Shayne Wissler
“… to understand is, above all, to unify.” – Albert Camus

A review of Harry Binswanger’s “How We Know: Epistemology on an Objectivist Foundation”

April 16 2014

The review below is also published at Amazon.


There are a number of standards one could use to review a book like this. For example, one might review it in terms of how well it corresponds to or builds on Ayn Rand’s philosophy. But I’ll be reviewing it relative to how well it corresponds to the truth.

The book has obviously taken many years of thought and effort, and it represents the “state of the art” in Objectivist epistemology, so regardless of how true it is, it is at least very noteworthy.


1. FOUNDATIONS

I mostly agreed with this chapter. As human beings we don’t have mystical access to knowledge, for us, all information we receive must come by causal means (i.e. sense data which reason processes.) Leaving aside relative quibbles, I generally agree with him regarding axioms – to dispute that we can know anything via the senses is to be absurd and self-refuting.


2. PERCEPTION

It is worth identifying the basic issue here: we all know that some things are as they appear, and some things are not. The question naturally arises: what is different in each case, and how can we reliably distinguish them? This is obviously a very important philosophic question. We also know the difference between what we might call “raw sense data” and sense-perception: sense data is the discrete units of information coming in from each (say) individual cell of the retina. Sense-perception on the other hand is the awareness of an external object, produced by some kind of processing of this sense data.

Summary of this chapter: Ayn Rand said that sense-perception was “the given,” (i.e. infallible) so it is. To underscore what is being asserted here: Binswanger is not merely claiming that the senses must be regarded as valid in a basic axiomatic sense, but that sense-perception always and in each particular case gives us infallible information.

The first thing one should think of when hearing a claim like this is the phenomenon of illusion, and Binswanger mentions a few, but he has a poverty of examples, and his examples are cherry-picked to make it appear as if perception is indeed infallible. What, after all, could be more self-serving in this connection than to pick the “stick appears bent in water” example, which demonstrates only that light is governed by the laws of physics? Why does he not pick examples that underscore actually interesting idiosyncrasies of the faculty of perception as such? E.g., consider the beautiful/ugly woman illusion, which reveals an interesting hysteresis phenomenon in your faculty of perception. Are we to regard the contradictory modes of perception as both being “infallible”? Binswanger completely ignores this sort of example, where it seems to be that the perceptual mechanism itself can be made to switch from one “interpretation” of the sense data to another, and at will.

The fact that we can paint accurate portraits proves that no perception is a “given”. We can take any visual perception and break it down into its various sensory elements, to whatever level of detail we wish. We can therefore analyze our perceptions, we don’t need to take them as “givens”, and in fact, we must not – not if we wish to be objective. (Take an art class, and one of the first things they tell you is to try not painting what you perceive, but what you actually see, i.e., you need to isolate the units of perception, the sensations.) We must always be open to verifying whether any perception of ours corresponds to reality or not. We must take responsibility for our sense-perceptions, just as we do for our emotions, especially since they are automatic integrations. This means not only being open to decompose a perception into various sense data but in looking at things from various points of view. (Binswanger would perhaps not be opposed to this process, but he would refuse to revise his abstract philosophical descriptions accordingly.)

Binswanger rigidly clings to a view of subconscious awareness referred to as “bottom up”. Yet, given phenomenon such as the ugly/pretty woman illusion, most psychologists recognize that there are both bottom-up and top-down aspects to sense-perception (in the ugly/pretty woman illusion, it is clear that our decision about how to look at the image changes the perception.) Rather than making an extended argument for the fact that perception isn’t merely a bottom up process, but is a synthesis of both “top” and “bottom” aspects, I will point out the philosophic issue, one that is not disputed by psychologists: that raw sense input (e.g. from individual cells in the retina) is the beginning of the causal chain that results in a percept. In other words, it is clear that, viewing the human being as a system that begins with discrete stimuli (such as light falling on the individual cells of the retina), we are able to somehow integrate what starts as nothing more than mere “pixels” of information an experience of real entities.

Contrary to Rand, it is not a “scientific” fact (i.e., irrelevant to philosophy) that we can in principle isolate sensations from percepts, it is a fact open to anyone who wishes to hold their thumb to a visible object and introspect on how a “perceptual given” can be analyzed into smaller elements of color – i.e., we don’t need to know that we have a retina or individual cells that receive and process light, we can directly observe from a partially blocked image that we can receive and can consider each portion of light independently. This is an important observation for philosophy, it sheds at least some light on the problem of appearance vs. reality and how it should be resolved.

The stubborn insistence that discrete sensations are not available to us leads to a dogmatic psychology analogous to that of a religious zealot: just as he takes his beliefs on faith as “a given” (from God), the Objectivist is likewise taught to take perceptions as “givens.” By itself, this error might not be so harmful, but to encourage an unquestioning habit of thought in one area can only worsen habits of thought in other areas.

What Binswanger’s view amounts to is an overreaction: the skeptic puts forth the example of how we can experience perceptual illusions as an argument against knowledge, and Binswanger thinks that by arguing that sense-perception is immune from error he has defended his view from skepticism, but because he is being absurd, he has only given the skeptic encouragement.

Ayn Rand was vague on exactly what she meant by perception, and I suspect that Binswanger himself would admit that he is interpreting Rand and could have made a mistake. However, in my experience, most Objectivists interpret her in the same way as Binswanger.


3. CONCEPT FORMATION

Binswanger begins with a very poor assumption: “To gain proper guidance in conceptualization, we must first know what concepts are. And to know that, we must understand how concepts are formed.” This is precisely the reverse of the truth: if we don’t know what they are, we can’t study them. Binswanger does implicitly indicate what he thinks they are: something that corresponds to the terms of a proposition. Beyond this, he gives no definition before plunging into the mechanics. The reader is left with no idea of what Binswanger thinks he is talking about, beyond what the reader himself might think he should be talking about.

Things get worse from there.

The natural assumption of a freethinker is that a concept can mean what we want, and that so long as we don’t contradict ourselves and can identify what we mean by a given word, then we have a concept. On its face, it is presumptuous to tell another person that, not only should he not mean that (and a rational person would of course always be open to criticism on that point), but he can’t mean it, as if concepts are only concepts if they fit the Objectivist notion of them. This sort of stipulation brings to mind a book review by Dorothy Parker: “This is not a novel to be tossed aside lightly. It should be thrown with great force.”

Does Binswanger address this natural stance? Perhaps he thinks he did, under the nominalist heading (the principled form of which he calls “truly a lunatic theory”), but the more he describes nominalism the less it looks like what I’m talking about here. So, while he gives only three major schools of thought on concepts (“realism”, “nominalism”, “Objectivism”) there is an obvious third school he ignores – a school, I might add, that would appeal to most normal people on a commonsense level. This school is nominalist in the sense that it allows a person the freedom to mean what they mean, but it is not nominalist in Binswanger’s sense because it allows criticism of what they choose to mean, so it is not the “subjectivist” school he defines nominalism as being.

(It is curious that the Objectivists err in the sense of defining what they disagree with out of existence, since they don’t make this mistake elsewhere. They understand that art they do not like is still art, and yet, if they come across a concept formed in a way they do not like, for them it’s not a concept. In fact, for them it’s even worse than not a concept – it has no place in their theory of mind whatsoever. They literally have no word for “a unit of meaning that was not created the way Ayn Rand said meaning should be.”)

As the chapter proceeds, it seems that Binswanger strives for a “one true conceptual hierarchy” – he wants to be a realist, who justifies himself by giving us the allegedly objective mechanism that justifies his realism. Like a technologist who is so enamored with the fine points of an arcane technology that he can’t see the forest in the trees, Binswanger tears into the nuts and bolts of his theory of conception, and we are apparently on the hook for knowing and practicing all this, lest we fall to the fate of the nominalist theory, which, he alleges, “deprives man of objective guidance in the crucial aspect of his life: how to form and use the concepts which his control over the course of his life depends.”

Hogwash. We don’t need Binswanger’s conceptual technocracy in order to be rational. I am reminded of John Locke, who wrote:

“But God has not been so sparing to men to make them barely two-legged creatures, and left it to Aristotle to make them rational, i.e. those few of them that he could get so to examine the grounds of syllogisms, as to see that, in above three score ways that three propositions may be laid together, there are but about fourteen wherein one may be sure that the conclusion is right; and upon what grounds it is, that, in these few, the conclusion is certain, and in the other not. God has been more bountiful to mankind than so. He has given them a mind that can reason, without being instructed in methods of syllogizing: the understanding is not taught to reason by these rules; it has a native faculty to perceive the coherence or incoherence of its ideas, and can range them right, without any such perplexing repetitions.”

What the Objectivist “similarity” technocracy amounts to is an equivocation. The law of identity is already implicit to concepts, thus a concept means what it means, and thus, any referent of a concept has at least something in common with the other referents, it is “similar” in that regard. This is an intrinsic consequence of the classical laws of logic on any theory of concepts. But in the other sense, Binswanger gives us a bunch of technobabble about what similarity is and how it works, and this over-specifies and over-regulates what should be your free, independent mind. I.e., there is an equivocation going on here between the law of identity, and this technocratic notion of similarity, the former giving a veneer of credibility to the latter.

Binswanger spurns nominalism as “subjectivist” (and indeed, as he defines nominalism, it is), but just as many would-be dictators castigate freedom as anarchy, he is really spurning liberty of thought as subjectivist, and offering his technocratic regime as a substitute for a rational and free mind.


4. HIGHER-LEVEL CONCEPTS

The subject of this chapter is a good one. In a mature vocabulary, there is a hierarchy of concepts, and this hierarchy is critical to knowledge; it enables us to efficiently make sweeping, powerful statements of truth, as in when we learn something about animals, then we also know it about dogs, flies, elephants, etc. But in this power comes the danger of making recklessly false statements as well. Ergo the importance to epistemology of this subject.

Measured by volume, most bits in this chapter are uncontroversial and true, covering such matters as grammar and lexicography (“An impressive array of English verbs use subtle differences to narrow ‘walking.’ Here are a few: ambling, striding, pacing, sashaying, strolling, limping, parading.”) Some might find the level of detail in Binswanger’s treatment fascinating, while others might find it excruciatingly tedious; but I suppose that neither group would find these bits false, nor particularly novel. Excepting the false theory of concepts these bits of truth are enmeshed within, it may serve well as a survey in the rich variety and interrelations of concepts.

The troublesome part is the novel part: the distinctively Objectivist theory of concept formation. Generally but not always, when he is talking about what various concepts mean and how they relate to one another, I find his descriptions to be correct, but when he is talking about the genesis of these concepts, I find him to be mistaken. It may be the case that the part he got from Ayn Rand is the part that has led him astray, but where she did not specify any particular conclusion, he has found the right answers.

He starts off with an odd procedure: treating the alleged mental processes of children as being relevant to adult epistemology. I don’t know what precedent there is in the history of philosophy for consulting children on how to think, except The Bible, which claims that adults should “become as little children.” In spite of the obvious danger that studying children as a source of epistemology leads to childish epistemology, he uses this little device throughout the book, and fairly intensively in this chapter, beginning it with the claim that children can’t form the concept “animal” without first having formed concepts like “dog”, “elephant”, “fly”, etc. Even if it were true that a 10-year-old could not possibly understand what an animal is without first knowing what these other animals are, it does not follow that an intelligent adult would have the same difficulty.

He is tacitly entangling two drastically different issues here: 1) the alleged temporal order of a child’s intellectual development (which is irrelevant to the subject of his book – adult epistemology); 2) the logical order an adult mind must in principle follow when forming concepts. As far as I can tell, he makes no case whatsoever for his assertions about what we must learn first; he merely asserts it and proceeds from there. It is as if he wishes to bypass a philosophical argument for his premise, and instead rely on what usually happens to be the case: children usually happen to know about dogs, flies, birds, etc., before learning the concept animal.

It almost goes without saying that the concept “animal” is broader than the concept “dog” (i.e., it includes not only “dog” but many other things). But the idea of a cognitive requirement such that we necessarily go through concepts like “dog” or “fly” in order to get to the concept “animal” is another thing entirely. He claims that a dog, a fly, and an elephant are not similar enough to directly form the concept “animal” from, that we must first see a bunch of different dogs, form the concept dog; see a bunch of different flies, form the concept fly; etc. Only after having done this can we, he alleges, proceed to the “high level” concept of “animal.” E.g., supposing that we had a world where no one had yet seen an animal (but where people were otherwise intellectually normal), his claim is that normal adults visiting (say) a zoo for the first time would not be able observe the entities (that we call animals) as all being self-moving, that eat food, etc., and thereby form the concept “animal” directly, without going through these intermediary concepts first. Evidently, he is under the impression that these normal adults entering a zoo for the first time would not be able to tell the difference between the dirt and rocks and buildings, and the animals, without having been walked through “bear”, “lion”, “zebra” ahead of time (or something along these lines).

Related to this is the very strange idea that we can only get to the concept “animal” by “abstracting from abstractions”. He elaborates on “abstraction from abstractions” with the claim that “To form wider concepts, one applies the same processes [one used to form concepts like “dog”, where the units are real dogs observed and abstracted from], but with earlier-formed concepts taken as the units to be integrated.”

I think one should pause and consider how awful this statement is.

What he is saying (but perhaps it is not what he really means to say?) is that one directly learns the concept “dog” by using real dogs as the units (this is a fine statement as far as it goes), but one reaches the concept “animal” by using the concept dog as the unit. I.e., a real dog is to your concept of it, as the concept “dog” is to the “higher-level” concept of animal.

But a concept is not a thing; it’s a mental product, it’s only in your head. I somehow think Binswanger wouldn’t go so far as to claim that the referent of the concept “animal” is the concept “dog”; really I think his analysis is just very confused, by virtue of his attempt to make the case for Ayn Rand’s “abstraction from abstractions” notion of conception.

(I consulted ITOE to see if Ayn Rand had erred quite this badly, and she has a sentence very nearly the same as Binswanger’s, but it’s in the context of a paragraph that at least suggests the following more plausible approach to forming concepts like “animal.” I.e., it is as if Ayn Rand had mixed some absurdity with something sensible, and Binswanger has stripped what was sensible from her explanation.)

Rather than his strange procedure actually being true, it’s more plausible that we can use the concept “dog”, “fly”, etc. to direct our attention to real things, and then we can abstract properties common to all these real things, from the real things. This theory directly contradicts the one Binswanger is putting forth, since it means that we can arrive at the basic concept of “animal” in various ways, one of them via a process somewhat like what he suggests (however, the idea of literally using concepts to abstract from, rather than from their referents, is hopeless), another via directly observing various different animals, without first forming narrower concepts of particular types of animals.

This theory does not imply that conceptual hierarchy is unimportant, it does not mean that we should not eventually arrive at a revised, integrated, scientific, and hierarchical knowledge of the kind “a dog is a mammal; a mammal is an animal …”, it only means that we don’t necessarily arrive at that ideal in a rigidly determined order, it means we can form any concept we want so long as we can actually sort into “belongs” and “does not belong”. Choosing the standard by which we sort such that we create an integrated hierarchy leads to unit-economy and powerful expression, but it does not mean these things get “further removed from perceptual reality” (as Rand puts it in ITOE). The only time things are “further removed from perceptual reality” is when they are, e.g. when things are too tiny or too far away to be observed directly.

It’s no wonder Binswanger asserted in the previous chapter that sense-perception is infallible: his epistemology doesn’t require continuing observation, correction, and refinement for all concepts; rather, you only need to check reality once, when you form “first level” concepts from “infallible” percepts, and then you simply dogmatically rely on these from then on, not needing to consult experience ever again. Furthermore, only children need to directly observe from reality; adults are at liberty to build conceptual castles in the air based only on the “first level” concepts they grasped when they were children. Perhaps this helps explain the apparent need Objectivists have in incessantly consulting children – they are trying to make up for having cut off their “higher level” concepts from reality. (I would sympathize somewhat if he disputed this characterization, but I submit that he would only be arguing with himself.)

What Binswanger has done is to confuse the final ideal form of the result (a scientific conceptual hierarchy) with the process by which it is formed. So just as in the previous chapter, where we observed that Objectivism has an overly rigid conception of conception, in this chapter we see that Objectivism has an overly rigid conception of the process of conception. I.e., concerning conception, Objectivism is, very ironically, concrete-bound.


5. PROPOSITIONS

I found the constant referencing of how children think in this chapter to be too grating, so I am skipping this chapter.


6. LOGIC: THEORY

This chapter starts off well, including a useful analysis of the laws of thought. The section on “context” is mostly good, but there is a bit that foreshadows a wrong conception of certainty, very distinctive to the Objectivist attitude regarding infallibility and error, but also very wrong and very harmful (I may revisit this issue later).

The section on “hierarchy” reveals a bit more about why he is incessantly bringing children into the discussion: he considers knowing “the necessary order of acquiring knowledge” to be of paramount importance to logic. I would certainly agree that if we are designing a curriculum, then this is of paramount importance, it is also something that the status quo gets very wrong and with very harmful and authoritarian results, but there is no importance whatsoever of temporal learning order to logic. The dependence is the reverse: logic should inform finding an efficient learning order.

For example, he insists that while a grandfather came into existence first, in order to grasp the concept “grandfather” we must first go through the concept “father.” For obvious reasons, it’s clear that a child would usually learn “father” before “grandfather”, and it is clear that if one is giving a systematic lesson, one should teach the child the concept “father” before teaching them “grandfather.”

But whether we must or must not learn these concepts in a given order is utterly irrelevant to logic. What is relevant is that the concept “grandfather” logically presupposes the concept “father” (it means “father of father/mother”), and that the concept “father” logically implies the concept “grandfather” (to know what a “father” is and to know that human beings die and reproduce, is to understand the inevitability of “grandfather”). What this means is that if someone denies either concept while using the other, we can logically accuse them of one kind of self-refuting ignorance or another. I.e., implication and presupposition are the actually important ideas here, which Binswanger obscures with his lengthy and unremitting insistence on the order in which we allegedly must learn.

I can imagine someone making the following objection to my view: If a person had really learned the concept “grandfather”, how could they not also understand the concept “father”? My answer is that the actual order of learning can be a great deal more messy and roundabout than the questioner supposes, and yet in spite of apparent chaos, it finally resolves into clear understanding. One should not mandate the specific process of learning (even while guidance informed by logic is helpful); one should focus on standards by which we can measure the result (logic). We are all individuals learning in our own different ways, and it’s presumptuous to create recipes of rigidly required learning order; what is important is that when we arrive at knowledge, we can tell that it in fact is knowledge – and this does require rigid standards. The perspective Binswanger chooses to rigidly specify and emphasize is irrelevant (learning order), and thereby diminishes what is actually important to rigidly adhere to (logic).

Binswanger also revisits his “can’t reach the concept ‘animal’ without going through concepts like ‘dog’” bit, claiming that it’s not “crow-friendly” to reach for the concept “animal” first. This is presumptuous. Perhaps he can’t do so, but it’s not his business to declare that others may not do so. It is not “illogical” to learn the same things in a different order than the one Binswanger prefers. (It may or may not be inefficient – but efficiency is a curriculum question not a logic question.)

This confusion of Binswanger’s is one major problem with this book (another major problem is his false theory of perception.)


7. LOGIC: PRACTICE

This is a very good and important chapter. Every high-school student should be given a mini-course on its contents, and given a refresher course again, in college. But since my purpose here is to give a critical review (Objectivist culture hypocritically damns criticism of itself, so it won’t be found from within), I leave it to others to elaborate on precisely why this chapter is so valuable.

There is only one small section in which he errs in a manner resembling the errors in prior chapters. This is at the end of the chapter, in the section on induction, where he reiterates the Peikoff/Harriman theory. I won’t get into why I find his problematic here since that would take this review far afield (induction is a topic I’ve written about elsewhere).

There are a few minor defects:

  1. A logical blunder strangely appears on p. 238: “[Optional concepts] represent ‘borderline cases,’ in the sense that they fall between mandatory concepts and invalid ones.” If Binswanger seriously means this then he doesn’t believe in the law of excluded middle, since a partly invalid concept is still invalid and should be refined such that it’s not invalid. A rejection of the law of excluded middle would make it hard to interpret his earlier remarks. Perhaps he got confused while writing and no editor brought it to his attention: he started off talking about cases where it was optional to form a concept, i.e., where either choice would be valid, not where one is “a little invalid”, which is an idea akin to being “a little bit pregnant.”
  2. Another error, less strange coming from an Objectivist, is where he blithely claims that whenever Bertrand Russell uses the term “freedom”, then his remarks are to be dismissed, since Binswanger has found one place where Russell improperly defined freedom. This kind of remark sheds some light on why Objectivism has such a rich history of insularity, ostracism, and splitting into factions. Bertrand Russell had a long and massively productive life, and was known to change his opinions over time, so Binswanger’s remark is plainly ridiculous.

Again, these errors constitute minor flaws in this chapter, it is still very much worth reading.


Notes:

  • Revised to include “Concept Formation” (4/17); “Higher-level Concepts” (4/23); “Logic: Theory”, “Logic: practice” (4/26).
  • I altered the original title and edited the original review for clarity.

A noteworthy remark from the Amazon review, highlighting a type of moral cowardice that is – very ironically – running amok in Objectivist circles in these past few decades:

Edwin A. Locke criticizes me:

This was not a useful review for me–too many snide comments–very unprofessional…

My reply:

I think it should be pointed out for non-insiders that Edwin Locke is an Objectivist lecturer and his moral indignation here is almost certainly a pretense.

Rand’s writing is replete with “snide” (as Locke puts it) remarks about other thinkers. Was Rand’s behavior “unprofessional”?

Rand was a moralist. When a thinker advocated something she found to be the cause of evil in the world, she’d say so. I’d say that if we want a better world, we need more of this, so long as the criticism is rational.

I presume Locke would agree when Rand does it to others, but not when someone else does it to his Objectivist peers, even when the motive is identical. This makes him a hypocrite. He’s conveniently hiding behind academic standards he himself rejects. If he denies this, he should kindly point the reader to where he has criticized Rand on this point.

Indeed, I’ve seen this hypocritical behavior from a number of other prominent Objectivists. They have institutionalized a rigid demand for extreme politeness toward themselves, particularly when faced by criticism from outsiders. I noticed this in the book that Chuck recommended above:

“we may on occasion publish writings [that disagree with our personal interpretations of Objectivism], so long as these writings are respectful of Rand and her work…”

You can choose to respect Rand above all else or respect the truth above all else, but you can’t do both. And why, one might ask, if a rule is good enough for Ayn Rand, would it not be extended to all thinkers generally? Why do we have to be especially polite to her ideas?

Consider what it would mean for someone who routinely crosses moral lines to stipulate that you must never be “rude” to him. What does this mean? It can only mean that you must never morally judge him. I.e., this demand that you be “polite” is simply him trying to get a blank check on morally reprehensible behavior. And when you institutionalize this sort of demand, you get an insular cult of sycophants, because it constitutes a de facto ban on the most impactful types of criticism.

It is especially shameful when Objectivists pretend to be prissy Victorian prudes when it comes to harsh criticism, given that it is completely contrary to how Ayn Rand conveyed her own philosophy.