But Here's a Try at Summarizing

As these inconsistent passages show, it isn't easy to determine Weizenbaum's position, but the following seem to be the book's main points:

1. Weizenbaum doesn't name any specific task that computers cannot carry out, because he wishes ``to avoid the unnecessary, interminable, and ultimately sterile exercise of making a catalogue of what computers will and will not be able to do, either here and now or ever''. It is also stated that human and machine reasoning are incomparable and that the sensory experience of a human is essential for human reasoning.
2. There are tasks that computers should not be programmed to do.

   Some are tasks Weizenbaum thinks shouldn't be done at all - perhaps for political reasons. One may quarrel with his politics, and I do, but obviously computers shouldn't do what shouldn't be done. However, Weizenbaum also objects to computer hookups to animal brains and computer conducted psychiatric interviews. As to the former, I couldn't tell whether he is an anti-vivisectionist, but he seems to have additional reasons for calling them ``obscene''. The objection to computers doing psychiatric interviews also has a component beyond the conviction that they would necessarily do it badly. Thus he says, ``What can the psychiatrist's image of his patient be when he sees himself, as a therapist, not as an engaged human being acting as a healer, but as an information processor following rules, etc.?'' This seems like the renaissance era religious objections to dissecting the human body that came up when science revived. Even the Popes eventually convinced themselves that regarding the body as a machine for scientific or medical purposes was quite compatible with regarding it as the temple of the soul. Recently they have taken the same view of studying mental mechanisms for scientific or psychiatric purposes.

3. Science has led people to a wrong view of the world and of life.

   The view is characterized as mechanistic, and the example of clockwork is given. (It seems strange for a computer scientist to give this example, because the advance of the computer model over older mechanistic models is that computers can and clockwork can't make decisions.) Apparently analysis of a living system as composed of interacting parts rather than treating it as an unanalyzed whole is bad.

4. Science is not the sole or even main source of reliable general knowledge.

   However, he doesn't propose any other sources of knowledge or say what the limits of scientific knowledge is except to characterize certain thoughts as ``obscene''.

5. Certain people and institutions are attacked.

   These include the Department of ``Defense'' (sic), Psychology Today, The New York Times Data Bank, compulsive computer programmers, Kenneth Colby, Marvin Minsky, Roger Schank, Allen Newell, Herbert Simon, J.W. Forrester, Edward Fredkin, B.F. Skinner, Warren McCulloch (until he was old), Laplace and Leibniz.

6. Certain political and social views are taken for granted.

   The view that U.S. policy in Vietnam was ``murderous'' is used to support an attack on ``logicality'' (as opposed to ``rationality'') and the view of science as a "slow acting poison". The phrase ``It may be that the people's cultivated and finally addictive hunger for private automobiles '' (p.30) makes psychological, sociological, political, and technological presumptions all in one phrase. Similarly, ``Men could instead choose to have truly safe automobiles, decent television, decent housing for everyone, or comfortable, safe, and widely distributed mass transportation.'' presumes wide agreement about what these things are, what is technologically feasible, what the effects of changed policies would be, and what activities aimed at changing people's taste are permissible for governments.