's actually worse than that. If you understand all the implications of Santayana's notion of essence, you will realize that Rand's approach to concepts --- i.e., her mania for "validating" concepts by proving they're "connected" to reality --- is all wrong. Concepts are symbolic all the way through --- they are scraps of meaning, not pieces of knowledge. They can be used to describe not merely reality, but fantasy, fiction, lies, counter-factuals, prognostications of the future, unproven hypotheses, and myth. They are not somehow "invalid" when not used to describe reals.
Santayana's philosophy also helped me understand what was wrong with the Objectivist metaphysics. Rand wanted to build a true "system" of philosophy starting with first principles, but she had no idea what she was getting herself into. She introduced her three "axioms" without fulling comprehending what that meant. An axiom is postulate that serves as a starting point for further arguments. Any system of thought based on axioms is a self-enclosed logical system, like geometry or mathematics. Ludwig von Mises' "praxelogy," in which he tried to deduce all of economics from the premise of human action, is such a system. Spinoza tried something along those lines in this Ethics. If Rand was serious about deriving a systematic philosophy from axioms, she would have needed to deduce her entire philosophy, or at least the main bulk of it, from her three axioms. She of course did no such thing. In fact, it would have been logically invalid for her to attempt anything along those lines, for the simple reason that two of her axioms, "existence exists" and "A is A," are tautologies, and nothing can be deduced from tautologies. David Kelley has admitted as much --- although he still tries to defend the axioms. But they are not really defensible. They are another example of Rand's philosophical illiteracy. She had no idea what she was doing, and she made a mess of it.
Consider what she did with the little mantra "A is A." She originally picked up the phrase from Isabel Paterson, who was fond of using it to emphasize the importance of logical thinking. Rand decided that she could turn it into foundational principle of her philosophy --- that is, her so-called "law of identity." But once again she is confused beyond all help. She didn't understand that not all forms of identity are the same. "A is A" is simply the identity of one essence or idea with itself. Any symbol of consciousness, taken as distinct unit, is identical to itself. Rand confuses this very simple and non-empirical form of identity with other types of identity, particularly the identity of existing things and their attributes. These forms of identity are potentially empirical and as such are not self-evident. Identifying the existence of an object, such as a dog or lamppost, is not A is A, but A is. Identification of attributes involves predication. Instead of A is A, predication posits A is B, as in the rose is red. Again, it's not self-evident that roses are red. It's not something that cannot be determined a priori, but requires doing some empirical leg work.
Such was my intellectual journey that enabled me to pinpoint Rand's main epistemological and metaphysical errors. With the issues of human nature, historical change, philosophical literacy, and the relation between concepts and reality well in hand, the only challenge left for me to figure out was how to approach the most notorious parts of the Objectivist philosophy, the ethics and the politics.