01 · The First Methods

THE SHORTEST PATH — A History of the Art and Science of Problem-Solving

A Scribe's Desk, Around 1650 BC

A man named Ahmes sits with a papyrus scroll nearly five meters long unrolled across his lap, copying out mathematics that was already two centuries old when he began. He tells us so himself, in the opening lines of the document: he is writing in the thirty-third regnal year of the Hyksos king Apophis, transcribing a text that goes back to the reign of Amenemhat III of the twelfth dynasty (British Museum; Wikipedia). The scroll he is producing will end up, three and a half thousand years later, in the hands of a Scottish lawyer named Alexander Henry Rhind, who buys it in Thebes in 1858; today the two largest fragments sit in the British Museum, cut apart at some point by antiquities dealers hoping to sell the halves separately (British Museum). We know it as the Rhind Mathematical Papyrus. Ahmes calls it, in his own opening words, something closer to "accurate reckoning for inquiring into things, and the knowledge of all things, mysteries... all secrets."

That is a large claim for a document that spends most of its length on bread, beer, and grain. The papyrus contains 84 problems, worked out one after another: how to divide loaves among men so each gets a fair share, how to convert one measure of grain into another, how to find the slope of a pyramid face given its height and base (British Museum). None of this looks like the mathematics we associate with "theorem." There is no proof in it, in the sense Euclid will later demand. What there is, problem after problem, is something almost more valuable to this book's history: a written, repeatable procedure, worked start to finish, that a second scribe — one who never met Ahmes and never will meet the original problem-setter — can pick up and execute correctly.

Look at Problem 24. A quantity, added to a seventh of itself, makes 19. Find the quantity. A modern student reaches for algebra: x + x/7 = 19, so 8x/7 = 19, so x = 133/8. Ahmes has no algebra — no symbol for the unknown, no equation to rearrange. Instead the papyrus does something that looks, at first, like cheating: it guesses. Try 7, since a seventh of 7 is a clean whole number. Seven plus a seventh of seven is eight. But the answer needs to be 19, not 8 — so the guess was off by a factor of 19/8. Scale the guess by that same factor — 7 times 19/8 — and the guess becomes correct: the true answer, expressed the way Egyptian scribes wrote fractions, comes out to 16 plus a half plus an eighth. This is the method later called false position — regula falsi in the Latin that would revive it more than two thousand years afterward — and its logic depends on nothing more than the fact that the equation is linear: whatever the initial guess produces, the final answer is off by the same proportion everywhere, so a single corrective multiplication fixes it in one stroke (Regula falsi, Wikipedia; MacTutor, Egyptian papyri).

Notice what has actually happened here. Nobody has "solved for x" in our sense. What Ahmes has done is compress an entire family of problems — any quantity plus any fraction of itself equals any target — into a fixed sequence of moves: guess something convenient, measure the error as a ratio, rescale. The genius is not in problem 24 alone. It is in the fact that problem 24 is written down as a method, separable from its numbers, so that a scribe training a thousand miles and a hundred years away can apply the same three moves to a different heap, a different fraction, a different target, and trust the answer. The route has become reusable. That, and not the number 16⅝, is what survived four thousand years to reach us.

Two Rivers, Two Habits of Mind

A few hundred kilometers to the east, and largely without contact with Egyptian scribal schools, Babylonian mathematicians were solving harder problems with an even more explicit sense that they were writing procedures, not just answers. The clay tablet now catalogued as YBC 7289 — a small disc, easily held in one hand, incised sometime between 1800 and 1600 BC — shows a square with its diagonals drawn in, and beside one diagonal a sexagesimal (base-60) numeral that works out to an approximation of the square root of two accurate to about six decimal places (YBC 7289, Wikipedia). Other tablets, grouped by historians under the heading "procedure texts," go further: they lay out, in prose instructions rather than worked numerical examples, the general steps for solving whole classes of problems — what we would now call quadratic equations, systems of two equations, canal-digging volumes — with the numbers in each problem serving as illustration rather than the point (MacTutor, Babylonian mathematics).

The distinction between the Egyptian and Babylonian habits is subtle but important for this book's argument. Egyptian texts, the Rhind papyrus included, read overwhelmingly as worked examples: solve this bread problem, then that one, each new problem a fresh occasion to apply a family of techniques the student is expected to notice by pattern-matching across dozens of cases. Babylonian procedure texts more often state the general steps first — "take this, do that, then this" — with a single concrete numerical case appended to show the abstract instructions actually run. The difference is the difference between a cookbook that only ever shows finished dishes and one that occasionally pauses to write down the technique of braising, once, so it can be pointed to instead of re-demonstrated every time. Both traditions "wrote the route," but the Babylonian tablets took one step further toward writing the route in a form abstracted from the particular number that happened to prompt it — a step toward what a modern reader would recognize as an algorithm, a word that would not exist for another two and a half thousand years, when it was coined from the name of the Persian mathematician al-Khwarizmi (his story belongs to chapter 05).

This matters for the thesis of this book because it is the earliest visible instance of a pattern that recurs at every later stage: a problem's difficulty shrinks the moment its solution procedure, not merely its solution, becomes an object that can be inspected, criticized, and improved by someone other than the person who first found it. A single answer helps one person once. A written method helps everyone, forever, with every future instance of the same shape of problem — provided someone bothers to write it down as a method rather than an anecdote.

Euclid's Machine

Around 300 BC, in the newly founded city of Alexandria, a mathematician about whom almost nothing personal is reliably known — not his birthplace, not his exact dates, barely even the certainty that "Euclid" refers to one person rather than a school — assembled thirteen books that would become the single most reprinted mathematics text in history: the Elements (MacTutor, Euclid; Euclid's Elements, Wikipedia). What makes the Elements the hinge of this chapter is not any single theorem inside it — most of the geometry was not new, gathered as it was from generations of earlier Greek mathematicians whose individual proofs are largely lost. What Euclid did that nobody had done at this scale before was turn the entire discipline of geometry into a single, explicit, checkable machine for producing new truths out of old ones.

The machine has three moving parts, and Euclid states all of them, in order, at the start of Book I: 23 definitions of the basic terms (point, line, angle...); five postulates, which are unproved starting assumptions specific to geometry (a straight line can be drawn between any two points; all right angles are equal to each other; and so on); and five further "common notions," unproved assumptions general enough to apply outside geometry entirely (things equal to the same thing are equal to each other) (Euclid's Elements, Wikipedia). Everything that follows — all 465 propositions across all thirteen books — is required to be derivable from only these starting points, using only two permitted construction tools (an unmarked straightedge and a compass) and strict logical inference. A proposition is not allowed to smuggle in a fact that "looks obviously true" from a diagram; it must trace back, step by step, to the postulates. You can read the opening definitions today essentially as Euclid wrote them, in the Greek text still preserved and studied at digital archives such as the Perseus Project (Perseus, Euclid's Elements, Book I).

This is the payoff of writing method down taken to its logical extreme. Ahmes wrote one procedure per problem. Euclid wrote the rules for generating and checking any procedure whatsoever within an entire field. Once the axioms and permitted moves are fixed and public, a claim is no longer "true because a respected geometer says so" — it is true if and only if a specific, traceable, reproducible chain of construction and inference connects it back to five postulates and five common notions. Any sufficiently patient reader, in any century, with no access to Euclid himself, can verify any proposition in the Elements by re-walking that chain. This is why the book survived, essentially unchanged in method, as the standard geometry textbook in the West for well over two thousand years, and why "Euclidean" still means, to a mathematician, not merely "about flat-plane geometry" but "built by the axiomatic machine."

The cost of building such a machine is a certain narrowing. Euclid's tools are fixed — straightedge and compass, nothing else — which is precisely why three problems the ancient world cared about (doubling the cube, trisecting an arbitrary angle, squaring the circle) resisted solution inside his system for two thousand years; they are not solvable with those tools, a fact only proved with nineteenth-century algebra. That narrowing is not a flaw. It is the design: a method earns the power to guarantee correctness by being explicit about exactly what it will and will not allow itself to do. Chapter 06 returns to this theme in its harshest form, when the same impulse — state precisely what is and is not permitted — produces a mathematics of provable impossibility.

The Two Archimedeses

A century after Euclid, on the island of Sicily, Archimedes of Syracuse left behind two utterly different accounts of how a mathematician actually finds things — and the gap between them is one of the most instructive in this book.

The first account is not by Archimedes at all. It comes from the Roman architect and engineer Vitruvius, writing in the preface to Book IX of his treatise On Architecture sometime in the first century BC — roughly two hundred years after Archimedes' death (Vitruvius, De Architectura IX, Perseus). King Hiero of Syracuse, the story goes, commissioned a golden crown and, suspecting the goldsmith of adulterating it with silver, asked Archimedes to determine the truth without damaging the crown. Stepping into a bath and noticing the water rise in proportion to the volume of his own submerged body, Archimedes supposedly realized he could compare the crown's volume (via displaced water) against its weight, catching the fraud — and, in his excitement, ran naked through the streets of Syracuse shouting "Eureka," "I have found it." The story is vivid, quotable, and almost certainly embellished: Vitruvius is our earliest surviving source, writing long after the fact, and modern commentators have pointed out that the water-displacement method as described would have been far too imprecise, given ancient measuring technology, to reliably detect the small amount of silver a goldsmith might realistically substitute (Scientific American, "Fact or Fiction?"). Whether or not it happened as told, the story survives because it captures something people badly want a discovery to feel like: sudden, embodied, and involuntary — a bath, not a proof.

The second account is by Archimedes himself, and it describes something almost the opposite of a flash of bathtub insight: a deliberate, two-stage discipline for doing mathematics. In a treatise called The Method of Mechanical Theorems, addressed as a letter to Eratosthenes — the polymath running the Library of Alexandria, whose own solution appears in chapter 07 — Archimedes reveals a technique he had been keeping semi-private. To find the area or volume of a shape (a parabolic segment, a sphere, a cylinder), he did not start with rigorous geometric proof. He started by imagining the shape sliced into infinitely many thin strips, and mentally weighing those strips against strips of a shape he already understood, balanced on an imaginary lever, using the law of the lever from his own work in mechanics (The Method of Mechanical Theorems, Wikipedia). This procedure is not rigorous by Greek geometric standards — it treats an area as if it were literally made of an infinite stack of one-dimensional lines, a move that would trouble mathematicians until the invention of calculus two thousand years later. But it works: it tells Archimedes, quickly and reliably, what the answer is. Only afterward did he go back and construct a fully rigorous proof of the same result using the accepted method of exhaustion, the laborious double-contradiction argument Euclid's tradition required. In his own words to Eratosthenes, he explains that he is sending the results first, without their proofs, precisely because finding the proof of a result you already know to be true is far easier than finding a true result and its proof simultaneously from nothing.

That is a completely different confession about method than the bath story offers, and it is arguably the more radical one: Archimedes is telling us, in his own hand, that mathematical discovery and mathematical rigor are two separate activities, performable in either order, and that doing the ugly, unrigorous, mechanically-assisted one first is not a shortcut to be ashamed of — it is a legitimate research strategy. The Method was itself lost for most of history and known only by title until 1906, when the Danish scholar Johan Ludvig Heiberg identified it underneath a thirteenth-century Byzantine prayer book — a palimpsest, a manuscript scraped and reused — now known as the Archimedes Palimpsest. Large portions remained illegible until a further imaging campaign between 1998 and 2008, using ultraviolet, infrared, and X-ray techniques at what is now the Walters Art Museum in Baltimore, recovered text hidden under mold damage and even forged medieval paintings added by a twentieth-century forger trying to increase the manuscript's resale value (Archimedes Palimpsest project). We are, in other words, still recovering Archimedes' record of his own method eighteen centuries after Vitruvius wrote down the more entertaining but less trustworthy story about his bath.

Ibn al-Haytham's Room

A thousand years after Archimedes, in Cairo, a scholar named Abu Ali al-Hasan ibn al-Haytham — known in Latin translation as Alhazen — was, by most accounts, confined to his own house for roughly a decade, from around 1011 to 1021. The reason given by several historians is almost too neat: he had rashly told the volatile Fatimid caliph al-Hakim that he could regulate the annual flooding of the Nile with a dam, was sent to survey the site near modern Aswan, and concluded on arrival that the project was impossible with the era's engineering. Fearing execution for the failure — al-Hakim's court had a documented record of brutal punishment for far smaller offenses — Ibn al-Haytham is said to have feigned madness to be spared, and was instead placed under house arrest until the caliph's death in 1021, at which point he was released as no longer legally responsible for a promise made while "insane" (Physics World, "The Scholar and the Caliph" — note that this source itself flags the account as reconstructed from incomplete records, calling it merely "likely"). What is much better attested is what he did with those ten years: he wrote a seven-volume treatise, the Kitab al-Manazir, known in the West as the Book of Optics (UNESCO Courier; Book of Optics, Wikipedia).

The Book of Optics addressed a question the ancient Greeks had never settled: does the eye see by sending rays out toward objects (the "extramission" theory associated with Euclid's own optical writings), or by receiving light that travels from objects into the eye ("intromission")? Ibn al-Haytham did not simply argue for intromission — he built physical devices to test it, describing his apparatus and procedures in enough concrete, reproducible detail that a reader with no access to him could set up the same experiment. His most famous device, a forerunner of the pinhole camera, used a small aperture in a screen between a set of candles and a viewing wall: light from each candle, he showed, cast a single, distinct, invertible spot on the far wall, corresponding to a straight line from that candle through the hole — evidence that light was moving from the object toward the observer along definite straight paths, not the reverse (UNESCO Courier; Book of Optics, Wikipedia).

This deserves the same caution this book applies elsewhere: calling Ibn al-Haytham the sole "father of the scientific method," a label popular accounts attach to him often, overstates a more gradual, distributed history that includes earlier Greek and later European experimentalists, and some historians of science push back hard on the tidier version (The Renaissance Mathematicus, "Ibn al-Haytham did not invent the Scientific Method"). What is not in dispute is narrower and, for this book's purposes, more important: the Book of Optics is a written record of an experimental procedure — apparatus, setup, prediction, result — precise enough that others who had never met its author could pick it up and re-run it. Translated into Latin around 1200 as De Aspectibus, it shaped the optics of Roger Bacon, Witelo, and eventually Kepler, centuries after its author's death — outliving him by exactly the mechanism this chapter keeps returning to: he wrote the route down clearly enough that it needed no further help from him to keep working.

Descartes' Unfinished Machine

By the early seventeenth century, a French soldier-turned-philosopher named René Descartes had noticed the same thing this chapter has been arguing and decided to attack it head-on: if geometry could be turned into a machine (Euclid) and if a scholar under house arrest could turn optics into a reproducible experimental procedure (Ibn al-Haytham), why should there not be one single method, stated once, that would work for discovering truth in any field at all — physics, medicine, ethics, mathematics alike?

Descartes made two attempts at writing this universal method down. The first, composed around 1628 and left unfinished, was the Regulae ad Directionem Ingenii — Rules for the Direction of the Mind. It was not published in his lifetime; the earliest edition appeared only in 1701, more than fifty years after his death (Rules for the Direction of the Mind, Wikipedia). In its fourth rule, Descartes names explicitly what he is chasing: a mathesis universalis, a "universal mathematics," which he describes as "a general science which explains all that can be sought concerning order and measure, restricted to no special subject matter" (Mathesis universalis, Wikipedia; MacTutor, Descartes). That is an extraordinary ambition stated plainly: not a better way to do geometry, but the master method underneath geometry, algebra, and every other exact science — the pattern Euclid had built for one subject, generalized to all subjects at once.

The Regulae stops abruptly, unfinished, partway through its planned three parts. Descartes' second and more famous attempt, nine years later, is more modest in scope but far more influential in form: the 1637 Discourse on the Method, written in French rather than the scholarly Latin, framed as something closer to autobiography than treatise. Descartes narrates his own dissatisfaction with everything he had been taught and distills his working method into four rules: accept nothing as true unless it presents itself so clearly and distinctly that there is no occasion to doubt it; divide every difficulty into as many parts as necessary to resolve it; conduct one's thoughts in order, beginning with the simplest objects and ascending by degrees to the more complex; and, throughout, make enumerations so complete, and reviews so general, that nothing is omitted (Discourse on the Method, Wikipedia).

Appended to the Discourse as one of three scientific essays illustrating the method in action was La Géométrie, in which Descartes showed how to attach a pair of numbers to every point in a plane and rewrite every geometric curve as an algebraic equation (La Géométrie, Britannica). That single technical result — coordinates — turns out to be so consequential for the argument of this book that it earns its own extended treatment in chapter 02, as the master case of changing a problem's representation rather than merely following a fixed procedure within one. For now, notice only the shape of the ambition that produced it: Descartes wanted a method that would make discovery itself mechanical, the way Euclid had made geometric proof mechanical. He never finished building it — the mathesis universalis remains, by his own account, a program rather than a completed system, and philosophers still argue about how much of it he actually believed was achievable. But the dream he left behind — that solving, in general, might be systematized rather than merely practiced — is the ambition every later chapter of this book, in its own century's vocabulary, keeps trying to make good on.

Method Becomes an Object

Step back from the four thousand years this chapter has covered and one pattern holds across Egypt, Babylon, Alexandria, Cairo, and Descartes' stove-heated room in Germany where he claims to have had his founding insight: in every case, someone took a way of solving problems that had previously lived only inside a single skilled person's head — a scribe's trick, a geometer's construction, a natural philosopher's experimental setup, a philosopher's private rules of thought — and wrote it down as an object separate from any particular problem it had been used to solve.

That act of writing is not mere record-keeping. It is the first and most fundamental kind of compression this book will describe. A method that lives only in one mind can only be applied by that mind, to problems that mind happens to encounter, using whatever the mind happens to remember correctly on a given day. A method that has been written down — named, with its steps made explicit, its assumptions stated, its scope of application at least implicitly bounded — becomes something that other people can inspect, test against new cases, criticize, and improve, without needing to reconstruct it from scratch or consult its inventor. Ahmes' false position, Euclid's postulates, Archimedes' lever-balancing heuristic, Ibn al-Haytham's pinhole apparatus, and Descartes' four rules are five different answers to the same underlying design question — how do we solve this class of problem, in general? — and every one of them survives, in some form, into the mathematics and science practiced today.

This groundwork makes the rest of this book possible. You cannot deliberately change the space a problem lives in — the master move this book spends its remaining chapters tracking — until you have noticed that the space has a shape worth examining, distinct from any one instance sitting inside it. Writing the route down, rather than just reaching the answer, is what first makes that shape visible.

The Move

Move 01 — Turn the method into an object.

• Write the route, not just the answer. When you solve something, the number or outcome is often the least valuable thing you produced. The sequence of moves that got you there — stated plainly enough that someone else could follow it without you in the room — is the reusable asset.
• Name it so it can be improved by someone who isn't you. A technique without a name lives and dies with its inventor's memory. A named, written procedure — "false position," "the axiomatic method," "the four rules" — becomes a target other people can criticize, test, and sharpen.
• Separate finding from proving. Archimedes' letter to Eratosthenes is explicit permission, twenty-two centuries old, to solve a problem by whatever ugly, mechanical, half-rigorous means gets you the right answer fastest — and only afterward go back and build the clean argument. Chasing rigor and certainty simultaneously, from a blank page, is usually the slower path to both.
• State your starting assumptions before you build on them. Euclid's postulates and Descartes' four rules both do this: they say, explicitly and in advance, what is being taken for granted, so that everything constructed afterward can be checked against a fixed, visible foundation rather than an unspoken intuition.

Next — Chapter 02 · The Reframe: a bridge-crossing puzzle so trivial Euler almost didn't bother, a schoolboy's shortcut whose true story is stranger than its legend, and a high jumper who solved "jump higher" by refusing to face the bar. Once method is an object, the next move is to change the space that object operates in.