Memex Dreamer, Wartime Science Boss
He built a mechanical equation-solver at MIT, ran America’s wartime science, then imagined the memex: a desk that remembers the trails you think along.
Engineer & Imaginer of the Memex
Vannevar Bush learned to make difficult problems tractable with instruments, then organised scientists to meet the demands of war. In 1945 he offered two different tools for the peace: the imagined memex for following ideas, and a public system for sustaining research.
Bush studied engineering at Tufts and returned to MIT in 1919 to teach electrical power transmission. The equations describing a power network were laborious to solve by hand. That difficulty made an instrument for working through them more than a mathematical curiosity.
At MIT, Bush and colleagues built the differential analyzer around rotating shafts and wheel-and-disc integrators. Harold Hazen began the project as Bush’s student; setting up a new problem required skilled hands to connect the machine’s parts. It found uses beyond electrical networks, in physics and ballistics. Claude Shannon later worked with the analyzer as a graduate student, bringing another kind of mathematics to its relay controls.
Bush left MIT leadership for the Carnegie Institution in 1938 and, during the war, directed the Office of Scientific Research and Development. The change from building a machine with colleagues to coordinating laboratories across the country gave his later question a practical edge: how could specialised knowledge be found and used by other people?
In As We May Think, Bush begins with the abundance of scientific records and the difficulty of finding the relevant result. His memex is an imagined desk of microfilm, viewing screens, controls and photographic capture. It was a proposal, not a machine he built or a description of the later web.
The crucial operation is simple to picture. A reader displays two records side by side, gives their connection a code, and adds that link to a named trail. The trail can be annotated, revisited and copied into another reader’s memex. An index retrieves a known item; a trail preserves a route through several items and some of the judgement that made the route useful.
Read the worked examples in the essay before looking for descendants in later computing. They ask what happens when a scholar can leave a traversable account of a line of thought, rather than only a citation at its end.
As We May ThinkStart here because the memex is easiest to understand in Bush’s worked examples: a reader links, annotates and shares trails through records. Notice that it is a hypothetical microfilm machine, not a description of the later web.Read the essayPresident Roosevelt asked Bush in 1944 how the wartime application of science could continue in peace. Science, the Endless Frontier, delivered to President Truman in 1945, proposed federal support for basic research in universities and other nonprofit institutions, together with scholarships and fellowships for developing scientific talent. Its committees also considered medicine and publication of scientific information. Bush’s report grew from a collective consultation, though he supplied the policy argument and direction.
That plan entered a political contest. Senator Harley Kilgore had advanced bills for a broader agency with more political accountability and geographic distribution of funds. Bush favoured scientific autonomy and selection by research merit. The National Science Foundation was established in 1950 after years of debate and compromise; it was not simply Bush’s report put into law.
The memex and the research foundation address different bottlenecks. One helps a reader make a path through records; the other decides how inquiry is supported and by whom. Bush died in 1974, leaving both proposals open to the same useful question: who gets to participate in making and finding knowledge?
Science, the Endless FrontierA proposal for supporting scientific research after the war.Read the PDF1945 · Vannevar Bush · Reprinted at Carnegie Mellon
Start here because the memex is easiest to understand in Bush’s worked examples: a reader links, annotates and shares trails through records. Notice that it is a hypothetical microfilm machine, not a description of the later web.Read the essay1945 · Vannevar Bush · NSF anniversary edition
A proposal for supporting scientific research after the war.Read the PDFVannevar Bush · Revisiting the associative information machine
Two decades later Bush revisits associative trails, voice annotation and shared records, showing what he retained and revised after digital computing had advanced.Read PDF1931 · Vannevar Bush · Original engineering paper
The original engineering paper explains an analogue machine that integrates differential equations through coupled mechanical components.View the paper1936 · Vannevar Bush · American Mathematical Society
Bush and colleagues explain how machines can transform mathematical analysis, a practical precursor to his later reflections on intellectual tools.Read the paper1929 · Vannevar Bush · Reprinted in The Essential Writings
Bush situates circuit calculation as an operational mathematical practice tied to the engineer’s actual problems.Read the work1932 · Vannevar Bush · Reprinted in The Essential Writings
A short address on research institutions and practical invention, useful for testing Bush’s own account of how organised science succeeds.Read the work
Claude ShannonA direct connection through the differential analyzer at MIT
Douglas EngelbartContinue from associative trails to a working augmentation system
