Brass Tacks The Cambridge Project

(The Cambridge Project is a new Defense Department-sponsored program which will provide a computer-based laboratory to behavioral scientists at MIT and Harvard with the purpose of promoting basic research in social science methodologies. Several social scientists from Harvard and MIT have already begun work under the sponsorship of the program, which is expected to provide some 7.6 million over a five-year period. Harvard has not yet decided whether or not it will become affiliated with the project. This article is a description and discussion of the Project's history and purposes.)

TOUNDERSTAND the Cambridge Project. we have to go back a few years, to the Pentagon's first disastrous brush with the new social science. In 1964 a program under the quaint name of Camelot was launched by the Special Operations Research Office of the U.S. Army. Camelot's purpose. according to an official description was

to devise procedures for assessing the potential for internal war within national societies; to identify with increased degrees of confidence, these actions which a government might take to relieve conditions which are assessed as giving rise to a potential for internal war; and to assess the feasibility of prescribing the characteristies of a system for obtaining and using the essential information needed for doing the above to things.

Translated back into English, Camelot was intchded to enlist social science data-gathering and model-building techniques in the service of America's global efforts to prevent social revolutions ("internal war"). The project was to concentrate on the Latin American countries, where left-wing insurgencies were getting to be a pretty scrious problem in the early 1960's, and a major field office was to be established in the region to co-ordinate data-gathering operations. The initial Camelot project was to be a three-to-four year undertaking with a total cost of about five million dollars, and in the early stages there was optimistic banter in some Washington circles about the possibility that the project would eventually mushroom into a $50 million extravganza extending into the distant future.

These bold designs aborted in the spring of 1965, when the project came under strong attack from left-wing and university groups in Chile. The project was soon widely characterized, not without a certain accuracy, as an espionage program designed to serve American imperialist policies in Latin America. The Dominican intervention of May 1965 cemented this feeling within Chile, and eventually the American Ambassador to Chile was moved to protest strongly to the State Department about what he felt to be the project's adverse effects on the U.S. position in the country. It was becoming apparent that if SORO itself was to be saved form the wrath of Defense Department officials who had doubted the value of social science research from the outset, then Camelot would have to be sacrificed, and in mid-1965, Secretary MacNamara announced the cancellation of the program. The episode concluded with a memorandum from President Johnson which gave the State Department the power to veto Government funding of any further social science research in foreign countries.


The years that followed have been increasingly lean ones for the scattering of bureaus within the Pentagon which concern themselves with social science research. The notion that the behavioral sciences could be of much use to the U.S. military has always been regarded with considerable skepticism by most ranking officials at the Pentagon, and after the Camelot disaster the job of selling the behavioral sciences was that much more difficult. This meant that such outfits as the Behavioral Science Program of the Defense Department's Advanced Research Projects Agency (ARPA) were increasingly hard put to justify their continued existence. What the Behavioral Science Program needed was a new largescale project that would produce usable and interesting results to impress authorities higher up in the Pentagon, and that wouldn't blow up in everyone's face as Camelot had. And so the Cambridge Project began to be born.

THE SECOND thread to our story begins at M.I.T. in the late fifties, where the revolution in computer technology was just getting underway. The great breakthrough that was being made at this time was in the development of computer programming techniques that allowed several different people to work on different tasks on a single computer simultaneously. The development of such "time-sharing" systems enormously improved the efficiency and usefulness of computers, and in 1963 ARPA agreed to fund a project which was intended to develop the full potentialities of the new time- sharing technology. This program (Project MAC) developed over several years a time-sharing system for the IBM 7094 computer, and by 1967 had brought this system to the point where the machine could be used simultaneously by 30 persons. This appeared to be the highest degree of development possible with the IBM 7094, and so the computer people got themselves a new grant and a new more advanced computer, and went off to see what they could do with that one. Thus M.I.T. was left with a rented IBM 7094 computer which would quickly become a liability to the Institute unless a new grant could be secured.

At this point, late in 1968 M.I.T.'s behavioral sciences mandarins under Professor Ithiel Pool began discussing the possibility that the Behavioral Science Program at ARPA could be interested in funding a vast new social science project to center around the then-idle IBM 7094 computer.

Pool's idea which was also pushed at M.I.T. by Professor J C. R. Licklider, was clearly just what the behavioral science people at ARPA needed to re-establish themselves with the Pentagon bureaucracy. It was not to be an information gathering project as Camelot had been, but would center instead on developing new ways of using and interpreting behavioral science data. Thus it entailed none of the diplomatic risks that had proved fatal to project Camelot (and almost fatal to the little social science bureaucracy within the Pentagon as well). At the same time the behavioral science officials at ARPA also believed that the M.I.T. project might convince the higher levels of the Pentagon research bureaucracy that the behavioral sciences could begin to approach the reliability and "hardness" of the natural sciences. Perhaps computers would work where foreign data-gathering had failed. In any case, there wasn't any harm in trying.

By last spring, Pool and Licklider were working closely with the behavioral sciences branch of ARPA in drawing up a proposal that the ARPA people would be able to sell on behalf of both M.I.T. and themselves. They had a tough job ahead of them: the project that they were working out would have to impress people in the Defense Department who didn't expect to be impressed by anything that the behavioral scientists and their ARPA friends could come up with. Specifically that meant John Foster, the Defense Department's top research official. Foster's scientific work has been concerned with thermonuclear bombs (he did his graduate work under Edward Teller), and while Cambridge's behavioral scientists seem to like Foster personally (he is something of a Strangelovian cowboy, with a fondness for zooming around at the controls of his own jet plane), it is very clear that Foster puts his faith in hardware, and has little appreciation of the new social science technology.

So the proposal that Pool, Licklider and their friends at ARPA were working on last spring was to be an attempt to convince the men who manage the Pentagon's research policies that 1) the behavioral science scould be developed to the status of a "hard" science, and 2' that such development would in turn concretely aid the Defense Department in achieving its goals at home and around the world.

THERE was something else. An M.I.T. project might have been all right from the viewpoint of the Behavioral Sciences Program, but the Program was really looking for something with a little more novelty than could be offered by yet another M.I.T.- ARPA contract. M.I.T. is the Defense Department's house whore, so although the content of the Pool-Licklider project might have been novel enough, the institutional arrangement could hardly be considered a breakthrough. But if Harvard could be persuaded to join the venture, the project would appear somewhat more new and exciting.

There were three reasons why Harvard was felt to be important to the success of the project. The first was simply the prestige of the Harvard name. A second reason was that Harvard's participation would enable ARPA to simplify and centralize its support of social science research in the Cambridge community.

The third reason is a bit more complicated. In the early sixties, the Pentagon's Defense Science Board invetigated the state of social science research and concluded. among other things, that one development which would help the social scientists along the road to developing a "real" science would be to organize themselvesinto research institutes along the lines of the natural science institutes which are found on many university campuses. Such development, it was felt, would increase interaction between social scientists and thus further the creation of an integrated discipline of "hard" social science. Shortly after this report was released, Licklider joined the ARPA staff and for a year and a half tried from Washington to encourage behavioral scientists to start forming such institutes. He didn't get much of a reaction: the top men in the social sciences still preferred their comfortable positions within University Departments, surrounded by coteries of graduate students and still able to associate with other scientists on an informal basis.

Nevertheless, the notion that the formation of social science institutes within the universities would represent a step toward the solidification of the behavioral sciences as a "hard" discipline appears to survive at the Pentagon, and so the Licklider/Pool/ARPA group had an additional reason for wanting to extend their project to cover all of Cambridge-although this didn't in itself amount to establishing a new center, it would bring large numbers of Cambridge social scientists together for the first time, and thus would be that much more pleasing to the natural scientists back in the Pentagon.