New ideas become simpler over time. This is most obvious in intellectual domains, where the work that introduces an idea never shows how simple it can be. The first draft, being new, is labored, and being new-fangled, is cautious.
This is less obvious in other domains. A late mechanical clock, though compact, looks far more complex than a room-filling medieval clock. But the idea is not the clock; it is the escapement. The painstaking blacksmith, evaluating materials, working and reworking them, test-fitting, filing and re-fitting – his efforts were more complex than the industrial procedures which allow an escapement to be made by someone who has no idea what one is. Likewise the modern computer looks far more complex, though compact, than the Cold War computer, big as a car, a room, a house. But the new idea was not the machine; it was the transistor, and now Shockley’s circuit, which took days to build, is printed by the millions in fractions of seconds.
Of course, as an artifact, the industrial clock is far more complex than the medieval, and the post-industrial computer is far more complex than the industrial. But artifacts are not ideas. Indeed the clearest definition of an idea, its clearest distinction from other kinds of constituent thought, is that the idea is the part that becomes simpler over time.
Ideas appear to obey a kind of conservation principle, one of complexity. In order for an idea to stand on its own, it must be complex in itself. In order for an idea to be simple, it must be embedded inside of a complex system. This is easy to understand for clocks and computers – as mechanisms and circuits get simpler, they get smaller and more fragile, and must be embedded in more complex, larger, more robust objects.
But simplicity is not a function of size. Consider guns: as the idea of driving a projectile with expanding gas got simpler, guns did get smaller – a path runs from the cannon that destroyed Byzantium’s walls to the concealable pistol – but they also got bigger – guns have been built to launch payloads into space. Or consider the internal combustion engine, which powers motor scooters as well as container ships.
This principle holds for all sizes of artifacts, and for all degrees of abstraction. Complex ideas in complex systems are possible, but perverse. Simple ideas in simple systems are also possible, but limited to the first steps of a technology. But for an idea to be useful, it must begin as a complex idea in a simple system, and end as a simple idea in a complex system.
Let me suggest some practical consequences.
1. By the time an idea has become simple enough to be generally understood it has usually ceased to be independently useful. Sometimes this is tautological: when everyone understands democracy, democracy already exists.
2. An idea is not a realization; a realization is not an idea. Few improvements are due to ideas; most are due to realizations. Someone realizes that step B could be eliminated by an alternative to step C; someone has the idea that the entire process is wrongheaded. To equate realizations and ideas both neuters useful but limited realizations by turning them into abstractions, and suppresses ideas by simplifying them prematurely. Treating realizations as ideas is how we get the anti-ideas of management theory. Losing ideas among mere realizations is how once-great company X is bankrupted by startup or foreign competitors, whose ideas inevitably turn out to have been screened as babies from company X’s torrential bathwater.
3. When an idea is new it may be unclear which part of the initial formulation is the idea. Often you must proceed with no more than a sense that your line of research contains a new idea somewhere in it. And even when the initial formulation is ready for use, use must sometimes be widespread and practical before the idea stands out.
Consider guns again. A submachine gun is a sort of hybrid of the rifle and the pistol. It uses pistol rounds in a rifle-sized frame. Since the gun is relatively heavy and the rounds are relatively low-powered, an individual can control the recoil when the weapon is fired on automatic. But the first submachine gun – the Thompson, that is, the Tommy gun – was not designed with this idea in mind.
One of its inventors had observed in his time on battleships that under the conditions of high pressure in the firing of a naval gun different metals would stick to one another. He called his observation (after himself) the Blish Principle of Metallic Adhesion and patented it as a way of dissipating recoil. In fact what makes recoil manageable is a heavy gun. But not only were the first Thompsons built with Blish’s slivers of brass in them, they continued to be built this way until the scale of wartime production eliminated the extra step. The weapon had been in service for two decades before the idea behind it became clear.
But enough complication. Surely I have given this idea enough complexity to start on.
