The Sciences of the Artificial (The MIT Press)

by Herbert A. Simon

The thesis is that certain phenomena are “artificial” in a very specific sense: they are as they are only because of a system’s being molded, by goals or purposes, to the environment in which it lives. If natural phenomena have an air of “necessity” about them in their subservience to natural law, artificial phenomena have an air of “contingency” in their malleability by environment.

Engineering, medicine, business, architecture, and painting are concerned not with the necessary but with the contingent—not with how things are but with how they might be—in short, with design.

we must be careful about equating “biological” with “natural.” A forest may be a phenomenon of nature; a farm certainly is not. The very species upon which we depend for our food—our corn and our cattle—are artifacts of our ingenuity. A plowed field is no more part of nature than an asphalted street—and no less.

Simulation may even take the form of a thought experiment, never actually implemented dynamically.

However, the Moniac, while useful as a teaching tool, told us nothing that could not be extracted readily from simple mathematical versions of Keynesian theory and was soon priced out of the market by the growing number of computer simulations of the economy.

How can a simulation ever tell us anything that we do not already know?

The difficulty of the design problem often resides in predicting how an assemblage of such components will behave.

No artifact devised by man is so convenient for this kind of functional description as a digital computer.

I wonder what people will think as they read this after quantum computers become the big thing?

veridicality

Of course the decision that is optimal for the simplified approximation will rarely be optimal in the real world, but experience shows that it will often be satisfactory.

real-world optimization, with or without computers, is impossible,

I retain vivid memories of the astonishment and disbelief expressed by the architecture students to whom I taught urban land economics many years ago when I pointed to medieval cities as marvelously patterned systems that had mostly just “grown” in response to myriads of individual human decisions. To my students a pattern implied a planner in whose mind it had been conceived and by whose hand it had been implemented. The idea that a city could acquire its pattern as naturally as a snowflake was foreign to them. They reacted to it as many Christian fundamentalists responded to Darwin: no design without a Designer!

Although the presence of uncertainty does not make intelligent choice impossible, it places a premium on robust adaptive procedures instead of optimizing strategies

Organizations can localize and minimize information demands just as markets do, by decentralizing decisions. Matters of fact can be determined wherever the most skill and information is located to determine them, and they can then be communicated to “collecting points” where all the facts relevant to an issue can be put together and a decision reached. We can think of a decision as produced by executing a large computer program, each subroutine having its special tasks and relying on local sources of information.

Uncertainty is especially troublesome when it involves expectations by one unit about what other units in the same organization will do.

Affected by their organizational identifications, members frequently pursue organizational goals at the expense of their own interests—that is to say, behave in a way that is altruistic from a personal standpoint.

Operations research and artificial intelligence have enhanced the procedural rationality of economic actors, helping them to make better decisions.

Viewed as a geometric figure, the ant’s path is irregular, complex, hard to describe. But its complexity is really a complexity in the surface of the beach, not a complexity in the ant.

Human beings, viewed as behaving systems, are quite simple. The apparent complexity of our behavior over time is largely a reflection of the complexity of the environment in which we find ourselves.

we should not look for great complexity in the laws governing human behavior, in situations where the behavior is truly simple and only its environment is complex.

But the remarkable fact was that, when masters and grandmasters were shown other chessboards with the same numbers of pieces arranged at random. their abilities to reconstruct the boards were only marginally better than the duffers’ with the boards from actual games, while the duffers performed as well or poorly as they had before.

this seems to relate to general sense that skill breeds literacy. reading chess boards as you'd read text or a drawing. prev ref to chunks seems to resonate here. recognise structures. data structure is compressed/compressible

Until connectionism has demonstrated, which it has not yet done, that complex thinking and problem-solving processes can be modeled as well with parallel connectionist architectures as they have been with serial architectures, and that the experimentally observed limits on concurrent cognitive activity can be represented in the connectionist models, the case for massive parallelism outside the sensory functions remains dubious.

i wonder if he still believes this?

The difference between the hardware of a computer and the “hardware” of the brain has not prevented computers from simulating a wide spectrum of kinds of human thinking—just

the theory of design is that general theory of search.

There is a constant competition between the elaboration of knowledge and its compression into more parsimonious form by theories.

Learning is any change in a system that produces a more or less permanent change in its capacity for adapting to its environment.

meaningful material is stored redundantly, so that if any fraction of it is forgotten, it can be reconstructed from the remainder.

Everyone designs who devises courses of action aimed at changing existing situations into preferred ones.

Design, so construed, is the core of all professional training; it is the principal mark that distinguishes the professions from the sciences. Schools of engineering, as well as schools of architecture, business, education, law, and medicine, are all centrally concerned with the process of design.

As professional schools, including the independent engineering schools, were more and more absorbed into the general culture of the university, they hankered after academic respectability. In terms of the prevailing norms, academic respectability calls for subject matter that is intellectually tough, analytic, formalizable, and teachable. In the past much, if not most, of what we knew about design and about the artificial sciences was intellectually soft, intuitive, informal, and cookbooky. Why would anyone in a university stoop to teach or learn about designing machines or planning market strategies when he could concern himself with solid-state physics? The answer has been clear: he usually wouldn’t.

The older kind of professional school did not know how to educate for professional design at an intellectual level appropriate to a university; the newer kind of school nearly abdicated responsibility for training in the core professional skill.

The professional schools can reassume their professional responsibilities just to the degree that they discover and teach a science of design, a body of intellectually tough, analytic, partly formalizable, partly empirical, teachable doctrine about the design process.

Design, on the other hand, is concerned with how things ought to be, with devising artifacts to attain goals.

The only trouble is that the computations required are astronomical (the number 10120 is often mentioned in this context) and hence cannot be carried out—not by humans, not by existing computers, not by prospective computers.

Chess eh...

Since there did not seem to be any word in English for decision methods that look for good or satisfactory solutions instead of optimal ones, some years ago I introduced the term “satisficing” to refer to such procedures. Now no one in his right mind will satisfice if he can equally well optimize; no one will settle for good or better if he can have best. But that is not the way the problem usually poses itself in actual design situations.

solving a problem simply means representing it so as to make the solution transparent.

Some of us even think there are not just two cultures but a large number of cultures. If we regret that fragmentation, then we must look for a common core of knowledge that can be shared by the members of all cultures—a core that includes more significant topics than the weather, sports, automobiles, the care and feeding of children, or perhaps even politics.

The ability to communicate across fields—the common ground—comes from the fact that all who use computers in complex ways are using computers to design or to participate in the process of design. Consequently we as designers, or as designers of design processes, have had to be explicit as never before about what is involved in creating a design and what takes place while the creation is going on.

We ask, “If we can go to the Moon, why can’t we... ?”—not expecting an answer, for we know that going to the Moon was a simple task indeed, compared with some others we have set for ourselves, such as creating a humane society or a peaceful world.

“As there is a degree of depravity in mankind which requires a certain degree of circumspection and distrust, so there are other qualities in human nature which justify a certain portion of esteem and confidence.”

The first generation of management information systems installed in large American companies were largely judged to have failed because their designers aimed at providing more information to managers, instead of protecting managers from irrelevant distractions of their attention.113 A design representation suitable to a world in which the scarce factor is information may be exactly the wrong one for a world in which the scarce factor is attention.

As of the mid-1990s the lesson has still not been learned. An “information superhighway” is proclaimed without any concern about the traffic jams it can produce or the parking spaces it will require. Nothing in the new technology increases the number of hours in the day or the capacities of human beings to absorb information. The real design problem is not to provide more information to people but to allocate the time they have available for receiving information so that they will get only the information that is most important and relevant to the decisions they will make. The task is not to design information-distributing systems but intelligent information-filtering systems.114

Numbers are not the name of this game but rather representational structures that permit functional reasoning, however qualitative it may be.

Weather prediction is difficult in good part because the course of meteorological events is highly sensitive to the details of initial conditions. We have every reason to think that social phenomena are similarly sensitive.

The heart of the data problem for design is not forecasting but constructing alternative scenarios for the future and analyzing their sensitivity to errors in the theory and data.

Design for distant futures would be wholly impossible if remote events had to be envisioned in detail. What makes such design even conceivable is that we need to know or guess about the future only enough to guide the commitments we must make today.

At microsocial levels of design it is tacitly assumed that the professional architect, attorney, civil engineer, or physician works for a specified client and that the needs and wishes of the client determine the goals of the professional’s work. In this model of professional activity the architect designs a house that meets the living requirements of the client, while the physician plans a course of treatment for the patient’s ailments. Although in practice matters are not so simple, this definition of the professional role greatly facilitates the development of technologies for each of the professions, for it means that consequences going beyond the client’s goals don’t have to enter into the design calculations. The architect need not decide if the funds the client wants to spend for a house would be better spent, from society’s standpoint, on housing for low-income families. The physician need not ask whether society would be better off if the patient were dead.

If eating the apple revealed to us the nature of good and evil, modern analytic tools have taught us how to detect good and evil in minute amounts and at immense distances in time and space.

There is no reason to suppose that a modern industrial society is more conducive to human happiness than the simpler, if more austere, societies that preceded it. On the other hand, there seems to be little empirical basis for the nostalgia that is sometimes expressed for an imagined (and imaginary) happier or more humane past.

We must be alert also to the possibility that rationality applied to a broader domain will simply be a more calculatedly rational selfishness than the impulsive selfishness of the past.