Objectivism Versus Rationalism and Empiricism, Part 1

Disclaimer: I’m not an expert on philosophy. I’m just a person trying to figure things out for myself, and speak for no one but myself.

This post is about Lecture Nine, "Objectivism Versus Rationalism and Empiricism", in Leonard Peikoff’s book Understanding Objectivism. I’ll go through the chapter, summarizing and adding my own thoughts, comments, or questions.

Peikoff says we're going to contrast Objectivism with rationalism and empiricism and starts off with a recap of rationalism and empiricism. He says that both rationalism and empiricism are two sides of the mind/body dichotomy, with rationalists rejecting the physical world (the body) and empiricists rejecting the mind. Objectivism says there's no dichotomy. There's no mystical/supernatural soul – consciousness is a part of nature/reality/the world, and only exists under certain conditions and in certain organisms. "[C]onsciousness is nothing but the faculty of perceiving the something that exists, which is physical nature." The mind acquires knowledge and defines values, and the body carries out the conclusions and value judgments of the mind. Both are indispensable.

Without a mind, man has no means of knowledge and no way to direct his actions or preserve his life. Without a physical brain and body, we can have no consciousness or ideas at all, let alone any way of carrying out our ideas in action. The two elements are two indivisible aspects of one harmonious, integrated entity.

Peikoff then turns to various issues he's discussed regarding rationalism and empiricism and then presents the Objectivist perspective on each point.

Relation of Ideas to Reality

Point one is the relation of ideas to reality.

Rationalism advocates ideas above reality; empiricism advocates reality above ideas, ideas are to be dispensed with. Objectivism says ideas are the means of knowing reality.

Peikoff makes a statement that I'm not sure about:

In essence and leaving out a lot of complexities, concepts are integrations of percepts; that’s the central point from the Introduction to Objectivist Epistemology; they have no content other than the percepts, the concretes, that they integrate.

He does heavily qualify this statement, but still, I take issue with it. I think Peikoff's claim re: "no content other than the percepts" is related to his pro-induction ideas. I disagree with Peikoff and think a concept needs to have some sort of guess/idea/explanation/notion about what unifies some percepts, which would mean that the concept has an idea in addition to the percepts. I think that this approach is actually consistent with Objectivist themes, since it emphasize the role of the mind and human cognition in coming up with the concepts that allow us to organize our thinking about the world.

Peikoff says concepts aren't a way of knowing another dimension (the rationalist approach) or arbitrary/subjective (the empiricist approach). They let us integrate concretes based on the nature of those concretes.

Peikoff says that if you have a health mental process, you should have a "continuous dual urge" where on the one hand you're straining in the direction of thought, abstractions, and concepts, and on the other hand you're straining in the direction of percepts, observations, and concretes. If you're in an abstract discussion, you should have an urge to concretize, illustrate, and apply before you get overwhelmed by abstractions. And if you're observing concretes, you should have an urge to conceptualize/integrate – "because the more the concretes are multiplied, the more the crow should be rearing its head".

The same applies to thinking about definitions. Definitions involve a narrowing of focus about the thing you're defining. You have to leave out a whole bunch of attributes of the actual thing you're looking at and just focusing on what's definitional in order to discuss things (like, say, the welfare state). So on the one hand definitions are necessary. But on the other hand, you have to avoid equating the thing you're discussing with its definition (which sounds like a mistake a rationalist would make – reducing a thing to a name for a concept about the thing).

Peikoff says that in general thinking is specialized. You think about one topic at a time. But you should have an urge to integrate that thinking to your other thinking. So there's a dual urge again: on the one hand, you should want to learn the details of certain specialized fields, but on the other hand, you should want to connect those details to the whole of your ideas. You don't want to compartmentalize.

Peikoff says that Rand's preference for writing her philosophy in fiction reflects wanting a union of concretes and concepts.

Induction vs Deduction

Peikoff says that induction and deduction both have a role, but induction is primary because it gives you the general principles, the application of which you can work out with deduction. Peikoff says deduction is fine but not the primary cognitive process.

Peikoff says that the problem of induction is fully solved (!) and "[t]here is nothing further to wait for as to the solution of the problem of induction." He admits, though, that there are issues in philosophy of science like "how you choose among competing theories, both of which seem to explain all the known facts". Indeed. But he thinks induction is solved for everyday people.

Peikoff admits you can't validate an inductive statement (like that all men are mortal) by observing every example of man. He says you can nonetheless take the "inductive leap" by means of "integration" and "delimitation". (BTW I think it'sa bit odd that he's purporting to lay out a solution to the problem of induction, which would be a big breakthrough, as point 2 of 8 in a lecture about mostly another topic.)

Peikoff says that the answer to the question of how many instances you need to enumerate of some instance before you validly generalize depends on your context of knowledge.

Take the obvious example: We see a thousand white swans, and we never come across an exception. With nothing further, you would be unjustified to generalize. It would be shaky, unconvincing, to conclude all swans are white. Why? Because if your knowledge has been developing normally, you would have presumably observed that color is generally a nonessential, that it’s a superficial attribute with little connection to the structure, nature, or actions of a given species. You can see that in human beings, who range from red, white, black, and so on, and yet are all still human. Once you see this across many species, you already have a principle to guide you with regard to the swan, which is: Be suspicious. Do not assume that ten thousand cases proves anything, because it’s only color. That other knowledge—if you integrate your observations of the swans to it—would undercut, undermine, jeopardize your potential generalization. On the other hand, “All men are mortal”—now there you may have observed only a thousand or a hundred. But it integrates with a huge number of other observations. You know that it integrates with what you observe about all living beings, in all categories—ages, animal, vegetable, plants, whatever the climate, the region, and so on. There is a powerful mass of data with which you could integrate this observation, which immensely strengthens it.

Peikoff's approach here basically seems like: think about your context of knowledge and see a generalization would violate the principles embodied in that knowledge. That's fine and reasonable, but doesn't really seem like induction to me. BTW one of the things you might know about humans is that skin color correlates to location (I believe this is related to melanin and protection from ultraviolet rays and the fact that some places get more sun than others, which makes more melanin and thus darker skin advantageous. That's my vague understanding though – I'm not a scientist.) Anyways, the application of one's observations of humans to one's observations on swans might thus depend on the location of one's observations of swans. Were all the swans you saw in Iceland? Do you happen to know if swan populations are dispersed across the globe, and at what latitudes? And have they been dispersed for a long time (long enough for different colors to be selected for) or was that dispersal more recent due to e.g. humans liking swans and bringing them different places? So basically in considering a potential generalization about swans, you'd consider your background knowledge and various potential criticisms of certain generalizations. And that's all well and good but not really induction.

Peikoff says that Rand said she'd be less secure in the generalization that all men are mortal if there wasn't a readily apparent aging process and instead people just seemed super healthy and then suddenly dropped dead. So the fact of an observable aging process is one of the things we integrate when coming up with generalizations about the mortality of men.

Peikoff:

Now someone is sure to say, “Isn’t the whole thing circular as you’ve presented it?” Because after all, I say, “The knowledge that man is mortal is strengthened, is established, by the tie-in to aging. And the conviction that aging will go on is based on, or partly based on, the observation of mortality.” So don’t we have an infinite regress? How can you use A to strengthen B, which then in turn strengthens A? And is not this a circle?

Peikoff says yeah it's circular but it's a different sort of circularity than saying “I am infallible” -> "Jim says so" -> "I know he's right, and I'm infallible." With the aging stuff, you're not arbitrarily making some claim, but connecting together things like men aging and men dying and coming up with an idea based on those observations. Peikoff:

If you remember our hierarchy exercise, each level led to the next and enabled us to grasp it, but the next in turn made the early one clearer, fuller, richer, more concretized. So don’t be afraid of what’s called “mutually propping up” knowledge—each element props up the others, and what gets carried on is the total. This occurs throughout knowledge; it’s really what we mean by the spiral.

I agree there's a difference between making some arbitrary claims not connected to observations and integrating claims connected to observations. I just don't think the latter involves anything that's reasonably described as induction.

Peikoff says that you can make big generalizations given a few examples if you have the right context of knowledge, which allows you to validate induction with a few instances. I agree that you don't need to accumulate tons of examples to come up with useful ideas that you can act on and treat as true, though I would not call that validating an induction.

Peikoff moves onto a subpoint about delimitation. He says that proper induction must always been delimited since we always acquire knowledge within a certain context. He says that every inductive conclusion needs to be contextualized by a statement like "Within the knowledge so far acquired, within the context of what we already know" or “Taking into account all of our observations, definitions, cognitions, inductions, and so on, so far, this is our generalization.” He says inductive conclusions aren't rationalist dogmas or out-of-context intrinsicist bolts from the sky. He says they aren't uncertain, though, and he'll explain later, but:

Here the point is that later knowledge never contradicts earlier knowledge, not if the earlier knowledge was acquired by the right method. All it does is specify the conditions of the earlier knowledge.

This implies that if you follow the right method you can never err, which seems blatantly false.

Peikoff says induction is just integration in a specified context and rationalists and empiricists are skeptical of it cuz they are compartmentalizers.

Peikoff says rationalists want to model human cognition on math and empiricists want to model it on some sprawling, unstructured science like statistics as misused by social scientists. Objectivism doesn't model its approach to human cognition along some favorite science but follows a principled approach.

To be continued.