Search This Blog

Saturday, October 23, 2004

Dawkins's Bad Idea: Memes, Genes, and the Metaphors of Psychology

In times past, if the devil didn’t get you, the vapors would, and if they didn’t, humours, poisons, bile or any number of fanciful entities would do you in. With Pasteur and 19th century biology, these agents of illness were replaced with microscopic organisms, and the invention of the disease model made it easy to attribute your aches and pains to malicious bacteria, viruses, or other little microscopic buggers. Of course, then as now, few people understand the actual biological processes that are responsible for disease, but the metaphors for disease do just fine, and have been duly incorporated into the common vernacular. Some may say that these metaphors have worked a bit too well, since they have made a Procrustean stretch to cover all sorts of behaviors, from alcoholism to gambling. Nonetheless, the incorporation of metaphors from the syntax of science does provide a correcting influence to common sense, which earlier had only recourse to metaphors that engaged evil spirits and deadly vapors to help explain the world.

Nowadays, man has access to a wealth of metaphors from modern science. Cancers, black holes, laser beams, and computer viruses have replaced the vitalistic metaphors that assigned causes to ethereal spirits, evil demons, or invisible ethers. Of course, the incorporation of the metaphors of the biological and physical sciences into common language does not entail the ability to map actual processes, but only suggest those processes. The metaphorical description of a cold and its viral causes does not equate with a biochemical or biological description that requires a strict syntax and data language all its own. Thus when we mix our metaphors by talking about rampaging viruses spreading like wildfire from person to person, we know that our description only suggests what viruses are doing in the large, not the actual processes that cause them to propagate and harm.

The mixing of metaphors from different data languages can be poetic insofar as it suggests the juxtaposition or correlation of causes and events, or it can be interpreted as literal insofar as it presumed to denote actual processes. But how do we know if ‘rampaging viruses’ are a literal or a figurative representation of the truth? The ability in science to distinguish the literal from the alliterative is the mark of good science, and good science writing. A physicist may talk about matter using the metaphors of billiard balls, time warps, and cosmic string, but the literalness of those concepts is intentionally undermined by a continuous restatement of the mathematical metaphors that belie their literal reality. So why does a physicist engage two sets of metaphors when he can participate in his science quite well without the need to postulate billiard ball atoms and the like? It is simply because ‘understanding’ requires it. On the one hand, common sense metaphors are easily understood through their appeal to conceptual domains that we readily perceive (e.g. up, down, fast, slow, hot, cold), but mathematical metaphors (e.g. E=mc2) correct for the tendency to make their existence literal. Well-written books that popularize ‘hard’ physical science all recognize the necessity to utilize two different sets of metaphors that correct the deficiencies of the other. Thus an understanding of the physical world can engage metaphors that are derived from our native experience and those that are derived from the abstract language of mathematics. Understanding consists in our ability to move from once set of conceptual metaphors (e.g. green grass, hot suns, expanding universes) to another (e.g. the calculus, non-Euclidean geometry). It is no less than our ability to shift between different languages that enables us to envision the world.

Sometimes however, two entirely different sets of conceptual metaphors may be quite similar in terms of the processes they describe, and proceed to confirm rather than contrast with one another. Newtonian ideas of acceleration, mass, gravity, force, etc. have long been assimilated into our popular lexicon because a Newtonian view of the universe coheres with our own naïve experience. The mechanical universe of Newton corresponds with common sense theories of the physics of cars, boats, apples, and other physical objects, and is much easier to understand and accept than other more accurate physical theories that are reflected in the conceptual metaphors of Einsteinian relativity and Quantum theory. However, the conceptual metaphors associated with relativity and the invisible quantum correct for the literal interpretation of Newtonian mechanics as a representation of reality, and have entailments (e.g. time travel, multiple universes, quantum indeterminacy) that are dramatically at odds with the Newtonian conception of a clockwork universe.

In the biological sciences, the Darwinian principles of natural selection have an import comparable to Newton, and the conceptual metaphors of evolution find an equal correspondence with common sense theories of human psychology. Common sense or ‘folk’ theories of psychology tell us that we are motivated by ordinary objects (e.g. cars, jewelry, money, sex) whose value we determine consciously, and either impel (as in eliciting drives and reflexes) or compel (as in rewards, reinforcers, or punishers) behavior. In a Newtonian sense, our lives revolve around the collection of ordinary objects that push and pull us to them from a distance, and populate but scarcely our psychological universe.

This common sense explanation of how behavior is selected bridges quite easily to conceptual metaphors that describe how biological entities are selected, and by implication to the behavior that is instantiated by those entities. Thus patterns of behavior that are elicited by instinctive events can be ultimately attributed to individual genetic influences that are objectified in the metaphors of the activities of individual genes. Similarly, the common sense notion that ideas are selected by some obscure competition between objective alternatives also finds an equal bridge to selectionist principles that are derived from biology. Thus, just as Newtonian physics and common sense physics seem to confirm each other, common sense psychology and Darwinian biology share similar metaphorical principles that explain respectively how behavioral and biological selections are made.

As with the blending of metaphors that saw the adoption of Newtonian terms into common sense physics, metaphors from biology and ‘folk’ psychology have also become commingled, and thus form a new explanatory framework for behavior that explains behavior as a Darwinian process. Thus genes become ‘selfish’, and ideas or memes become ‘contagious’. But these are only two different levels of thinking, and do not implicate the metaphorical schemes that explain the actual neurological processes that underlie behavior.

Darwinian or sociobiological explanations strongly imply that the molar processes of cumulative selection that led to bumblebees and human beings are isomorphic with neural or ‘molecular’ processes of the mind that lead to the selection of behavior. But an implication is not a demonstration, as a sociobiological explanation merely establishes a similarity between the metaphors of common sense and natural selection. The overriding question is not whether genes and memes represent a good metaphorical bridge between common sense and biological explanations for behavior, but whether they provide an equally good metaphorical representation of the biobehavioral processes that instigate behavior. The question is analogous to that posed by Quantum physicists to the common sense and academic views that extended the metaphors of Newtonian physics to the molecular world of the atom. The answer to that question was not a reaffirmation of the perspective of atoms as mere baby solar systems, but of the creation of an entirely new science that was equally rooted in the empirical tradition of science. That science was quantum physics.

A biobehavioral explanation of behavior represents the mapping of the actual neurological processes that comprise behavior to the patterns of information or environmental contingencies that parallel and elicit them (Donahoe and Palmer, 1993). This information is in turn mediated by somatic events that are perceived as emotion (Damasio, 1996), activating neurological events that comprise attentional processes (Donahoe and Palmer, 1993), and nativistic (i.e.inborn or genetic) sensitivities to certain abstract patterns of information (Bolles, 1976). Whereas a biobehavioral explanation is neurally realistic because it ties behavior to actual neural and informative events, a sociobiological explanation is neurally unrealistic, and merely substitutes neurological processes with Darwinian metaphors.

Biobehavioral science, which is also known as 2nd generation cognitive science (Lakoff, 1999), or theoretical behaviorism (Staddon, 1990), is like evolutionary psychology entirely informed by evolutionary principles. However, it is more rigorously empirical because of its insistence on ultimately observing or reliably inferring the neural processes that intervene between information and behavior. Biobehavioral and evolutionary psychology are represented by entirely different sets of conceptual metaphors that are respectively entailed by molecular (small scale processes and time frames) and molar (large scale processes and time frames) processes. The polarity of these metaphors is remarkable, and can be reduced to the following contrasting principles.

Evolutionary psychology and folk psychology share the implicit presumption that decision making is generally based on the conscious and disembodied appraisal of ordinary objects that lead to the maximization of our self interest. In contrast, biobehavioral psychology has demonstrated that most reasoning is not conscious but nonconscious (Lewicki, 1992; Greenwald, 1992), and is guided by embodied non-verbalized somatic (Damasio, 1994) and neural activation events (Donahoe and Palmer, 1997) that are ‘just as cognitive as any other perceptual image’ (Damasio, 1994). Because nonconscious embodied reasoning is computable but not directly accessible by conscious reasoning, we often find that our conscious reasoning about what is a ‘best outcome’ conflicts with our nonconscious determination of ‘best outcome’. Thus there is no univocal or self-consistent locus of value (Lakoff and Johnson, 1999). (Remember this next time you reluctantly try to get up in the morning.) Values are due to the binding of many information streams that are mediated by disparate neural and somatic processes, and motivate behavior in real time as they are perceived, and not when the physical or objective entity that denotes such information is attained. (In other words, it’s the thought that counts) Thus value is not found in some conceptual object like a meme, but in separate threads of information that are mediated by the mind and body that individually have salience to an individual and are perceived presently, independently, virtually, and for the most part nonconsciously. In other words, the concept of ideas as compartmentalized memes leads us to find value in the obvious topographical or ‘surface’ aspects of an idea, whereas it is the non-obvious abstract properties that are actually selected. Because value resides in information that is often incognizant to us, it cannot be subject to the economic models that are based on the rationing of value according to some single utilitarian measure, or the mathematical decision models such as game theory that conceptualize value simplistically as being no more than material wealth.

The definition of a meme as an independent conceptual object is ultimately not simple, but simplistic, since it does not denote the web of informative relationships between behavior and and the environment that is denoted by consciously and nonconsciously by the mind and body proper. For example, the concept of the sport of football is a well traveled meme to be sure. Football represents a rather involved information pattern that has infected the minds of young men nationwide, and football games, commentary, and assorted chatter has parasitized not only the minds of people, but the network airwaves, the written media, and many unwilling housewives. But is a football game an indivisible meme like entity, or is it somewhat different than the sum of its parts? Actually, the ‘meme’ of football is not a singular information pattern that replicates like a strand of DNA, but rather emerges from a web of separate patterns of information that are mediated not only by consciously perceived information but by neural and somatic activating processes that we otherwise call emotion. The meme of football is not just a compendium of rules, but comprises the memory of the somatic responses that occur while watching (excitement, depression), the natural feeling of elation that occurs with a high state of alertness, the virtual extension of control over all those partisans of the losing team, the constantly changing and stimulating prediction error that occurs as one play after another unfolds, the smell and taste of hot dogs and beer, the camaraderie of friends, and so on. The meme of football is in other words a web of perceptual relationships that is volatile and constantly changing. Moreover, different aspects of the meme football may be present in one circumstance, and not present in another. Watching your team lose at a hostile stadium on a rainy day is a whole lot less rewarding than if you were watching your team at home while among friends.

The most important distinction contrast between evolutionary psychology and bio-behavioral psychology is that bio-behavioral psychology denotes value not in the assimilation of ideas or memes, but in changes in the relationships between memes, or behavioral discrepancies (Donahoe and Palmer, 1993). To explain this, we must understand first how a meme does not reproduce.

Although a meme represents a self replicating packet of information, unlike a virus it possesses no internal instructions that secure its influence on behavior, let alone its retention in memory. Memes or ideas take root in memory because they are rehearsed, and they are rehearsed because of their contingent relationship to a myriad other ideas that comprise the stimulus context of a behavior. This idea of contingency is critical to the methodology of modern behaviorism, and underlines the fact that it is not ideas alone that motivate, but the dependencies between ideas. The meme of a fishhook for example hardly comes to mind until it is perceived as part of a means-end (memes-end?) expectancy. We think of fish hooks because of the fish it can catch, but to even think about fishing one must also think about the time, place, and equipment that allows one to fish. If any of these events fail to take place, there is hardly a need to think about fishing, or for that matter the meme of fishhooks. An atomized universe of memes does not implicate the contingent relationship between ideas that secures the rehearsal and retention of a ‘good idea’. Behavior is elicited not by individual memes but by global maps of means-end expectancies that are constantly changing, and are in general non-consciously perceived. However, what causes us to think about fishhooks, fishing trips or other ideas is the fact the relationships they denote are selected and are mentally rehearsed. We constantly think about a fishhook as it winds its way from our tackle box to the end of our fishing line because in every moment the relationship between the fishhook and the line changes, and it is the change that gains our attention. Thus, we select not only memes, but also the abstract relationships between memes as they are moderated by our thoughts and overt behavior. Ultimately, as Alexander the Great found out when he wept upon having no more new worlds to conquer, what motivates is not the end, but in the traveling.

Unfortunately, Darwinian and common sense models can no more describe the molecular ‘environment-behavior’ relationships that comprise human motivation than a weatherman’s description of an impending cold front describes how a snow storm forms over your head. We would err in using a molar analysis (cold fronts) to describe molecular process (the formation of clouds) because the inherent processes implied by storm systems and storm clouds are different. Likewise, the human brain is a massively parallel biological computer, and metaphors from information processing are far more apt than biological metaphors that liken ideas to viruses and their spread to contagion, let alone the metaphors from common sense that posit a disembodied objectified reasoning. The lack of ‘fit’ of Darwinian and common sense metaphors to bio-behavioral science does not invalidate the selectionist principles that inform all of the sciences. But it does point out the level confusion that occurs when a set of principles from one level of understanding (biology) are invoked not just to explain another (an in the juxtaposition of the metaphors of cosmic string and mathematics) but to embody another. In other words, because the uses of memes and genes is not corrected by an understanding of how the brain as a neural system actually works, the metaphorical conception of memes and genes can easily be seen as not just figuratively real, but literally real.

The ultimate danger in assigning a literal reality to the means-end rationality imposed by utilitarian memes and genes is that it implies that we implicitly know what is in our best interest. Furthermore, the convergence of the metaphors of common sense and Darwinism reinforce the idea that value is objectified and is a limited commodity, and must be allocated to those who are most fit to achieve it. In this way, a meme world becomes ‘mean’ world, wherein our memetic impulses robotically drive us forward to achieve our goals, with the long term survival of our genes and memes being the only necessary outcome.

In contrast to this cold and sterile vision, bio-behavioral psychology defines value not as a scarce commodity, but in the creation of information that is consumed virtually. That is, if value is denoted in abstract informative relationships between ideas, then it is prospectively unlimited, and is constrained not by our inability to manufacture physical things, but by our ability to create and perceive information. But perception requires the skills that enable us to mentally model the world, from the implications of the cheers of a crowd in a football game to the thoughts of a proud parent. To experience the world is to model it, and that is nothing more than empathy. Universal empathy allows us to expand and enhance the rewards we perceive, but it also constrains our behavior due to the virtual penalties (e.g. shame, embarrassment) we perceive. Cultures that understand that value derives from the development of empathy will take an entirely different course than the materialistic societies that posit value as the accumulation of objects. Indeed, information is more economically produced by a societal exaltation of sports, art, literature and music than by the manufacture of a new prestige automobile. Ironically, the lasting legacy of a psychology that is informed by evolutionary principles is not the amoral world driven by the erroneous metaphors of selfish genes and infectious memes, but by the evolutionary mandate of an expanding empathy, and our innate interest in the cultivation of beauty.

Bolles, Robert C. (1976) Theory of Motivation. 2nd ed. New York: Harper and Row
Damasio, Antonio R. (1994) Descartes Error: Emotion, Reason, and the Human Brain. Avon

Greenwald, A.G. (1992) Unconscious Cognition Reclaimed, American Psychologist, 766-775
Lewicki, P., Hill, T., & Czyewska, M. (1992) Nonconscious acquisition of information, American Psychologist, 47, 796-801
Donahoe, J. W. and D. C. Palmer (1993) Learning and Complex Behavior, Allyn and Bacon
Donahoe, J. W., D. C. Palmer, and Jose E. Burgos (1997) The Unit of Selection: What do reinforcers reinforce?, Journal of the Experimental Analysis of Behavior, 67, 259-273
Donahoe, J. W. (1997) Neural-Networks Models of Cognition, J. W. Donahoe and V. Packard Dorsel (Eds.)
Lakoff, George, and Johnson, Mark (1999) Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought, Basic Books
Staddon, John (1993) Behaviorism , Duckworth


(the best introduction to bio-behaviorism comes from Shull’s article on the website of the Journal for the Experimental Analysis of Behavior, and the scholarly commentary also at the JEAB website that discussed this new school of behaviorism. An understanding of metaphor and how it heavily influences ideas in evolutionary psychology can be found on the many web sites that discuss the work of the cognitive linguist George Lakoff, and in particular his new book: Philosophy in the Flesh: The Embodied Mind and its challenge to Western Thought)

No comments: