Friday, September 28, 2007

The Failure of the Hierarchy to Justify Abstract Reasoning

In Anon's most recent reply, he focuses largely on the issue of explaining why "the concepts we need to learn first are just the ones whose referents have the most obvious perceptual similarities." I'll leave it to cognitive scientists (including those mentioned by Anon, along with others) to determine the degree to which this part of the theory accords with reality. In the meantime, I want to focus at what I consider the chief point at issue: the hierarchy as a theory of justification. As Anon admits, the theory attempts "to show how abstract reasoning connects to perceptual evidence." Does it really? But how can that possibly be? At the very most, it only shows how the concepts in use among propositions relate to perceptual evidence. However, abstract reasonings are used to formulate theories; and under no circumstances is it plausible to conclude that a theory can be justified merely by showing "how one's higher-level concepts (and associated propositions) derive from lower-level concepts." Knowledge about very complex aspects of reality — knowledge of economics or politics, for instance — is expressed in theories, and theories are more than the sum of their concepts or the logic of their propositions. It is entirely possible that a theory could be made up of valid concepts and pristinely logical reasoning yet still fail to accord with important facts. The reason for this is quite simple: social theories are (and must be because of the complexity of the subject matter) based on simplified premises. Economics, for instance, seeks to understand what would happen under simplified conditions never realized, but often closely approached in practice. Knowledge of reality involves achieving human scale, and that will sometimes involve a certain degree, hopefully within tolerable limits, of cognitive inadequacy. Sometimes the best we can hope for in a given branch of knowledge is merely to achieve a tolerable degree of plausibility.

Rand does not appear to have accepted the notion that theoretical knowledge of subjects like economics and politics might be characterized by a unspecifiable degree of uncertainty. She presented her own notions on these matters as if they were quite obviously the certain conclusions of reason, which any rational and moral person would accept as a matter of course. And she appears to have believed that her validation of concepts, of which the hierarchy theory (along with measurement omission) is a critical part, removed any problems related to the possible distorting effects of condensation. Nor, sad to say, does she appear to have confronted the problems of the actual simplification that takes place in the process of devising the premises of theories. The hierarchy of knowledge does not, and cannot, address this issue, since the problem has nothing to do with whether the concepts (or even the propositions) are "connected" to reality, since the way in which any really sophisticated theory "connects" to reality is far too complicated to be tested by tracing its concepts back to their perceptual concretes. Theories "model," in greatly compressed form, certain aspects of reality. When we talk about theories or concepts or propositions being "connected" to reality, we are speaking metaphorically, not precisely or literally. Theories don't mirror reality; they reduce an immense complexity to human scale, so that the mind can get some inkling of that complexity. Theories of complex phenomena are not proved or corroborated or justified by tracing their concepts back through a hierarchy to their "perceptual base." Perhaps that would be possible with something really simple, like a theory of furniture. But it's not in the least possible if we are talking about the theory of the fall of the Roman Empire, or an economic theory of tariffs, or a theory of the circulation of elites within a socio-political system.

So how do we know if these theories are true if you can't connect them back through the "hierarchy" to the perceptual base? Science uses rigorous empirical testing to corroborate the conclusions of a theory. No need to trace anything back to the perceptual base from which reasoning proceeds; no, simply test the conclusions! Much simpler, and far more convincing. But there are situations when, because of the inability to run experiments and isolate variables, such empirical tests can hardly be considered conclusive. In economics, for instance, one can't provide entirely convincing empirical tests "proving" that, under all conditions, tariffs will reduce the total output of an economy. The most you can hope for in fields of study where empirical tests yield ambiguous results is to attempt to determine, on the basis of abstract reasoning from simplified premises, what is likely to be true or plausible. Certainty (or near certainty), whether of a contextual variety or otherwise, does not appear possible in such disciplines.

Now, the belief in an epistemological theory that promises what it can't deliver will likely lead to cognitive mischief. The danger is that the theory will give its proponents an unjustified confidence in "abstract reasoning." Even if the intention of the hierarchy is merely to justify so-called "justified" speculation, if the theory does not deliver on its justificationist promises, it will likely lead its proponents unwittingly down the primrose path of rationalist speculation. Isn't this what we find in Objectivist thought — unjustified speculation run rampant, yet all done with the most naive sort of cognitive innocence? Rand herself made any number of assertions about matters of fact that are grossly implausible, and yet she seems to have made them with an entirely serene and untroubled epistemological conscience. For example: "Man is born with an emotional mechanism, just as he is born with a cognitive mechanism; but, at birth, both are 'tabula rasa.'" "Emotions are the automatic results of man’s value judgments integrated by the subconscious." "Emotions are not tools of cognition." “[A man’s] body will always follow the ultimate logic of his deepest convictions.” "A man is a ... being of self-made soul." These statements are not presented as possible conjectures or yet-to-be-proved hypotheses, but as veritable certainties. Some of the statements are very likely false. You would be hard pressed to find any scientist acquainted with the relevant evidence who believes in the blank slate model of human nature advocated by Rand. Nor are Rand's views of emotions any better supported by scientific evidence, as Damasio, among others, has discovered. Her statements relating to human nature are, in the light of evidence compiled in genetics and cognitive science, immensely implausible. Her justificationist epistemological theories have here let her down! Instead of leading her to truth and enlightenment, they have mislead her into error and delusion!

I don't see among Rand or her followers any appreciation for how difficult it is to determine the veracity of certain types of truth claims. Too much confidence is placed in reasonings based on vague generalities or insufficiently detailed understanding of the subject at issue. The consequence is a philosophy characterized by an overabundance of implausible and semi-plausible assertions. My suspicion is that Rand's justificationist theories, including her hierarchy theory, by giving her an unwarranted confidence in her own reasoning abilities, served merely to fortify her rationalistic and dogmatic tendencies.

8 comments:

Anonymous said...

You write:

"However, abstract reasonings are used to formulate theories; and under no circumstances is it plausible to conclude that a theory can be justified merely by showing 'how one's higher-level concepts (and associated propositions) derive from lower-level concepts.' Knowledge about very complex aspects of reality — knowledge of economics or politics, for instance — is expressed in theories, and theories are more than the sum of their concepts or the logic of their propositions. It is entirely possible that a theory could be made up of valid concepts and pristinely logical reasoning yet still fail to accord with important facts."

I'm starting to notice a pattern of straw-man arguments coming from this site. Why assume that the view is that a theory can be justified MERELY by showing how its concepts derive from lower level concepts. Of course that's not sufficient. But it is necessary. A theory that does not contain meaningful concepts doesn't even stand the chance of being true. Just because a theory of concepts does not make for an entire theory of justification does not mean that it isn't an important part of such a theory.

What's more, understanding the meaning of abstract concepts helps us to know what kind of evidence would justify our theories. A justified theory as expressed by some fundamental proposition, like all propositions, is essentially a justified application of a concept to some subject matter. How do we know if the theory of evolution by natural selection is correct? We need to know that the concept "inherited" is correctly applied to biologically adaptive traits. What does it take to know if some trait is inherited? "Inherit" is an abstract concept: it designates a process over time that cannot be directly observed. To know how it correctly applies, we need to know the reason we formed it in the first place. We need to know what differentiates an inherited trait from a learned or conditioned trait.

You also write:

"It is entirely possible that a theory could be made up of valid concepts and pristinely logical reasoning yet still fail to accord with important facts. The reason for this is quite simple: social theories are (and must be because of the complexity of the subject matter) based on simplified premises. Economics, for instance, seeks to understand what would happen under simplified conditions never realized, but often closely approached in practice. Knowledge of reality involves achieving human scale, and that will sometimes involve a certain degree, hopefully within tolerable limits, of cognitive inadequacy. Sometimes the best we can hope for in a given branch of knowledge is merely to achieve a tolerable degree of plausibility."

First, Rand would say that if a theory is made up of valid concepts but doesn't accord with the facts, it isn't logical. Logic is a matter of both deduction and induction, and the practice of justifying a theory by reference to the facts is surely a matter of inductive logic (which is in turn conditioned by the logic of concept-formation). But second, I fail to see the relevance of the fact that some theories use simplified assumptions to model complex realities. Assuming that this is a valid description of some sciences--and that our knowledge is advanced by modeling simplified aspects of reality when we cannot yet capture the complex ones--how does this relate to the question of the justificatory relevance of proper concepts? First you attack the relevance of theory of concepts by saying that a theory with proper concepts might not accord with reality. But then you say that valid sciences (by your account) diverge from reality when it simplifies. So whose side are you on?

Next you say that Rand does not accept that some sciences may involve degrees of uncertainty. I don't know where you're getting that from or why, once again, it relates to the epistemological relevance of a theory of concepts. It's perfectly consistent to think that a theory of concepts is an indispensible part of a theory of justification, and also think that some theories are not justified to a the degree of perfect certainty. These are just entirely different issues and you are running them together in a bizarre way. You present no evidence to justify the claim that she thinks her theory of concepts removes the possibility of uncertainty from reasoning. And indeed you can't, because she didn't.

I don't fully know what a theory of concepts has to say about the process by which science "simplifies" the premises of its theories, if that is a valid description of what science does. But why should a theory of concepts should have to say anything about that? I do, however, imagine that it is precisely our ability to form abstract concepts that allows some of the valid "simplifications" that you may have in mind from the history of science. Think about Galileo's imaginary frictionless surfaces that were useful in his thinking about dynamics. By abstracting away from the known effects of friction, Galileo was able to form the concept of "inertia," which was essential to overcoming the Medieval concept of "impetus" and to making advances in classical mechanics. So many scientific revolutions come about because of conceptual revolutions. Justifying new theories is a matter of justifying the application of new concepts. This has nothing to do with "mirroring" reality. You're right that it has to do with "reduc[ing] an immense complexity to human scale, so that the mind can get some inkling of that complexity." But that is precisely the task accomplished by concepts, as Rand's principle of unit-economy makes evident.

As for your point about the rationalization of philosophic speculation, I don't know where it's coming from. Objectivism views philosophic reasoning and scientific reasoning as of a piece. See Leonard Peikoff's "Objectivism Through Induction" and "Induction in Physics and Philosophy." The Objectivist theory of ethics is no different in principle from a physical theory of inertia. Just as Galileo had to justify the concept of "inertia" to validate his theory of mechanics, a theory of ethics has to validate the concept of "value." Both "inertia" and "value" are abstract and need to be abstracted in the proper order, connected to the proper evidence, if they are to yield valid theories. If Rand made assertions that you take to be contrary to what some scientists say, then either she was simply mistaken in her reasoning or the scientists are. The problem has nothing to do with her theory of concepts.

Incidentally, your aversion to the Objectivist theory of emotion, which I've seen both on this blog and in your book, is a particularly odd bone to pick when it comes to assessing Objectivism's theory of human nature. That's because the cognitive view of emotions is hardly distinctive to Objectivism. It has been held by Aristotelians for decades (see Magda Arnold), and more recently by the highly successful discipline of cognitive-behavioral therapy, and a number of other more theoretical branches of cognitive science, viz.:

http://plato.stanford.edu/entries/emotion/#5

Perhaps cognitivism about the emotions is not entirely uncontroversial, but it's also not so at odds with reality that no scientists take it seriously.

Anonymous said...

Well said, Anon.

I would also add that Rand acknowledged the presence of certain automatic mechanisms in human nature and the potential difficulty in overcoming them. The content in this site makes it seem like Rand thinks we consciously choose every single facet of our thought processes.

Anonymous said...

"That's because the cognitive view of emotions is hardly distinctive to Objectivism. It has been held by Aristotelians for decades (see Magda Arnold)...,"

Then if Objectivism is wrong/right on that issue, so are the Aristotelians.

"...and more recently by the highly successful discipline of cognitive-behavioral therapy..."

CBT operates from the premise that thoughts, feelings and behavior are all interconnected and influence each other. This belies the concept of reasoning able to fully divorce itself from feeling and emotion. It says nothing about feelings coming ONLY from propositional attitudes, just the specific pathologies it treats (mood and anxiety disorders).

"...and a number of other more theoretical branches of cognitive science...Perhaps cognitivism about the emotions is not entirely uncontroversial, but it's also not so at odds with reality that no scientists take it seriously."

Emphasis on the 'theoretical'; the link you provided raises multiple, effective objections to the pure cognitivist theories, and in fact the article itself says it is heavily criticized. The current majority of opinion and empirical evidence still seems to point away from a strictly propositional attitude model of emotions.

Anonymous said...

Yeah, like I said, the cognitive view of emotions is not uncontroversial. And there are criticisms of it which an advocate would need to deal with. My point was that this web site treats the theory as if it was simply an intellectual fabrication by AR without any reference to empirical evidence. The fact that some scientists take it seriously, independent of ideological orientation, is evidence that this isn't necessarily true.

Anonymous said...

According to the article you referenced, however, cognitive theory of emotions seems to be the realm of philosophers only, not scientists. No empirical evidence was offered for it, nor were any actual scientists i the field mentioned; even if they were you would need to show that there are more than just a few (after all, I am sure you can find at least one geologist who thinks the earth is flat).

Anonymous said...

As for your point about the rationalization of philosophic speculation, I don't know where it's coming from. Objectivism views philosophic reasoning and scientific reasoning as of a piece. See Leonard Peikoff's "Objectivism Through Induction" and "Induction in Physics and Philosophy."
I highlighted this little quote from the above reply by the principal respondent in this discussion to note that the citations are available ONLY in audio form, and are thus not amenable to independent, critical analysis. Even a blog post is better than that!
I can't see how anyone posturing as a credible scholar or interpreter of Objectivism can do so on the basis of cites from oral sources. It's also a good reason in serious academic work to discount contributions from that community - they simply don't publish their work. Tara Smith notwithstanding.

Daniel Barnes said...

Anon:
>I'm starting to notice a pattern of straw-man arguments coming from this site.

Well, this may be the case, but underneath her supremely confident rhetorical style Rand regularly wrote in a vague, confused, and self-contradictory way. (For example, in her "Ethics of Emergencies" essay she says one thing on one page, then quite the opposite on the next, without blinking an eyelash)

As a result, if one makes a critical claim about Rand, her work can usually be mined to produce a suitably vague contrary remark, and the critic then accused of creating a "straw man" (or perhaps "dropping context").

Anon:
>Logic is a matter of both deduction and induction, and the practice of justifying a theory by reference to the facts is surely a matter of inductive logic (which is in turn conditioned by the logic of concept-formation).

Having debated various Objectivists at length over the years, we at this site also notice a pattern of argument coming at us. I'll call this pattern the retreat to esoteric definition.

Take your above, which makes a number of frankly odd claims - for example, that "logic is a matter of both deduction and induction."

Now, "logic" is generally considered the rules governing a valid inference. Further these deductive rules famously clash with what is generally called "induction." Yet here we find this all casually tossed together in a sentence as if it was all the most given and self-evident thing in the world. We also hear, with some suprise, that by this tossing-together we effortlessly arrive at a what all philosophy has been looking for since the Greeks - a "justified" theory, one that we can have "absolute certainty" about.

Puzzled, we critics press further. We are then informed that we have the incorrect concept of "logic" - that it is not merely about "valid inference", but about something described as "non-contradictory identification". Further what we (and everyone else) mean by "induction" is also not the true meaning of the term, which is in fact something like "observational-integrative generalisation". Finally, it turns out we critics don't know what "justified" truly means either; for a theory is, it turns out, "justified" and "certain" in Objectivism if you believe it to be true today even if it turns out to be false tomorrow. And if we still have any questions, the answers are supposedly all provided by Leonard Peikoff in one of his 20 hour audio lecture series, so we need to then listen to that to fully grasp how these thorny issues are forever resolved by Objectivist doctrine.

By this stage, its clear that Objectivists are almost talking an entirely different language to their critics - and in fact just about everyone else. This use of seemingly ordinary terms, but that turn out to have vague and highly esoteric definitions behind them, is one of the main barriers to criticising Objectivism. For it is all too easy to use common terms in an obscure fashion, then accuse critics of creating "straw men." It is also all to easy to make certain arguments, then retreat to these obscurities and claim a lack of agreement as to "fundamentals" rather than the initial arguments themselves. Perhaps most importantly of all, while Objectivists see this line of retreat as a return to the vital fundamentals on which all else rests, as critics we tend to see it as a Brer Rabbit ducking into a briar patch of mere verbiage and word-games.

For if Objectivists insist on their own special definitions of basic terms - even of that most fundamental critical tool, logic itself! - to support their doctrines, it can blame outsiders for supposedly misrepresenting said doctrines, nor for being highly suspicious of such a marking-your-own-homework approach in the first place. In fact such insistence makes such problems almost inevitable.

At any rate, given the above, isn't it about time Objectivists took some responsibility for the "pattern of strawmen" that supposedly emerge from critics of Rand's doctrines? They might also pause to consider that this insistence, much as it might shield their doctrines from straightforward criticism, might also be hampering its spread to the wider world.

Daniel Barnes said...

Errata above:
"...it can blame outsiders.." should read "...it can hardly blame outsiders..."