Appendix G — Historical Notes

Calculus was not invented in a single flash of genius. It was assembled over more than two thousand years, by people who often could not have explained what they were really doing, and who would have been astonished to see the clean machinery you have been using throughout this book. The story is one of slow accumulation punctuated by two extraordinary decades in the 1660s and 1670s, followed by a century and a half of brilliant but uneasy progress, and finally a great housecleaning in the 1800s that put everything on solid ground. This appendix tells that story roughly in order, and points to where each idea is developed in the chapters.

Ancient Roots: Measuring the Curved with the Straight

The deepest problem calculus solves is also the oldest: how do you measure something that curves? You can lay rulers end to end along a straight road, but the area inside a circle, or under an arc, resists that kind of direct counting. The Greeks attacked this with breathtaking ingenuity.

Eudoxus of Cnidus (c. 390–337 BC) gave the first rigorous tool, the method of exhaustion. The idea is to trap an unknown curved quantity between two known quantities — inscribed and circumscribed polygons, say — and then make the gap between them as small as you please by using more and more sides. This is a limit in everything but name; it is the ancestor of the $\varepsilon$–$\delta$ machinery you met in Chapter 3. Eudoxus used it to prove, for instance, that the volume of a cone is one-third the volume of the cylinder around it.

Archimedes of Syracuse (c. 287–212 BC) carried the method to heights no one would match for nearly nineteen centuries. Using exhaustion, he found that the area enclosed by a parabola and a chord is exactly $\tfrac{4}{3}$ of the area of an inscribed triangle — a result that, in modern terms, is a definite integral. He bracketed $\pi$ between $\tfrac{223}{71}$ and $\tfrac{22}{7}$ by comparing the circle to inscribed and circumscribed 96-gons, and he derived the surface area and volume of the sphere, the result he reportedly asked to have carved on his tomb. In a manuscript lost until 1906 (The Method of Mechanical Theorems), Archimedes even revealed that he first discovered such results by imagining figures balanced on a lever and sliced into infinitely many thin pieces — essentially summing infinitesimals — and only afterward dressed the findings in the rigor of exhaustion. He was, in spirit, doing integration two millennia early.

The thread did not run only through Greece. In China, Liu Hui (3rd century AD) used an inscribed-polygon method to approximate $\pi$, and later Zu Chongzhi (5th century) pushed it remarkably far. In Kerala, India, Madhava of Sangamagrama (c. 1340–1425) and his school developed infinite series for $\sin$, $\cos$, and $\arctan$ — including what we now call the Gregory–Leibniz series for $\pi$ — anticipating power-series ideas you meet in Chapter 23 by roughly three centuries. (Madhava's exact dates and which results are his rather than his successors' are debated by historians; the work survives mostly through later commentaries.)

Precursors: The Seventeenth-Century Stirring

After a long quiet, the early 1600s saw an explosion of partial techniques across Europe — methods for areas, tangents, and maxima that worked but lacked a unifying idea.

Johannes Kepler (1571–1630), better known for planetary motion, computed volumes by imagining solids sliced into infinitely many thin laminae (famously, the optimal proportions of wine casks). Bonaventura Cavalieri (1598–1647), a student of Galileo's circle, systematized this in his Geometria indivisibilibus (1635): an area is "made of" infinitely many parallel line-segments, a volume of infinitely many planes — his method of indivisibles. Cavalieri's principle (two solids of equal corresponding cross-sections have equal volume) is still taught and still useful.

Pierre de Fermat (1607–1665) developed a method of adequality around 1636–1638 for finding maxima, minima, and tangents. He would perturb a variable by a small amount, simplify, and then set the perturbation to zero — exactly the gesture of a derivative, computed before derivatives officially existed. Lagrange and Laplace later called him the true first inventor of differential calculus, and Fermat's tangent and optimization methods foreshadow Chapters 4 and 9.

René Descartes (1596–1650), in La Géométrie (1637), fused algebra and geometry into analytic geometry — curves became equations, equations became curves. This is the language the whole book is written in; without coordinates there is no graph of $f(x)$ to take a slope of. John Wallis (1616–1703), in his Arithmetica Infinitorum (1656), pushed infinite processes boldly, producing his infinite product for $\pi$ and the systematic use of fractional and negative exponents, and popularizing the $\infty$ symbol.

The most direct herald was Isaac Barrow (1630–1677), Newton's teacher and the first Lucasian Professor at Cambridge. In his Lectiones Geometricae (1670), Barrow stated, in geometric language, a result equivalent to the inverse relationship between tangent-finding and area-finding — that differentiation and integration undo each other. He had, in geometric clothing, a version of the Fundamental Theorem of Calculus (Chapter 14). What he lacked was the algebraic engine and the notation to make it a method rather than a theorem about particular curves.

A flag on attribution: It is sometimes said that James Gregory (1638–1675) proved the Fundamental Theorem in 1668. Gregory's Geometriae Pars Universalis does contain results closely related to it, and Barrow's lectures contain another version. Historians generally credit these men with geometric precursors, while reserving the unified, computational FTC for Newton and Leibniz. Treat any single clean "first proof" date for the FTC with caution.

The Invention: Newton and Leibniz

By 1660 the pieces were lying everywhere on the table. Two men, independently, swept them into a single system.

Isaac Newton (1643–1727) did his foundational work during the plague years of 1665–1666, when Cambridge closed and he retreated to the family farm at Woolsthorpe — his annus mirabilis. He developed what he called the method of fluxions: quantities ("fluents") flow in time, and their rates of flow are "fluxions," written with the dot, $\dot{x}$ — notation this book still uses for time derivatives (see the notation table). Crucially, Newton understood the inverse relationship between fluxions and the areas they generate, and he wielded infinite series with total command. He used all of this privately for decades, applying it in his Philosophiæ Naturalis Principia Mathematica (1687) to derive the laws of motion and universal gravitation — though, famously, the Principia presents its physics in classical geometric form, not in the fluxional calculus he actually used to find the results.

Gottfried Wilhelm Leibniz (1646–1716), a philosopher, diplomat, and polymath, came to calculus a few years later, working in Paris around 1673–1676 under the influence of Christiaan Huygens. Leibniz thought hard about notation as a form of reasoning, and his choices were inspired. He introduced the integral sign $\int$ — an elongated "S" for summa, a sum — and the differential $d$, giving us $dy$ and $dx$ and the quotient $\dfrac{dy}{dx}$. He published first: a paper on differential calculus in Acta Eruditorum in 1684, and on the integral calculus in 1686. His notation made the chain rule look like cancellation and made the Fundamental Theorem look almost like a definition.

Why did Leibniz's notation win? Because it does work for you. The symbol $\dfrac{dy}{dx}$ behaves, in many manipulations, like the fraction it resembles; $\int f(x)\,dx$ openly advertises that an integral is a sum of little pieces $f(x)\,dx$; the $d$ operator composes cleanly. Newton's dots are economical for time derivatives in physics — and survive there — but they do not scale to the rich symbolic algebra that calculus became. Continental Europe adopted Leibniz's symbols and raced ahead; Britain, loyal to Newton, lagged for a century partly because its notation was harder to compute with.

That loyalty hardened into one of the ugliest feuds in the history of science. The priority dispute pitted Newton and his allies against Leibniz, with each side accusing the other of plagiarism. The Royal Society convened a committee in 1712 to adjudicate — but Newton was the Society's president and, it later emerged, anonymously drafted much of its supposedly impartial report. Leibniz died in 1716 under a cloud, his reputation damaged; the bitterness outlived both men. The modern verdict is settled and generous: the two invented calculus independently, by different routes, and we use Leibniz's notation with Newton's dot kept for time and Lagrange's prime, $f'(x)$, added later.

The Eighteenth Century: Power Before Rigor

With the machinery in hand, the 1700s became a century of dazzling, sometimes reckless, productivity. Mathematicians computed first and worried later.

The Bernoulli family of Basel — especially the brothers Jakob (1655–1705) and Johann (1667–1748), and Johann's son Daniel (1700–1782) — were among Leibniz's earliest and most effective champions. They solved the catenary, the brachistochrone, and a flood of differential equations, and effectively launched the calculus of variations. Johann tutored a young man who would become the most prolific mathematician who ever lived.

Leonhard Euler (1707–1783) is the presiding genius of the era. He wrote so much that his collected works still are not fully published. He standardized much of the notation you take for granted: the function notation $f(x)$, the symbol $e$ for the base of natural logarithms, the symbol $i$ for $\sqrt{-1}$, the use of $\Sigma$ for summation, and he popularized $\pi$. He solved the Basel problem, proving $\sum_{n=1}^{\infty} \frac{1}{n^2} = \frac{\pi^2}{6}$. And he discovered the astonishing bridge between exponential and trigonometric functions, $$e^{i\theta} = \cos\theta + i\sin\theta,$$ which at $\theta = \pi$ collapses to the most celebrated equation in mathematics, $e^{i\pi} + 1 = 0$ — the anchor result this book builds toward and derives in full in Chapter 24. Euler, blind for the last seventeen years of his life, dictated work to the end.

Joseph-Louis Lagrange (1736–1813) reformulated mechanics in the analytic, coordinate-free form that bears his name and tried to put calculus on a rigorous footing by basing it on power series (his prime notation $f'(x)$ is his enduring gift). Pierre-Simon Laplace (1749–1827) applied calculus to the heavens in his monumental Mécanique Céleste, and to chance in his work on probability. The century's results were magnificent — and built on a foundation that, on close inspection, was alarmingly soft.

The Nineteenth Century: The Great Rigorization

The soft spot was the infinitesimal. What, exactly, was $dx$? Sometimes it was treated as a nonzero quantity (so you could divide by it), and a moment later as zero (so you could discard it). The Anglican bishop George Berkeley skewered this in The Analyst (1734), mocking infinitesimals as "the ghosts of departed quantities" and pointing out, fairly, that the physicists laughing at religious mystery were trafficking in mysteries of their own. For a century the criticism stood essentially unanswered, because calculus worked, and working trumps explaining — until the explanations could no longer be put off.

Bernhard Bolzano (1781–1848), a Bohemian priest and mathematician working in relative isolation, gave (in 1817) a definition of continuity and a proof of the intermediate value theorem free of geometric hand-waving, and constructed a continuous-but-nowhere-differentiable function — but his work was little known in his lifetime.

The decisive figure was Augustin-Louis Cauchy (1789–1857). In his Cours d'Analyse (1821) and later lectures, Cauchy made the limit the central concept and defined continuity, the derivative, and the integral in terms of it. He gave the first systematic, broadly adopted rigorous treatment of calculus. His limits still leaned on phrases like "approaches indefinitely," but the conceptual reorganization was the breakthrough: derivatives and integrals became things you prove things about, not just compute.

Karl Weierstrass (1815–1897) supplied the final precision. Working first as an obscure schoolteacher, he forged the $\varepsilon$–$\delta$ definition of a limit — the statement that for every $\varepsilon > 0$ there exists a $\delta > 0$ such that $|x - a| < \delta$ implies $|f(x) - L| < \varepsilon$. With this, every appeal to "infinitely small" could be eliminated; Berkeley's ghosts were banished. This is the definition you met head-on in Chapter 3, and it is the bedrock under every theorem in the book. Weierstrass also exhibited a function that is continuous everywhere yet differentiable nowhere, a shock that proved how much intuition needed the discipline of proof.

The integral got its own rigorous footing from Bernhard Riemann (1826–1866), whose definition via sums over partitions — the Riemann integral of Chapter 13 — made precise what it means for a function to be integrable. Beneath all of this lurked a question even more basic: what are the real numbers? Richard Dedekind (1831–1916) answered with "Dedekind cuts," constructing the reals from the rationals so that the line truly has no gaps (the completeness that the Intermediate and Mean Value Theorems silently rely on). Georg Cantor (1845–1918), building the theory of infinite sets, showed there are different sizes of infinity and gave another construction of the reals. By around 1900, calculus — now properly called analysis — rested on definitions that would survive every later scrutiny.

Modern Threads: Generalization and Ubiquity

Rigor did not end the story; it opened new doors.

Henri Lebesgue (1875–1941), in his 1902 thesis, redefined integration by partitioning the range of a function rather than its domain — the Lebesgue integral. It integrates far wilder functions than Riemann's, behaves beautifully under limits, and became the foundation of modern probability and much of functional analysis. Élie Cartan (1869–1951) and others developed differential forms, the language in which the gradient, curl, divergence, and the classical integral theorems all reveal themselves as faces of a single result: the generalized Stokes' theorem, $\int_{\partial M} \omega = \int_M d\omega$, the grand unification this book reaches in Chapter 38. John von Neumann (1903–1957) and the functional-analysts recast calculus in infinite-dimensional spaces, the setting that quantum mechanics demanded.

Meanwhile calculus quietly conquered every quantitative field. It was always the native language of physics, but the same derivatives and integrals now drive engineering (control systems, signal processing, structural analysis), economics (marginal analysis, optimization, growth — Track B of your portfolio), biology (population and epidemic models like the SIR system of Chapter 19 — Track A), statistics (the normal-curve integral of Chapter 13 — and the basis of continuous probability), and computing. The arrival of digital computers in the 1950s automated numerical integration and the solving of differential equations. In our own century, automatic differentiation — the algorithmic computation of exact derivatives through code — became the engine of machine learning: the gradient descent you first met in Chapter 6 and develop through Chapter 30 is, at bottom, the chain rule run at enormous scale across the parameters of a neural network. Calculus is not a finished antique. It is the working mathematics of the present.

A closing word on calculus education. For most of its history, calculus was the province of a tiny elite. Today it is a rite of passage for millions of students across the sciences and increasingly in data-driven fields. The tension Berkeley identified — between computing fluently and understanding why — is still the central challenge of teaching it, which is exactly why this book insists on presenting every major idea three ways: intuitively, computationally, and formally. You stand at the end of a very long line of people who wanted to measure the curved with the straight, to capture change with precision, and who refused to be satisfied with answers they could not justify.

A Compact Timeline

A Note on Whose Names Stuck

Many ideas in calculus carry the wrong name, or no single right one. The Fundamental Theorem was implicit in Archimedes, geometrically present in Barrow and Gregory, and only made into a method by Newton and Leibniz. The Mean Value Theorem cannot be pinned on one person; Cauchy, Bolzano, and others all contributed. The "Gregory–Leibniz" series for $\pi$ was known to Madhava's school centuries earlier. Stigler's law of eponymy — the wry observation that no scientific discovery is named after its original discoverer — applies almost perfectly here. History is messier than any textbook's tidy attributions suggest. But the deeper lesson is robust and worth carrying with you: calculus was not handed down whole. It was built, argument by argument, over centuries, by people who measured the curved with the straight and refused to accept what they could not prove.