The systemic model of individual and social learning processes given in the previous section may be used to model the processes of technological development and diffusion in society as a whole and in particular industries. In economic terms it bridges between the micro-economic processes of the individual inventor, researcher, product innovation, development engineer, marketer, salesman, and so on, and the macro-economic processes of the development of a major industry. It brings the basic knowledge processes together with the psychological factors that affect the attraction of different phases of product development to people with differing personalities.
The most significant factor in determining the large-scale effects of the detailed learning process described are the information flows required to establish high-level skills in a technology. Learning from experience requires increasing information flows to build up each level of knowledge in the hierarchy. The inductive process at one level is dependent on the results of the inductive process at the level below and hence the rate of development of knowledge at the higher levels of the hierarchy is initially extremely slow. However, were it not for the higher-level inductive processes the lower levels would become overloaded with excessive detail very rapidly. The modeling hierarchy represents an evolutionary adaptation to a world in which a small number of generative principles can account for an extremely wide range of experience.
As learning accumulates at the higher levels it makes information acquisition at the lower levels decreasingly necessary. When a level is able to account for all the information it receives then no surprise flows up from it and any information flow from it can be discarded. Our civilization is itself a learning system with this overall structure and the societies, organizations and people within it form linked substructures. Viewed in this way, the flow of information through language may be seen as a way of passing knowledge between the substructures, and the accumulation of knowledge in libraries and databanks may be seen as a way of preserving it across time. Language and the knowledge construct are the side-effects of a learning civilization based on a distributed system of autonomous, interacting learning units.
If we examine sub-structures of knowledge that are somehow complete in that they give an adequate basis for a certain class of actions, then the same considerations apply. At the beginning the learning process is very slow because it has no foundation other than creative action, that is search procedures and experimentation directed by curiosity, the drive for novelty. When one person makes a creative breakthrough their knowledge is at level one and can be communicated to others only informally by aiding them in mimicking the breakthrough procedure. The originators have little basis for communicating through language except in aiding mimicry by saying what they do as if they were observing themselves. The communication of knowledge in this way after breakthrough leads to a phase of replication.
When more experience is gathered of breakthrough-like situations, inductive inference at level two enables rules to be stated which generalize the experience. Natural language plays an important role here because it facilitates the expression of generalizations as apparently vague statements using fuzzy quantifiers. The explanations that can be given based on rules appeal to past experience, and this phase is one of empiricism. The rules are a linguistic form of reinforcement stating that it is good to do some things and bad to do others, and their empirical content is shown in that "why?" questions have to be answered in terms of "because we did it and it did, or did not, work."
As further experience leads to the accumulation of the rules, their numbers become excessive in terms of memory and application. Fortunately, they also make possible the inductive inference of generative models that reproduce the rules but are far more condensed in form. This condensation involves encoding the information in a form that is usually remote from experience and requires the development of technical languages for its expression. As these languages are not directly related to experience it becomes important to be very precise in their use in order to communicate correctly. The explanations that can be given of the rules generated are now based on technical languages and appeal to models. This phase may be termed one of theory because of the disconnection between experience and technicality. Note that the theory is not deep from a scientific point of view in that it does not appeal to fundamental laws or basic principles, but it is adequate from a technological point of view in that it is concise and correct.
The next phase is complex because it has multiple branches. An adequate theory solves the problem within a given domain and one effect is to reduce research interest in that domain as noted by Crane (1972). However, as a variety of theories develop, the learning process at level four becomes significant leading to interest in foundations, reductionism, and the use of one theory as a metaphor to speed up the development of another. This metaphorical process has occurred many times in extending human knowledge and is embedded in the structure of language (Lakoff & Johnson 1980). In technological terms, as the use of the models enables more and more action to be subsumed under a more and more concisely encoded rules it becomes feasible to incorporate these encoded rules in mechanisms and hence to enter a phase of automation.
Finally, the technology enters a mature phase where the theory is well-established and well-founded in relation to other areas. The learning process continues at level five as a search for further condensation into basic laws, and at level six where the complete process of knowledge acquisition can now be examined to draw out systemic principles that are applicable to wide domains of human processes.
This sequence from breakthrough through replication to empiricism, theory, automation and maturity (BRETAM) is fundamental to the analysis of the development of major new technologies and their associated industries, such as computing and genetic engineering. The following sections analyze the phenomena associated with it in greater detail.
A simple mathematical model may be applied to the learning process described in the previous section by noting that it involves two conflicting processes: needing to have knowledge in order to acquire more; and slowing down in the acquisition as there is less and less left to acquire. This can be approximated to the first order as a systemic process in which the rate of growth is related to: the amount already accumulated; and also to the amount remaining to accumulate. The differential equation of such processes is dominated by a term of the form:
where a is a rate constant and b is the peak attainable. Integrating this gives:
where
is a constant of integration, which is the standard symmetric S-shaped logistic
curve. If other phenomena occur which perturb the dynamics of Equation 1 then
the symmetries of Equation 2 may be distorted but the basic shape remains the
same.
The logistic curve has been found to be a useful phenomenological model of the introduction of any new knowledge, technology or product which follows a "learning curve" in which growth takes off slowly, begins to climb rapidly and then slows down as all the information has been assimilated. Such curves arise in many different disciplines such as education, ecology, economics, marketing and technological forecasting (Van Dujin 1983, Stoneman 1983). It has also been noted in many disciplines that the qualitative phenomena during the growth of the logistic curve vary from stage to stage in a similar sequence to the BRETAM analyzed above (Crane 1972, De Mey 1982).
From Equation 2, if we take the amount of knowledge necessary for a mature technology to be 1, then the 0.1, 0.25, 0.5, 0.75 and 0.9 points are equally spaced in time along the logistic curve. These points can be roughly correlated with the significant phases in knowledge acquisition analyzed previously as shown in Figure 6 where the BRETAM sequence is superimposed on the logistic learning curve. When less than 0.1 of the knowledge is available little can be achieved and we are in a breakthrough period where most projects attempted will fail through lack of some critical information or skill. However, at around the 0.1 level some form of recognizable breakthrough becomes possible and we move into a phase of replication in which the breakthrough activity is noted and mimicked widely with little understanding. This leads to increasing experience and knowledge based upon it and when the 0.25 level is reached sufficient information is available for a phase of empiricism in which design rules are formulated based on experience. At the 0.5 level sufficient information has been accumulated for deeper patterns to be discerned, and a phase of theory commences in which causal rules are derived that allow further experience to be predicted and design rules to be derived. At the 0.75 level the theories have become accepted as the normative basis for further progress and an automation phase commences in which they are used for design and to search out the residual knowledge. Above the 0.9 level a phase of maturity is entered when the knowledge is expected to be available and used routinely.
Figure 6 Phases of learning superimposed on a logistic learning curve
This is a very tidy description of what is essentially a random process with many local peculiarities but strong global trends. One problem with using it predictively is that the asymptotic final level can only be estimated in retrospect, and attempting to determine the form of a logistic curve from data on the early parts is notoriously sensitive to error (Ascher 1978). However, fitting logistic curves to historic data gives a very precise account of the development of major technologies such as the successive substitutions of one form of energy production for another (Marchetti & Nakicenovic 1979).
The BRETAM learning sequence shown in Figure 6 represents the phases of knowledge acquisition involved in a specific technology as a linear sequence. However, it does not make apparent the impact of developments in one technology on those in another or the changes in people, locations and applications involved in the technology. The development of one technology is often the prerequisite for a breakthrough in another so that the BRETAM sequences of related technologies tend to be linked together. These linkages become strongly emphasized because of the personalities and roles of the people involved at different stages of the sequence and the socio-economic effects of a developing technology.
It has been noted previously (Gaines & Shaw 1984d) that major differences in people result from different emphasis on the levels of the epistemological hierarchy. The core constructs (Kelly 1955) that determine the dynamics of the individual and manifest as "drive" and "motivation" may be predominantly at one level of the hierarchy. Some individuals are concerned primarily with experience not assimilable at level one, with extreme novelty where appropriate constructs do not yet exist. They are inventors of modes of experience. Their emphasis on novelty may make it difficult for them to operate in tight organizational structures and it is difficult to manage the promotion of breakthroughs. These tend to occur unexpectedly by their very nature.
Some individuals are primarily concerned with the assimilation of novel constructs at level one. They are the generators of new experience and gravitate towards research positions where facilities are available to replicate breakthroughs and refine them. Some individuals are primarily concerned with the assimilation of experience at level two. They are innovative designers who formulate rules for product development and gravitate towards companies willing to undertake the risk of product innovation. Some individuals are primarily concerned with the assimilation of theory at level three. They design products ranges that are well-founded on principles enabling integrated systems to be developed and supported as long-term product lines. They tend to gravitate towards industry majors who take a strategic position towards supporting a large customer base requiring product continuity.
These personal and industry phenomena lead to the successive stages of the BRETAM sequence being associated with different individuals, functions within companies, and types of company. For those concerned with any phase of the BRETAM sequence there is the equivalent of a switch of attention when a particular technology enters the next phase. The breakthrough person will look for other breakthroughs but probably never find another. The research team will switch to alternative projects as will product innovation teams and product line teams. Thus social dynamics are superimposed on the basic learning sequence.
The overall effect is that BRETAM sequences for a sequence of technologies, each of which is a prerequisite for the breakthrough of the next, tend to get stacked in a regular formation as shown in Figure 7. The base technology which initiates the tier is a prerequisite to the next level technology, which is a prerequisite to the technology above, and so on. The focus of attention in research, product innovation, and so on, shifts from one technology to another in each successive phase of the learning sequence:
Figure 7 Tiers of learning curves defining generations of technology
From a macroscopic viewpoint the generic technology from which all the others derive defines an industry, and the phases of the learning curves define generations of technology within that industry. Note how product innovation occurs about two generation periods after the initial breakthrough, and product lines are three generation behind the breakthrough. This is the delay between invention and innovation noted by Schumpeter (1939) and linked to the phenomena of Kondratiev long-waves in the economy (Freeman 1984).
In terms of the overall industry, knowledge acquisition proceeds as a succession of learning curves each building upon one another, and it is the envelope of these that we see as the improving performance of the technology in time. This is what has happened in the base technology underlying computing where relays were replaced by vacuum tubes in the 1940s, by transistors in the 1950s, by integrated circuits in the 1960s, and so on. Each new technology has had its own learning curve, and hardware-based performance parameters such as speed of operation have shown a steady improvement based on the envelope of these learning curves (Ayres 1968).
Note how the horizontal BRETAM sequence becomes replicated in a vertical BRETAM structure representing the stages of development of all the differing technologies in a given generation. This replication is complete by the fifth generation which is often seen as the one in which the industry itself reaches maturity. For example, Drucker (1978) compares the development of the electricity industry with that of the information industry, making the analogy that "information is energy for mind work." He remarks that if it had been fashionable to speak of generations of generators then there "would have been a `fifth-generation' or `sixth-generation' generator before there was any widespread use of electric power." It was the invention of the light bulb that led to the use of electricity as a universal form of energy.
Information technology is interesting in providing examples of well-documented learning processes that exhibit the phenomena of Figure 7 (Gaines & Shaw 1984c). For example, the electronic computer is only one component in an elaborate and highly differentiated infrastructure. This infrastructure has grown through a succession of generations of computers, each of which represents a major change in computing technology (Gaines 1984b). During the eight year span of each computing generation, revolutionary changes have taken place that correspond in magnitude to those taking place over some seventy years or more in the aircraft industry. The first generation commenced in 1948 and we are now in the fifth generation spanning 1980 through 1987.
There is a simple systemic model of the development of the computer industry. The precursors necessary to the birth of a new industry are a social need and a technology capable of satisfying it. The need allows a market to be generated. The technology allows products to be created for that market. For computers the enabling technology was electronics. The key operational concepts of a digital computer date back at least to Babbage, Lovelace and Boole in the nineteenth century (Goldstine 1972). However, the mechanical technology of their time could not support a digital computer with adequate operations, storage capacity and speed.
The social need for computation in Babbage's time was navigation and its support through astronomical calculations themselves supported by the calculation of tables of logarithms. The enabling technology of electronic devices dates to the early 1900s with De Forest's development of the triode vacuum tube (Shurkin 1984). However, it was not until the outbreak of World War 2 in 1940 generated the need for rapid calculation of ballistic trajectories that the social requirement for computation became sufficiently urgent. Mechanical device technology in the ball and disk differential analyzer had been taken to its limit and was inadequate. The pressures created by the urgent need to find an alternative technology supporting computation led to the use of vacuum tubes in the design of ENIAC by Mauchly and Eckert (Stern 1981).
Thus the initial breakthrough for computing was in electronic device technology (EDT) and it is interesting to see how well a simple, one-dimensional learning curve model for the substitution of different electronic technologies accounts for generation changes in computing:
This definition of generations in terms of hardware works well for the zeroth through second generations because of the distinct qualitative changes involved. However, as Rosen (1983) notes in his analysis of computer generations it blurs thereafter and "we are faced with an anomalous situation in which the most powerful computers of 1979-1980, the CDC Cyber 176 and the CRAY 1, would have to assigned to the second and third generations, respectively, while the most trivial of hobbyist computers would be a fifth-generation system." The reason for this anomaly is that the substitution effects of one form of technology for another along the horizontal BRETAM curve are gradual and do not generate hard boundaries. The enabling effects of changing technologies giving the vertical BRETAM sequence are far more dramatic: the change from mechanical to electronic devices made it possible to store programs as data and enabled the use of computers as a general-purpose tool and then the development of programming language compilers; the transistor made reliable operation possible and enabled routine electronic data processing and then interactive timesharing; integrated circuits reduced costs to the level where computers became commonplace and made possible the personal computer dedicated to a single user.
The pattern is one of improvements in the basic technology allowing new developments to occur in the architecture and applications of systems based on it. What is missing from the model of computing developing through successive generations of electronic device technology (EDT) is the infrastructure that has been built above this. A complete account of the development of the computing industry must be two-dimensional with breakthroughs in key system technologies also represented--a version of Figure 6 specifically for computing.
Figure 8 shows the BRETAM tiers for the computing industry. The breakthrough in EDT leading to the zeroth generation is placed at 1940 about the time of the Atanasoff and Berry experiments with tube-based digital calculations. Experience was gained with COLOSSUS and ENIAC during the next eight years leading to a spate of empirical designs in the 1956-63 period, theoretical foundations for logic minimization and automata theory, automation with computer-aided design tools, and culminating into maturity with the 4004/8008 microprocessor chips in 1971/1972. The number of devices on a chip follows Moore's law in doubling each year through the second and third generations, and has continued to double every 1.5 years through the fourth and fifth generations (Robinson 1984). Automation has reached the extreme level where silicon compilers allow a designer to implement his ideas directly in EDT with little further human intervention (Fields 1983). However, our knowledge of EDT is now largely accumulated and future developments will be quantitative in speed, size and cost rather than qualitative in enabling new breakthroughs. Radical changes will come only from the way in which we use EDT or from alternative basic technologies such as volumetric rather than planar devices, optical processing or organic devices (Tucker 1984).
Figure 8 Eight generations of computers and their underlying technologies
The first breakthrough generating a computing infrastructure was the introduction of the stored program concept which led to the transition from the ENIAC to the EDVAC designs. The key concepts were discussed by Mauchly in his paper of 1947 and the first implementations were the BINAC and EDSAC machines in 1949. Mauchly (1947) recognized the significance of stored programs in enabling the machine instruction set to be extended, noting that subroutines create "a new set of operations which might be said to form a calculus of instructions." This was the key conceptual breakthrough in computer architecture, that the limited functionality provided directly by the hardware could be increased by stored programs called as subroutines or procedures, and that the hardware and these routines together may be regarded as a new virtual machine. This is the foundation of the development of a variety of forms of virtual machine architectures (VMAs, Weegenaar 1978) that separates out computing science as a distinct discipline from other areas of electronic applications. The use of subroutines to give access to arithmetic and operating system capabilities was followed by the development of machine architectures dependent on traps to procedures emulating missing hardware and led to theories such as those of semaphores, Petrinets and the logic of databases underlying diverse architectural concepts. In the fifth generation era the virtual machine technology has become a mature foundation for advanced system design in human factors and knowledge engineering (Gaines 1984c).
The next level of breakthrough was to bridge the gap between machine and task through the development of problem-orientated languages (POLs). Their foundations in the first generation were subroutine libraries providing VMAs closer to the problem requirements, notably floating point arithmetic and mathematical functions. Work on the design of FORTRAN in 1954 and its issue in 1957 marks the beginning of the second generation era with languages targeted to specific problem areas of business data processing, text processing, database access, machine tool control, and so on. A 1968 paper on the coming fourth generation notes that "programming today has no theoretical basis" and calls for a scientific basis in the next generation (Walter, Bohl & Walter 1968). Sure enough the theory linking POLs to the underlying VMAs developed during the fourth generation era, for example, that of abstract data types and initial algebras (Goguen, Thatcher & Wagner 1978). In the fifth generation era the application of experience, design rules and theory to the automation of software production has become the top priority (Balzer, Cheatham & Green 1983).
The binary VMA-POL relationship bridges the gap between the computer and its application. However, the computer industry is based on the triadic relationship between computer, application and person, for example, programmer, system analyst or user. The next level of breakthrough was to bridge the gap between the computer and the person with the development of interactive computers. The move from batch-processing to direct human-computer interaction (HCI) was made in 1963/1964 with the implementation of MIT MAC, RAND JOSS and Dartmouth BASIC systems (Gaines & Shaw 1983a). The study of such systems led to design rules for HCI in the 1970s (Hansen 1971) and theoretical foundations have started to emerge in the 1980s (Gaines & Shaw 1984e). The improvement of HCI is a major priority in the Japanese fifth generation development program (Karatsu 1982).
The third element in the triadic relationship between computer, application and person is that between the application and the person. It is one of knowledge-processing, the human capability to store information through its inter-relations and make inferences about its consequences. The breakthrough in knowledge-based systems (KBS) dates from the development of DENDRAL (Buchanan, Duffield & Robertson 1971) for inferring chemical structures from mass-spectrometry data and MYCIN (Shortliffe 1976) for the diagnosis of microbial infections in the early 1970s. It led to a spate of expert system development in the fourth generation era of the 1970s (Gevarter 1983), and pragmatic design rules for knowledge engineering in the current fifth generation era (Hayes-Roth 1984). The utilization of their massive VLSI production capability (Gaines 1984a, Galinski 1983) for the support of KBS through LISP (Bawden et al 1979) and PROLOG (Clark & Tarnlund 1982) machines is the other major priority in the Japanese fifth generation development program (Moto-oka 1982).
Defining the upper levels of the infrastructure becomes more and more speculative as we move into the immediate past of our own era and look for evidence of learning curves that are at their early stages. It is reasonable to suppose that the level above the representation and processing of knowledge in the computer will be that of its acquisition, breakthroughs in machine learning and inductive inference systems (IIS). Two breakthroughs in this area have been Lenat's AM learning mathematics by discovery (Davis & Lenat 1982) and Michalski's inductive inference of expert rules for plant disease diagnosis (Michalski & Chilausky 1980). In the current fifth generation era machine learning has become a highly active research area still in its replication phase (Michalski & Carbonell 1983).
One may speculate that the growth of robotics will provide the next breakthroughs in which goal-directed, mobile computational systems will act autonomously to achieve their objectives. The breakthrough into the sixth generation era commencing in 1988 will be one of autonomous activity systems (AAS). It is possible to see the nascent concepts got this breakthrough in the adoption of the goal-directed programming paradigms of logic programming languages such as PROLOG. When, in a robot, a goal specification is expanded by such a programming system into a sequence of actions upon the world dependent on conditions being satisfied in that world, then the behavior of such a system will deviate sufficiently from its top-level specification, yet be so clearly goal-directed, as to appear autonomous. However, to achieve significant results with such systems we need to add perceptual acts to the planning structures of a language such as SIPE (Wilkins 1984) and develop logic programming languages that cope with the resulting temporal logic (Allen 1984) --in these developments the sixth generation breakthrough will come to be recognized.
One may speculate further that interaction between these systems will become increasingly important in enabling them to cooperate to achieve goals and that the seventh generation era commencing in 1996 will be one of socially organized systems (SOS). The social reflective equilibrium model outlined in the earlier part of this paper gives some indication of the need to embed knowledge science, applying to both human and machine intelligence, within a socio-cultural framework. It is reasonable to suppose that current theoretical work in this area will form the basis of a technological breakthrough within the next decade.
However, it is also reasonable to suppose in the light of past forecasting failures in computing technology that these speculations will be greatly in error. The projected breakthroughs may not occur or may occur much earlier. The recognized breakthroughs may be in completely different areas. It is even possible that building an adequate forecasting model based on the premises of this paper may undermine the very processes that we model. If we come to understand the dynamics of our progress into the future then we may be able to modify the underlying process--to make the next steps more rapidly when the territory is better mapped.
The lines of invention, research, product innovation and product lines for the development of the computing industry are shown on the BRETAM infrastructure in Figure 9, and it is interesting to relate them to events in each generation. In particular note that the delay between invention and product innovation in the computer industry is about 16 years, and that between invention and product lines is about 24 years.
Figure 9 Invention, research, innovation and product
lines through
generations of computers
In the zeroth generation (1940-47):
In the first generation (1948-55):
In the second generation (1956-63):
In the third generation (1964-71):
In the fourth generation (1972-79):
In the fifth generation (1980-87):
In the sixth generation (1988-95):
In the seventh generation (1996-2003):
It is interesting to examine the Japanese fifth and sixth generation computing programs in the light of this analysis:
Figure 8 shows the generations of computers and all the associated technological structure linked into a consistent eight year cycle. The literature on computer generations specifies a variety of starting dates and time scales for the generations, anywhere between five and ten years. Withington's (1974) analysis which separates hardware, software, applications and organizational effects, allocates eight years to generations two through four but puts the starting dates two years later than ours. The boundaries of each generation are naturally somewhat fuzzy and depend on whether one counts from the initial availability or the widespread use of the technology.
The model of the development of computing presented here is partially based on a collection of key events in computing. When this collection was started it seemed reasonable to regard "generations" as convenient demarcations of the history of computing and to be dubious about there being any underlying cyclic mechanism. However, as more data were gathered the discrete nature of each era became more and more pronounced, both subjectively in relation to recollections of the ethos of those eras, and objectively as in relation to the critical events at the boundaries.
One interpretation of the timing of the generations is that they relate to the underlying business cycle of capital equipment purchase, the Juglar cycle originally identified in the nineteenth century (Juglar 1889) and apparent as a definite eight year cycle through to the current times (Lewis 1978). The peaks in the Juglar cycle correspond to our eight year boundaries from 1940 through 1980; the USA statistics show the 1964 peak as being in 1967 due to purchases to support the Viet Nam war but the next peak is in 1972 on target showing the dominating effect of world trade cycles rather than national events (Van Dujin 1983).
It is possible to speculate further on the causal linkages between the trade cycles and innovation in computing. The slump in purchasing following a peak leads to cut-backs in research and development investment so that the next generation of products tends to be based on the innovations made prior to the preceding peak. There are also longer term trends and cycles in invention and innovation that underlie the development of complete waves of new technology over periods of about fifty years, some six Juglar cycles.
Marchetti (1981) projects the current long-wave as having a learning curve in inventions from the 0.1 level in 1951 to the 0.9 level in 1989, and in their application as product innovations from the 0.1 level in 1987 to the 0.9 level in 2003. This corresponds to the creative era of computing being the first through fifth generations and the social impact era as being the sixth and seventh generations. The exodus of KBS researchers from the universities to industry is symptomatic of this transition (Business Week 1984). It is also analogous to the transition to widespread impact between the fifth and sixth generations of technology in the electricity industry quoted previously from Drucker (1978)
The biological application of information technology which has led to the development of the genetic engineering industry is a rather more recent phenomenon than computing and it is not yet possible to establish the detailed infrastructure of the industry. However, there is historical documentation on the development of both the science of molecular biology (Allen 1975, Judson 1979, Glass 1982, De Pomeray 1985 Gribbin 1985,) and the resultant industry (Cherfas 1982, Prentis 1984, Nossal 1985, Kahn 1985), and already a number of significant parallels have emerged.
Miescher first isolated DNA in 1869 and one of his students, Altmann, named it as nucleic acid. Miescher also speculated on the mechanism of heredity and in 1892 suggested that the information was carried in large organic molecules where the repetition of a few chemical configurations could encode a complex message, "just as the words and concepts of all languages can find expression in twenty-four to thirty letters of the alphabet", but named proteins rather than DNA. This parallels Babbage's attempt to develop a mechanical computer in 1834 and Lovelace's development of the concept of programming it. The concepts of the early information technology pioneers were remarkably advanced but the mechanical and chemical sciences of the nineteenth century could not support them.
Turing's (1936) paper on computable numbers gave scientific foundations to the notion of computing well before the machinery to reify it was developed. This is paralleled by Schrodinger's (1944) book on the nature of life which analyzed the foundations of genetic encoding in molecular structures well before the decoding of DNA. In both cases these theoretical developments were a great stimulus to the scientific and technological pioneers to give meaning to the theory through practice, to invent or discover a physical model for the mathematics.
The breakthrough in molecular biology was Watson and Crick's discovery of the double helix model of DNA in 1953. This gave the architecture of the genetic process and is comparable to the Mauchly and Von Neumann concepts of computer architectures in 1946. It took a further thirteen years for the details of the genetic code to be worked out, and hence the programming mechanism by which DNA specifies proteins to be understood. This brought molecular biology to an advanced state as a predictive science by the mid 1960s. The breakthrough into a normative technology came in 1970s with the development of cutting enzymes that allowed the structure of DNA to be changed and controlled, and hence genetic engineering to begin to emerge as an industry. In the 1980s the conceptual models to mediate between the specifications of required organic structure and the genetic structures to produce them are being developed--the problem-orientated languages of biological information engineering.
The next stages of development in information technology, both in computing and biology, are ones that emphasize the recursiveness of the model presented in this paper. As computing developments the techniques of artificial intelligence, and genetic engineering develops techniques for changing the structure of life, the human species is modifying its knowledge acquisition processes both through the artefacts that mediate it and through the brains that have been its primary dynamics. It has been suggested in this paper that there is a single system underlying all these phenomena: the anticipatory processes of a living system targeted on its survival. The model proposed is not yet adequate in detail to carry the full weight of this global conjecture, but it is already useful in providing a rationale for developments in information technology, and suggesting future directions.
Information technology may be viewed as autonomously impacting society. However, it may also be viewed as a product of society, created through its needs and molded by its values. From this second point of view, the development of information technology may be modeled as a succession of social learning processes arising out of the need to cope with an increasingly complex world. This paper has given a system-theoretic accounts of the epistemological processes underlying knowledge acquisition, and has shown that these apply to both individual human behavior and social development processes. Such accounts are applicable to the upper levels of the hierarchy of autonomous systems to provide models of socio-economic behavior.
The knowledge acquisition model has been applied to the development of information technology, and used to account for past events and predict future trends in relevant industries such as computing and genetic engineering. Underlying all developments in information technology is a tiered succession of learning curves which make up the infrastructure of the relevant industries. The paper has provided a framework for the industries based on this logical progression of developments. It has linked this empirically to key events in the development of computing and genetic engineering. It has linked it theoretically to a model of economic, social, scientific and individual development as related learning processes with a simple phenomenological model. It has used this model to account for past developments in information technology and extrapolated it to predict future trends. Only now in the fifth generation and beyond are computing systems becoming able to interact with people at the levels of the higher processes of the mind. Thus, their major impact is yet to come.
As a parting remark, let us note that the understanding of the processes of knowledge, induction and problem-solving that are necessary to the development of future generations of computers is equally applicable to human development. The tools of humanity change the world and humanity itself. If we view the development of computing as a symptom of a changing basis for civilization then we can begin to gain some glimpses of the future. If we view it as autonomous technology causing problematic and unwanted change then the future will continue to be opaque.
Financial assistance for this work has been made available by the Natural Sciences and Engineering Research Council of Canada.
Allen, J.F. (1985). Towards a general theory of action and time. Artificial Intelligence, 23(2), 123-154 (July).
Allen, G. (1975). Life Science in the Twentieth Century. New York: John Wiley.
Ascher, W. (1978). Forecasting: An Appraisal for Policy-Makers and Planners. Baltimore, Maryland: John Hopkins University Press.
Ayres, R.U. (1968). Envelope curve forecasting. Technological Forecasting for Industry and Government: Methods and Applications. pp. 77-94. Englewood-Cliffs, New Jersey: Prentice-Hall.
Balzer, R., Cheatham, T.E. & Green, C. (1983). Software technology in the 1990's: using a new paradigm. Computer, 16(11), 39-45 (November).
Bawden, A., Greenblatt, R., Holloway, J., Knight, T., Moon, D. & Weinreb, D. (1979). The Lisp machine. Winston, P.H. & Brown, R.H., Eds. Artificial Intelligence: An MIT Perspective. pp. 343-373. Cambridge, Massachusetts: MIT Press.
Bell, D. (1973). The Coming of Post-Industrial Society. New York: Basic Books.
Blum, A. & McHugh, P. (1984). Self-Reflection in the Arts and Sciences. Atlantic Highlands, New Jersey: Humanities Press.
Boose, J.H. (1985). A knowledge acquisition program for expert systems based on personal construct psychology. International Journal of Man-Machine Studies 20(1), 21-43 (January).
Brod, C. (1984). Techo Stress: The Human Cost of the Computer Revolution. Reading, Massachusetts: Addison-Wesley.
Brown, G.S. (1969). Laws of Form. London: George Allen & Unwin.
Buchanan, B.G., Duffield, A.M. & Robertson, A.V. (1971). An application of artificial intelligence to the interpretation of mass spectra. Milne, G.W.A., Ed. Mass Spectrometry Techniques and Applications. New York: John Wiley.
Business Week (1984). Artificial intelligence is here. Business Week(2850), 54-62 (July).
Carney, T.P. (1984). Instant Evolution: We'd Better Be Good At It. Middlesex, England: Penguin Books.
Cherfas, J. (1982). Man-made Life: An Overview Of The Science, Technology And Commerce of Genetic Engineering. New York: Pantheon Books.
Clark, K.L. & Tarnlund, S.-A., Eds. (1982). Logic Programming. London: Academic Press.
Crane, D. (1972). Invisible Colleges: Diffusion of Knowledge in Scientific Communities. Chicago: University of Chicago Press.
Davis, R. & Lenat, D.B. (1982). Knowledge-Based Systems in Artificial Intelligence. New York: McGraw-Hill.
De Bono, E. (1979). Future Positive. London: Maurice Temple Smith.
De Mey, M. (1982). The Cognitive Paradigm. Dordrect: Reidel.
De Pomerai, D. (1985). From Gene to Animal. Cambridge: University Press.
Dizard, W.P. (1982). The Coming Information Age: An Overview of Technology, Economics and Politics. New York: Longman.
Drucker, P.F. (1978). The Age of Discontinuity. New York: Harper & Row.
Ellul, J. (1964). The Technological Society. New York: Vintage Books.
Fields, S.W. (1983). Silicon compiler cuts time to market for DEC's MicroVAX I. Electronics, 56(21), 47-48 (October).
Freeman, C., Ed. (1984). Long Waves In The World Economy. London: Francis Pinter.
Gaines, B.R. (1977). System identification, approximation and complexity. International Journal of General Systems, 3, 145-174.
Gaines, B.R. (1980). General systems research: quo vadis ?. Gaines, B.R., Ed. General Systems 1979. Vol. 24, pp. 1-9. Kentucky: Society for General Systems Research.
Gaines, B.R. (1983). Precise past--fuzzy future. International Journal of Man-Machine Studies, 19(1), 117-134 (July).
Gaines, B.R. (1984a). A framework for the fifth generation. Proceedings of National Computer Conference, 53, 453-459. Arlington, Virginia: AFIPS Press.
Gaines, B.R. (1984b). Perspectives on fifth generation computing. Oxford Surveys in Information Technology, 1, 1-53.
Gaines, B.R. (1984c). From ergonomics to the Fifth Generation: 30 years of human-computer interaction studies. Proceedings of INTERACT'84: First IFIP Conference on Human-Computer Interaction, 1, 1-5.
Gaines, B.R. (1986). Sixth generation computing: a conspectus of the Japanese proposals. ACM SIGART Newsletter, No.95, 39-44 (January).
Gaines, B.R. & Shaw, M.L.G. (1981). A programme for the development of a systems methodology of knowledge and action. Reckmeyer, W.J., Ed. General Systems Research and Design: Precursors and Futures. pp. 255-264. Louisville, Kentucky: Society for General Systems Research (January).
Gaines, B.R. & Shaw, M.L.G. (1983a). Dialog engineering. Sime, M. & Coombs, M.J., Eds. Designing for Human-Computer Interaction. pp. 23-53. London: Academic Press.
Gaines, B.R. & Shaw, M.L.G. (1983b). Is there a knowledge environment?. Lasker, G., Ed. The Relation Between Major World Problems and Systems Learning. pp. 27-34. Society for General Systems Research (May).
Gaines, B.R. & Shaw, M.L.G. (1984a). The Art of Computer Conversation: A New Medium for Communication. Englewood Cliffs, New Jersey: Prentice Hall.
Gaines, B.R. & Shaw, M.L.G. (1984b). Hierarchies of distinctions as generators of system theories. Smith, A.W., Ed. Proceedings of the Society for General Systems Research International Conference. pp. 559-566. California: Intersystems.
Gaines, B.R. & Shaw, M.L.G. (1984c). Generations of computers: modeling and forecasting. Possible Worlds, 1(4), 3-16.
Gaines, B.R. & Shaw, M.L.G. (1984d). Emotion: a systemic development of the personal construct model. Proceedings of Personal Construct Network Conference (June). Cincinnati.
Gaines, B.R. & Shaw, M.L.G. (1984e). Dialog shell design. Proceedings of INTERACT'84: First IFIP Conference on Human-Computer Interaction, 1, 344-349.
Gaines, B.R. & Shaw, M.L.G. (1985). Three world views and system philosophies. Banathy, B.H., Ed. Proceedings of the Society for General Systems Research International Conference. pp. 387-396. California: Intersystems.
Galinski, C. (1983). VLSI in Japan: the big leap forward, 1980-1981. Computer, 16(3), 14-21 (March).
Gevarter, W.B. (1983). Expert systems: limited but powerful. IEEE Spectrum, 18, 39-45.
Glass, R.E. (1982). Gene Function. Berkeley: University of California Press.
Glover, J. (1984). What Sort Of People Should There Be?. Middlesex, England: Penguin Books.
Goguen, J.A., Thatcher, J.W. & Wagner, E. (1978). An initial algebra approach to the specification, correctness and implementation of abstract data types. Yeh, R., Ed. Current Trends in Programming Methodology. pp. 80-149. Englewood Cliffs, New Jersey: Prentice-Hall.
Goldstine, H.H. (1972). The Computer from Pascal to Von Neumann. New Jersey: Princeton University Press.
Goodman, N. (1973). Fact, Fiction and Forecast. Indianapolis: Bobbs-Merrill.
Gribbin, J. (1983). In Search Of The Double Helix: Quantum Physics And Life. New York: McGraw Hill.
Hall, E.T. (1959). The Silent Language. New York: Doubleday.
Hansen, W.J. (1971). User engineering principles for interactive systems. AFIPS Proceedings of the FJCC. Vol. 39, pp. 523-532.
Hayes-Roth, F. (1984). The industrialization of knowledge engineering. Reitman, W., Ed. Artificial Intelligence Applications for Business. pp. 159-177. Norwood, New Jersey: Ablex.
Hume, D. (1739). A Treatise of Human Nature. London: John Noon.
Judson, H.F. (1979). The Eighth Day Of Creation: The Makers Of The Revolution In Biology. New York: Simon & Schuster.
Juglar, C. (1889). Des Crises Commerciales et Leur Retour Periodique. Paris: Librairie Guillaumin.
Kahn, C. (1985). Beyond the Helix: DNA and the Search for Longevity. New York: Times Books.
Karatsu, H. (1982). What is required of the fifth generation computer--social needs and impact. In Moto-oka (1982) pp. 93-106.
Kawanobe, K. (1984). Present status of the fifth generation computer systems project. ICOT Journal, 5, 13-21.
Kelly, G.A. (1955). The Psychology of Personal Constructs. New York: Norton.
Klir, G.J. (1976). Identification of generative structures in empirical data. International Journal of General Systems, 3, 89-104.
Klir, G.J. (1985). Architecture of Systems Problem Solving, New York: Plenum Press.
Kohak, E. (1984). The Embers and the Stars: A Philosophical Enquiry into the Moral Sense of Nature. Chicago, Illinois: University of Chicago Press.
Krimsky, S. (1983). Genetic Alchemy: The Social History Of The Recombinant DNA Controversy. Cambridge, Massachusetts: MIT Press.
Lakoff, G. & Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press.
Lewis, W.A. (1978). Growth and Fluctuations 1870-1913. London: George Allen & Unwin.
Luhmann, N. (1979). Trust and Power. Chichester: John Wiley.
Marchetti, C. (1981). Society as a learning system: discovery, invention and innovation cycles revisited. Report RR-81-29 (November). Laxenburg, Austria: International Institute for Applied Systems Analysis.
Marchetti, C. & Nakicenovic, N. (1979). The dynamics of energy systems and the logistic substitution model. Report RR-79-13 (December). Laxenburg, Austria: International Institute for Applied Systems Analysis.
Mauchly, J.W. (1973). Preparation of problems for EDVAC-type machines. Randell, B., Ed. The Origins of Digital Computers: Selected Papers. pp. 365-369. Berlin: Springer-Verlag.
Michalski, R.S. & Chilausky, R.L. (1980). Knowledge acquisition by encoding expert rules versus computer induction from examples--A case study involving soyabean pathology. International Journal of Man-Machine Studies, 12, 63-87.
Michalski, R.S. & Carbonell, J.G., Eds. (1983). Machine Learning: An Artificial Intelligence Approach. Palo Alto, California: Tioga.
Moto-oka, T., Ed. (1982). Fifth Generation Computer Systems. Amsterdam: North-Holland.
Mowshowitz, A. (1976). The Conquest of Will: Information Processing in Human Affairs. Reading, Massachusetts: Addison-Wesley.
Nossal, G.J.V. (1985). Reshaping Life. Cambridge: University Press.
Pask, G. (1975). Conversation, Cognition and Learning. Amsterdam: Elsevier.
Pask, G. (1980). Developments in conversation theory--Part I. International Journal of Man-Machine Studies, 13(4), 357-411 (November).
Pope, S. (1984). Conceptual synthesis: beating at the ivory gate?. Smith, W., Ed. Systems Methodologies and Isomorphies. pp. 31-40. California: Intersystems.
Prentis, S. (1984). Biotechnology: A New Industrial Revolution. New York: George Braziller.
Rawls, J. (1971). A Theory of Justice. Cambridge, Massachusetts: Harvard University Press.
Robinson, A.L. (1984). One billion transistors on a chip?. Science, 223, 267-268 (January).
Rosen, S. (1983). Generations, computer. Ralston, A. & Reilly, E.D., Eds. Encyclopedia of Computer Science and Engineering. pp. 657-658.
Schumpeter, J.A. (1939). Business Cycles. New York: McGraw-Hill.
Schutz, A. & Luckmann, T. (1973). The Structures of the Life-World. London: Heinemann.
Shackle, G.L.S. (1955). Uncertainty in Economics. Cambridge University Press.
Shaw, M.L.G. (1980). On Becoming a Personal Scientist: Interactive Computer Elicitation of Personal Models of the World. London: Academic Press.
Shaw, M.L.G. & Gaines, B.R. (1981). The personal scientist in the community of science. Reckmeyer, W.J., Ed. General Systems Research and Design: Precursors and Futures. pp. 59-68. Society for General Systems Research (January).
Shaw, M.L.G. & Gaines, B.R. (1983a). A computer aid to knowledge engineering. Proceedings of Expert Systems 83. pp. 263-271. London: British Computer Society.
Shaw, M.L.G. & Gaines, B.R. (1983b). Computer systems in the knowledge environment: a new medium. Lasker, G., Ed. The Relation Between Major World Problems and Systems Learning. pp. 35-41. Society for General Systems Research (May).
Shaw, M.L.G. & Gaines, B.R. (1984). Fifth generation computers as the next stage of a new medium. AFIPS Proceedings of National Computer Conference. Vol. 53, pp. 445-451. Arlington, Virginia: AFIPS Press.
Shaw, M.L.G. & Gaines, B.R. (1986). A framework for knowledge-based systems unifying expert systems and simulation. Birtwistle, G., Ed. AI, Graphics and Simulation, to appear (January). La Jolla, California: Society for Simulation Research.
Shortliffe, E.H. (1976). Computer-Based Medical Consultations: MYCIN. New York: Elsevier.
Schrodinger, E. (1944). What is Life?. Cambridge: University Press.
Shurkin, J. (1984). Engines of the Mind: A History of the Computer. New York: W.W.Norton.
STA (1985). Promotion of Research and Development on Electronics and Information Systems That May Complement or Substitute for Human Intelligence. Tokyo: Science and Technology Agency.
Stern, N. (1981). From ENIAC to UNIVAC: An Appraisal of the Eckert-Mauchly Computers. Bedford, Massachusetts: Digital Press.
Stich, S.P. & Nisbett, R.E. (1984). Expertise, justification and the psychology of inductive reasoning. Haskell, T.L., Ed. The Authority of Experts. pp. 226-241. Bloomington, Indiana: Indiana University Press.
Stoneman, P. (1983). The Economic Analysis of Technological Change. Oxford: University Press.
Toffler, A. (1980). The Third Wave. New York: Bantam.
Tucker, J.B. (1984). Biochips: can molecules compute?. High Technology, 4(2), 36-47.
Turing, A.M. (1936). On computable numbers, with an application to the entscheidungsproblem. Proceedings of the London Mathematical Society Series 2, 42, 230-265.
Turkle, S. (1984). The Second Self: Computers and the Human Spirit. New York: Simon & Schuster.
Van Dujin, J.J. (1983). The Long Wave in Economic Life. London: George Allen & Unwin.
Walter, C.J., Bohl, M.J. & Walter, A.B. (1968). Fourth generation computer systems. AFIPS Proceedings of Spring Joint Computer Conference, 32, 423-441.
Weegenaar, H.J. (1978). Virtuality and other things like that. Proceedings of IEEE COMCON, CH1388-8/78, 287-293.
Weizenbaum, J. (1976). Computer Power and Human Reason: From Judgement to Calculation. San Francisco: W.H.Freeman.
Wiener, N. (1950). The Human Use of Human Beings. Cambridge, Massachusetts: Riverside Press.
Wilkins, D.E.. (1984). Domain-independent planning: representation and plan generation. Artificial Intelligence, 269-301.
Withington, F.G. (1974). Five generations of computers. Harvard Business Review, 99-108.
Wojciechowski, J.A. (1983). The impact of knowledge on man: the ecology of knowledge. Hommage a Francois Meyer. pp. 161-175. Marseille: Laffitte.
Wojciechowski, J.A. (1986). Social psychiatry in the man-made world. American Journal of Social Psychiatry. In press.
Zadeh, L.A. (1965). Fuzzy sets. Information and Control, 8, 338-353.