Notes for Herbert A. Simon The Shape of Automation for Men and Management
Key concepts: Gresham's Law of Planning.
Related theorists: .
Introduction: Computers and Automation
THE TECHNOLOGICAL DIMENSION
THE ECONOMIC DIMENSION
States position as technological radical and economic conservative.
(xii) I am a technological radical and an economic conservative.
(xiii) We choose among experts by forcing the experts to disclose how they reached their conclusions, what reasoning they employed, what evidence they relied upon. Having made this disclosure, they have exposed their assumptions and reasoning to the scrutiny and challenge of other experts. We do not have to be championship boxers to referee a fight.
ORGANIZATION OF THE ESSAYS
Basic computer ontology privileges running program, exuding debunking source code fetishism by technology pundits.
(xv) In the computer field, the moment of truth is a running program; all else is prophecy.
The Long Range Economic Effects of Automation
FACTS OF THE MATTER
(3) The widely held notion that productivity, driven by automation, is rising at breakneck speed is simply false.
Questions whether a highly educated workforce is really needed to operate a highly automated economy based on examples of western European and Japanese work forces: does this contribute to our becoming stupid?
(4) There is, in fact, no evidence that a highly automated economy could not be operated efficiently with an educational profile like that of the western European or Japanese work forces—that is, with far less formal education than now prevails in the United States.
A GLUT OF GOODS AND LEISURE?
MECHANIZATION, OR WHY THE HORSE DISAPPEARED
THE PRICE OF CAPITAL
REAL WAGES IN EQUILIBRIUM
THE EFFECT OF TECHNOLOGICAL CHANGE
Technological advance will raise real wages unless scarcity of capital causes rising interest rates.
(13) In sum: so long as the rate of interest remains constant, an advance in technology can only produce a rising level of real wages. The only route through which technological advance could lower real wages would be by increasing the capital coefficient (the added cost being compensated by a large decline in the labor coefficient), thereby creating a scarcity of capital and pushing interest rates sharply upward.
THE PATH OF EQUILIBRIUM
TECHNOLOGICAL CHANGE AND REAL WAGES
(18) The economic analysis shows that, for the kind of economy under consideration, technological change, whether capital-saving or labor-saving or both, is bound to increase real wages.
SKILLED AND UNSKILLED
A FINAL WORD ON THE HORSE
Human symbiosis with machines more profitable than with horses; however, machines will not abandon humans.
(25) The mean, limited in number, found that they could earn higher real wages in symbiosis with machines than in symbiosis with horses, and they abandoned their old friends for the new. So long as the supply of computers responds to market forces—and they do not impose birth control on themselves—they will be in no position similarly to abandon man. The rewards of technological progress will still be his, not theirs.
Will the Corporation be Managed by Machines?
(27) Perhaps the automated executive of the future has a great deal in common with the automated worker or clerk whom we can already observe in many situations today.
Picture society after new equilibrium settled.
(28) First of all, we must predict what is likely to happen to the job of the individual manager, and to the activity of management in the individual organization. Changes in these patterns will have secondary effects on the occupational profile in the economy as a whole. Our task is to picture the society after it has made all these secondary adjustments and settled down to its new equilibrium.
PREDICTING LONG-RUN EQUILIBRIUM
The Causes of Change
Learning deeper understanding of complex information processing and skills to write programs are causes of change, such as thinking devices; the trend will not reach programmer kings any more than the invention of writing created scribal kings. Equivocating management and computing invites humanities study of all these terms in all their language expressions run times, also reminding (in another recently read passage perhaps not this author) us that the only real objects are running programs reflects upon another passage read in this run time virtual reality PHI.
(30) The new knowledge consists in a fundamental understanding
of the processes of thinking and learning or to use a more neutral
term, of complex information processing. We can now write programs
for electronic computers that enable these devices to think and
(30) Within the very near future—much less than twenty-five years—we shall have the technical capability of substituting machines for any and all human functions in organizations. Within the same period, we shall have acquired an extensive and empirically tested theory of human cognitive processes and their interaction with human emotions, attitudes, and values.
(32) By the operation of the market place, manpower will flow to those processes in which its productivity is comparatively high relative to the productivity of machines, it will leave those processes in which its productivity is comparatively low. The comparison is not with the productivities of the past, but among the productivities in different processes with the currently available technology.
THE NEW TECHNOLOGY OF INFORMATION PROCESSING
The Nearly Automatic Factory and Office
The Occupational Profile
Another Approach to Prediction
Human advantage retained in use of brain as flexible, general-purpose problem-solving device, flexible use of sensory organs and hands, and use of legs, whereas competitive advantage as energy machine diminished; note Darwinian emphasis on flexibility that Malabou critiques.
(38) Thus, man's comparative advantage in energy production has been greatly reduced in most situations—to the point where he is no longer a significant source of power in our economy. He has been supplanted also in performing many relatively simple and repetitive eye-brain-hand sequences. He has retained his greatest comparative advantage in: (1) the use of his brain as a flexible general-purpose problem-solving device, (2) the flexible use of his sensory organs and hands and (3) the use of his legs, on rough terrain as well as smooth to make this general-purpose sensing-thinking-manipulating system available wherever it is needed.
Flexibility in Automata
Environmental Control a
Substitute for Flexibility
(42) The stabilization of the environments for productive activity will reduce or eliminate the need for flexible response at many points in the productive process, continuing a trend that is as old as multicellular life.
Man as Man's Environment
Summary: Blue-collar and Clerical Automation
Prediction by Simon that rapid automation under full employment with stable skill profile will make workplace happier and more relaxed, most people being in sales: critiques of global capitalism instead describe an erosion of the middle class aided by ERP and communication technologies.
(45) Under the general assumptions we made—rapid automation, but under full employment and with a stable skill profile—it will be a “happier” or more relaxed place than it is now; perhaps more of us will be salesmen.
THE AUTOMATION OF MANAGEMENT
Decision Making by Computer
Managers supervise and solve well-structured and ill-structured problems; prediction that middle management activities will be completely automated and the workforce diminished.
(47) Managers are largely concerned with supervising, with solving
well-structured problems, and with solving ill-structured
(47) But there is reason to believe that the kinds of activities that now characterize middle management will be more completely automated than the others, and hence will come to have a somewhat smaller part in the whole management picture.
Some Other Dimensions of Change
(48) Most automation calls for increased technical skills for maintenance in the early stages; but the farther automation proceeds, the less those who govern the automated system need to know about the details of its mechanism.
Failed prediction of extinction of programming occupation by self-programming techniques, although at the interface level human knowledge requirements to use computers has diminished.
(48-49) Similarly, we can dismiss the notion that computer programmers will become a powerful elite in the automated corporation. It is far more likely that the programming occupation will become extinct (through the further development of self-programming techniques) than that it will become all-powerful. More and more, computers will program themselves; and direction will be given to computers through the mediation of compiling systems that will be completely neutral so far as content of the decision rules is concerned. Moreover, the task of communicating with computers will become less and less technical as computers come—by means of computing techniques—closer and closer to handling the irregularities of natural language.
THE BROADER SIGNIFICANCE OF AUTOMATION
Predictions resembling Marxist utopia as automation proceeds: developing human science, alternatives to work and production as social goals, reformulating place in universe.
(50) Three of them in particular, I think, are going to receive a great deal of attention as automation proceeds: developing a science of man, finding alternatives for work and production as basic goals for society, and reformulating man's view of his place in the universe.
A Science of Man
Expects rapid advances in teaching and dealing with human maladjustment, as if the human remains static in the process of technological advance.
(51) We may expect very rapid advances in the effectiveness and efficiency of our techniques of teaching and our techniques for dealing with human maladjustment.
Man in the Universe
The New Science of Management Decision
A. THE EXECUTIVE AS DECISION MAKER
DESIGN, AND CHOICE IN DECISION MAKING
(54) Decision making comprises three principal phases: finding occasions for making a decision; finding possible courses of action; and choosing among courses of action.
DEVELOPING DECISION-MAKING SKILLS
EXECUTIVE RESPONSIBILITY FOR ORGANIZATIONAL DECISION MAKING
AND NONPROGRAMMED DECISIONS
(59) Most of the programs that govern organizational response are not as detailed or as precise as computer programs. However, they all have the same intent: to permit an adaptive response of the system to the situation.
(60) The four-fold table below will provide a map of the territory I propose to cover.
B. TRADITIONAL DECISION-MAKING METHODS
TRADITIONAL TECHNIQUES FOR PROGRAMMED DECISIONS
TRADITIONAL TECHNIQUES FOR NOPROGRAMMED DECISIONS
Use of training and planned experience to improve nonprogrammed organizational decision making.
(66) The processes of learning have been as mysterious as the processes of problem solving. But improvement there is. We are thus able, in a crude way, to use training and planned experience as a means for improving nonprogrammed decision making in organizations.
Notes Greshams Law of Planning in which programmed activity drives out nonprogrammed activity like the Freudian ego over the id; provisions must be made to maintain nonprogrammed decision making responsibilities: does this contribute to our becoming stupid?
(67) An important principle of organization design that has emerged over the years has been called facetiously “Gresham's Law of Planning.” It states that programmed activity tends to drive out nonprogrammed activity. If an executive has a job that involves a mixture of programmed and nonprogrammed decision-making responsibilities, the former will come to be emphasized at the expense of the latter. The organizational implication of Gresham's Law is that special provision must be made for nonprogrammed decision making by creating specific organizational responsibilities and organizational units to take care of it.
C. NEW TECHNIQUES FOR PROGRAMMED DECISION MAKING
Operations research extends management decision-making techniques developed for military needs to natural scientists.
(69) Operations research is a movement that, emerging out of the military needs of World War II, has brought the decision-making problems of management within the range of interests of large numbers of natural scientists and, particularly, of mathematicians and statisticians.
(71) Whatever the specific mathematical tool, the general recipe for using it in management decision making is something like this:
1. Construct a mathematical model . . . 2. Define the criterion function . . . 3. Obtain empirical estimates . . . 4. Carry through the mathematical process.
(71) In any decision-making situation where we apply this recipe successfully, we have, in fact, constructed a program for the organization's decisions.
Program for making decisions by applying four steps of management decisions leads to mathematicians aphasia, pretending the problem was always the simplified abstraction reached so that the program can be executed: point made by Hayles concerning cybernetics; does this contribute to our becoming stupid?
(72) This leads to an ailment that might be called mathematician's aphasia. The victim abstracts the original problem until the mathematical intractabilities have been removed (and all semblance of reality lost), solves the new simplified program, and then pretends that this was the problem he wanted to solve all along.
ENTER THE COMPUTER
THE REVOLUTION IN PROGRAMMED DECISION MAKING
Automated factory will operate on automated office, ERP.
(76) The automated factory of the future will operate on the basis of programmed decisions produced in the automated office beside it.
D. HEURISTIC PROBLEM SOLVING
Revisit human problem solving techniques for poorly structured tasks; AI based on heuristics rather than grand algorithms (Edwards, Golumbia).
(77) Nevertheless, when we run out of ideas for handling poorly structured problem-solving tasks, it seems plausible to examine more closely the processes used by humans who have handled such tasks—not always efficiently, to be sure—for several millennia.
UNDERSTANDING HUMAN PROBLEM-SOLVING PROCESSES
THE SIMULATION OF HUMAN THOUGHT
Nonnumerical Symbol Manipulation
Flexible and Adaptive Response
Assumption that human thinking is governed by programs that are like machine programs leads to computational model of mind and belief that computer programs can be written to simulate human thought.
(81) In solving problems, human thinking is governed by programs that organize myriads of simple information processes—or symbol manipulating processes if you like—into orderly, complex sequences that are responsive to and adaptive to the task environment and the clues that are extracted from that environment as the sequences unfold. Since programs of the same kind can be written for computers, these programs can be used to describe and simulate human thinking.
A GENERAL PROBLEM-SOLVING PROGRAM
General Problem Solver uses complex structures of familiar simple elements to solve problems, putatively modeling how the mind works by buying into assumption that simple elements are mental programs.
(82-83) The secret of problem solving is that there is no secret. It
is accomplished through complex structures of familiar simple
elements. The proof is that we can simulate it, using no more than
those simple elements as the building blocks of our programs.
(83) From the standpoint of human simulation, perhaps the most interesting program of this kind is one labeled GPS (General Problem Solver).
SOME OTHER HEURISTIC PROGRAMS
The Geometry Theorist
The Checker Player
Another Learning Program
THE AUTOMATION OF NONPROGRAMMED DECISION MAKING
IMPROVING HUMAN DECISION MAKING
Prospect of building aids to human thinking based on understanding of human thinking.
(92) All of these aids to human thinking, and many others, were devised without understanding the process they aided—the thought process itself. The prospect before us now is that we shall understand that process.
E. ORGANIZATIONAL DESIGN: MAN-MACHINE SYSTEMS FOR DECISION MAKING
SOME COMMENTS ON AUTOMATION
Automation and Unemployment
The Pace of Automation
The Composition of the Labor Force
The Routinization of Work
Desire for principle of moderation from Berlyne work on curiosity: interest of people and rats thrives in zones of manageable complexity, problems comprehensible in deep structure but unfamiliar in detail; thus, routine can be a welcome refuge.
Routine is a welcome refuge from the trackless forests of unfamiliar
(97-98) The work on curiosity of Berlyne and others suggests that some kind of principle of moderation applies. People (and rats) find the most interest in situations that are neither completely strange nor entirely known. . . . The pleasure that the good professional experiences in his work is not simply a pleasure in handling difficult matters; it is a pleasure in using skillfully a well-stocked kit of well-designed tools to handle problems that are comprehensible in their deep structure but unfamiliar in their detail.
FUNDAMENTALS OF ORGANIZATIONAL DESIGN
The Hierarchical Structure of Organization
(99) I can suggest at least two reasons why complex systems should generally be hierarchical.
for hierarchy in complex systems: appear in evolutionary processes
and require less information transmission among their parts.
(99-100) 1. Among possible systems of a given size and complexity, hierarchical systems, composed of subsystems, are the most likely to appear through evolutionary processes.
(100) 2. Among systems of a given size and complexity, hierarchical systems require much less information transmission among their parts than do other types of systems.
Size of the Building Blocks: Centralization and Decentralization
Before emergence of distributed control as an organizational design, option is how far to decentralize.
(102-103) But centralizing and decentralizing are not genuine alternatives for organizations. The question is not whether we shall decentralized, but how far we shall decentralize.
Need for infrequent outside intervention in high speed data-processing systems; intervention takes form of system design and programming, away from operations, as reflected in development of timesharing systems.
(106) Since processing steps in an automated data-processing system are executed in a thousandth or even millionth of a second, the whole system must be organized on a flow basis with infrequent intervention from outside. Intervention will take more and more the form of designing the system itself—programming—and less and less the form of participating in its minute-by-minute operation.
Authority and Responsibility
Expect managerial jobs to shift toward rationalization and impersonalization; note predominant role of spreadsheet model of business highlighted by Golumbia.
(108) If a couple of terms are desired to characterize the direction of change we may expect in the manager's job, I would propose rationalization and impersonalization.
A FINAL SKETCH OF THE NEW ORGANIZATION
Simon, Herbert A. The Shape of Automation for Men and Management. New York: Harper Torchbooks, 1965. Print.