Simulmatics Corporation

~ the goodbye ~

Ed Greenfield incorporated the Simulmatics Corporation in New York on 18 February 1959. He was a Madison Avenue advertising executive rather than a scientist, and the name he picked was a weld of simulation and automatic. The product was a program his staff called the People-Machine, and the sales line was that it could tell a client what a population would do before the population did it.

The first real contract was with John F. Kennedy's 1960 campaign. The People-Machine ran on IBM 704 hardware in FORTRAN and sorted the American electorate into 480 categories, built out of demographics so specific they read like photo captions: working class white protestant women in northern cities. For each category it predicted a response to each position Kennedy might take. The firm wanted the analysis to recommend speaking up for civil rights, and the analysis recommended speaking up for civil rights.

After the election a Harper's Magazine article revealed the campaign's use of the company. It was written by a Simulmatics employee working under the cover of a freelance writer, and nobody disclosed that at the time. The piece set off a national argument about whether an election had just been manipulated by a computer, and Simulmatics read that argument as advertising. It went public in 1961 and started taking contracts on automobile traffic and the effectiveness of ad campaigns.

Jill Lepore's reading, in her 2020 book If Then, is that most of what the People-Machine produced was common sense wearing a punch card. That is the part worth knowing about a company that went bankrupt fourteen years before I was born. The machine was never the product. The confidence was.

~ the chronicle ~

~ the people-machine ~

The 480 categories were the whole invention. Nothing about the hardware was unusual for 1960 and the statistical method was within reach of anyone with survey data and patience. What Simulmatics did was package the arithmetic as a machine with a name, so that the client got a computer's answer where a consultant's opinion used to go.

Eugene Burdick turned it into fiction before the company was five years old. He had done data work on Adlai Stevenson's 1952 campaign alongside Greenfield, and Greenfield asked him to join Simulmatics at its founding. Burdick declined, partly on ethical grounds, and then wrote The 480 instead. The company cooperated with the book and opened its confidential files to him, expecting good publicity. Publication in 1964 produced the opposite.

~ saigon ~

In 1966 the Pentagon hired Simulmatics through ARPA to work on pacification in South Vietnam. The job was to measure what Vietnamese villagers wanted and feared, then build propaganda and policy on top of the measurement. The company opened an office in Saigon.

Most of the original staff would not go, so the firm hired researchers who were less qualified and less carefully vetted. Almost none of them spoke Vietnamese, so students from the University of Saigon were used as the interface to every interview subject. Thematic Apperception Tests and the Cantril Scale were applied to rural farmers in search of a standard Vietnamese psyche. In Dinh Tuong and Chau Doc provinces the researchers recruited villagers to report on the sentiments of their own neighbors, which built a large database and a network of informants.

Complaints accumulated and the Saigon branch came under review. In late 1967 ARPA terminated every Simulmatics project and the office closed.

~ the long echo ~

The Kerner Commission paid Simulmatics $221,000 to analyze how television and newspapers had covered the riots of 1967. The content analysis covered fifteen cities, with employees sorting television footage into discrete categories by hand and entering the result on punch cards, under a rule where a burning building counted as a riot action only if somebody was shown lighting it. The finding was that 4.8 percent of coverage showed rioting in progress while 75 percent showed the aftermath, and that the coverage had not spread the riots. The commission's published report leaned on that chapter heavily.

In 1969, after debating Ithiel de Sola Pool about the war at MIT, Noam Chomsky published The Menace of Liberal Scholarship in the New York Review of Books and named both Pool and Simulmatics in it. Pool left the chairmanship of MIT's political science department. Molotov cocktails were thrown at his house and a bomb was placed in the political science building. Simulmatics was bankrupt by 1970, and the Wikipedia article about it today lists Cambridge Analytica under See also.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »