The transcript of a 1965 talk that Brodey gave to an audience of psychotherapists. Offers some insights as what drew a psychiatrist such as Brodey to computers. Original in the library of Harvard’s School of Design. Apologies for the annotations.
My secretary and I argued about the spelling of grammar. She insisted that G-R-A-N-D-M-A was more appropriate. I am sure you would enjoy a discussion of Grandma's relation to courtship and marriage–it’s more familiar. But by grammar I mean the grammar of handwavings, smiles, grunts, words that grow as the family learns a style of changing to match each other’s changings. This style evolves the family beyond what its members knew before they met.
Modern technology seeks to imitate this style of evolving. Modern technology aims to help the student evolve previously unknown styles of learning, we would like to construct our teaching devices to adapt their teaching to match the student’s way of evolving his learning beyond the style he learned in his family.
The problem is quite specific. In building a manual control for a 2000 mile per hour jet, we would like the exchange between the control lever and the man to be as rich as the telling handshake which responds to all the information in each person's experience. To imitate a handshake the hand held lever must change characteristics in delicate response to the whole system's changing touch. The control lever may feed back its information to the loop by changing its drag to facilitate a manual move toward an unseen radar detected target. The pilot's judgement joins this responsively displayed observation to his. We are modeling the handshake as a guidance device: a handshake of cordial hello may signal by a most delicate shift in the evolving action that another hand is waiting, and one turns to view the husband of the lady.
Hands together evolve a way of speaking. This level of self organizing sensitivity is necessary for controlling the behavior of man and his new tools. Our new technology demands more use of natural human information processing and control skill than previously conceived. There is a new need to know which does not have a moral base.
The design of life support systems for space requires that we specify and gain control of a new kind of relationship: the space suit’s heat regulator must adapt not only to the pilot’s heat but to the change in the pilot’s heat which results as the space suit’s heat is changed. Shades of bed time. The time delay before the space suit answers the pilot’s changing need is only one way of holding the system in dynamic but stable equilibrium. There are many.
There was no way to study this grammar of ongoing changing until the computer. New computer-based control speed allows us to move into a resonant synchrony more and more delicately matched to particular human rhythms. The ability to lock in step and vary just beyond the human changes is not the same as measuring by smoothing out variations. Time delay stabilization simplifies observation by averaging out the changings over the period of delay; it also destroys the fine graining at the growth edge of change. Measurement by matching and unmatching with delicately timed movement is a way of gaining knowledge of living process.
Like our old machines, some families average the world. They prevent escalation of any change into a new experience by long time delay before awareness of accumulated change; then they make corrections. They live in an informationally poor world. A world of less frequent and delicate surprises.
Other families live in richer variation. They touch ongoing change. They responsively resonate with their informational environment. Their style of structuring their living world grows moment to moment. There is a timing to growing that they intuitively know; they are not overwhelmed by the richness they live in. We are beginning to be able to formally describe growth in action.
But where are the words with which to define suitable experiments and to test intuitive psychological notions? We have learned by apprenticeship and long experience to question the relevance of theory in a way that does not kill the action to be studied.
You will wonder when I intend to speak on grandma.
The richness that a clinician and some grandmas know intuitively comes from experiencing courtship and marriage as a means of growing change together. Lacking a notation I must continue to talk by using man-man and man-machine as metaphors for each other.
The relationship in a man-machine system may simply involve a youngster filling out a machine generated exercise with a zero or a cross. A teacher’s exercise may be as simple. A teacher may make communication richer by generating the exercise in a way that evolves the potential for growth in this conversation. The child displays his readiness, his fatigue, and his questions, and the teacher or teaching device makes this conversation richer by using the student's response as a guide to change even as the child is responding. The machine or teacher can predict the likely course of action once it is begun. We all finish each other's sentences when the conversation is carrying its speakers. In such exchanges it is the use of small variation from prediction that produces those variations which tip the unstable moments towards evolving the surprises.
A space man and his computer, like student and teacher or two scientists sharing observations, are joined in a problem-solving dialogue by necessity. So is the couple in a courtship and the courtship is for testing the likely evolution of their problem-solving style.
But by now you will be sure I have lost touch with a clinician's reality, have turned into a machine—or that I am just teasing a receptive audience.
So now to courtship and marriage.
Girl and boy meet. Each displays to the other's senses not only eyes moist or less, darting or dull, but eyes changing as each other's change. They entwine in a conversation aimed at nothing. They are simply talking. They would explain the course of their conversation differently to whomever they told. As they talk they change what they are saying to match each other's words in a way that resonates so that together they find their ideas, and their dialogue involves them in its creation. They speak sometimes averted, sometimes sensing the resonance of a shared format and delighted. They know the capacity to organize their forward course is different. This is the uniqueness which may bring about a marriage. The night before with different partners and so many times they had spoken in the same tongue about the same topic but with a different grammar; then the dialogue did not grow of its own accord.
What is this resonance that grows small variation through predictive stability toward a self-sustaining conversation which sweeps across levels joining ideas to being in touch to physiological changes which involve the total organism; and all this may begin in the evolving grammar of a handclasp or an eyes meeting. The problem-solving and creative power of this kind of growing has many applications.
In building devices for remote exploration on a planet where man cannot go, we wish to achieve a strategy for teaching our machines to create patterns from what was previously unknown by the earth-bound programmer. Our machines presently learn to perform tasks that are usually considered to require thought. The strategy for creating relevance is less well known than proving. Are there rules? We need at first to know the rules which will allow us to move into this new territory. Of course, at first our devices will be crude. They will have the properties that we can learn by imitating the way the children are. They too build what was irrelevant into pattern.
The new pressure and the need to know develops visibly out of the requirements of our space program. Unmanned space explorers—the Voyager and Mariner flights for example—crystallize new technical problems long known but amorphous. Devices must be built to take soil samples and to examine them for their different characteristics. They identify the differences by preprogrammed chemical tests but then must map the location of the specific samples and to examine them for their different characteristics. They decide where they are most likely to find what they are seeking, for it is not possible to search through every grain of soil. Similar devices are being built to move intelligently across a planet surface and to search for living organisms. Such explorers must be taught a strategy of learning which does their experience—information previously stored and processed.
We are seeking in many ways to build devices that can imitate our patterns of learning, even our skill in learning how to learn. There is a new need to find out how we learn and perceive rather than just what we know. The study of a family as our primary teaching institution has become a technical necessity. This motivation is different than study in order to relieve sickness.
Different families use different strategies for learning how to learn. Some families establish very little family learning, or adaptive capacity. These families do not have a life of their own in addition to the life of their component people. It is a legal family but does not have the growth and change potential of a living organism. It is an assembly of people which, like the old-style assembly of component machines, can produce a product but has no capacity to evolve a better product or to meet a changing market. It requires individual intervention. This family does not repair its own family organization. It is not like the conversation that catches its participants into going beyond the individual skills. Our stimulus-response, cause-effect, grammar pressures those families which are already borderline further into this kind of dead assembly. Such a system must stay nearly perfect, it is often preoccupied with this need for perfection. Variations do not produce matching variation which sustain the system's function. Obsolescence is the only possible result when there is no rising to the challenge of change. In these families, there is no redundancy of role or of control. There is not the requisite detailed variety which can be used to evolve further small variations to match the new requirements. Booster of safety resources designed to ameliorate the ineffectiveness of this kind of family are soon used up: strain does not eventually initiate self-perpetuating and corrective growth.
The usual family has a viability which arises from its capacity to metabolize change in a way that allows shift and change to occur between its resonating parts without destroying the stability which slowly paces more long-term growth as needed to match long-term environmental shift. A way of describing the vitality of living systems such as we clinicians use in making our judgments needs to be specified; for these rules can be tested within the framework of our new technology.