Professor of Computer Science at Stanford University

Dear Mr. President:

A President can only do a very few of the many things proposed for the improvement of American society. I want to make one positive proposal for which Presidential action is appropriate and to issue one warning.

The weakest area of science is social science, specifically as it evaluates the effects of changes in social institutions. These institutions include education, taxation, criminal justice, structure of government, the institutions that support scientific research. Proposed changes in these and other institutions need to be evaluated by systematic experiment with clear criteria of success and proper statistical evaluation. The experiments need enough publicity all the way through so that the results are heeded.

There are experimental programs today, but they are mostly too dilute. Thus every state or city or school district wants its share of experimental money, and there isn't enough money to experiment everywhere. Social experiments need to concentrate the money in a few places and provide enough scientific observers and statistical resources.

Because Congress and administrations often respond to pressures to divide the money "fairly", Presidential leadership can play a big role in concentrating experimental programs.

Now here's my warning; it may be unnecessary. The scientific community, worldwide as well as in America, is like other communities and given to fads and taboos. These taboos have prevented some subjects from being researched or even discussed. These subjects include the genetics of behavior and intelligence. When a program fails, the possibility that genetic differences are involved is not allowed to be mentioned. Another taboo in the scientific community that studies energy is nuclear energy. Nuclear energy provides the one guarantee that those aspects of American society that depend on high use of energy including personal mobility can be sustained indefinitely, i.e. beyond the supply of oil

John McCarthy
Professor of Computer Science at Stanford University
Pioneer in Artificial Intelligence