Schismogenesis, then, is learned: if is as much an acquired habit as is the nonschismogenic behavior characteristic of Bali. Yet it seems so fundamental that we are forced to ask what learning itself consists of, if it so inseparably links cognition and emotion (eidos and ethos). What does it mean to learn something, to "know" something? After the Macy Conferences on cybernetics, Bateson made this question the subject of his next major investigation.
Bateson began his study of learning theory with an ostensibly nonsensical question: Is there such a thing as a "true error"? More broadly, is there such a thing as a true ideology? Ideologies are cultural things learned in cultural contexts, but they usually work for the cultures that believe in them. The Balinese believe certain things about the world that to us, or to the Iatmul, seem almost inconceivable. Bateson had regarded cumulative interaction as an inherent trait, but Bali showed him that an entire nation could learn to do something quite different. Furthermore, Balinese society was far more stable than Iatmul or Western European society, and thus in some sense its "crazy" premises had to be more true. Put in this way, the crucial question became, How are ideologies (perceptions, world views, "realities")and emotive patterns (dominance/submission, succor/dependency) formed within the mind of an individual or his society? In response, Bateson followed Benedict's notion of configuration and returned to the concept of the "grammar," or code. Individuals and societies are organized entities; they are "coded" in a certain way that is coherent, that makes sense in both emotional and cognitive terms. Since it was this process of coding which rendered them stable (so long as the code continued to work), it was essential to explain that process more fully.18
Down to some point in the mid-1960s, learning theory was dominated by the behavioral model, which is most commonly associated with J.B. Watson and B.F. Skinner. The real grandfather of such work was Ivan Pavlov, who had managed to immortalize a dog by getting it to salivate when he rang a bell. What Pavlov did was to set up a context of association. Repeatedly, bell-ringing was followed by food, until the sound of the bell alone was enough to trigger the animal's entire gastronomic response. In one of Skinner's experiments, a rat learned to press a bar and thereby release a pellet of food. Skinners rat had to contend with a set of rules different from those that confronted Pavlov's dog, but again a context of (causal) association was central: event occurs, food appears. Furthermore, all of these experiments involved a progressively faster rate of learning on the part of animals. Dog and rat quickly caught on to the rules of the game. After a number of trials, the dog did not need meat to salivate; he had learned what the bell meant. Similarly, the rat discovered that the food pellet was no accident, and began to spend much of its time pressing the bar.
What is going on in .such experiments? What do the terms "learn" and "discover" mean, as I have just employed them? Bateson uses the term "proto-learning" to characterize the simple solution of a problem. Bell rings, or bar is presented. The Pavlovian situation requires a passive response, the Skinnerian situation a more active one, but there is still a problem to be solved in each case: what does this phenomenon require of me (dog, rat), and what does it lead to? Solving such a specific problem is proto-learning, or Learning I. "Deutero-learning," or Learning II, Bateson defines as a "progressive change in [the] rate of proto-learning." In Learning II the subject discovers the nature of the context itself, that is, he not only solves the problems that confront him, but becomes more skilled in solving problems in general. He acquires the habit of expecting the continuity of a particular sequence or context, and in so doing, "learns to learn." There are, furthermore, four contexts of positive learning, as opposed to negative learning, in which the subject learns not to do something. There are the two already described, Pavlovian contexts and those of instrumental reward; and there are also contexts of instrumental avoidance (e.g., rat gets an electric shock if it doesn't press the bar within a certain time interval) and of serial and rote learning (e.g., word B is always to be uttered after word A). So proto-learning is the solution of a problem within such contexts, and deutero-learning is figuring out what the context itself is -- learning the rules of the game.
Character and "reality" have their origins in the process of Learning II; indeed, character and reality prove to be inseparable. A person trained by a Pavlovian experimenter would have a fatalistic view of life. He would believe that nothing could affect his state, and for such a person reality might well consist of deciphering omens. A Skinnerian-trained individual would be more active in dealing with his or her world, but no less rigid in his or her view of reality. Western cultures, notes Bateson, operate in terms of a mixture of instrumental reward and avoidance. Its citizens deutero-learn the art of manipulating everything around them, and it is difficult for them to believe that reality might be arranged on any other basis. The link between fact and value is (a) that such acquired perceptions are also acquired character traits, and (b) that they are purely articles of faith. In other words, to take (a) first, any bit of learning, especially deutero-learning, is the acquisition of a personality trait, and what we call "character" (ethos, in Greek) is built on premises acquired in learning contexts. All adjectives descriptive of character, says Bateson -- "dependent," "hostile," "careless," and so on -- are descriptions of possible results of Learning II. The Pavlovian-trained person not only sees reality in fatalistic terms; we might also say of him or her, "She is fatalistic," or "He is a passive type." Most of us raised in Western industrial societies have been trained in instrumental patterns, and therefore we do not ordinarily notice these patterns: they constitute our ethos. They are "normal," and thus invisible. In especially egregious cases, we will say, "He's only out for himself" -- a character description that is at the same time an epistemology. Dominant, submissive, passive, self-aggrandizing, and exhibitionistic -- all are simultaneously character traits and ways of defining reality, and all are (deutero-) learned from early infancy.
The second point, that these "realities" are articles of faith, raises the issue of the "true ideology." If you have been raised with an instrumental view of life, you will relate to your social and natural environment in that way. You will test the environment on that basis to obtain positive reinforcement, and if your premises are not validated, you will probably not abandon your world view, but classify the negative response, or lack of response, as an anomaly. In this way you remove the threat to your view of reality, which is also your character structure. Neither the witch doctor nor the surgeon gives up magic or science when his methods fail, as they often do. Behavior, says Bateson, is controlled by Learning II, and molds the total context to fit in with those expectations. The self-validating character of deutero-learning is so powerful that it is normally ineradicable, usually persisting from cradle to grave. Of course, many individuals go through "conversions" in which they abandon one paradigm for another. But regardless of the paradigm, the person remains in the grip of a deutero-pattern, and goes through life finding "facts" that validate it. In Bateson's view, the only real escape is what he calls Learning III, in which it is not a matter of one paradigm versus another, but an understanding of the nature of paradigm itself. Such changes involve a profound reorganization of personality -- a change in form, not just content -- and can occur in true religious conversion, in psychosis, or in psychotherapy. These changes burst open the categories of Learning II itself, with magnificent or hazardous results. (We shall deal with Learning III at greater length below.)
It should be clear, then, that the union of fact and value, which modern science denies in principle, occurs quite naturally in Bateson's analysis of learning. A system of codification, he says, is not very different from a system of values. The network of values partially determines the network of perception. "Man lives by those propositions whose validity is a function of his belief in them," he writes. Or as he says at a later point, "faith is an acceptance of deutero-propositions whose validity is really increased by our acceptance of them."
But what
is character structure? If it was an error to reify ethos in New Guinea, Bateson realized, it was no less fallacious to treat a character trait as a thing. Adjectives descriptive of character are really descriptions of "segments of interchange." They are descriptions of transactions, not of entities, and the transactions involved exist between the person and his or her environment. No person is "hostile" or "careless" in a vacuum, despite the contrary contention of Pavlov, Skinner, and the whole behavioral school. Clearly, Learning II is equivalent to the acquisition of apperceptive habits, "apperception" being defined as the mind's perception of itself as a conscious agent. Such habits can be acquired in more than one way, and the behaviorist is wrong to believe that habit is formed only through the repeated experience of a specific kind of learning context. "We are not concerned," writes Bateson,
with a hypothetical isolated individual in contact with an impersonal events stream, but rather with real individuals who have complex emotional patterns of relationship with other individuals. In such a real world, the individual will be led to acquire or reject apperceptive habits by the very complex phenomena of personal example, tone of voice, hostility, love, etc. Many such habits, too, will be conveyed to him, not through his own naked experience of the stream of events, for no human beings (not even scientists) are naked in this sense. The events stream is mediated to them through language, art, technology and other cultural media which are structured at every point by tramlines of apperceptive habit.19
The psychology laboratory is probably the last place to learn about learning, just as the physics laboratory is the last place to learn about light and color. Both Skinner and Newton were guilty of narrowing the context to the point that they could have precise control over the trivial. If you wish to find out about learning, contends Bateson, study individuals in their cultural context, and study especially the non-verbal communication that goes on between them. Deutero-learning proceeds largely in terms of what he would later call "analogue," as opposed to "digital" cues. It is in this arena that we shall, he believed, find the source of our character "traits" and our cognitive "realities."
To enlarge on this for a moment, digital knowledge, which expanded rapidly after Gutenberg s time, is verbal-rational and abstract. For example, a word has no particular relationship to what it describes ("cow" is not a big word). Analogue knowledge, on the other hand, is iconic: the information represents that which is being communicated (a loud voice indicates strong emotions). This kind of knowledge is tacit, in Polanyi's sense, and includes poetry, body language, gesture and intonation, dreams, art, and fantasy. Pascal and Descartes had debated this distinction between style and nuance on the one hand, the measurement and geometry on the other. Although at first glance these two forms of knowledge may seem irreconcilable, Bateson chose to believe that Pascal was right when he wrote that the heart had its reasons which reason did not perceive. Perhaps it was time for scientists to start formulating some cardiac algorithms.
Bateson recalls that it was in January 1952, while watching monkeys playing at the Fleishhacker Zoo in San Francisco, that he realized that their play (the monkeys' captivity notwithstanding) could provide a foothold on the whole area of nonverbal commumcation. The resulting article, "A Theory of Play and Fantasy," argued: (1) that play between mammals dealt with 'relata,' rather than manifest content, and in this way was very similar to primary-process material (or dream and fantasy) in its structure; (2) that although it was not familiar to our conscious minds, such material was subject to the analysis of formal logic, specifically the rules of paradox described by Russell and Whitehead in their classic work, "Principia Mathematica" (1910-13); (3) that since humans were mammals, our own learning -- and therefore our character and world view -- depended on such material; that what we called "personality" and "reality" were formed by a (deutero-) learning process that permeated our environment and taught us, in ways that were subtle but definite, certain allowable patterns that the culture labeled "sane"; and (4) that conversely, insanity (ostensible lack of coherence of personality and world view) probably involved the inability to manipulate the relationship between conscious and unconscious according to the deutero-propositions of a particular cultural context.
The theoretical starting point for Bateson's research here was Russell and Whitehead's "Theory of Logical Types." In itself, the theory simply states that no class of objects, as defined in logic or mathematics, can be a member of itself. Let us, for example, conceive of a class of objects consisting of all of the chairs that currently exist in the world. Anything we customarily term "chair" will be a member of that class. But the class itself is not a chair, any more than a particular chair can be the class of chairs. A chair, and the class of chairs, are two different levels of abstraction (the class being the higher level). This axiom, that there is a discontinuity between a class and its members, seems trivially obvious, until we discover that human and mammalian communication is constantly violating it to generate siginficant paradoxes.
One of the most famous of these paradoxes is known as "Epimenides' Paradox," or the "Liar's Paradox" (see Chapter 5, note 30). It might be presented as in Figure 14:
We see the problem at once. If the statement is true, it is false, and if false, it is true. The resolution lies in the Russell-Whitehead axiom. The word "statements" is being used in both the sense of a class (the class of statements) and as an item within that class. The class is being forced to be a member of itself, but since this situation is not allowable according to the formal rules of logic and mathematics, a paradox is generated. The statement itself is being taken as a premise for evaluating its own truth or falsehood, and thus two different levels of abstraction, or logical types, are being scrambled.
Now the truth is that neither human nor mammalian communication conforms to the logic of "Principia Mathematica." In fact, all meaningful communication necessarily involves metacommunication -- communication about communication -- and is therefore constantly generating paradoxes of the Russellian type. Let us take human communication first. Suppose I announce to you, just as we embark on some particular action or conversation, "This is play." The message I am conveying is, "Do not take the following seriously." What does the phrase really mean? "This is play," says Bateson, can be translated into the statement: "These actions in which we now engage do not denote what those actions for which they stand would denote"; or, since "stand" and "denote" mean the same thing here, the translation can be rendered: "These actions, in which we now engage, do not denote what would be denoted by those actions which these actions denote." If I give my lover a playful nip, the nip denotes a bite, but it does not denote what would be denoted by an actual bite. It is not an act of aggression, and I express this by using the act to comment negatively on itself. But neither this behavior, nor the statement "This is play," are allowable in formal logic. The "translated" sentence is a good example of the Liars Paradox: the word "denote" is being used in two degrees of abstraction, which are incorrectly treated as though they were on the same logical level (one is allowed to contradict the other). Both the nip, and the statement "This is play," set up a frame which is then allowed to comment on its own content.
This discussion returns us to the Fleishhacker Zoo, and the question of what we can learn from monkeys. Metacommunicative messages are logically inadmissable, because they are frames that comment on their own content. This point is clear enough in the case of a verbal statement, such as "This is play," and in fact we are constantly checking the frame of reference in ordinary discourse: "What are you really saying?" "Do you mean that?" "You've got to be kidding!" and so on. But, says Bateson, although we are capable, unlike monkeys, of spoken or written metacommunication, we are like them in the following crucial sense: the vast majority of metamessages remain implicit. "I love you," I say absentmindedly to my lover who just walked into the room in search of my attention or affection, while my body language and tone of voice say, "Leave me alone so I can finish writing my chapter on Bateson." As fo
r our mammalian cousins, they are limited by the lack of language such that they can refuse or reject an action, but not negate or deny it. Two dogs meet, and neither wishes to fight. They are unable to say, "Let's not fight." Being friendly does not solve the problem either, because it is a positive statement that omits any "discussion" of fighting, rather than specifically deciding against it. So the dogs bare their fangs, stage a mock brawl, and then stop. The message exchanged: the nip is not the bite, or "These actions, in which we now engage, do not denote . . ., etc." Play is a phenomenon in which actions of play denote other actions of not-play; and like dogs or monkeys, we exchange such messages all the time. In fact, says Bateson, on the human level we have evolved some very complex games based on a deliberate confusion of map and territory. Catholics say that the wafer is the body of Christ, a sacrament. Protestants say it is like the body of Christ, a metaphor. Millions have been killed in war, tortured or burned to death over just this issue, and millions continue to die for this or that flag -- bits of cloth which are much more than metaphors in the eyes of the soldiers who march under them.
The Reenchantment of the World Page 24