IOE, Chapter Seven
The Cognitive Role of Concepts
by Raymie Stata
Date: 9 Mar 1992
Forum: Moderated Discussion of Objectivist Philosophy
Copyright: Raymie Stata
In this chapter, Rand describes the principle of unit-economy, which explains the enormous power of the conceptual level of consciousness (relative to the perceptual level); the file folder analogy, a metaphor for the role of concepts in cognition; and some guidelines for conceptualization.
The sections below are split into "summary" and "discussion" subsections. In the summary subsection I try as best I can to describe only Rand's theory, although I use some examples of my own. In the discussion section, I shamelessly mix my own ideas in with hers (caveat emptor!).
The "crow" limitation and unit-economy
Rand notes that consciousness is limited in the number of existents that it can grasp at any one time. (Rand illustrates this point by referring to an experiment with crows; thus, it is known as the "crow epistemology.") For example, we can attend individually to our dog Spot, or perhaps even to Spot and Rover simultaneously, but when confronted with 101 dalmatians, they become a blur, a collective mass of "many" dalmatians.
The power of concepts, says Rand, lies in "unit-economy," their ability to reduce the many to the one, their ability to _condense_ a vast range of awareness into a small number of units. We might not be able to attend individually to each of 101 dalmatians simultaneously, but using concepts we can grasp that each one will wag its tail when it's happy.
Rand holds that unit-economy is the essence of the conceptual level of awareness, ie, the causal factor that explains their ability to overcome the limits of the perceptual level. Concepts vastly expand our awareness by condensing huge amounts of perceptual data into a handful of mental units. Nouns such as "dalmatian" allow us to grasp with a single unit an infinite number of perceptual units. The counting numbers such as "101" allow us to grasp with a single unit a number of concretes that, at the perceptual level, would be an undifferentiated mass. Concepts for unobservable existents such as molecules and galaxies allow us to grasp these existents without constantly referring to the body of observations that point to their existence. The concepts "parent," "friend," "lover," and "enemy" allow us to grasp with a single unit relationships that have hundreds, often subtle aspects manifest at the perceptual level.
Concepts as open-ended file folders
Rand observes that a concept serves as a repository, or "file folder," of knowledge about a group of existents. In this analogy, the file folders (concepts) have three components. First, they have referents, the existents subsumed by the concept. Second, they have content, the knowledge accumulated regarding the characteristics of the referents subsumed. Third, they have a label (definition), which condenses the file's content and gives us an instantaneous grasp of the nature of the folder's referents (this third component comes from Peikoff's description of the analogy [OPAR, p 105]).
Concepts are open-ended, ie, they are repositories for *all* the knowledge pertaining to their referents, including the yet-to-be-discovered. The open-endedness of concepts has both static and dynamic aspects. Concepts are static in that their referents are fixed: the existents subsumed by a concept never changes from discovery to death, or even from generation to generation.
Concepts are dynamic in that their definitions and content are changing. The evolution of definitions was discussed in chapter five of ITOE. The growth of the content of concepts is discused in this chapter: as information is learned about the referents of a concept, it is integrated in with the previously learned knowledge.
The theory that concepts function as open-ended repositories of knowledge is an important discovery of Rand's, a discovery whose importance is often overlooked in favor of her theory of measurement omission. However, this aspect of concepts is at least as important in understanding the how reason works, perhaps even more so. Below are some examples that illustrate how this aspect of concepts shapes various important features of the conceptual level of consciousness.
Rand points out that the open-endedness of concepts makes possible learning, research, scientific specialization and the accumulation of knowledge over generations. The dynamic aspect of concepts allows an individual to accumulate more and more knowledge about a class of existents, and allows different individuals to work independently to make new discoveries about a class. The static aspect of concepts assures that the new knowledge and discoveries can be easily integrated into a coherent system of knowledge.
Rand indicates that the file-folder quality of concepts is largely responsible for the unit-reduction concepts effect. Concepts blend into a single mental unit an entire repository of knowledge. This is an enormous amount of information, all available to us "as a context" when we recall into our consciousness a single unit . Rand calls the psychological process that allows this instant recall "automatizing," the function of which is "to make certain connections automatic, so that they work as a unit and do not require a conscious precess of thought every time they are evoked"  (notice how this description is phrased in terms of unit-economy). It is one of the most powerful abilities of the human brain.
The role of concepts as file folders explains the "spiral" quality of learning. Let me illustrate by speculating about how the concept "marriage" might evolve within a person's mind. Some of the first higher-order concepts a child forms are probably those pertaining to the special relationships within the nuclear family, such as the relationship between himself and his parents and the relationship between his parents. The concepts pertaining to this latter relationship include "friendship" and "marriage." At this early stage they are imprecise and based on emotions; they don't even have names, and thus are not full concepts. But they are there, and they serve as repositories for observations.
And these observations are deposited, day after day, probably without much conscious effort. As the child learns new things about the relationships between people and about how society is structured, his concept of marriage grows in content and sophistication until, when he is an adult, "marriage" seems like a very complicated, high-level concept. "How can one understand marriage," he might ask, "without understanding romance and property and law?" The answer is that our first understanding of marriage is a crude one based on emotion, but the concept is open-ended, and as we learn about romance or property, we spiral back to "marriage" and deposit this new information into it.
This spiral pattern of learning repeats itself in countless cases. Fairness and honesty, the quantitative aspects of problem solving, the role of observation in discovery---we face these issues in our earliest life experiences, and spiral back to them again and again until our concepts for them become some of our most complicated ones. This spiral pattern of learning emerges because concepts are open-ended repositories for knowledge; we can form a concept with just the barest inkling of the true nature of the concept's referents, and having done this, we can add more and more knowledge to that concept without practical bound.
Guidelines for conceptualizing
When should we integrate a group of existents into a concept? Rand's short answer is "follow the requirements of cognition (and the principle of unit-economy)." Her slightly longer answer is given in the form of a razor: "concepts are not to be multiplies beyond necessity---nor are they to be integrated in disregard of necessity."
Rand also answers the question by example. One example is a list of four general cases for which forming concepts is mandatory: (a) the perceptual concretes we deal with daily, (b) new discoveries of science, (c) new man-made objects, and (d) complex human relationships. Her other examples are concrete. She gives some situations in which forming a new concept would be mandatory---justice, law, and marriage (area (d) from above), and rational spiders from Mars---and some situations in which new concepts would be incorrect---24 year-old blondes, and black swans.
Rand acknowledges cases for which forming new concepts is neither mandatory nor incorrect, cases for which options exist. She describes this area as consisting "predominantly of subdivisions that denote subtle shades of meaning, such as adjectives which are almost, but not fully, synonymous," and says this is the "special province of literary artists," who use such subtle distinctions allow an enormous eloquence of expression.
Another optional area Rand acknowledges is that of classifying borderline cases, existents whose characteristics are equally balanced between the referents of two different concepts. In these situations, there is "no cognitive necessity to classify them under either (or any) concept."
Implicit in Rand's discussion are four guidelines for deciding when to form a concept. These four guidelines are all aspects of the following dictum: form concepts when and only when they are needed. Examining these aspects in detail indicates how the principle of unit-economy guides good thinking.
1. One must integrate existents into concepts when necessary.
This rule governs Miss Rand's identification of the four mandatory areas of concept formation. When one deals regularly with a group of existents on a daily bases, thinking about these existents in terms of visual images or lengthy verbal phrases is clumsy at best. The principle of unit-economy demands that we integrate these existents into a concept so we can refer to them easily and so we have a repository for storing knowledge about them.
2. One must not integrate groups of existents into concepts arbitrarily.
Rand illustrates the problem with violating this rule by pointing to "certain schools of biology and psychology, whose false definition of the concept 'learning' has led to attempts to equate the 'behavior' of a piece of magnetized metal with the 'behavior' of man."
Violating this rule nullifies the basic function of concepts as repositories for knowledge, since there are no common characteristics to deposit into the concept. Needless to say, this prevents concepts from serving their function of reducing the units. Rand goes further, and says that arbitrary groupings obliterate the meaning of the concepts so integrated because it obliterates the essential differences between them. This, she says, is the basis of anti-concepts: when we integrate, say, generosity and self-sacrifice into the concept of altruism, we are not only stuck with the false concept of altruism, but we obliterate the meanings of generosity and self-sacrifice in the process.
3. One must subdivide concepts when necessary.
Rand illustrates this rule with the example of giant, rational spiders from Mars. The point of this example, I think, is that we start with a group of existents known as the "rational animals," which seems to be coextensive with human beings, and then discover a new group of rational animals: spiders from Mars. At this point, we are required to break the set of rational animals into two concepts---human beings and Martian spiders---rather than try to integrate the spiders into the single concept of human being.
The operative principle here is economy of cognitive effort. If we did try to integrate these existents into a single concept, much of the knowledge stored in it---eg, number of legs, preferred diet, arrangement of internal organs---would be bifurcated: one case for the Martians and another for the Earthlings. Such a complication would surely "overload the crow," and thus violate the principle of unit-economy. There are essential differences between the two groups, differences that require special study and specialized knowledge, and having two concepts facilitates this study and simplifies the knowledge.
4. One must not subdivide concepts arbitrarily.
Rand illustrates this rule with two examples: it would be wrong to invent special concepts for 24 year-old blondes or black swans. The problem with unnecessary subdivisions is that they "lead to senseless duplication of cognitive effort:" everything of significance discovered about the units of the subdivision would apply to the wider group as well, a waste of cognitive effort. Also, every time we used the needless concept, we would also have to use the wider concept, another inefficiency. A piece of knowledge should be filed under as few concepts possible, rather than separately filed under many concepts.
Rand seems to allow more freedom when subdividing adjectives than when subdividing nouns. It's not exactly clear why. One reason Rand gives is that the existence of a wide range of overlapping but subtly different concepts facilitates unit-economy by allowing an enormous compactness of expression. One might speculate another reason is that nouns, more that adjectives, are the study of continual study and learning over a life time, and therefore the mental "cost" of forming overlapping nouns is higher than that of overlapping adjectives.
In Chapter 5, Rand observes that one's cognitive context can influence one's definitions, and I'm sure a similar point holds here. Specialists in the sciences and people with other specialized needs have to make distinctions finer than the layman needs. However, rule 4 reminds us that there is a danger when we do this. When we make fine distinctions, we must be careful to think at the right level of abstraction, to think in terms of the most abstract subdivision that still permits specialized study. An unfortunate example of failing to avoid this danger comes from my technical education. My Professors often get caught up in the specialized subdivisions with which their courses deal. The bad results are manifold; for example, the big picture, the overall purpose, is missing, and thus the material begins to float; more relevant to this discussion, fruitful connections to other fields and material we might already know from other courses are not made.
It is interesting to think about the relationships between the four rules above, the principle of unit-economy, and rule of fundamentality. Rand herself says "It is the principle of unit-economy that necessitates definition of concepts in terms of essential characteristics." The rule of fundamentality, then, is the means to the end of unit-economy. The four rules above are also necessitated by the principle of unit-economy (I've tried to indicate how violating the rules would violate unit-economy). Also, the "must-not rules" (2 and 4 above) are closely related to the rule of fundamentality in as much as violating them would violate fundamentality.
 ITOE (2ED), pg 65. This notion of making a collection of knowledge available as "a context" is intriguing. I'm not sure exactly what Rand means by it. Peikoff defines context as "the sum of cognitive elements conditioning an item of knowledge" [OPAR, 123]. Thus, to say that recalling a concept makes its content available as a context seems to mean that by recalling the concept integrates its content into the current "context" of our mind, ie, the knowledge currently conditioning what we are thinking about.
I find the phrase "as a context" evocative of the following image. When we recall a concept, its content (our accumulated knowledge of the characteristics of its referents) does not immediately flood our mind---if it did, it would overload the crow. Instead, the content is brought to the fringe of our awareness, with relevant bits of it ready to spring forward when it might be relevant.
 "Art & sense of life," RM (pb), pg 36; quoted from "Automatization," ARLex (pb), pg 43.
Find Enlightenment at enlightenment.supersaturated.com.