The Strange Persistence of QWERTY

The keyboard you are reading this on uses a layout designed in 1873 for a mechanical typewriter that no longer exists, to solve a problem that disappeared more than a century ago. It is one of the strangest survivals in the history of technology, and the reasons for its persistence say more about how systems get locked in than about how they get designed.

The original problem

Christopher Latham Sholes was a Wisconsin newspaper editor and amateur inventor. In the late 1860s he was tinkering with a writing machine — what would become the Remington No. 2 typewriter. The early prototypes had a serious mechanical flaw: when typists pressed two adjacent keys quickly, the typebars would jam against each other before they could swing back into place.

The fix was not to redesign the typebars — that came later — but to redesign the layout. Sholes spread the most common letter pairs apart on the keyboard, so that typists physically could not hit them in rapid succession with the same hand. TH goes to opposite hands. ER goes to opposite hands. The letters that English uses most often are scattered across the layout in a way that looks random and is actually anti-ergonomic on purpose.

By 1873 the layout was finalized roughly as we know it. By 1878 the QWERTY arrangement was standard on every Remington. By the early 1900s, the mechanical problem it was solving had been engineered out of typewriters entirely. Typebars were redesigned, key actions improved, and jamming was no longer the bottleneck.

The layout, however, stayed.

The Dvorak experiment

In 1936, August Dvorak, an educational psychologist at the University of Washington, patented a layout he believed was demonstrably better. The Dvorak Simplified Keyboard puts the most common letters on the home row, places vowels under the left hand and the most common consonants under the right, and balances the workload so that around 70% of typing happens without leaving the home row (versus around 32% for QWERTY).

Dvorak ran experiments through the U.S. Navy in the 1940s and reported that typists trained on his layout were faster, more accurate, and less fatigued. The studies were criticized later — Dvorak owned the patent, the methodology was loose, and the speed gains for trained typists were modest. But even skeptical replications find that Dvorak is at least somewhat better for sustained typing, and it is certainly easier on the hands.

None of this mattered. By the time Dvorak was patented, QWERTY had thirty years of installed base, millions of trained typists, and a complete training infrastructure. Switching meant retraining the typist, replacing the typewriter, and convincing schools to teach a new system. Dvorak made the better mousetrap. The world kept the worse one.

The economics of lock-in

QWERTY became the textbook example for economists studying what they called path dependence — the tendency for early decisions to lock in outcomes that later look indefensible.

Paul David's 1985 paper "Clio and the Economics of QWERTY" laid out the argument. Three forces hold a suboptimal standard in place once it gets going. First, technical interrelatedness: typewriters were built for QWERTY, so QWERTY typists were the most valuable, so businesses bought QWERTY typewriters, so manufacturers built QWERTY typewriters. Each part reinforced the others. Second, economies of scale: the more QWERTY existed, the cheaper QWERTY became, in both hardware and training. Third, quasi-irreversibility of investment: every typist had spent hundreds of hours on QWERTY muscle memory. That investment was a sunk cost they would lose by switching.

The result is that even when a better technology exists, switching costs exceed the gain for any individual actor. Everyone is rationally locked into the worse option. This is not a market failure in the strict sense — markets work fine within the standard — but it is a kind of historical accident that economic theory had not previously had a name for.

QWERTY became shorthand for the whole phenomenon. It is now invoked to explain why the world uses VHS over Betamax, why the U.S. drives on the right and Britain on the left, why English uses non-metric units, why JavaScript is the language of the web, why nuclear power plants are built around enriched uranium rather than thorium. Some of these analogies are stronger than others, but the framework is the same.

The complication: maybe QWERTY isn't that bad

In 1990, the economists Stan Liebowitz and Stephen Margolis published a counter-argument titled "The Fable of the Keys." They re-examined Dvorak's claims and found them weaker than the legend suggested. They cited modern studies in which the speed difference between QWERTY and Dvorak, for fully trained typists, is small — perhaps 5%. They argued that QWERTY's persistence may simply reflect that it is good enough, not that it is locked in despite being inferior.

This is a real complication. The lock-in story is too tidy. QWERTY may have started as an awkward solution to an extinct problem, but it is not actively painful. The vast majority of typists who try Dvorak go back to QWERTY within weeks, not because of training friction but because the gain is too small to bother.

What QWERTY shows us is not that markets fail to optimize, but that the standard for "good enough" is much lower than designers like to imagine. Once a tool is workable, the bar for replacement is not "is this an improvement" but "is this enough of an improvement to justify retraining everyone, replacing all the hardware, and breaking the muscle memory of three billion typists."

That bar is very high. Almost nothing clears it.

The contemporary tax

The smartphone keyboard inherited QWERTY because that's what people knew, even though the original mechanical reason — preventing physical typebar jams — has nothing to do with a touchscreen. Mobile predictive text and swipe input have started to relax the layout's grip; on a phone, you barely "type" at all anymore. But the visible arrangement remains.

The interesting question is whether voice input, gesture-based input, or direct neural interfaces will finally do what Dvorak couldn't — make QWERTY irrelevant by removing the keyboard from the path. If you don't type, you don't care what layout you're not typing on. The layout problem dissolves rather than gets solved.

Until then, every keyboard you touch is a fossil of the 1873 mechanical workaround. The world has rebuilt itself around computers a hundred times since then, and through every revolution the layout has stayed the same.

The lesson, if there is one

The design of physical and digital systems is full of decisions that look arbitrary because they were made for reasons that no longer apply. The metric system competes with imperial. UTC competes with local time. UTF-8 competes with everything that came before it. In each case the question is not which system is better in isolation but which system is too entrenched to dislodge.

Designers who hope to build the next standard should remember: being better is not enough. Being better by enough to overcome a hundred years of hardware, training, and habit — that is the bar. QWERTY is not on your keyboard because it is good. It is on your keyboard because nothing has been good enough, for long enough, to displace it.