藏身中的功课

The details he gave

Grothendieck’s account of hiding occupies only a few lines. Writing more than four decades later, he recalled that most children at the Secours Suisse home were Jewish and that local police warnings before Gestapo roundups sometimes sent groups of two or three into the woods for one or two nights. He also recalled that the children did not fully understand the danger. No independent record concerning him has yet corroborated this account.11Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 14.

Every element in that account defines a limit. There was warning, but the officer or officers are not named. There were woods, but no location is given. There were small groups, but no companion is identified. There was a short duration, but no particular date. There was mortal danger, understood more clearly by the adult writer than, in his telling, by the children at the time. A biographer can preserve all five elements and still refuse a scene.

The refusal matters because concealment invites narrative machinery. Darkness, weather, footsteps, whispered instructions, and the fear of discovery can be supplied from imagination or borrowed from another person’s testimony. Such details would make the episode vivid by making it someone else’s. Grothendieck did not say whether a given departure occurred by daylight, whether it rained, what they ate, or whether German agents came close. The woods should therefore remain the woods of a sourced sentence, not a stage.

The statement does not authorize a general theory that refugee children were unaware of danger. It records only how Grothendieck later characterized the understanding of his small groups.

The account establishes a bounded sequence: a warning preceded departure in small groups, and the children later returned. It does not identify the warning chain, preparation, route, or timing.

To say that local police warned them is not to say that “the police” protected Jews. French police institutions enforced antisemitic policy and took part in arrests. Particular officials could use privileged information against that purpose. The same uniform or office could belong to the machinery of persecution and provide an opening within it. Grothendieck’s memory preserves the opening, not the name.

The children’s return is as important as departure. Hiding did not replace school; it interrupted it. The Collège program resumed, textbooks remained, classmates made noise, and teachers continued. A life composed only of raids would be false to duration. A life composed only of lessons would be false to condition. The chapter must hold both temporal scales at once: occasional acute danger and the long ordinary work of being an adolescent pupil.

Reading ahead

At the Collège Cévenol, Grothendieck later wrote, what struck him most was how little his classmates cared about what they were learning. The judgment may contain the impatience of a mature mathematician looking back at ordinary pupils. He contrasted them with himself: at the beginning of each school year he devoured the textbooks, expecting at last to encounter genuinely interesting things. He then spent the rest of the year as usefully as he could while the program was delivered through the terms.22Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 14–15.

The image is not of a child refusing books. It is of a reader exhausting their intended sequence. A textbook normally parcels knowledge according to a school calendar. Chapter one precedes chapter two; exercises follow examples; the teacher controls pace. Reading the whole book early separates access from schedule. Once the content is visible as a whole, the pupil can compare parts, see repetition, and notice what has been omitted.

Grothendieck’s disappointment arose partly because exercises resembled one another. Repetition has a legitimate pedagogical use: a method becomes reliable through practice. His deeper objection was that the problems seemed to fall from the sky one after another without saying where they came from or where they led. They belonged to the book rather than to him.33Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 15.

A “book problem” can be excellent mathematics. The distinction he drew was experiential, not a universal classification. A problem becomes one’s own when a constraint is felt as necessary: a gap in understanding, a pattern that demands explanation, a question whose answer changes what one can see. Assignment and ownership can coincide, but they do not automatically.

He spent time during lessons doing mathematics problems, eventually finding the supplied exercises insufficient. His later prose acknowledges that this independence was not wholly virtuous attention. He was using class for something other than the class. An account that calls this pure intellectual courage would erase the teacher’s perspective and the social obligations of a classroom. His self-direction included disregard.

The La Guespy note, whose provenance is incomplete, contains further personal descriptions; a single account of this kind cannot support a portrait of his character.

A triangle from three sides

One question he remembered as natural began with a triangle. If the lengths $a$, $b$, and $c$ of its sides are known, then—apart from position and reflection—the triangle is determined, provided the lengths satisfy the triangle inequalities. Therefore, he reasoned, its area should be expressible as a function of $a$, $b$, and $c$.44Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 15.

The reasoning begins with determination. Three arbitrary positive numbers do not always form a triangle: each must be smaller than the sum of the other two. When they do, side–side–side congruence says that any two triangles with those side lengths are congruent. Area cannot vary between them. A formula ought to exist.

The classical answer is Heron’s formula. Let the semiperimeter be

$$s=\frac{a+b+c}{2}.$$

Then the area $K$ satisfies

$$K=\sqrt{s(s-a)(s-b)(s-c)}.$$

No angle or altitude appears in the formula; it converts three boundary lengths directly into area.

One derivation begins with the familiar area formula

$$K=\frac12 ab\sin C$$

and the law of cosines

$$c^2=a^2+b^2-2ab\cos C.$$

Solving for $\cos C$, using $\sin^2 C=1-\cos^2 C$, and factoring yields Heron’s expression. This derivation relies on trigonometric tools. Geometric derivations are also possible. The surviving source does not identify which route Grothendieck used, whether he found Heron’s formula independently, or whether he later recognized the question in a book.

That boundary is important. He wrote that he sought an explicit formula; he did not claim priority for a rediscovery in this passage. The biographer should not elevate a remembered school problem into an unpublished theorem. Its significance lies in the direction of curiosity. Instead of accepting area as something computed from a base and an altitude selected by a diagram, he asked what the intrinsic side data determined.

The word “intrinsic” must be used cautiously at this age. Grothendieck later made intrinsic and coordinate-free formulations central to his work. The school question can be described in modern language as intrinsic because it uses lengths belonging to the triangle rather than its placement in a coordinate plane. There is no evidence that the adolescent used that terminology or possessed the later conceptual program.

Still, the problem contains a real movement from procedure to structure. A textbook might give several kinds of triangles and several calculation recipes. Grothendieck’s question asks for an invariant of the object determined by minimal data. It is a mathematically meaningful question independent of what followed in his career.

Six edges and a tetrahedron

He then asked the three-dimensional analogue: to determine a tetrahedron’s volume from its six edge lengths. He remembered struggling and believed he eventually succeeded. No school notebook or formula has been located.55Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 15.

The question is subtler than the triangle case. Six positive numbers do not automatically form a tetrahedron. Each triangular face must satisfy triangle inequalities, and the faces must fit together in three-dimensional space with nonnegative volume. A flat configuration has volume zero; incompatible distances define none.

A modern compact answer uses the Cayley–Menger determinant. Label the vertices $0,1,2,3$ and let $d_{ij}$ be the distance between vertices $i$ and $j$. Then

$$288V^2= \det\begin{pmatrix} 0&1&1&1&1\\ 1&0&d_{01}^2&d_{02}^2&d_{03}^2\\ 1&d_{01}^2&0&d_{12}^2&d_{13}^2\\ 1&d_{02}^2&d_{12}^2&0&d_{23}^2\\ 1&d_{03}^2&d_{13}^2&d_{23}^2&0 \end{pmatrix}.$$

The determinant packages the dependence of volume on pairwise distances. Positive value gives a nondegenerate tetrahedron, zero a flat one, and a negative result signals that the proposed distance data cannot be realized as an ordinary Euclidean tetrahedron.

This formula is a modern illustration, not a reconstruction of Grothendieck’s work. A school solution might choose a face as base, compute its area, derive the height from edge lengths, and use $V=\frac13(\text{base area})(\text{height})$, or it might introduce coordinates. Grothendieck’s retrospective wording conveys confidence mixed with imperfect recall, not a recoverable derivation.

The pair of problems—triangle and tetrahedron—also teaches the limit of analogy. The two-dimensional formula has a compact symmetric product. The three-dimensional relation is substantially more complicated. Moving up a dimension requires conditions beyond adding another variable to the same recipe. This difference would matter in the years after 1945, when Grothendieck criticized treatments that stopped at three dimensions.

No source says that he generalized the school problem to arbitrary dimension while at Le Chambon. His later manuscript specifically assigns the sustained effort on definitions of measure to 1945–1948 at Montpellier. The school years supplied dissatisfaction and examples; the university years supplied prolonged work. A chronological biography must not move the later research backward to make the adolescent seem complete.66Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 15–16.

Friedel’s classroom

The teacher Grothendieck remembered most favorably at the Collège was Monsieur Friedel, who taught natural history and physics. In his later account, Grothendieck praised Friedel’s ability to identify and convey an essential idea rather than recite inventories of facts, formulas, and dates.77Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 14–15; Grothendieck, La Clef des Songes (1987), 102.

In the same retrospective account, Grothendieck described classroom noise that eventually covered Friedel’s voice. No independent school record has yet corroborated the episode, and Friedel’s full identity and personnel file remain unknown.88Grothendieck, La Clef des Songes (1987), 102–3 n. 68.

This memory complicates an idealized picture of the Collège. It records disorder within an institution associated with moral courage and educational reform, but it does not establish the motives of Friedel or the pupils.

Grothendieck later recalled that Friedel gave extracurricular talks at the children’s home, including one on evolution in March 1944. The separate personal material in that passage remains outside this account’s factual claims because it lacks independent corroboration.99Grothendieck, La Clef des Songes (1987), 102.

Grothendieck’s contrast between the noisy classes and attentive extracurricular audience suggests that setting and group dynamics affected attention. This is his later interpretation rather than an independently reconstructed scene.

Grothendieck later attributed significant philosophical consequences to the evolution talk. That belongs to Chapter 11, where the retrospective theology can be treated on its own terms and separated from scientific claims. Here Friedel’s role is educational. He represented, in Grothendieck’s memory, the rare teacher who loved what he taught and sought the essential idea.

Recipe and definition

The deepest dissatisfaction Grothendieck remembered from school concerned not a missing formula but a missing definition. His books gave procedures for lengths, areas, and volumes, yet offered no serious account of what the length of a curve, area of a surface, or volume of a solid was. He promised himself that he would fill the gap when he had time.1010Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 15.

For a line segment, length seems unproblematic. Coordinates $x$ and $y$ give $|x-y|$. For a polygon, perimeter is the sum of side lengths and area can be reduced to triangles. A rectangular box has volume length times width times height. The problem appears when boundaries curve, surfaces bend, or sets have irregular structure.

A curve can be approximated by inscribed polygonal paths. Finer partitions raise the question whether the sums of segment lengths approach a limit independent of the chosen refinement. For a sufficiently regular parametrized curve $\gamma(t)$, this becomes

$$L(\gamma)=\int_a^b \|\gamma'(t)\|\,dt.$$

That formula presupposes an integral, differentiability, and a parametrization that preserves the intended length. Curves without derivatives or with infinite length show the boundary of the elementary formula.

Area and volume raise parallel questions. A region can be approximated by rectangles and a solid by boxes. Inner approximations may miss boundary pieces; outer approximations may include too much. If the two limiting contents agree, the region is well behaved in the relevant sense. Irregular sets expose the limits of elementary definitions.

Nineteenth- and early twentieth-century analysis developed several answers. Jordan content works for bounded sets whose boundaries are sufficiently tame. Lebesgue measure enlarges the collection of measurable sets and organizes approximation through countable operations. Lebesgue measure is a framework for assigning size consistently under translation, countable additivity, and limiting processes, rather than a longer formula for familiar solids.

The source does not establish what measure theory Grothendieck knew at the Collège. He later wrote that a Montpellier professor told him of Lebesgue’s earlier work, but that university episode belongs after 1945. The famous account of his independent work belongs after 1945. The wartime evidence establishes the question and the promise, not the completed theory.

The distinction between recipe and definition is philosophical as well as technical. A recipe computes size for a prepared class of examples. A definition specifies what kind of size is being assigned, how it behaves, and where it ceases to apply. Recipes can be correct without revealing their domain. A definition can expose why several recipes are instances of one structure.

Grothendieck’s retrospective narrative places this school dissatisfaction at the opening of a survey of his work. That authorial continuity does not establish causation. His later mathematical program developed through university study, Paris, Nancy, collaborators, problems, and existing theories. The adolescent question is an antecedent he chose to remember.

Hiding and abstraction

It is rhetorically tempting to connect concealment in the woods with abstraction in mathematics. Both might be described as leaving an exposed position for a safer general setting; both might be made into a story of invisibility. Such writing would aestheticize danger and misuse mathematics. No source links the experiences, and the analogy contributes no factual understanding.

The two lines should instead remain adjacent. On some occasions, a warning interrupted life and the children hid. Across months and years, Grothendieck pursued questions of geometry and measure. A human life can contain both without one symbolizing the other. The point of chronological biography is not to make every simultaneous event secretly identical.

The absence of direct connection does not require emotional sterility. Studying under threat has a human meaning. It shows that curiosity persisted and that institutions preserved time in which it could operate. The credit is distributed: to the pupil, to teachers including Friedel, to La Guespy, to relief organizations, and to residents whose actions made return possible. Mathematics did not save him from a roundup. People did.

The memoir does not say that Grothendieck solved geometry to forget the war. Assigning that motive would turn a biographical trope into a private fact. His later account identifies sustained interest in the problems, not psychological refuge, as his explanation.

The program and the person

School programs serve collective purposes. They sequence material for a class, prepare examinations, and make credentials portable. A refugee pupil especially needed a recognized credential. The baccalauréat later enabled Grothendieck to enter the University of Montpellier. His impatience with the program therefore existed alongside dependence on its result.

His later prose describes the program as repeating from term to term. His reported response was to read ahead, identify gaps, and work on self-generated questions over long periods. This style also led him to neglect assigned subjects and, later at Montpellier, to attend lectures only intermittently. Between 1945 and 1948, he failed an examination after a numerical error and extended his course of study.1111Grothendieck, Récoltes et Semailles (Université des Sciences et Techniques du Languedoc and Centre National de la Recherche Scientifique, 1986), 16.

At the Collège, the school’s ethical history can overshadow its curriculum. Yet passing through education required both. The institution sheltered and certified. Its founding ideals made room for international and refugee students; its ordinary academic structures linked them to French higher education. Moral courage and bureaucratic credentialing were not opposites here.

The title “Lessons in Hiding” therefore has two meanings, but they should not be merged. There were lessons received while living under a concealed or protected identity, and there were intervals of literal hiding. The first occupied years; the second, in his account, nights. The longer education included language, science, geometry, social conflict, and observation of adults. The shorter precautions kept that education possible.

From formula to measure

The school question can be formulated through properties rather than constructions. Suppose a function $\mu$ assigns a nonnegative size to suitable subsets of space. For volume or area, one may reasonably demand:

1. $\mu(\varnothing)=0$;
2. congruent sets have equal size;
3. if $A$ and $B$ are disjoint, $\mu(A\cup B)=\mu(A)+\mu(B)$;
4. a unit cube has size $1$;
5. sizes behave well under appropriate limits.

Finite additivity handles a finite dissection. Countable additivity,

$$\mu\!\left(\bigcup_{n=1}^{\infty}A_n\right) =\sum_{n=1}^{\infty}\mu(A_n)$$

for pairwise disjoint measurable sets, supports much of modern analysis. It allows one to build complicated sets from countably many pieces and control limits.

The words “suitable” and “measurable” carry the real difficulty. No assignment satisfying all attractive properties can apply naively to every subset while preserving the structures one wants. A theory of measure must specify a collection of measurable sets, commonly a sigma-algebra, closed under complements and countable unions. The definition solves some problems by setting a principled boundary.

This is far beyond the reported Collège curriculum. It is included to show what lay inside the apparently simple request for “a serious definition.” The child’s dissatisfaction was not answered by writing a longer list of formulas. It opened toward analysis, set operations, invariance, and limits.

Grothendieck’s independent Montpellier work would lead him into territory related to Lebesgue integration, but later sources caution against the legend that he recreated the entire theory in isolation. His ideas emerged from partial education, conversations, and available books as well as his own reconstruction. Chapter 13 must examine that claim with mathematical manuscripts and university records. Here the endpoint is a promise made to himself.

The philosophy of a problem

Calling a problem one’s own can suggest possession, but Grothendieck’s later mathematical practice often emphasized discovery rather than ownership. In the school passage the pronoun marks origin and responsibility. The problem does not arrive only because an authority assigns it. It arises because the relation between known data and unknown quantity becomes visible.

For the triangle, the logic is:

• the side lengths determine the figure;
• area belongs to the figure;
• therefore area should depend on the side lengths;
• find the dependence.

For measure, the logic is:

• calculation rules presuppose that length, area, and volume mean something;
• the books do not say adequately what;
• therefore reconstruct the notion.

These are generative problems. Each begins not from a request to apply a named theorem but from pressure within concepts. A curriculum can teach theorems; teaching how such pressure is noticed is more difficult.

Friedel’s remembered strength was precisely to identify the crucial idea from which the rest followed. This links teacher and pupil without making Friedel a mathematical mentor he was not. The science teacher modeled an explanatory priority: essence before inventory. Grothendieck valued him despite the failed classroom and despite later disagreeing with how scientific accounts understood chance and creation.

The educational philosophy should not be generalized from one exceptional student into a policy of leaving pupils alone. Grothendieck’s later isolation at Montpellier cost time and left gaps. Collaborative environments at Nancy and IHÉS would be decisive. Self-generated questions need languages, criticism, and communities. The Collège passage records one pole of learning, not the whole.

An adolescent among adolescents

Grothendieck’s later account resists a sanctified school portrait. He wrote that he worked on his own material during lessons, judged classmates uninterested, and witnessed severe disorder around a teacher he admired. These are one adult witness’s retrospective judgments, not independently verified character claims about the pupils.

His protected status also differed from that of the adults protecting him. He followed warnings and hid; he did not design the regional rescue system. He observed solidarity; he did not possess a survey of its participants. The later famous name should not move him into the center of other people’s decisions.

At the same time, being protected did not eliminate agency. He later recalled reading ahead, forming questions, and pursuing geometry during lessons. Refuge was lived time rather than a blank interval between danger and later achievement. The strongest evidence is specific: textbooks read early, a disorderly class, and questions about a triangle and tetrahedron.

The word “child” also requires care. He entered Le Chambon around fourteen and approached sixteen in 1944. French relief institutions classified residents as children; modern readers may imagine much younger dependants. He was an adolescent able to walk, study at secondary level, and reflect, while still legally and materially dependent on adults. Both capacities and vulnerability belong in the frame.

Coordinates as one possible answer

Heron’s formula is symmetric and elegant, but a pupil need not discover it in that form to solve the triangle problem. Coordinates offer another route and show how a question can have several valid proofs—the issue already present in the Mende examination.

Place one vertex at $A=(0,0)$, a second at $B=(c,0)$, and let $C=(x,y)$. If the remaining side lengths are

$$AC=b,\qquad BC=a,$$

then

$$x^2+y^2=b^2$$

and

$$(x-c)^2+y^2=a^2.$$

Subtracting the first equation from the second gives

$$-2cx+c^2=a^2-b^2,$$

so

$$x=\frac{b^2+c^2-a^2}{2c}.$$

Then

$$y^2=b^2-x^2.$$

The area is $\frac12 c|y|$. Substituting and factoring eventually yields Heron’s formula. The sign of $y$ distinguishes the two mirror placements of the same triangle; the absolute value makes their areas equal.

This derivation illustrates determination concretely. Choosing coordinates fixes position artificially, but the final area no longer depends on that choice. The method uses an external frame to compute an intrinsic quantity.

The source does not identify whether Grothendieck knew coordinate geometry at the time, used trigonometry, constructed altitudes, or found a known formula. The precision window deliberately presents alternatives rather than reconstructing one. Its purpose is to show why his question was answerable within school mathematics and why the route still mattered.

For a tetrahedron, coordinates scale up. Set

$$A=(0,0,0),\quad B=(c,0,0),\quad C=(x_C,y_C,0),$$

using three edge lengths to determine the base triangle. Let the fourth vertex be $D=(x_D,y_D,z_D)$. Its distances to $A,B,C$ yield three equations. Subtractions turn some quadratic terms into linear relations for $x_D$ and $y_D$; the remaining equation determines $z_D^2$. The volume is

$$V=\frac13 K_{\triangle ABC}|z_D|.$$

This is a plausible school-level strategy if the necessary algebra was available. It is not evidence of his actual solution. The Cayley–Menger determinant given earlier packages the same distance geometry symmetrically and generalizes beyond three dimensions.

The coordinate calculation also exposes realizability. If the computed $y_C^2$ or $z_D^2$ is negative, the proposed distances do not fit an ordinary Euclidean triangle or tetrahedron. If $z_D=0$, the four points are coplanar and volume is zero. Thus the formula tests whether the input data define the object as well as computing its size.

That test may have contributed to the problem’s attraction. The book could ask for the volume of a pictured solid whose existence was assumed. Beginning from six numbers asks first whether there is a solid at all. Again, this is an analysis of the mathematical question, not a report of the adolescent’s stated philosophy.

From three dimensions to $n$

Grothendieck later complained that his Montpellier books used single, double, and triple integrals while avoiding dimensions above three. The school questions already form a sequence from triangle to tetrahedron. It is natural for a modern reader to ask what comes next.

An $n$-simplex has $n+1$ vertices. Every pair of vertices determines an edge, giving

$$\binom{n+1}{2}=\frac{n(n+1)}{2}$$

pairwise distances. Under appropriate consistency conditions, these distances determine the simplex up to rigid motion and reflection. A generalized Cayley–Menger determinant gives its squared $n$-dimensional volume.

This statement belongs to the mathematical horizon of the remembered problem. It should not be attributed to Grothendieck at sixteen. His source explicitly moves the work on general volume to Montpellier. The distinction between a question’s logical extension and a person’s historical knowledge is central to responsible mathematical biography.

Higher-dimensional volume challenges visual intuition. A four-dimensional simplex cannot be pictured directly in ordinary space, but its distances and algebraic relations can be handled. Definitions become more valuable precisely when pictures fail. A recipe tied to pyramids and spheres in three dimensions does not automatically explain what “volume” means in $n$ dimensions.

Measure theory supplies one kind of generality: Lebesgue measure on $\mathbb{R}^n$ for every positive integer $n$. Differential geometry supplies notions of volume on curved spaces. Probability interprets total measure one. Functional analysis develops integration in abstract settings. Grothendieck’s later work would engage spaces far removed from Euclidean solids, but those histories should not be collapsed into the school promise.

The progression does explain why “serious definition” was more demanding than a formula. If length, area, and volume are separate entries in a table, a new dimension needs another entry. If they are manifestations of a common measure concept, the dimension becomes a parameter in a unified structure.

There are limits to unification. Curve length, surface area, and ambient volume may use measures of different dimensions. A curve in the plane usually has two-dimensional area zero while possessing positive one-dimensional length. A surface in three-dimensional space may have volume zero and positive area. Calling all of them measure without qualification hides the need to specify dimension and construction.

This distinction would later be expressed through Hausdorff measure and related tools, not through one elementary formula. It shows the depth of the gap the young pupil noticed even though he did not yet possess its language.

Textbooks as historical objects

Grothendieck’s criticism refers retrospectively to the mathematics books used at school, but the editions have not been identified. Finding them would turn a general memory into a testable curricular claim. A French geometry manual used at the Collège in 1943 might define area differently from a German book he had read earlier; wartime shortages might have kept older editions in use.

A textbook audit should establish:

• the titles in Collège inventories or pupil lists;
• the program and examination series followed by his class;
• where the books introduce congruence, Heron’s formula, solids, and integration;
• whether definitions are absent, informal, or postponed;
• whether exercises follow the repetitive arrangement Grothendieck later remembered.

His dissatisfaction remains valid as testimony even if a book contains a paragraph labeled “definition.” A formal sentence can fail to answer the conceptual question a reader has. Conversely, an audit may show that the mature account sharpened the omission.

The books are also image candidates. A correct cover, contents page, or measure passage could anchor the documentary window more truthfully than a generic blackboard. Rights may be easier to clear if the edition is old, but public-domain status must be assessed by jurisdiction, author, and edition.

Marginalia would be stronger still. A volume owned by Grothendieck, if preserved, could show what he marked. No such schoolbook has been located.

The teacher’s role also depends on the book. A program may have required standard proofs for examination purposes. The Mende teacher’s insistence could reflect marking conventions rather than inability to judge, though Grothendieck interpreted it personally. School archives might reveal official model solutions and help reconstruct the institutional pressure around the episode.

This does not invalidate his remembered principle. A nonstandard proof should be assessed mathematically. It does remind the biographer that teachers work within credential systems and that the pupil’s view is one position in the classroom.

Time protected, time interrupted

Mathematics requires stretches of attention, but not necessarily tranquil lives. Grothendieck later wrote that when a question held him, he stopped counting hours and days. That retrospective account supports sustained concentration but does not establish how often danger interrupted it.

The children’s home created protected time through material organization. Meals, beds, supervision, and school access removed tasks the adolescent could not have secured alone. Local warnings protected time by interrupting it: leaving for the woods sacrificed a night in order to preserve the person who might return to the problem.

This relation helps avoid calling mathematics an escape. The rescue network, not mathematics, moved him away from immediate danger. His memoir shows that mathematical questions occupied his attention across institutions; it does not describe a psychological escape.

The same questions could survive a change of place. A triangle did not depend on whether it was considered at the Collège or later in Mairargues. That portability differs from a home, which is materially located and socially maintained. Mathematical objects can offer continuity without providing shelter.

Attention also carries social costs. Working during lessons withdrew attention from a teacher and classmates. The La Guespy note, whose provenance is incomplete, cannot support further claims about his character.

The school’s routine, which he found tedious, was itself a temporal protection. Terms proceeded, examinations approached, and ordinary expectations continued. Routine can feel oppressive to an advanced pupil and stabilizing to a community under threat. Both functions belong in the account.

When Grothendieck later wrote that the camp taught what books could not, he distinguished lived time from textual knowledge. At Le Chambon he also learned things through books that immediate experience could not supply: geometry beyond the visible setting, scientific history across billions of years, problems independent of political borders. The opposition between life and books was never complete.

Problem ownership and pedagogy

Grothendieck’s distinction between the book’s problems and his own can be read too quickly as a rejection of pedagogy. A teacher cannot wait for every necessary concept to arise spontaneously in every pupil. Textbook exercises expose students to structures they might not encounter and build techniques required before deeper questions can be handled.

The criticism targets motivation and relation. An exercise may be externally assigned and still become one’s own if its purpose becomes visible. Conversely, a difficult exercise can remain alien if the pupil sees only instructions. Good pedagogy moves between the two.

The triangle question offers an example. A book may present Heron’s formula for substitution. A teacher can instead ask why three sides should determine area, let pupils explore constructions, compare formulas, and identify impossible triples. The content is the same; the problem’s origin becomes intelligible.

Grothendieck’s style was unusually self-propelling. Educational policy designed around him alone would fail many pupils. His recollection should not be used to claim that instruction is harmful or that gifted children need no teachers. Friedel mattered precisely as a teacher who conveyed an essential idea.

Nor did self-generated work guarantee novelty. A natural problem may have a classical solution. Discovering it independently can be educationally important and historically nonoriginal. The young pupil’s ownership concerns the act of knowing, not priority.

This distinction will matter throughout the biography. Grothendieck later entered fields with established literatures, renamed and reorganized concepts, and worked with collaborators. Questions could be deeply his while answers also belonged to a mathematical inheritance. Biography must separate creative possession from exclusive authorship.

The classroom has a social dimension that the solitary account minimizes. A teacher must allocate attention among students, maintain order, and meet an examination schedule. Grothendieck could read ahead because a book existed and because others kept the institution operating. His retrospective account records impatience within that institutional setting.

The Collège’s reform ideals may have allowed more independence than a conventional lycée, but the source does not describe special provision for him. No mentor is reported assigning advanced work. His private program grew alongside, not necessarily through, formal differentiation.

An educational portrait should therefore include three agents: the learner who forms a question, the teacher who creates access and criticism, and the institution that preserves time and credentials. In Grothendieck’s wartime account, the first is most audible because he wrote it. The other two remain necessary.

Mathematics during war, not mathematics from war

Chronology places the triangle, tetrahedron, and measure questions inside wartime refuge. Causal language must be stricter. “During” is established by his memoir. “Because of” is not.

War affected conditions of study. It displaced him into French schools, separated him from parents, created dependence on relief, interrupted nights through hiding, and limited access to books and experts. These conditions shaped which teachers and materials were available.

They do not explain the mathematical content by themselves. Heron’s formula and Euclidean distance geometry arise from properties of triangles and tetrahedra. Measure theory arose from nineteenth-century analysis, integration, and set theory. The questions have intellectual histories independent of his biography.

No inspected source supports claims that institutional rupture produced self-reliance, that uncertain civil categories caused an interest in definitions, or that hiding encouraged abstraction. Grothendieck placed these wartime school memories near the start of his mathematical autobiography. That editorial placement suggests an authorial claim of continuity, not a documented psychological mechanism.

This rule prevents both romanticization and compartmentalization. The mathematics should not float in a chapter as if war were a backdrop; protection and schooling made the work possible. War should not be credited as a teacher of mathematics; people and concepts did the teaching.

The phrase “genius forged by adversity” is therefore excluded. Adversity destroyed lives and education far more often than it produced celebrated work. Alexander’s survival and opportunity were contingent achievements of rescue.

Reconstructing one school year

A rigorous reconstruction would combine several records. The Collège register could show class level and dates. A timetable could identify hours with Friedel and mathematics teachers. Textbook lists could reveal available content. La Guespy logs could record absences or warnings. Examination papers could show what Alexander actually wrote.

These sources may not all survive. Their absence should be recorded by type rather than collapsed into a general claim of ignorance. Two manuscripts provide substantial retrospective evidence, while particular contemporaneous school documents remain missing.

Oral histories from other pupils could corroborate classroom noise, Friedel’s talks, and warning practices. They must be used for shared events and attributed experience. Another pupil’s fear in the woods cannot be assigned to Alexander, but a dated lecture remembered by several people can strengthen chronology.

School personnel files might identify Friedel’s full name and later appointment. His scientific and religious background could clarify the evolution talk. Care is needed not to infer lecture content from his biography alone.

The correct textbooks could also establish the terminology for triangle congruence, whether they gave Heron’s formula, and how they defined curve length. A page-level comparison would turn his critique into intellectual history rather than anecdote.

Mathematical reconstruction has its own limit. Even with a notebook, an editor should not silently repair errors and publish a polished solution as his. Facsimile, diplomatic transcription, and mathematical commentary must be separated. A teenager’s incomplete attempt can be more historically valuable than a perfected result.

The image program should reflect those layers: object, transcription, explanation. An original diagram by the book’s design team must be labeled as such. A source diagram must preserve citation. Grothendieck’s hand must never be simulated.

The role of error

The remembered school episodes are often told as conflicts between Grothendieck’s insight and other people’s limitations. Mathematics also advances through his own errors, incomplete attempts, and revisions. No notebook survives here to show them.

The poor mark at Mende may have reflected teacher rigidity, a flaw in the proof, or both. Grothendieck remained certain of its validity, but certainty is not independent verification. An account of the episode should preserve that possibility unless the script is found.

His tetrahedron recollection is similarly tentative: his later account recalls difficulty and probable eventual success. The wording acknowledges blurred memory. It does not present a polished formula. Failed approaches may have occupied most of the work.

Error matters educationally because a self-generated problem supplies no answer key. One must test special cases. A proposed triangle-area formula should vanish when $a+b=c$, remain symmetric in side labels, and scale by a factor of $\lambda^2$ when all lengths scale by $\lambda$. A tetrahedron-volume formula should scale as $\lambda^3$ and vanish for coplanar points.

Dimensional analysis is a simple check. Area must have units of length squared. In Heron’s formula, the four factors under the square root each have units of length, producing length squared after the root. For the Cayley–Menger expression, determinant scaling yields the required sixth power for $V^2$.

Such checks do not prove a formula, but they expose mistakes. They also show how first-hand judgment is disciplined by invariance and consequence rather than confidence alone.

Grothendieck’s later career included sweeping programs that required collaborators to test details and write proofs. The school portrait of solitary certainty should not become the only model of his mathematics. Error correction is social as well as individual.

Three notions of “natural”

Grothendieck called the triangle and tetrahedron questions natural. The word can mean at least three things. A question may arise naturally from given data; it may fit a historical development; or it may feel inevitable to a particular thinker.

The triangle question is structurally natural because side lengths determine congruence class and area is an invariant. Historically, the formula was ancient. Personally, it attracted Alexander more than the listed exercises.

The measure question is natural in a different way. Calculation rules use length and volume as if already known, so asking for a definition exposes a logical presupposition. Historically, the question had driven major developments before his birth. Personally, he encountered it without adequate access to that history.

Calling a problem natural can conceal education. What appears inevitable after a concept has been formed may be invisible before someone points it out. Textbooks, teachers, and traditions shape which gaps a pupil notices.

Grothendieck’s talent included treating familiar words as unresolved. That does not mean everyone who accepted school definitions lacked curiosity. Different mathematical practices require different levels of foundation at a given time.

No later metaphor for Grothendieck’s method is imported into these school calculations. The triangle and tetrahedron questions could be attacked directly; the broader measure project belongs to Montpellier.

Explanation at three levels

The chapter contract asks mathematical exposition to move from human problem to faithful intuition to optional precision. Grothendieck’s school memories show why the levels should not be collapsed.

At the human level, a pupil encounters formulas without origins. The motivating issues are recognizable without notation: side lengths determine a triangle and should therefore determine its area; calculations of volume presuppose a definition of volume.

At the intuitive level, determination and approximation do the work. Side lengths fix shape. Small rectangles approximate a region. Finer partitions should improve an estimate. This level makes the direction visible and can also conceal difficulty: not every approximation converges, and not every boundary behaves.

At the precision level, conditions and formulas appear. Triangle inequalities determine whether sides are possible. Heron’s formula computes area. Cayley–Menger determinants test distance data. Sigma-algebras and countable additivity define measure on an appropriate domain.

Each level has a failure mode. Human narrative can romanticize curiosity. Intuition can mistake a picture for proof. Precision can bury the motivating question under notation. A biography of mathematics needs all three because the historical actor moved among them unevenly.

The surviving testimony occupies mainly the first two levels. Grothendieck gives questions and confidence, not the school derivations. The precision supplied here belongs to the modern narrator. No equation in this chapter is a quotation from his notebook.

Compression introduces another danger. “He discovered the area formula and volume formula” would be shorter and less accurate. The primary text says he sought the formulas and believes he eventually reached the tetrahedron answer. “He invented measure theory” would be false. “He noticed a definitional gap and later worked on it” preserves chronology.

Mathematical readers may find the school questions elementary beside later algebraic geometry. Their importance is biographical, not technical rank. An elementary question can reveal a mode of attention. Non-specialist readers should not be told that elementary means trivial; the foundations of familiar operations can be deep.

The chapter’s mathematical density is therefore deliberate. It shows that the remembered dissatisfaction had content, while the repeated caveats protect against retrospective attribution. Technical explanation is not decoration and not proof of genius. It is the means by which a claim about “questions of volume” becomes checkable.

Why elementary mathematics belongs here

A long biography of Grothendieck will later contain schemes, cohomology, topoi, motives, and higher categorical ideas. It may seem disproportionate to spend pages on triangle area. The early problem serves a different function from the later theories.

It is one of the few pieces of mathematics he explicitly located in wartime schooling. Omitting it would leave war context rich and intellectual development vague. Inflating it into precocious research would be equally false.

Elementary mathematics also permits exact exposition to general readers. The relation among side data, congruence, and area can be followed without advanced prerequisites. Readers can follow the move from a supplied exercise to a generative question instead of receiving a general claim of originality.

The tetrahedron then shows a real increase in complexity, and measure opens toward abstraction. The sequence models the book’s three-layer method while remaining chronological.

Later chapters should not repeat the full derivations. They may refer back to the “measure question” and add new evidence.

Above all, technical accuracy protects biographical restraint. Once the actual questions are stated, there is no need for vague superlatives. The reader can see what was elementary, what was difficult, what was remembered, and what remains unverified.

Formulas, assumptions, and historical attribution can be checked without validating a psychological analogy. Heron’s formula and Cayley–Menger belong to established mathematics; the claim that Alexander pursued the questions belongs to testimony; any claim about his exact solution would require the solution itself.

For literary review, the equations create a pacing risk. If they interrupt danger at the wrong moment, mathematics can appear insensitive; if moved to an appendix, his intellectual life disappears from the chapter. Placing them after the school memories keeps chronology and allows readers to skip precision without losing narrative.

For image review, original diagrams must avoid handwriting fonts or aged paper that suggest a facsimile. The visual distinction between documentary object and explanatory reconstruction is part of truthfulness.

For continuity review, later chapters should distinguish the adolescent’s question of measure from the adult’s vast geometric work. The same word “space” will acquire new technical meanings. Cross-references must not imply that a scheme is a generalized solid or that topos theory answers the school volume problem.

This protects readers from a common explanatory shortcut: every later abstraction presented as a larger version of one childhood intuition. Mathematical development branches, abandons problems, and incorporates other people’s work. The school episode is an opening, not a master key.

Worked examples and boundaries

The modern formulas become clearer when tested on cases that separate existence, degeneracy, and ordinary shape. Side lengths $1,1,1$ determine an equilateral triangle with positive area. Lengths $1,1,2$ reach the boundary of the triangle inequality and produce a degenerate figure with area zero. Lengths $1,1,3$ do not determine a Euclidean triangle at all. Heron’s expression reflects each case through the sign or vanishing of its factors.

A regular tetrahedron with all six edges equal to $1$ has volume $\sqrt{2}/12$. Four coplanar points can satisfy the triangle inequalities on every face and still produce tetrahedral volume zero. Pairwise face checks are therefore necessary but do not alone guarantee a nondegenerate solid; the three-dimensional compatibility condition matters.

Measure supplies a different boundary. The interval $[0,1]$ has Lebesgue measure $1$. The rational numbers inside it are countable and have measure zero even though they are dense: every subinterval contains rational points. Density and measure answer different questions. This is a true example of why visual crowding cannot serve as a definition of size.

The indicator function of the rationals on $[0,1]$ is another useful test. It equals $1$ on rational points and $0$ on irrational points. Its Lebesgue integral is $0$ because the rational set has measure zero, while the function is not Riemann integrable in the ordinary sense because every interval contains both rational and irrational points. The example shows one historical role of the broader measure framework: it handles functions and limiting behavior beyond elementary area recipes.

These are modern examples supplied for explanation. No source shows that Grothendieck used them at the Collège or in his earliest Montpellier work. They give real instances and failure points, while the biographical claim remains narrower: he later remembered asking for definitions adequate to length, area, and volume.

The terminology must remain equally explicit. Heron’s formula is a modern standard name for an ancient result; the Cayley–Menger determinant is a later formalism for distance geometry; sigma-algebra, outer measure, and Lebesgue integration belong to mathematical developments already available before Grothendieck entered university. None should be placed in the adolescent’s vocabulary without a notebook or textbook.

The historical role of the examples is therefore explanatory rather than proprietary. They show what kinds of conditions a general definition must control: existence, degeneracy, dimension, countable decomposition, and limiting behavior. They do not establish what Grothendieck proved or whether his private work was original.

This distinction is especially important for priority language. To ask a classical question independently is not to discover the classical answer first. To reconstruct part of a theory without access to its literature can be intellectually significant without becoming a priority claim. Part III must compare definitions and proofs before assigning originality.

The examples also separate three verbs that biographies often blur. A student may calculate an instance, derive a known formula independently, or define a framework that covers new cases. Each claim requires different evidence. Grothendieck’s memoir supports the questions and a tentative memory of success with the tetrahedron; it does not identify which of those mathematical achievements occurred.

The missing notebook therefore controls both technical and historical claims. Without it, no formula can be attributed to the pupil, no error can be repaired on his behalf, and no priority can be assigned. Modern notation remains explanatory apparatus only.

Correct normalization of the Cayley–Menger determinant, conditions for measure, and the contrast between Riemann and Lebesgue integration are mathematical questions; whether Grothendieck used them is a separate historical question.

The notation must also remain local to the window. Each symbol is defined beside first use, and no temporary symbol should be assumed in later chapters. The formulas require a rendering check as well as semantic review, because a correct idea can still be corrupted by conversion into DOCX or PDF during the final build process.

Return from the woods

After a warning, the small groups returned from the woods. Grothendieck gives no ceremony of return. School and home resumed. The absence of drama in the telling may reflect how a precaution became part of life, or it may reflect the compression of later memory. The narrative should not decide.

He returned with questions unfinished by the curriculum: how to judge a proof on its reasoning, how side lengths determine area, how distances determine volume, and how volume itself should be defined. These questions differed in scale. Some had classical formulas. The last demanded a theory.

Grothendieck later dated Friedel’s evolution talk to March 1944. Hanka’s post-Rieucros chronology is not narrated because no person-level movement or release file has been inspected.

The questions of length, area, and volume crossed into his postwar study. The woods did not answer them, and neither did the war. A protected school let an adolescent whom French authorities had treated as foreign continue asking. His precise legal status remains unresolved. Survival preserved the question.