The unifying theme of this work has been the search for a deeper theory of structural emergence. I have judged this task to involve specifying how local changes lead to globally significant results and how local changes within the individual’s mind relate to his social context and to the organization of the individual’s mind on the large scale, My focus on interactionaist, empirically based structural models of natural learning has led me to put different tasks, methods, and ideas in the foreground than are common in most of the contemporary cognitive sciences [note 1]. The accompanying Table 6.1 summarizes the primary conclusions from this work under headings appropriate to the analyses where they are developed.
THEME BASED SUMMARY TABLE: 1
THE DEVELOPMENT OF OBJECTIVES
ASSUMPTIONS: a more “human” rather than “goal-driven” model of behavior
guidance is needed
QUESTIONS: where do objectives come from ? ASSERTIONS: objectives come from elaboration of recent achievements
(one’s own or others’)
objectives come from demonization of failed elaborations abstracted failures become demons capable of symbolic
realization; abstracted successes become generalized tools.
THEME BASED SUMMARY TABLE: 2
THE PROGRESSIVE CONSTRUCTION OF MIND
ASSUMPTIONS: task domain derived structures develop from interactions between what is already in the mind and the particular details of experience. these structures compete among themselves to interpret subsequent problems by deforming them to fit what they can recognize as objects to which they apply. QUESTIONS: how do complex mental skills develop ? can we disentangle the relations between prior knowledge and the circumstances of the experience in the formation of new knowledge ? if diversity comes from the world, whence comes coherence ? ASSERTIONS: complex skills develop from the integration of task-specific knowledge and its application through problem deformation. control element insertion can “explain” developing coherence. case study can be fruitfully applied to a new level of cognitive analysis through tracing the learning of the particular individual.
THEME BASED SUMMARY TABLE: 3
THE EQUILIBRATION OF COGNITIVE STRUCTURES
ASSUMPTIONS: major “stylistic” differences in thought exist, e.g. concrete versus formal QUESTIONS: how does “formal” thought relate to “concrete” thought ? what is it like to learn something really new ? how does new knowledge effectively replace competing old knowledge ? ASSERTIONS: Piagetian stages do not develop from completion of predecessors: formal thought is an epistemologically separate system of knowledges which achieves late dominance over the concrete system by chance in some cultures. Task specific knowledge has an association-like character but is different from stimulus response couples in this major sense: behaviorists’ associations are solutions responding to problem specific cues; on balance, active knowledge is more like solutions looking for problems to solve. Inhibition of obsolete knowledge (local reorganization via affect): new knowledge ultimately suppresses old knowledge through various techniques of “distancing” obsolete knowledge from one’s self image. dialogue
THEME BASED SUMMARY TABLE: 4
THE ARTICULATION OF COMPLEMENTARY OF ROLES
ASSUMPTIONS: two person games are a more complex problem environment than spontaneous learning about addition or taking instruction, worth investigating and able to be invesitgated QUESTIONS: how does knowledge of interactive problem situations get built up ? how does knowledge about people and things interact ? how does pattern oriented, associational/classificatory thinking turn into serial deductive thinking; how do these relate to more systematic thought and to penetrating understanding of a problem domain ? how does one get a major new idea ? ASSERTIONS: the desire for social interaction is the motive for multi-role
mental play which in its turn is cognitively productive; one’s model of the other derives from the obsolete self
serial deduction is the new application of adequate old knowledge to new problems; the only other alternatives are instruction and the top-down creation of a structure which
could ab initio be able to “see” a new problem as different from others already encountered.
major new ideas, like minor ones, are accidents prepared for by the establishment of individually adaptive precursors.
THEME BASED SUMMARY TABLE: 5
ASSUMPTIONS: there is no necessity for assuming that the uniformitarian hypothesis of mind holds; but non-uniformities can be based on other models of division than on simple “hardware” division (right versus left brain halves); the functional division of mind in service of the body parts is more significant than physical lateralization of hemispheres. QUESTIONS: why do some experiences act with force in enhancing the coherence of mind while others do not ? ASSERTIONS: concrete structures are close to the body-specific ancestry and are quite disparate; more abstract structures are less ancestry-constrained than the early predecessors and can emerge as elements of a general semantic network; thus the mediating structures of coordinating schemata give way to the nuclei of microview clusters in more abstract cognitive networks.
In summary, the principles applied and the conclusions drawn from these analyses can be categorized under the two general headings of structure and process.
STRUCTURE: Principles and Conclusions
I began with the primacy of self-construction through individual activity in a resistant environment, not the shaping of a resistant albeit marginally plastic creature by an active environment. This primary activity is the ground of elaboration through which new goals are generated and knowledge becomes more broadly and precisely applicable. Behavior and the “self” are emergents of self-activating cognitive processes; cognitive styles are surface emergents of interactions of large scale clusters of such cognitive structures. Denying the existence of central, universally applicable capabilities (such as Sussman’s “hacker”) committed us to seek local, stepwise changes in cognitive structures even though some steps may be below the threshhold of manifest behavior. The development of mature skills through interactions of imperfect but adaptively effective predecessors requires that those intermediate cognitive structures be at least metastable. The distinctness of computational primitives of the diverse components of mind and their possibilities of relation through a kinesthetic foundation of some of the major sub-systems imply that multi-modal descriptions of mind should be initially preferred unless they can be demonstrated to be inadequate. The uniformitarian description of mind should be that which requires support by argument. Within an assumed competitive structure of mind, the emergence of hierarchy is what must be explained. Within an assumed disparate structure of mind, the emergence of any coherence is what must be explained.
Basic Assumptions of Descriptions
For describing both states of knowledge and transformations between them, I have found it useful and compatible with data of The Intimate Study to assume of cognitive structures that:
- they originate through particular experiences and have a derived, highly specific character.
- they may be sub-dominant: this means they may exist and function but only become manifest in behavior under very special circumstances, such as experimental interventions.
- they are disparate; that is, structures developed from experiences in different task domains do not communicate among themselves except as a consequence of separately formed linkage through organizational development.
- they are competitive complexes of:
* specific memories derived from particular experiences.
* specific procedures derived from imitation (of uncertain use in most cases until the scope of constraints on their application is worked out): the Perspective of a cognitive structure is the collection of pattern matching, value-specifying demon procedures which judge its fitness for a problem; the Functions of a cognitive structures is the network of transformations of slot values activated by the filling of the structures slots.
* demonic objectives derive from failed elaborations of successful procedures; the collection of such objectives is the structure’s Agenda.]
- the class of situations upon which a structure can function is progressively uncovered and the definition of the structure is refined incrementally through the many interactions of the structure with situations to which it variously applies.
Assuming that organization is fundamentally competitive means disparate structures compete among themselves to solve problems as each can. Further, this permits re-organization to be seen as taking place through the more effective competition of later-developed cognitive structures which can and do thus function as logical ancestors to temporally preceding structures. Further, if the primitives of structures derive from movement primitives of fundamentally different body-parts, this implies:
- structures descending with compatible primitives comprise a cognitive mode; such structures may be interrelated more easily than those of different modal systems.
- direct descendents of coordinating schemata are central in advancing coherence of the mind.
Interactions With Others
Society presents individuals with frameworks for interpreting situations and scripts for activities which include categorizations (descriptions of what’s what) and with central issues resolved on conventional or negotiated bases. In process terms, these issues provide nodes of decision at which the individual can later assert different (often more flexible) decisions; this is a functional insertion of a control element in a network. Simulation of another person becomes possible only with the structures of the self. This implies that dialog-like activities which take place within one mind involve developing control over subsystems originally used differently.
PROCESS: Principles and Conclusions
I assumed initially that problem solving began as a race among structures competing to solve presented problems through the use of well-known results and familiar procedures. The first Piagetian aspect of processing was represented through the possibility of invocation of ancestral knowledge at need. Equally important was the notion — equivalent to the idea that experience is interpreted through imperfect structures — that on encountering a novel situation, structures deform problems to models derived from past experience and, when all past models fail, the mind goes into constructive mode, piecing together a new solution from fragments of earlier models and specific failure-provided guidance.
Problem Generation as the Key to Cognitive Development
The mind generates new problems for itself. This is a central aspect of mind which permits its progressive development. New problems are proposed first through elaboration of problems currently solvable successfully, next through activation of demonic objectives created from earlier failed elaborations. I have named this proposition “The Unified Generalization Proposal”: when elaboration leads to the successful application of a working procedure to a new problem, this implies the effective generalization of the procedure; when the failure of a procedure on a problem leads to its suspension and later reactivation with possible symbolic realization, the broader class of objectives which the procedure may serve constitutes a goal-related analog of procedure generalization.
Beyond the fundamental association of one thing with another, learning is dominated by the cleavages of cognitive structures. Insulation is the cutting apart of structures; amalgamation is their joining together into new organizations. There are at least four amalgamation processes: the elevation of control, the correlation of perspectives, cross-modal coordination, and cluster nucleation. The coherence of related domains of knowledge is achieved either through linkage of formerly disparate structures or through the reorganization of prior experience via the development of temporally later, but logically antecedent structures when they are competitively more effective. Two insulation processes (there may be others) are progressive discrimination and the repression of obsolete perspectives. Learning by example, by instruction and by discovery are typically of the progressive discrimination sort because the hypotheses about what are the relevant kinds of things to attend to are given, either by programming, as in Winston’s work, by society through instruction, or by ancestral experiences in human learning through discovery. Humor takes part in the repression of obsolete perspectives, as does assignment of a less preferred role to another, for both serve to separate the image of the self as an agent from the use of specific structures derived from one’s authentic experience.
Interactions With Others
People learn by taking instruction from other people in their environment because they enjoy being engaged with them. Such formation of simple associations is homely binding. But society is intrusive as well as engaging. By forcing new questions on structures developed for other purposes, other people engender changes of behavior which may be so remarkably different in effect as to manifest “changes of stage.” When deprived of social engagement, the individual, to redress that loss, recreates the interaction by simulating his former colleagues; this in turn leads to elevation of control and the increased lability of procedures (generalization).
CONCLUDING COMMENTS – Interaction, Creativity and Adaptation
The mind has a structure such as I describe BECAUSE the interaction of multiple ancestral structures, channelled by action and consequences within the task domain, provides the creativity for cognitive adaptation. The adaptivity of cognition is what unites the development of mind with the evolution of all living things, for the mind evolves through the interaction of the individual’s emerging objectives and opportunities. Those emergent opportunities are constrained by reality. More importantly, some interactions are so powerful that creatures who seize the opportunities they present completely dominate all competition (language and social organization is one such interaction). In this sense, some interactions create a new dimension of being — as flight literally brought a new dimension of life to our planet.
Under such a view, doing psychology is more like a combination of epistemological analysis and history than physics. This argues such psychology should not be expected be a science in the narrowest sense of THAT term — however much we might hope to describe, even model, the processes of change of specific configurations of ideas.
CONCLUDING COMMENTS – Questions and Methods
In the study of human learning, as in evolution and cultural development, we should address the kind of specific questions Kurt Lewin proposed fifty years ago: “what are the interactions of individual tendencies and environmental opportunities, and what are powerful synergies among environmental laws which permit such opportunities to exist.” It is possible to answer such questions. Even though the concreteness of experiences undergirding the skillful and deceptively abstract processes of mature thought are hidden by the processes of amalgamation, they may still be traced through the details of thought, for the particularity of knowledge gives away the path of its genesis and thus reveals the processes of learning. A more detailed appreciation of human learning will contribute to our understanding in general how is it possible for an idea to become embodied in a thing.