The Geometer Who Moved Mountains — and Followed Napoleon Into the Sand

Gaspard Monge: The Genius Who Transformed Warfare with Geometry!

How a fortification secret and a memoir about shifting piles of dirt made Gaspard Monge indispensable to modern engineering, French statecraft, and — two centuries later — artificial intelligence

BLUF

Gaspard Monge (1746–1818) invented descriptive geometry, the disciplined method for representing three-dimensional objects in two-dimensional projections that still underlies every engineering and architectural drawing; founded differential geometry; helped build the metric system and the École Polytechnique; and served the French Republic as Minister of the Navy before attaching himself, with near-total devotion, to Napoleon Bonaparte. His clearest legacy, however, was one he barely noticed: a slim 1781 memoir on the cheapest way to move a pile of earth defined what is now called the Monge problem of optimal transport — a field that lay largely dormant for 160 years, then exploded into one of the most active areas of modern mathematics, yielding Fields Medals in 2010 and 2018 and now powering economics, meteorology, image processing, and machine learning. Monge's life is a case study in how mathematical clarity and political judgment can diverge completely in the same mind.

A pile of sand, a hole in the ground

In 1781 a French military engineer stood before the Académie des Sciences in Paris with a question that sounds almost childish: given a heap of soil and a hole to fill, what is the least wasteful way to carry the dirt from one to the other? He wrote it up as Mémoire sur la théorie des déblais et des remblais — a memoir on the theory of cuts and fills — and then went on with a spectacularly busy life. The paper was essentially forgotten for a century and a half.

It should not have been. That "sandcastle problem," as mathematicians now cheerfully call it, is the seed of optimal transport theory, and its author, Gaspard Monge, spent the rest of his career being underestimated in exactly this way — brilliant at the thing in front of him, oblivious to which of his ideas would outlast empires. To understand how a provincial merchant's son became the geometer of the French Revolution, and then the willing courtier of a dictator, you have to start with a wall.

Beaune to Mézières: the fortification that had to be hidden

Monge was born on 9 May 1746 in Beaune, in the Burgundian wine country, the son of Jacques Monge, a merchant of modest means, and Jeanne Rousseaux. (A minority of reference works give 10 or 14 May; the date accepted by the standard biographical record is 9 May.) He was educated by the Oratorian priests, whose schools offered an unusually broad diet of mathematics and natural science, and by seventeen he was already teaching physics at the Collège de la Trinité in Lyon.

His break came from a drawing. In 1764 the young Monge produced a large-scale, meticulously surveyed plan of his hometown, inventing his own observing instruments to do it. An engineering officer who saw the plan recommended him to the École Royale du Génie at Mézières, France's elite military engineering school, where Monge was taken on — pointedly — as a mere draftsman, not a gentleman-cadet.

Then came the problem that made him. Military engineers of the era had to design defilement: the art of shaping fortifications so that no enemy position, wherever it stood, could see or fire directly on the defenders. The standard solution was a grinding, hours-long arithmetical procedure. Monge replaced it with a geometric construction and returned an answer so fast that the commandant at first refused to believe he had done the work. When the method was checked, its power was obvious — and the French army classified it. Monge's graphical technique was kept as a military secret for years.

That secret was the germ of descriptive geometry.

Making three dimensions teachable

Monge's insight was to project a solid object onto two mutually perpendicular planes — think of the plan view and the elevation view — so that spatial problems about volumes, intersections, and shadows could be solved with ruler-and-compass constructions on flat paper. In modern notation, a point $P=(x,y,z)$ in space is carried to its two projections

P \;\longmapsto\; \underbrace{(x,\,y)}_{\text{horizontal plane (plan)}} \quad\text{and}\quad \underbrace{(x,\,z)}_{\text{vertical plane (elevation)}},

with the shared $x$ -coordinate — the "recall line" of the draftsman — locking the two views into a single rigid description of the object. Generalized and systematized, this became the discipline every engineer now knows as orthographic projection: the grammar of the mechanical drawing.

Its practical importance is hard to overstate. Before Monge, translating a three-dimensional design into buildable instructions was an artisanal, error-prone craft. After him, it was a teachable, checkable method. For anyone who has ever read a fabrication drawing or a systems layout, Monge is the reason the front view, side view, and top view line up.

He did not stop at drafting. In Application de l'analyse à la géométrie, built from his lectures, Monge introduced the notion of lines of curvature on a surface in three-dimensional space and treated partial differential equations as geometric objects — the moves that earned him the title father of differential geometry. At each point of a smooth surface the two principal curvatures $\kappa_1,\kappa_2$ are the eigenvalues of the shape operator $S=-\,dN$ (the derivative of the unit normal $N$ ),

S\,\mathbf{v}_i = \kappa_i\,\mathbf{v}_i,\qquad K=\kappa_1\kappa_2=\det S,\qquad H=\tfrac{1}{2}(\kappa_1+\kappa_2),

where $K$ is the Gaussian and $H$ the mean curvature. A line of curvature is a curve whose tangent always points along a principal direction — equivalently, a curve satisfying Rodrigues' condition

dN + \kappa\,d\mathbf{r} = 0,

so that the surface normal turns in step with the position vector as one moves along it. He viewed algebraic analysis not as an end in itself but, in his own framing, as the script for a moving geometric spectacle: reality as space in transformation. That philosophy — geometry first, analysis as its notation — put him at odds with the more formalist mathematical establishment of Lagrange and Laplace, and it is precisely the instinct that produced the optimal-transport memoir.

Revolution, the meter, and a difficult year as Navy Minister

Elected to the Académie des Sciences in 1780, Monge was, by the time the Bastille fell in July 1789, one of Paris's leading scientists, with published work spanning geometry, partial differential equations, metallurgy, capillarity, optics, and chemistry. He was also an ardent republican.

His revolutionary service ran on two tracks. On the technical side, he sat on the Commission of Weights and Measures, the body — alongside Borda, Lagrange, Laplace, Condorcet, and Lavoisier — that built the metric system, defining the meter as one ten-millionth of the meridian quadrant from the North Pole to the equator,

1\ \text{m} \;\equiv\; \frac{1}{10^{7}}\times\bigl(\text{pole-to-equator meridian arc}\bigr),

a length fixed by geodesy rather than by any king's forearm. It is one of the Revolution's most durable exports; you are using it every time you read a millimeter.

On the political side, the new Republic made him Minister of the Navy in September 1792. He held the post for roughly eight months, through the period that included the execution of Louis XVI, before resigning in April 1793, worn down by the factional impossibilities of the job. Historians treat this chapter unsentimentally: as an administrator of a navy in revolutionary chaos, Monge was not a success.

When the Terror passed and the Académie was abolished, Monge poured himself into education. He was central to founding the École Centrale des Travaux Publics in 1794 — soon renamed the École Polytechnique — and became its instructor in descriptive geometry and later its director. His lectures at the parallel École Normale were compiled by his student Hachette and published in 1799 as Géométrie descriptive, the book that carried his method to the world. Polytechnique remains, to this day, one of the templates for the modern technical university.

The Egyptian seduction

Then came Napoleon. Sent to Italy in 1796–97 on a commission to select — which is to say, expropriate — art treasures for France, Monge met the young general Bonaparte and was, by every account, dazzled. Napoleon reportedly said Monge loved him the way a man loves a mistress, and the relationship reshaped the rest of Monge's life.

In 1798 Monge joined Napoleon's expedition to Egypt as one of its scientific savants, and was made president of the Institut d'Égypte in Cairo — the learned society whose surveys and drawings produced the monumental Description de l'Égypte. The military campaign was a catastrophe after Nelson destroyed the French fleet at the Battle of the Nile, but the scholarly enterprise reshaped European Egyptology.

Back in France, Monge accepted the honors Napoleon rained on him: Grand Officer of the Legion of Honor (1804), President of the Senate (1806), and Comte de Péluse (1808) — a title, fittingly, borrowed from an ancient Egyptian city. A committed republican in principle, he served an emperor in practice, a contradiction his critics have never let rest.

The reckoning was brutal. When Napoleon fell and the monarchy was restored, Monge was stripped of his honors and expelled from the Institut de France in 1816, politically persecuted and, by some accounts, in fear for his life. He died in Paris on 28 July 1818. The royalist government forbade any public tribute — whereupon the students of the École Polytechnique defied the order, marched to the cemetery, and honored their teacher anyway.

The idea he didn't know he'd had

Here the story loops back to that forgotten pile of sand.

Monge's 1781 déblais et remblais memoir asked for a transport map $T$ : a rule assigning every particle of the source distribution $\mu$ to a spot in the target $\nu$ , minimizing total cost. In modern form,

\inf_{T:\,T_\#\mu=\nu}\ \int_X c\bigl(x,\,T(x)\bigr)\,d\mu(x),

where the constraint $T_\#\mu=\nu$ (" $T$ pushes $\mu$ forward onto $\nu$ ") says no mass is lost or created, and Monge's own cost was simply the distance moved, $c(x,y)=|x-y|$ . Stated for continuous distributions rather than discrete heaps, this is a genuinely hard problem — the map $T$ appears inside the constraint, which is badly nonlinear, and a minimizing $T$ need not even exist. For 160 years it went essentially nowhere.

The thaw came in 1942, when the Soviet mathematician Leonid Kantorovich — unaware of Monge's work at the time — reformulated the question as a problem in linear programming, allowing mass to split. Instead of a rigid map $T$ , he optimized over transport plans $\pi$ — joint distributions with marginals $\mu$ and $\nu$ :

\inf_{\pi\in\Pi(\mu,\nu)}\ \int_{X\times Y} c(x,y)\,d\pi(x,y).

Because $\Pi(\mu,\nu)$ is convex and the objective linear in $\pi$ , the intractable problem becomes a well-posed (and dualizable) one — the crucial move that made the theory usable. He later connected his version explicitly to Monge's, and the modern subject took the name the Monge–Kantorovich problem. (Kantorovich shared the 1975 Nobel Memorial Prize in Economics for the broader body of optimization work.) In 1987 Yann Brenier linked optimal transport to fluid mechanics and the geometry of measure-preserving maps, and the field caught fire.

It has since become one of the crown jewels of contemporary mathematics — and the honors make the point. Cédric Villani won the 2010 Fields Medal in work deeply tied to optimal transport and wrote the field's standard texts. Alessio Figalli won the 2018 Fields Medal, per the International Mathematical Union's official citation, for contributions to optimal transport and its applications to partial differential equations, metric geometry, and probability — including landmark results on the Monge–Ampère equation. This is the thread that ties the whole story back to Monge himself. For the quadratic cost $c(x,y)=\tfrac12|x-y|^2$ , Brenier's theorem (1987) says the optimal map is the gradient of a convex potential, $T=\nabla\varphi$ . Feeding that into the conservation-of-mass (change-of-variables) relation, with densities $f$ for $\mu$ and $g$ for $\nu$ , gives the Monge–Ampère equation

\det\!\bigl(D^2\varphi(x)\bigr)=\frac{f(x)}{g\bigl(\nabla\varphi(x)\bigr)},

a fully nonlinear PDE in the Hessian $D^2\varphi$ that carries Monge's name and governs everything from soap-bubble shapes to the movement of weather fronts. Proving when its solutions are smooth was precisely the kind of result that anchored Figalli's medal.

For an engineer, the payoff is the applications. The distance between two probability distributions defined by optimal transport — the Wasserstein distance —

W_p(\mu,\nu)=\left(\inf_{\pi\in\Pi(\mu,\nu)}\int |x-y|^{p}\,d\pi(x,y)\right)^{1/p},

is now a workhorse across signal processing and machine learning: comparing images, training generative models (the "Wasserstein GAN" takes its name and its loss function directly from $W_1$ ), aligning datasets, quantifying how far one distribution sits from another. Unlike pointwise measures such as Kullback–Leibler divergence, $W_p$ respects the underlying geometry — it knows that two sharp, slightly shifted peaks are close, which is exactly the property a detection or estimation problem wants. Monge's question about moving dirt cheaply turns out to be the same question as "how different are these two data sets, and what is the most efficient way to morph one into the other?" As recently as 2025, Figalli and collaborators were still publishing sharp new results on transport distances aimed squarely at high-dimensional computation. The man who hid fortifications from cannon fire wrote, almost as an aside, one of the foundational problems of the age of AI.

SIDEBAR — Afterlife in physics: the toolkit, not the theorems

It is tempting to file Monge among the founders of mathematical physics alongside Euler and Lagrange. The precise version is subtler and, in some ways, more interesting: Monge did not write the governing equations of any physical theory — he built the geometric machinery by which those equations are solved. His fingerprints appear wherever a field problem is cracked by tracking geometry through space.

Compressible flow and the method of characteristics. Monge's geometric theory of first-order partial differential equations — the object now called the Monge cone, with its characteristic curves and strips — is a taproot of the method of characteristics. That method is the native language of hyperbolic PDEs, and hyperbolic PDEs are compressible gas dynamics: shock waves, supersonic nozzle design, and Riemann-problem solvers all inherit the picture of a solution surface threaded along characteristic curves. Cauchy, Lagrange, and Charpit built the analytic side; the geometric conception is Monge's.

Ray optics — the one real bridge to electromagnetics. The vector-calculus backbone of Maxwell's equations (divergence, curl, the flux and circulation theorems) descends from Green, Gauss, Stokes, and later Heaviside and Gibbs — not from Monge. But the eikonal equation of geometric optics, $|\nabla S| = n$ — the equation of wavefronts and rays — is the canonical fully nonlinear first-order PDE solved by the Monge-cone apparatus. Since geometric optics is the high-frequency limit of Maxwell's equations, this is exactly the regime of ray tracing, the geometric theory of diffraction, and physical optics: the characteristics of the eikonal equation are the rays, their envelopes are the caustics, and the curvature of the wavefront is described by the very lines-of-curvature surface theory Monge invented. He is foundational for the ray-and-wavefront description of electromagnetics, not for its field equations. (A thinner thread: his theory of families of surfaces fed into the orthogonal curvilinear coordinate systems, developed by his successor Lamé, that are standard for EM boundary-value problems.)

Ideal fluids, again by way of optimal transport. In a striking closing of the loop, Vladimir Arnold showed (1966) that the incompressible Euler equations are geodesics on the group of volume-preserving diffeomorphisms, and Yann Brenier then showed that projecting onto those maps is an optimal transport problem — Monge's own 1781 question, now serving as the variational skeleton of ideal fluid flow. The semigeostrophic equations of meteorology, solved via Monge–Ampère, are the geophysical-fluids version of the same idea.

The honest verdict: foundational for the mathematics of fluid dynamics and ray optics; a bystander to the field equations of electromagnetism. Monge geometrized analysis — and physics is written in PDEs — so he turns up wherever geometry is the solvent, and is absent where the physics is carried by vector fields he never touched.

Coda: clarity and loyalty

Monge's rehabilitation, when it came, was total. In December 1989, for the bicentennial of the Revolution, his remains were transferred to the Panthéon. His name is among the 72 inscribed on the Eiffel Tower; a crater on the Moon bears his name; a statue stands in Beaune. The republican-turned-imperial-courtier is now firmly in the French national pantheon of scientists.

And yet the uncomfortable question his biographers keep returning to is the one worth ending on. Monge could see three-dimensional space with a clarity almost no one before him possessed — could make the invisible geometry of a fortification or a curved surface snap into legible order. That same mind bound itself, without visible reservation, to a military dictatorship it had every republican reason to distrust. The most absolute spatial lucidity, it turns out, offers no protection at all against political blindness. Genius in one dimension guarantees nothing in another.

SIDEBAR — How France remembers him today

The Bourbon Restoration stripped Monge of his honors and expelled him from the Institut de France in 1816. The modern French Republic has reversed that verdict about as completely as a state can.

The Panthéon (1989). For the bicentennial of the Revolution, the state transferred Monge's remains to the Panthéon — the deliberate act by which France declares a person a national hero. The timing reclaimed him as a son of the Republic he served before Napoleon, not the emperor he followed after.

A living institution. The École Polytechnique he founded in 1794 is still a French state school, supervised by the Ministry of the Armed Forces and, since 2019, a founding member of the Institut Polytechnique de Paris. It treats Monge as the founder — he appears on its centennial coin beside Carnot, Lamblardie, and Prieur — and the honor is current, not merely commemorative: the school runs a Gaspard Monge Visiting Professor Program for international researchers, with the 2026/2027 call open as of this writing.

Everyday reverence. His name is among the 72 engraved on the Eiffel Tower. Central Paris carries a major rue Monge and a Place Monge (with its own métro station) near the original Polytechnique site; statues stand in Paris and in Beaune; a lunar crater bears his name; and the postal service issued a Monge stamp in 1990. France also remains the institutional guardian of the metric system he helped design — the International Bureau of Weights and Measures sits at Sèvres, outside Paris.

What the official memory leaves out. The commemoration is curated tightly around the geometer, the founder, and the metric reformer. His role in the Napoleonic cultural plunder of Italy and Egypt appears nowhere on the pedestal — and because recent French restitution debates (the 2018 Sarr–Savoy report and its aftermath) have centered on African colonial-era objects rather than Napoleonic seizures, Monge has never become a target of that reckoning. The state honors the clarity and quietly forgets the confiscations.

A note on sourcing

This article is drawn from standard scholarly and institutional sources: encyclopedic biographies (the University of St Andrews MacTutor archive, Encyclopædia Britannica), university and academic-society materials, and peer-reviewed and preprint literature on optimal transport.

Where dates differ across sources (notably his birthday), the article follows the consensus of the specialist biographical record.

The equations are written in LaTeX and render in any MathJax- or KaTeX-aware Markdown viewer (GitHub, Obsidian, Typora, Jupyter, Pandoc, most static-site engines). In a plain text editor they will appear as source between $ delimiters.

Verified sources

O'Connor, J. J., and E. F. Robertson. "Gaspard Monge." MacTutor History of Mathematics Archive, University of St Andrews. https://mathshistory.st-andrews.ac.uk/Biographies/Monge/
"Gaspard Monge, count de Péluse." Encyclopædia Britannica. https://www.britannica.com/biography/Gaspard-Monge-comte-de-Peluse
"Gaspard Monge." Wikipedia. https://en.wikipedia.org/wiki/Gaspard_Monge
"Gaspard Monge." EBSCO Research Starters. https://www.ebsco.com/research-starters/history/gaspard-monge
Ashworth, William B. "Gaspard Monge — Scientist of the Day." Linda Hall Library. https://www.lindahall.org/about/news/scientist-of-the-day/gaspard-monge/
"Gaspard Monge." Napoléon & Empire. https://www.napoleon-empire.org/en/personalities/monge.php
"How the French Revolution created the metric system." National Geographic. https://www.nationalgeographic.com/history/history-magazine/article/french-revolution-toppled-king-forged-metric-system
Villani, Cédric. Topics in Optimal Transportation. Graduate Studies in Mathematics 58, American Mathematical Society. https://bookstore.ams.org/gsm-58
Kolouri, S., et al. "Optimal Mass Transport: Signal Processing and Machine-Learning Applications." IEEE Signal Processing Magazine (PMC full text). https://pmc.ncbi.nlm.nih.gov/articles/PMC6024256/
"Optimal transport theory: from sandcastles to artificial intelligence." The Oxford Scientist. https://oxsci.org/optimal-transport-theory/
International Mathematical Union. "Fields Medal 2018 — Alessio Figalli citation" (official PDF). https://www.mathunion.org/fileadmin/IMU/Prizes/Fields/2018/Figalli-Citation.pdf
"Alessio Figalli wins the 'Nobel Prize of Mathematics.'" ETH Zürich (press release). https://ethz.ch/en/news-and-events/eth-news/news/2018/08/fields-medal-for-figalli.html
"ERC grantee receives 2018 Fields Medal for Mathematics." European Research Council. https://erc.europa.eu/news-events/news/erc-grantee-receives-2018-fields-medal-mathematics
"The Fields Medal 2018: Alessio Figalli." Plus Magazine, University of Cambridge. https://plus.maths.org/content/test-2-0
Hartnett, Kevin. "A Traveler Who Finds Stability in the Natural World." Quanta Magazine. https://www.quantamagazine.org/a-traveler-who-finds-stability-in-the-natural-world-20180801/
"Alessio Figalli." Wikipedia. https://en.wikipedia.org/wiki/Alessio_Figalli
Rout, Litu, et al. (and related literature). "Transport-based analysis, modeling, and learning from signal and data distributions." arXiv. https://arxiv.org/pdf/1609.04767
"Gaspard Monge, Founder of the School (1746–1818)." École Polytechnique — 225 Stories. https://225.polytechnique.fr/en/225-stories/gaspard-monge.html
"Gaspard Monge Visiting Professor Program." École Polytechnique. https://www.polytechnique.edu/en/gaspard-monge-visiting-professor-program
"École polytechnique." Wikipedia (current governance and Ministry of Armed Forces supervision). https://en.wikipedia.org/wiki/%C3%89cole_polytechnique
"The Golden Slope of French Geometry — A Celebration of Monge's Life." Institute of Mathematics and its Applications (IMA). https://ima.org.uk/9681/historical-notes-the-golden-slope-of-french-geometry-a-celebration-of-monges-life/
"Gaspard Monge's mausoleum." Wikipedia. https://en.wikipedia.org/wiki/Gaspard_Monge%27s_mausoleum
"Monge cone." Wikipedia (Monge cone, Monge axis, characteristics, and the eikonal equation as the simplest fully nonlinear first-order PDE). https://en.wikipedia.org/wiki/Monge_cone
"Method of characteristics." Wikipedia. https://en.wikipedia.org/wiki/Method_of_characteristics
Gallouët, T. O., and Q. Mérigot. "A Lagrangian Scheme à la Brenier for the Incompressible Euler Equations." Foundations of Computational Mathematics (Arnold's geodesic interpretation; Brenier's optimal-transport polar decomposition). https://link.springer.com/article/10.1007/s10208-017-9355-y
Brenier, Y. "Optimal Incompressible Transport and the Euler Equations" (lecture notes, IHP). https://pcombet.math.ncsu.edu/brenier.pdf

Prepared as a Scientific American–style feature. Facts verified against the sources above as of July 2026.

Search This Blog

Science, Time, and Spirit