Swarthmore College, Psychology 28: Introduction to Cognitive Science: Spring 1995
Professor Dufour

Teleological Concept Formation

This view of the embodied mind takes John Searle's position that "mental phenomena are caused by neurophysiological process in the brain and are themselves features of the brain." Edelman and Rand share this view that physiological processes cause mental phenomena, but do not constitute them.

The point here is that one's consciousness is not objective, i.e., its nature is not independent of one's consciousness. Its nature is the nature of one's consciousness. Hence, consciousness is irreducible to objective physical entities. Consciousness involves subjective personal mental states. Reducing consciousness to objective existents such as a neuron firing makes it objective. But the very nature of consciousness is that it is subjective.

The phenomenon of consciousness exists whether I am conscious of it or not. Yet, the phenomenon of consciousness itself, is conscious, i.e., in order for consciousness to exist, something must be conscious. Consciousness is irreducible to neuron firing. We can observe neurons firing, but that does not mean we are observing a conscious entity. The neurons are not conscious in themselves. Consciousness is a result of certain patterns of firings and neural connections, etc. Searle compares this to water in a wave. We can see a water wave, but we can't reduce the wave to the behavior of the individual water molecules in the wave.

Much of cognitive science tries to treat states of consciousness as if they were objectively observable. This approach leads to functionalism and the idea that the mind is a disembodied computer-like information processor of mental representations. It treats perception as the input of categorized information to the brain for processing. This discussion will show that we do not perceive a categorized world. The mind categorizes perceptions and gives them meaning based the relationship between the object and subject's life.

For example, poker players have concepts of, and can recognize flushes, straits, and full houses, etc. Why? They need those arrangements of cards to "survive", and/or flourish in their environment of the poker game. They do not readily notice other sets of cards, e.g., prime numbers of the same suit. If they were valuable, then they would categorize accordingly.

Analyzing Darwinian theory, Harry Binswanger notes shared characteristics of biological functions and conscious purposes. Organs and faculties have value-significance for the organism and exist for this purpose. (Vestigial structures notwithstanding, comment, September, 2001) and I will approach the human faculty of concept formation from this perspective.

The biological theory I will use in this paper is the Theory of Neuronal Group Selection (TNGS), as proposed by Gerald Edelman. I will discuss this theory as it applies to concept formation. From the biological basis, I will investigate the Objectivist theory of concepts formulated by Ayn Rand furthered by David Kelley.

Fundamental to the TNGS is what Edelman calls Neural Darwinism: the brain is affected by two types of selection: natural and somatic. Edelman introduces a few biological terms when speaking of evolution. One is "recognition,...the continual adaptive matching or fitting of the elements of one physical domain to novelty occurring in elements of another, more or less independent physical domain, a matching that occurs without prior instruction." The point is that no information is transferred from the two domains: an organism and its environment, i.e., the mind is not an information processor. "Evolution works by selection, not by instruction." Consciousness appeared as a result of natural selection or organisms over generations, and through somatic selection during an organism's life, recognition systems such as the immune system and, as Edelman proposes, the brain. Humans beings are animals adapting to their environments, one method is concept formation.

The TNGS is Edelman's theory of the brain as a selectional recognition system. Any such theory, he says, must be is biologically based and bridge the gap between psychological and physiological processes. This theory includes concept formation.

In describing the TNGS, I will start at the neural level of the human brain. The first tenet of the TNGS is that of Developmental Selection. Here, populations of neurons engage in biological competition. Populations of variant neural groups comprising neural networks in a brain region arise by somatic selection. These neural networks comprise the primary repertoire.

From the primary repertoire, Experimental Selection, consisting strength change of synapse population, yields the secondary repertoire. During behavior, biochemical processes selectively strengthen and weaken synaptic connections. This mechanism, underlying memory and other function, selectively carves out a variety of functioning circuits. A set of these circuits is called the secondary repertoire. Mechanisms involving the primary and secondary repertoires are sometimes intermixed.

Reentrent mapping is the third tenet of TNGS. It describes how the events of the first two stages connect psychology to physiology, and emergent brain areas in evolution coordinate with each other to yield new functions. Such functions require that the primary and secondary repertoires form neurally connected maps. Reentry is the process by which brain maps interact. Maps can have functions corresponding to color, motion, orientation, etc.

This is a theory of neuronal group selection because selection and reentry mapping occur with groups of neurons. No individual neuron has the properties alone that it shows in a group.

Edelman's proposal on concept formation remains to be tested, but is deduced from his well tested TNGS. He cites three fundamental higher brain functions: perceptual categorization, memory, and learning, the last depending on the first two. Also required for learning is a connection to value systems mediated by brain regions other than those carrying out categorization. These regions are the value-laden brain structures (hypothalamus, various midbrain nuclei) regulate the body's homeostatic, appetitive, and consummatory needs. I must emphasize here that consciousness is an active process, and to varying extents, concept formation requires conscious attention to percepts their relationships.

Learning is the result of neural linkages operating between global mappings* and the brain's value centers. Physiological systems have set points, which when not yet satisfied, are in a condition of expectancy. Learning connects categorization to behaviors have adaptive value under expectancy conditions. It is a result of behavior leading to synaptic changes in global mappings satisfying these set points.

Edelman discusses memory in much biological detail. He cites the hippocampus as responsible for linking categorizations of immediate and stored perceptions. This, and other structures including the cerebellum, basal ganglia, arose through evolution, and yielded new brain functions and types of memory. These structures, called the organs of succession, deal with timing, succession in movement, and memory establishment. They order the perceptual categorizations performed by the cerebral cortex.

An organism with a conceptual faculty, Edelman says, changes his behavior in reaction to the identification of an entity. There are relational recognitions connecting one perceptual categorization to another. They need not be related or be in the animal's perceptual range. Such abilities developed in evolution before speech. According to the TNGS, the emergence of conceptual abilities depends on the evolution of specialized brain areas (probably the frontal, temporal, and parietal cortices).

These brain operations involve the brain's constructing maps, not of external stimuli, but its own activities. They categorize parts of past global mappings according to modality, movement, and relationships between perceptual categorizations.

The frontal cortex is "a prime example of a conceptual center in the brain," Edelman writes. It contains reentrently connected and functionally segregated maps subserving sensory modalities and motor responses. Its connections to the basal ganglia and the limbic system relate value categorization and sensory experiences. Values affect conceptual memories.

The Objectivist theory of concepts is compatible with the TNGS. Consciousness is the interaction between our senses and entities resulting in our awareness of what we perceive. As a state of awareness, consciousness is an active process consisting of differentiation and integration. These two processes, we will see, play a complimentary role in concept formation.

Concepts as Mental Entities

One forms a concept by isolating groups of percepts based on perceived similarities (sharing a characteristic) distinguishing them from other existents. "A concept is a mental integration of two or more units possessing the same distinguishing characteristic(s), with their particular measurements omitted." One then selects a perceptual symbol, a word, to designate the concept.

Elaborating on her notion of a concept, Rand specifies that a concept is the product of a process of integration. The integration of elements involved in the process of concept formation results in a mental entity, a phenomenon of consciousness, a concept.13

That a concept is a mental entity does not mean it has a physical form. It is a mental state. Kelley writes that a concept, unlike a percept, "is not a conscious cognitive state with a content that can be described phenomenologically." He concludes that a theory of what concepts are and how they are formed area package deal. Applying these ideas to Edelman's neuroscience, a concept is a mental state resulting from the brain's mapping its own activities.

Concept Formation

The perception of similarity is the basis of concept formation. Similar objects posses the same characteristic(s), but in a different measure or degree. Such commensurable characteristics possess a common unit of measurement. For example, things can have a similar color, an attribute measured by its wavelength. Rand designates the commensurable characteristic as the "Conceptual Common Denominator (CCD)." The term measurement omission means that of commensurable characteristic's quantitative dimension must exist in some quantity, but can exist in any quantity.

David Kelley notes that the perception of differences among objects facilitates the perception of their similarities. Two objects, A and B, will appear similar to each other in respect to their difference to a third object, C. Actually, A and B are less different from each other than either are from C. Since A and B are similar, they differ quantitatively in their CCD measurement. Both differ from C in qualitatively, as C lacks the commensurable characteristic that A and B share.

Yet, Kelley stresses, the "distinction between qualitative and quantitative differences is itself quantitative-a difference in degree of difference." If the CCD is "hue," one may discriminate objects with qualitative differences in color, but all such colored objects will have a quantitative difference in wavelength if the CCD is changed to "opaque/clear." Categorization depends on one's context. As the size of the CCD increases, e.g., from red to color to opaque/clear, the differences between objects alternate between qualitative and quantitative.

Biologically, the first stage of concept formation corresponds to the brain's capacities for perceptual categorization. The TNGS provides a biological basis for perceptual categorization, the connection between physiology and psychology. The functions and activities of one map are connected and correlated to those of another through the strengthening of reentrent signaling between neural groups two different maps. This forms a "classification couple" consisting of two maps receiving independent signals from the world.

Perceptual categorization occurs through a global mapping: the coupling of outputs of multiple maps reentrently connected to the animal's sensory-motor behavior. Such sensory-motor activity over the whole mapping selects neuronal groups giving the appropriate output or behavior, resulting in categorization.

Abstraction is the second stage of concept formation. After one differentiates objects from others based on similarity, the cognitive process of abstraction yields a concept. From an awareness of distinct and particular objects, one forms a concept subsuming them that is abstract and universal. Measurements omitted, the concept applies to all entities satisfying the CCD(s).

Concepts are abstract (rather than determinant) because, as a result of measurement omission, they treat classes of objects as if they were identical. Concepts are also universal (rather than particular), meaning that they subsume an indefinite number of entities under the concept's meaning. Abstraction entails regarding units of a group of similar objects as identical.

Abstraction involves two modes of attention. In the differentiating stage of concept formation is to distinguish measurements of a CCD in objects. The integrative element in the first mode of attention is the awareness of their dimension of similarity, the commensurable characteristic. We perceive such similarities. The second mode of attention is measurement omission. Upon awareness that the units of the different objects are the same, we omit measurements of these units and focus on the common aspects of the objects. "Cognitively, a concept allows for us to separate the way in which two objects are the same from the way in which they are different (even though they are the same and different in the same respect).

Cognitive Role of Concepts

Rand cites the human conceptual faculty as the source of our cognitive power. Our reasoning ability is our distinctive means of survival, and concept formation is essential to it. Guiding it is the principle of unit economy, the reduction of vast amounts of information to a minimal number of units. The conceptual faculty allows people to overcome the limited range of what one can be conscious of at a given time. "Conceptualization is a method of expanding man's consciousness by reducing the number of its content's units-a systematic means to an unlimited integration of cognitive data...[C]oncepts represent condensations of knowledge, which make further study and the division of cognitive labor possible."

What determines the concepts we form? Existents vary in their descriptive complexity, the frequency of their use, and the value of their further study. Many existents may have descriptive complexity and/or frequent use. But their value to a person will determine whether he forms concepts to maximize unit economy.

The TNGS holds that an evolutionarily selected internal criterion of value constrains the domain of categorization. Brain regions regulating bodily functions exhibit this value system. "Categorization manifests itself in behavior that appropriately fulfills the evolutionarily selected requirements of such life-supporting physiological systems." Edelman holds categorization to be value based. It is epigenetic, i.e., the criterion of value evolved through natural selection, but the experience dependent categorizations rise through somatic selection of neuronal groups. Epigenetic phenomena occur only if certain previous events have taken place.

Both Edelman and Kelley cite Rosch's work dealing with family resemblances and typicality of entities in categories. The typicality phenomenon reflects a subject's ability to recognize new instances of an already formed concept. Humans form categories based on their values, but also on their culture and environment.

Definitions of concepts are also value based. The purpose of a definition is to help man organize and interrelate his concepts and to distinguish them from each other. Defining a concept consists of determining "the characteristic(s) of its units upon which the greatest number of their known characteristics depends, and which distinguishes the units from all other known existents" in the context of one's knowledge. These characteristic(s) are termed "essential," yet are not intrinsic to entities, i.e., metaphysical. They are epistemological, d are based on objective properties of existents.

Despite its definition, the meaning of a concept is still all the characteristics of the existents it subsumes. For example, using "rational animal" as the definition of man means that man is an entity, including all of its characteristics. In the present context, humans are "most fundamentally distinguished from all other entities by that fact that it is a rational animal." The meaning of the concept "man" still includes all other characteristics.

In his analysis of Rand's theory of concepts, Wallace Matson gives an example of how we arrived at the definition of the element "sulfur." First, sulfur may have been defined by its color and odor, and then as knowledge of chemistry grew, the definition became "a nonmetallic element with the atomic weight of 32.06." Since there could be a substance with the same characteristics as sulfur but with a different atomic weight (isotopes), this definition was not correct. Yet, with the development of the periodic table, sulfur was described by its electron configuration, and had the atomic number 16. Sulfur became defined by its atomic number. The designation accounts for all of sulfur's properties, which can be deductively derived from its definition.

This illustrates that a concept is an open-end classification including yet-to-be discovered characteristics of the existents it subsumes. If they were not, people would not be able to expand their knowledge and build upon previous experience, an essential scientific progress. Further, definitions are not arbitrary, the values humans must pursue, technology, in this case, determine them. These aspects of concepts and their definitions connect them to values.

A further implication of our value based categorizing is that science is not value free. The nature of science entails its being value laden (and that is a good thing). Humans build scientific theories through value based categorization and conceptualization. The theories we make, and the patterns we see in the world are based in our quest to flourish.

In this sense, the mind is not a computer. Humans do not see the world in precut categories with metaphysical essences. From the Objectivist point of view, essences are epistemological. They give our thinking unit economy. Edelman agrees with the idea of epistemological essences. We are not Turing Machines, and the world is not a piece of digital tape with innate meaningful categories. Our basic means of awareness is on the perceptual level. We categorize percepts according to our values.

In this discussion I have accounted for a teleological basis for concept formation. Human beings form concepts to survive and actualize their potential according to their nature as rational animals. This goal was the driving force behind every step of the process, from categorization, abstraction, and the assignments of words and definitions. The mind's embodiment is essential for understanding it, as its functions serve to sustain its existence.

References

Edelman, Gerald. 1992. Bright Air, Brilliant Fire: On the Matter of the Mind. New York: Basic Books.

Kelley, David. 1984. A Theory of Abstraction. Cognition and Brain Theory 7(3 &4): 329-357.

Matson, Wallace, 'Rand on Concepts,' The Philosophic Thought of Ayn Rand, Douglas J. Den Uyl and Douglas B. Rasmussen, ed. Chicago, University of Illinois Press, 1986.

Mozes, Eyal. 1995. Reality of Mind. Objectivity 2(1):93-107.

Piekoff, Leonard. 1979. The Analytic-Synthetic Dichotomy. Introduction to Objectivist Epistemology. New York, New American Library.

Rand, Ayn. 1990 [1979]. Introduction to Objectivist Epistemology. Expanded 2nd ed., H. Binswanger and L. Piekoff, editors. New York: Meridian, Penguin Books

(characteristic(s) are termed "essential," yet the essences are not intrinsic to entities, i.e., metaphysical. They are epistemological,