The Tyranny of History

Here is a question that sounds trivial and isn’t: why is your keyboard laid out the way it is? The honest answer is that in the 1870s a man named Christopher Sholes arranged the letters partly to slow typists down — to stop the metal type-bars of early mechanical typewriters from clashing and jamming when common letter-pairs were struck in quick succession. It worked. The machine sold. And then something strange happened: the reason for the layout evaporated — electric typewriters, then computers, have no bars to jam — but the layout didn’t. It’s still on the phone in your pocket. Roughly every keyboard on Earth inherits a design choice made to solve a problem that has been dead for a hundred years.

Nobody re-decides this each morning. You don’t wake up and choose QWERTY over a faster layout; you inherit it, because everyone you’ll ever share a keyboard, a document, or a job with already uses it, and the cost of being the one person who types differently swamps whatever speed you’d gain. That’s the whole phenomenon in miniature. The outcome we’re all stuck with wasn’t selected because it’s best. It was selected early, and then it fed on itself until it couldn’t be undone.

The one idea to take away

Before we unpack it across the course, here’s the entire model compressed into one sentence:

The phrase doing the secret work is more attractive the more it’s used. Ordinary goods don’t behave that way — the tenth person to buy a sandwich doesn’t make the sandwich better for the eleventh. But a standard does: the tenth person to adopt a file format, a language, a currency, a social network, or a keyboard makes it more valuable to the eleventh, because now there’s one more person to exchange with. That single twist — economists call it increasing returns to adoption — inverts the usual story where the best product wins. When returns increase, whoever gets ahead early gets further ahead, and the winner is decided not by quality but by the accidents of who showed up first. Path dependence is the name for outcomes shaped by that sequence; lock-in is the name for the trap it leaves you in.

Notice what you just did: you predicted a real market outcome without knowing a single technical fact about either format, purely from the shape of the feedback. That portability — same logic, wildly different domains — is exactly why path dependence earns a place in the latticework rather than staying a footnote about cassette tapes.

Watch a market lock in — and watch it pick the wrong winner

Reading “the early lead gets amplified” is one thing; watching it happen, and happen differently each time, is another. Below is a market with two competing standards, A and B. Crucially, B is the genuinely better one — a new adopter, all else equal, mildly prefers it. Each newcomer picks a standard based partly on its quality and partly on how many others already use it, and the slider sets how much that second factor matters — the strength of increasing returns.

Start with the strength at 0 and run it: adoption tracks pure quality, so the better standard (B) keeps a steady edge and nothing locks in — the market “gets it right.” Now crank the strength up past 1× and run it again. A tiny, random early lead — which standard happened to grab the first few adopters — now snowballs, and the market tips hard onto one standard and freezes there. Hit New run a few times at high strength and watch the punchline: sometimes the better standard wins, and sometimes an early accident locks the whole market onto the worse one. Same settings, different history, different winner.

Three things are worth holding onto while you play. First, the slider moves smoothly but the outcome flips at a threshold — below a certain strength the market stays open and quality-driven; above it, it reliably locks. That’s your old friend critical mass, now operating on history. Second, the winner is contingent, not optimal: at high strength, re-running with the same settings hands victory to different standards, and the worse one wins a big share of the time — because the early lead is an accident and the accident is what gets amplified. Third, try the external shock: once a market is deeply locked, even a large jolt of adopters onto the better-but-trailing standard often just gets swallowed as the incumbent’s loop reasserts itself — a first taste of why lock-in is so hard to break, which is the whole back half of this course.

Why this model earns a place in the latticework

Path dependence is a thinking tool, not a QWERTY trivia night, because the same shape governs an absurd range of things you actually care about:

• Technology standards. QWERTY, VHS, the width of railway tracks, whether your country drives on the left or the right, the dominance of a programming language or an operating system — most were set by early, contingent choices and are now held in place by everyone else’s choices, not by ongoing superiority.
• Markets and platforms. A marketplace, social network, or app store is worth more the more people are on it. That’s increasing returns, and it’s why these markets tip toward a single winner and then defend that position — a moat is lock-in seen from the incumbent’s side.
• Institutions and law. Legal systems, corporate structures, city layouts, and national policies calcify around early decisions. “We’ve always done it this way” is often literally true — and literally the reason it can’t easily change.
• Your own life. The career you’re in, the skills you invested in, the city you settled in, the habits running on autopilot — each was easier to start than it will be to switch, because you’ve been accumulating advantages (and sunk costs) along that particular path. Path dependence runs at the scale of a single résumé.

Learn the shape once and you can read all of them with the same three questions: what’s the reinforcing loop, how strong are its increasing returns, and is the thing it’s locking in actually the best option — or just the earliest?

The map of the course

Six short teaching lessons, then one exam you can’t undo. The route:

1. What Path Dependence Really Claims — the precise model, and the crucial line between the trivial version (“the past influenced the present,” true of everything) and the real one (“a small early accident got amplified into a large, persistent, hard-to-reverse outcome”). Contingency versus efficiency; why “first” and “best” and “what won” are three different things.
2. The QWERTY Story — and the Honest Debate — the canonical example, told straight, and then the academic brawl over it (David versus Liebowitz and Margolis). Because a model you only ever see confirming itself is a superstition; you’ll learn where QWERTY is strong evidence and where the story is thinner than it’s told.
3. The Four Engines of Increasing Returns — what actually makes an option get better as it spreads: set-up (fixed) costs, learning effects, coordination / network effects, and self-fulfilling expectations. Learn to name which engine is running in any given lock-in.
4. Lock-In, Switching Costs & Standards Wars — why a whole market stays frozen on a worse standard when no single player can profitably switch first (a coordination trap), and a tour of the great standards wars: VHS/Betamax, driving sides, rail gauges, metric versus imperial, programming ecosystems.
5. Institutions, Careers & the Paths We Can’t Leave — path dependence out of the lab: why organisations, laws, and even your own résumé and skills lock in, and how to think about the paths you’re standing on.
6. How Lock-In Breaks — and Where the Model Lies — the honest limits. How locked-in systems do get overturned (shocks, discontinuous leaps, coordinated switching, sponsorship), and the three ways the model lies: the just-so-story trap, mistaking mere causation for lock-in, and assuming lock-in is forever.

Then a Final Exam — graded, one question at a time, one-way: once you answer, it locks. (Rather fitting for a course on irreversibility.) No back button, no retries, 70% to pass.

How to use this course

One rule does most of the work: guess before you peek. When you hit an exercise, commit to an answer before revealing anything — the small sting of being wrong is what burns the idea in. And play with the lock-in explorer above until the “same settings, different winner” punchline feels obvious in your hands; a contingency you’ve watched happen sticks far better than one you’ve only read about.

Next up: lesson 1, where we pin down exactly what path dependence claims — and, just as importantly, what it doesn’t.