QWERTY Was Built for Telegraphs, Not Sabotage

In American Morse code, the letter Z is four short pulses [S1]. A receiver clattering away at a Sholes typewriter in 1872 would hear those four taps and have to decide, in real time, whether the sender meant Z or the much commoner pair S-E [S1]. So when Christopher Latham Sholes laid out his keyboard, he put S next to both [S1].

That is not the story you were told.

The story you were told is that QWERTY was engineered to slow typists down so mechanical typebars would stop jamming — a tidy parable about how bad design can outlive its reason for existing. It is wrong on every load-bearing detail: wrong about who the early customers were, wrong about what Sholes was solving for, and wrong about the famous experiment that supposedly proved the layout inefficient. The one part of the keyboard myth that didn't happen is the part everyone knows.

Start with the iteration record. Sholes filed his first typewriter patent in October 1867 [S1]. In November 1868 he reversed the second half of the alphabet, N through Z, right to left [S1]. In April 1870 he yanked A, E, I, O, U, and Y up to the upper row [S1]. The arrangement we'd recognize as QWERTY-ish appeared just before Remington began production in 1873, and Sholes patented that version in 1878 [S1]. Six years of accretion. Koichi and Motoko Yasuoka, who did the archival work on these patents at Kyoto University, describe the layout as having "accidentally grew into QWERTY" through a chain of conflicting user requirements [S1] — closer to a fossil than a blueprint.

Who were those users? Not the office secretary of cinematic stereotype. The earliest heavy buyers of Sholes' machine were telegraph operators, who received Morse pulses by ear and had to transcribe them into plain English on the fly [S5]. The Yasuokas trace the early Type-Writer keyboard directly to the Hughes-Phelps Printing Telegraph and conclude it "was developed for Morse receivers" [S1].

This is where the jamming story collapses on its own logic. A Morse receiver has to match the sender's transmission speed in real time [S1]. Deliberately crippling the operator's hands would have made the machine useless for its heaviest early customer [S1]. You don't sell a chainsaw with a brake on it to lumberjacks.

What Sholes actually optimized for was Morse disambiguation. American Morse — distinct from the International Morse that came later — renders Z as four short pulses, "· · · ·" [S1]. Receivers routinely misheard those four taps as the digram S-E, which is far more frequent in English [S1]. The Yasuokas' conclusion: "S ought to be placed near by both Z and E on the keyboard for Morse receivers to type them quickly" [S1] — which is exactly where S sits in QWERTY [S1]. Z is not exiled to the bottom corner because it's rare. It's parked next to S so an operator who guessed wrong could correct without moving a hand [S1].

So where did the slow-you-down story come from? Not from Sholes [S1]. His patents and correspondence don't make the claim [S1]. The retrofit emerges decades after the layout was set, once QWERTY was the entrenched standard and people had reason to want an explanation for why a standard might be bad [S1]. The most influential version of "bad" arrived in 1944.

That year, Lieutenant Commander August Dvorak ran a U.S. Navy study on 14 typists and reported his own layout, the Dvorak Simplified Keyboard, produced speeds about 74% faster than QWERTY after roughly 52 hours of retraining [S4]. The number is still cited in business-magazine listicles. What's usually left out is that Dvorak was the Navy's top time-and-motion analyst, personally owned the patent on the layout he was evaluating, and was selling that layout to the Navy that funded the work [S4]. Liebowitz and Margolis later argued the experiments were poorly designed and biased in Dvorak's favor [S2].

In 1956 the General Services Administration commissioned Earle Strong to run a controlled cost-benefit test of switching federal typists to Dvorak [S2][S4]. Strong found no advantage, and possibly a small QWERTY edge [S2][S4]. The result was never seriously replicated in Dvorak's favor [S2].

The story had legs anyway, because economists needed it to. In 1985 Paul David published "Clio and the Economics of QWERTY" in the American Economic Review, defining path dependence as outcomes shaped by "temporally remote events, including happenings dominated by chance elements rather than systematic forces" [S3]. QWERTY was his canonical example — a worse design locked in by accidents of history [S3]. The framework was load-bearing for a whole literature on market lock-in, and it needed a vivid case [S3].

Five years later, Stan Liebowitz and Stephen Margolis published "The Fable of the Keys" in the Journal of Law and Economics [S2]. They re-examined the Navy study, the Strong GSA study, and the ergonomic literature, then walked through the historical record of layouts that actually competed against QWERTY when QWERTY became dominant [S2]. Their conclusion: "QWERTY is about as good a design as any alternative" [S2].

That conclusion is still contested. Dvorak advocates argue Liebowitz and Margolis cherry-pick studies, dismiss the 1944 data too quickly, and underweight finger-travel and home-row measurements that favor Dvorak [S6]. The narrow Dvorak-vs-QWERTY question isn't settled.

But the jamming origin story is. Sholes' patents don't make the claim [S1]. His correspondence doesn't make the claim [S1]. The early-customer base would have rejected a machine designed to be slow [S1]. The Yasuokas call the slow-down explanation "nonsense" outright [S1].

Which leaves the meta-question, the one that actually matters. Why does the wrong story keep winning?

Partly because it's a better story. "Standard turned out fine, actually" doesn't fit on a podcast pull-quote. "Civilization is stuck with a bad layout because of 1870s manufacturing constraints" does. Partly because economists wanted a textbook case of lock-in, and QWERTY's history was old enough and weird enough to volunteer for the role [S3]. And partly because we are vastly more interested in stories about inefficiency than in stories about boring competence.

The keyboard under your hands is not a monument to lock-in. It's a fossil of a workflow that hasn't existed for a hundred years — a tool shaped, with surprising care, for people who decoded electrical pulses into English at the speed of a human ear [S1][S5]. The four dots of American Morse Z are still there, ghosted into the shape of your home row [S1].

You're typing on an answer to a question nobody asks anymore. That is a stranger and better thing than being stuck with a mistake.