Skip to content

Vannevar Bush and the Memex

Abstract

Vannevar Bush (1890–1974) ran American wartime research as head of the Office of Scientific Research and Development, chaired the committee that oversaw the early Manhattan Project, and in 1945 wrote the report that led to the National Science Foundation. He is remembered in computing history primarily for a 1945 essay describing a device he called the Memex: a hypothetical desk-mounted information machine that would let users store, retrieve, and navigate between documents through associative links. The Memex was never built. But the concept of linking documents by association rather than by hierarchy or sequential order directly influenced J.C.R. Licklider, [[Douglas_Engelbart_and_the_Mother_of_All_Demos|Doug Engelbart]], Ted Nelson, and Tim Berners-Lee, making Bush, through the force of a single speculative article, one of the intellectual ancestors of the World Wide Web.

Vannevar Bush
Vannevar Bush, 1938. Image: Harris & Ewing, public domain, via Wikimedia Commons.

The Engineer Who Organized American Science

Vannevar Bush was born in 1890 in Everett, Massachusetts. He submitted his doctoral thesis in April 1916 and received a doctorate in engineering jointly from MIT and Harvard; when his adviser Arthur Kennelly demanded more work, Bush refused and the department chairman overruled Kennelly. In 1919 he joined MIT’s Department of Electrical Engineering, where he built analog computing devices.

Bush’s most important pre-war engineering achievement was the Differential Analyzer (1931), an analog mechanical computer that solved differential equations by using sets of rotating discs and integrating gears. It was used for ballistic tables, physics calculations and engineering problems, and copies were built in Britain, Norway, Japan and elsewhere in the US. Its successor was larger in every respect: the Rockefeller Foundation gave $10,000 for planning in 1935 and $85,000 for construction in 1936, and when Bush left for Washington his MIT colleague Samuel Caldwell finished the Rockefeller Differential Analyzer. According to Bush and Caldwell’s own description it had about 2,000 valves, several thousand relays, 150 motors, some 200 miles of wire and a weight of about 100 tons, yet its mathematical core was still eighteen disc integrators, as in 1931. It was demonstrated at MIT on 13 December 1941, six days after Pearl Harbor, worked in secret on war problems, was revealed to the public on 29 October 1945 and was shut down in 1954, a dead end overtaken by electronic digital machines.

Bush became vice president of MIT and Dean of Engineering in 1932, and President of the Carnegie Institution of Washington on 1 January 1939. By 1940, when it was clear that the United States would eventually enter World War II, Bush had positioned himself as the obvious person to run the scientific mobilization.

The War: Organizing Scientific Power

On 12 June 1940, Bush met President Roosevelt and presented a single page describing a government agency that would coordinate military research and give civilian scientists authority to direct war-relevant projects. Roosevelt approved it in about fifteen minutes, and the National Defense Research Committee (NDRC) was formally established on 27 June 1940.

Bush understood something that most military officers did not: that modern war would be won by technology as much as by manpower, and that producing the right technology required giving scientists authority to set research agendas rather than merely executing engineering specifications. This was a radical proposition in 1940. The success of the model would prove it correct.

Under Bush’s direction, the NDRC (and its successor, the Office of Scientific Research and Development, created in 1941) funded: radar development at MIT’s Radiation Laboratory (which became the single most important technological factor in Allied victory in the air and at sea), proximity fuse development that made artillery far more effective, penicillin mass production that saved hundreds of thousands of lives, blood plasma research, and dozens of other projects. Bush started the Manhattan Project, secured it top priority and chaired its Military Policy Committee, while operational management went to General Leslie Groves of the Army and the scientists themselves.

The war-time science program Bush organized was one of the most successful research mobilizations in history. It also created the template for peacetime science policy: federal funding of university research through competitive grants, distributed to researchers chosen by peer review rather than government administrators. This model (codified in the Endless Frontier report Bush submitted to Truman in 1945) created the NSF, shaped NIH’s grant-making, and remains the basic structure of American academic science.

“As We May Think”: The 1945 Essay

The essay grew out of the Rapid Selector, a microfilm search machine that Bush and assistants including Claude Shannon worked on at MIT in the late 1930s, in which film strips raced past photocells. A first attempt to publish his ideas failed in 1939. In July 1945, as the war ended, The Atlantic published Bush’s essay “As We May Think.” It was the most influential piece of speculative technology writing of the 20th century.

Bush began by observing that the accumulation of human knowledge had outpaced the tools for storing and retrieving it. Scientists were drowning in literature. Important work was being duplicated because researchers couldn’t find what had already been done. The existing tools for organizing knowledge (library catalogues organized by subject headings, alphabetical indices) imposed hierarchical classifications that forced users to guess which category the information had been filed under, even if they knew exactly what they were looking for.

Bush proposed an alternative organizing principle: association. The human mind, he wrote, operates by association. When you think of one concept, related concepts appear, not because they were filed nearby in some catalogue but because the mind has learned that they are connected. A good memory system, Bush argued, should work the same way.

He described the Memex:

“Consider a future device for individual use, which is a sort of mechanized private file and library. It needs a name, and to coin one at random, ‘memex’ will do. A memex is a device in which an individual stores all his books, records, and communications, and which is mechanized so that it may be consulted with exceeding speed and flexibility. It is an enlarged intimate supplement to his memory.”

The Memex took up about a page and a half of the eight-page article, which had no illustrations. The familiar picture of it, a desk with screens, keyboard and camera, comes from the condensed version Life printed in September 1945, drawn by the artist Alfred Crimi. The Memex as Bush envisioned it was a desk with two screens, a keyboard, and a collection of microfilm, the state-of-the-art storage medium of 1945. Users would store documents and create trails: chains of associative links between documents that represented a line of research or thought. These trails could be shared with other users. A researcher studying, say, antibiotic resistance could follow trails created by other researchers, seeing not just the documents they had consulted but the connections they had seen between them.

The Memex was never built, though Engineering Research Associates delivered a working Rapid Selector to the US Department of Agriculture in 1949, holding 72,000 frames on 600 metres of film. The technology Bush imagined (microfilm storage, dry photography for rapid copying, mechanical indexing) was wrong. The concept of associative links between documents was right.

The Hypertext Connection

Ted Nelson, who coined the term “hypertext” in 1965, explicitly credited Bush’s Memex as the inspiration. Doug Engelbart, whose “Mother of All Demos” in 1968 showed collaborative hypermedia, read “As We May Think” in 1945 as a navy radar technician in the Philippines and named it as an influence on his research. Tim Berners-Lee’s 1989 Web proposal cited Nelson, not Bush, but in 1995 Berners-Lee spoke at the symposium marking the essay’s fiftieth anniversary.

The Postwar Years: From Power to Decline

Bush’s influence peaked in 1945. He never built the working relationship with Truman that he had had with Roosevelt, and after chairing the new Research and Development Board in 1947–48 he lost his place at the centre of defence science.

His Endless Frontier report, which proposed a National Research Foundation with strong autonomy from political control, was reshaped in Congress: Truman pocket-vetoed a first NSF bill in August 1947 because it gave the agency too little accountability to the President, and the NSF that emerged in 1950 was more government-directed than Bush had wanted. Bush was not chosen to head it.

Bush led the Carnegie Institution until 1955 and then returned to Massachusetts, writing books (Modern Arms and Free Men, 1949; Science Is Not Enough, 1967), serving on corporate boards, and watching the computing revolution he had helped catalyze develop in directions he had not anticipated. His Differential Analyzers were replaced by digital computers; the ENIAC and its successors operated on principles different from anything he had built. He was not a digital computing pioneer; his contributions were organizational and conceptual.

He returned to the idea in “Memex Revisited” in 1967 and died on 28 June 1974, two decades before the World Wide Web made the associative linking concept he had described in The Atlantic into the primary navigational structure of human knowledge on a global scale.

Legacy: The Idea That Outlasted the Technology

Bush’s legacy operates on two levels. The institutional legacy is the framework of federal research funding that created postwar American technological supremacy: the NSF, NIH’s extramural grant program, and the university research system that trained the scientists who built the digital age.

The intellectual legacy is the Memex concept, which seeded a line of thinking about human-computer interaction and information retrieval that ran from Bush through Licklider through Engelbart through Nelson to Berners-Lee. The specific insight (that information should be organized by association between ideas rather than by hierarchical classification) was correct in 1945 and remains the dominant metaphor for how we navigate digital information today. Every hyperlink is a small Memex trail.

📚 Sources