Once the focus of hot political controversy, the Cambridge Project has disappeared from public view in the past two years. Most people, if they remember it at all, have no idea that the Project continues, supported by a million and a half Defense Department dollars a year.
Conceived in 1969 as a joint Harvard-MIT project to develop computer methodology for the social sciences, it became a political target because of its Department of Defense (DOD) funding. In the uneasy political atmosphere of that fall, the Faculty debated whether to join the Project on an institutional basis while radical pickets, petitions, and demonstrations attacked it as a tool of U.S. government hegemony at home and abroad. While radicals attacked it as counter-revolutionary in intent, liberals accepted it as "value-neutral" basic research but questioned the ethical implications of DOD funding.
After months of tortuous hearings, subcommittee reports, and committee meetings, the Faculty voted that Harvard as an institution would not participate in the Project but that individual faculty members could. Two years, $4 million, and several dozen research contracts later, Harvard is as deeply involved as if the Faculty hadn't rejected the proposal--except that the President cannot appoint Harvard representatives to the Project's two ineffectual advisory committees. The Project goes on, but by opting for a low profile and avoiding publicity, it has been all but forgotten.
The purpose of the Project is the development of "on-line" computer capabilities for social science. Under the old "batch processing" method, the user would have to wait several hours or days for results from a computer. But "on-line" (or "interactive" or "time-sharing") computing means that many users can almost simultaneously obtain immediate results. To the social scientist, on-line computing enables formulation of each question based on the computer's answer to the previous question. The advantages of such a method for military decision-makers especially in a crisis situation, are evident. Researchers working under the Project's auspices are developing further elaboration of computer techniques to deal with the special problems of the social scientist: complex correlation and causal chains; many variables, none of which can be held constant; and textual data to be analyzed for thematic content.
These methods can be developed by the computer expert working abstractly. But, the Project's originators argue, methods most closely suited to the needs of social scientists can best be developed by the social scientists themselves, as they work with real data on a problem they are actually trying to solve. Thus the line between basic and applied research blurs, when applied to the Cambridge Project. While Project's stated goals may be widely-applicable techniques of modeling, data reduction, and statistical analysis, it grants funds to Harvard and MIT professors for what is also substantive research. Critics of the Project disagree as to which aspect is more useful to the Defense Department.
Two years ago, radical attacks on the Project focused on the specific substantive research aspects. The original funding proposal submitted to the Advanced Research Projects Agency (ARPA) of the Defense Department argued that the project would bolster projects (some already underway) such as work on "problems of the underdeveloped countries and on the conditions of stability in these areas so marked by turmoil" (emphasis theirs); analysis of texts rather than actual observation of movements and of countries (as in parts of the communist world where direct access is not possible; "the collection and processing of more than 750,000 pages of documents and analyses on contemporary communist and radical movements"; analysis of "several thousand detailed interviews with Viet Cong (which) will be released by Rand;" studies of "the problem of stability and disorder (by) cross-national comparisons of the performances of national governments;" "detailed analyses of the sociological and psychological bases of attitudinal change."
Under what conditions do peasants' protests become violent?"
Further, the funding proposal cited the kinds of data collections with which CAM would work: "world economic statistics," "U.S. economic statistics," "international armament expenditures and trends," "public opinion polls from all countries," "characteristics of local conflicts and limited was crises," "data on youth movements," "mass unrest and political movements under conditions of rapid social change," "peasant attitudes and behavior," and many more.
SDS, NAC (the November Action Coalition, SDS's now-defunct rival on the left), and Afro argued that the "abstract technology" aspect of the Cambridge Project was just a blind for counter-insurgency research. An SDS booklet read as follows:
"These kinds of data and much more will be processed for storage in the memory bank of a computer (one console of which will be located in the Pentagon.) Then operations will be developed to correlate and use them. To do what? To do things like estimate the number of riot police necessary to stop a ghetto rebellion in City X that might be triggered by event Y because of communications pattern K given Q number of political agitators of type Z; to plan a coup in country A where government B correctly assesses the needs of its people to be C and D and among whom the trend in public opinion is becoming more and more favorable according to indicators F-G..."
This radical critique proved overly simplistic, for almost none of the specific research projects cited in the original proposal have materialized under the Cambridge Project. Some had ended by the time the Project was underway; some sponsors objected to the Project's DOD funding; some had never intended to participate in the project. And in fact the funding proposal, on careful rereading, only mentions them as examples of the type of research that might benefit by the project. A common argument in defense of the proposal is that writing a grant proposal to the government is a game. The object is to make your particular academic interest sound as beneficial as possible to the potential source of funds; without necessarily intending to produce what the source wants.
Whether the specific data and research projects in the funding proposal reflected the intention of the Project's framers or an attempt to coax funds from the DOD is comparatively insignificant. The importance of the Cambridge Project lies in the "abstract technology" dismissed as a front by radicals. The Defense Department may never use specific programs produced by the Project, but it will use the technical expertise developed in the process of writing them. The funding proposal explicitly states the conceptual tools the DOD needs for its behavioral science problems:
Even an offhand list of behavioral science topics of interest to DOD rapidly becomes long, including, for example, leadership, organizational communications, personnel, training, policy analysis, public attitudes, morale, the psychology of deterrence, the psychology of bargaining, adjustment to foreign cultures, selection, allocation and assignment, man-machine communication, combat effectiveness, area knowledge, economic resources, and manpower utilization...Most of them are problems in which our understanding has been inhibited partly by the inadequacy of present modes of data management, analysis, and modeling.
Thus, when six professors at the Business School developed "a large simulation model of a competitive market" so that "750 M.B.A. candidates" could play a run-the-economy game, the validity of the particular model would be less important to the DOD than the understanding of information gathering and exchange for crisis-decision-making. This understanding would not be tied to the particular situation but would have a broad usage at the DOD and other government agencies as well. The analysis of Professor Griffith's collection of documents on comparative communism would be valuable in itself but even more valuable would be a technique for semantic analysis and thematic recognition by computers. The CIA, for example, collects enormous amounts of raw intelligence data from which meaningful patterns must be inferred. The sheer bulk of data makes the process too laborious to be done by hand. Programming a computer to recognize important themes and sift out relevant data would greatly reduce the human workload. The FBI could also use such a technique for tapped phones: instead of having an agent listen to all the conversations, a computer equipped to recognize spoken words could monitor the phone and print out conversations relating to themes it had been programmed to recognize as important.
The funding proposal documents further uses of computer techniques for DOD problems:
Read more in News
STATUESQUE, BUT IMMORAL