The Long Silence

A letter to Daniel Quillen dated February 19, 1983, opens Pursuing Stacks—À la poursuite des champs. Dated entries continue through November 12, and the University archive identifies the manuscript family as cote 134.11Grothendieck, À la Poursuite des Champs (Pursuing Stacks), opening and last-dated entry; Maronne and Toën, Le temps des réflexions, 19–21, 33–34.

The title “Long Silence” is therefore true only in a restricted sense. Grothendieck had not resumed ordinary journal publication, yet dated pages document months of concentrated work.

The distinction is the chapter’s governing fact. A bibliography can show a gap. An archive can fill the gap with forms the bibliography was not designed to count.

Four days from reluctance to a new work

On February 15 Grothendieck called himself “out of the game” while raising a technical disagreement; on February 19 the Quillen letter began the manuscript. No private conversion is inferred.22Grothendieck and Brown, Correspondance Grothendieck–Brown, February 15 and March 7, 1983, pp. 17–19; Grothendieck, À la Poursuite des Champs (Pursuing Stacks).

Strict and weak

The February 15 and 19 documents frame one disagreement: candidate models had to retain higher coherence while supporting homotopy theory. Quillen was the addressee because of Homotopical Algebra. The documents identify the problem that started the dossier without supplying a settled answer.33Grothendieck and Brown, Correspondance Grothendieck–Brown, February 15, 1983, pp. 17–19; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), opening Quillen letter; Daniel G. Quillen, Homotopical Algebra, Lecture Notes in Mathematics 43 (Springer-Verlag, 1967), 1–64.

The logbook form

The archive exhibition describes Pursuing Stacks as the striking example of a new style. Dated entries, digressions, corrections, and reports of daily mathematical life interrupt an axiomatic-developmental line. Grothendieck later contrasted this with a polished manuscript that hides the path of discovery.44Maronne and Toën, Le temps des réflexions, 17–21; Grothendieck, Récoltes et Semailles, ed. Carmona (1985), “Une lettre, ” p. L17.

The contrast should not be exaggerated. He continued to value formal axiomatic exposition and taught it. Scratchwork still lay behind the logbook. Spontaneity was composed; not every hesitation entered the text.

The new style retained motivation, false starts, and correction inside the composition record rather than presenting only final statements.

It also imposed costs. Length expanded. Readers had to separate current claims from abandoned ones. Personal digressions could obscure the mathematical architecture. An editor had to typeset the work and decide how to preserve chronology without sacrificing navigation.

The form was a mathematical and literary experiment in recording discovery alongside results.

Against and beside Bourbaki

Grothendieck had learned mathematical writing in a culture that valued definitions, propositions, proofs, and an impersonal order from general to particular. EGA and SGA differed from Bourbaki while sharing the ambition to build a dependable architecture in which later results followed cleanly.

The logbook did not reject rigor. It changed the order of exposure. A false start could appear before the definition that repaired it. Motivation and doubt entered the main text. The first person and date returned.

The advantages are pedagogical. Readers see why a notion is needed and which tempting simplification fails. They can distinguish discovery from verification. The author’s authority becomes less oracular because errors remain visible.

The dangers are equally real. A dated page may cease to represent the current claim, narrative momentum can replace proof, and length can make verification harder.

Grothendieck later accused polished formal writing of hiding the movement of thought. The 1983 dossier records that alternative compositional choice without adjudicating later editorial practice.

The manuscript therefore stands beside Bourbaki rather than against it. It tests a complementary genre. Grothendieck still anticipated eventual formal redaction for parts of the work, even while insisting that the process itself be shown.

University once a week

In the May 23 letter, giving his vacation address, Grothendieck remarked that outside the vacation he was at the University only once a week. The sentence is a precise but limited snapshot of 1983. It should not be generalized to all his Montpellier years.55Grothendieck and Brown, Correspondance Grothendieck–Brown, May 23, 1983, pp. 43–44.

It does show how the professorship then operated geographically. His correspondence came from Les Aumettes near Mormoiron. The campus was a recurring weekly destination rather than his daily intellectual residence.

No domestic scene is inferred from the address. The archive permits us to say where mail went, not what happened around the writing table.

Marginal, supported, and connected

The May letter establishes continued salary, affiliation, and a once-weekly campus snapshot while composition proceeded largely elsewhere. It does not establish a general schedule for earlier years or a motive for distance.66Grothendieck and Brown, Correspondance Grothendieck–Brown, May 23, 1983, pp. 43–44.

Quillen’s opening

The first manuscript witness is epistolary rather than an undated chapter: Grothendieck addressed Daniel Quillen on February 19, 1983, and that text opens Pursuing Stacks, Montpellier archive cote 134. His letter to Ronald Brown four days before had placed a technical disagreement beside his claim to be largely outside current machinery. The calendar establishes only an external sequence—a problem under discussion, followed by the start of a new written dossier. It supplies no evidence for an inner conversion during the interval.77Wikiquote, “Alexandre Grothendieck,” 1983, 17–19; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 19–21.

Quillen’s role is defined by the document addressed to him and by his earlier Homotopical Algebra. The chapter may say that Grothendieck chose Quillen as addressee for a foundational letter. It does not say that Quillen commissioned the work, endorsed its program, or replied. In September Grothendieck told Brown that he had received no line from Quillen and allowed for the possibility of lost faculty mail. That later statement proves only Grothendieck’s reported nonreceipt at that date; it does not settle whether a reply was written or survives elsewhere.88Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 65–69; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Quillen, Homotopical Algebra, 1–64; Centre for Grothendieckian Studies, “Correspondence”.

The opening question can be named without teaching its answer. Grothendieck objected that available strict structures did not capture the full homotopical behavior he wanted and sought a sufficiently flexible foundational language. This identifies the disagreement that prompted the letter; it neither defines the proposed structures nor claims that the opening solved the problem.99Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 17–19; Grothendieck, À la Poursuite des Champs (Pursuing Stacks).

The custody claim also remains bounded. The archive cote identifies the manuscript object; the Brown correspondence independently dates Grothendieck’s sending and correction of portions during 1983; the archive exhibition supplies a later scholarly account of the sequence. These witnesses corroborate composition without becoming one complete critical edition. Every page-level claim must say which witness carries it. The opening letter’s date is secure; its entire textual stemma, any draft before February 19, and the history of Quillen’s copy remain open archival questions.1010Grothendieck and Brown, Correspondance Grothendieck–Brown, 17–97; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 19–21.

The archive record describes a six-page Quillen letter at the opening of the dossier. That extent belongs to the letter witness, not to the full manuscript, its transcription, or the later printed volume. The date, addressee, cote, and opening role are secure; leaf-by-leaf mapping among witnesses remains unverified.1111Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 19–21.

The February 19 leaf performs two documentary functions at once. As correspondence, it has an addressee and anticipates response. As manuscript opening, it begins a text that soon exceeds ordinary letter scale. Those functions should not be fused into a claim that Quillen collaborated on the work. Brown’s March 7 acknowledgment of receiving a copy confirms that the opening was already being shared with a named correspondent, while leaving Quillen’s own receipt and reaction unresolved. The chapter records the letter becoming a manuscript; it does not turn the addressee into an editor or the copy into broad circulation.1212Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 19–20; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Centre for Grothendieckian Studies, “Correspondence”.

The three February–March objects do not have the same status. The February 15 item is a letter to Brown in an edited correspondence; the February 19 item is the dated Quillen opening preserved with the manuscript family; the March 7 item is Brown’s acknowledgment of a copy. Together they establish question, composition, and a named reader. They do not establish that the copy Brown saw was identical in pagination or corrections to the archive witness eventually cataloged as cote 134.1313Grothendieck and Brown, Correspondance Grothendieck–Brown, 17–20; Grothendieck, À la Poursuite des Champs (Pursuing Stacks).

A dated composition log

The dated spine begins on February 15 with the Brown letter and on February 19 with the Quillen opening. It then jumps in the opened correspondence to May. On May 10 Grothendieck expected a first packet sent for copying to have reached Brown and noted that a few pages from a later part were not yet included. On May 23 he sent four omitted pages and withdrew an assertion carried in the notes. On May 25 Brown acknowledged receipt of 191 pages. These dates show a manuscript growing in packets and corrections; they do not fill the unrecorded days between them.1414Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 42–45; Centre for Grothendieckian Studies, “Correspondence”.

June adds a feedback state. On June 14 Grothendieck thanked Brown for corrections, said that several months of reflection had made some comments more intelligible, and discussed changes in his understanding. The letter can establish learning and response because the author states them. It cannot be used to assign every change in the manuscript to Brown or to reconstruct an undocumented editorial collaboration. A correction in correspondence becomes part of the composition history only when it can be tied to a dated claim or page, not because the correspondent knew the subject.1515Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 45–49.

September supplies another precisely bounded interruption. The archive study reports that on September 7 Grothendieck recorded a telephone conversation in which Luc Illusie invalidated an important working hypothesis. The manuscript returned to that line on September 26. The interval is documented as a break in that argument. It is not evidence of mood, exhaustion, or the absence of work on other pages. The proper chronology has four parts: a call as recorded, a hypothesis rejected, a dated gap in that line, and a dated resumption.1616Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 134, 490; Maronne and Toën, Le temps des réflexions, 21.

The last controlling date is November 12, the endpoint assigned to the principal 1983 logbook sequence by the archive record. “Last dated” is the exact predicate. It does not mean final redaction, completed architecture, last mathematical thought on the subject, or last letter concerning the material. The record after November includes correspondence, while the manuscript’s planned continuation remained open. A gap after the last dated page is evidence only of an endpoint in the surviving sequence; it cannot by itself explain why composition changed or stopped.1717Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 88–92; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 20–21, 33–34.

The chronology can be represented as manuscript states rather than a continuous diary. State A is the February question and opening. State B is the May packet expected and then acknowledged. State C includes the May correction and the June response to feedback. State D records the September hypothesis failure and later resumption. State E is the last dated November page. The letters and archive account connect those states, but they do not establish the precise word count or subject of every intervening page. This structure prevents the prose from saying “then” where only “by this date” is known. It also distinguishes a correction date from the date on which the corrected manuscript page, if any, was rewritten.1818Grothendieck and Brown, Correspondance Grothendieck–Brown, 17–92; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 19–21.

Gaps remain visible at day and month level. The opened file does not provide a complete calendar for March and April, a daily sequence through summer, or a page-by-page map from September 26 to November 12. Later accounts may describe broad phases, but this chapter does not convert them into dates unsupported by the documents cited here. An archival log can be expanded only by dated leaves, colophons, correspondence, or a critical apparatus that maps pages to states. The present log is valuable precisely because every listed checkpoint has such a witness and every interval remains an interval.1919Grothendieck and Brown, Correspondance Grothendieck–Brown, 17–92; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 19–21.

December correspondence belongs after, not inside, the principal sequence. Letters dated December 7 and 8 show continued mathematical communication, but they do not move the manuscript’s last-dated endpoint beyond November 12 unless a manuscript leaf does so. This is a useful test of event type: a later letter may discuss, correct, or plan work without being a page of the work itself. The chronology therefore has an open tail—composition state ending at its last dated leaf, related correspondence continuing afterward. Calling the manuscript “open” preserves precisely that distinction.2020Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 88–92; Grothendieck, À la Poursuite des Champs (Pursuing Stacks).2121Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 20–21.

The log also distinguishes event date from report date. The September 7 call is known through a manuscript entry described by the archive study; the September 26 return is another dated manuscript state. The historian cites both witnesses rather than assuming that every recollection was written on the day it describes. A critical edition must verify that relation page by page.2222Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 134, 490; Maronne and Toën, Le temps des réflexions, 21.

Brown’s June 22 letter answers Grothendieck’s long June 14 response and continues the exchange after the acknowledged corrections. It is a new correspondence state, not evidence that a revised packet had already been sent or that Brown’s explanations entered a particular manuscript page. The date proves continued technical dialogue while leaving physical redaction to the manuscript witnesses.2323Grothendieck and Brown, Correspondance Grothendieck–Brown, 49–55; Centre for Grothendieckian Studies, “Correspondence”.

The September 6 letter belongs to architectural history rather than to a total-page count. Grothendieck described the first large part as taking far longer than the month or two he had once expected, discussed material still to be organized, and named projected supporting matter. The letter proves that the plan was still being revised before the September 7 correction and well before the November endpoint. It cannot prove that every projected chapter or apparatus item was drafted.2424Grothendieck and Brown, Correspondance Grothendieck–Brown, 65–69.

The December 7–8 exchange consists of letters and includes a presentation text connected with the planned series. These objects document mathematical communication, planning, and exchange in a different medium from manuscript leaves.2525Grothendieck and Brown, Correspondance Grothendieck–Brown, 88–92; Grothendieck, À la Poursuite des Champs (Pursuing Stacks).

An April 12 letter supplies an intermediate composition state that the later packet total alone cannot show. Grothendieck said that the reflection begun after the Quillen letter had quickly changed direction, that nearly two hundred typewritten English pages existed, and that he hoped to finish a first batch within another month or two. He invited Brown to read the notes before publication for intelligibility. The page estimate and forecast are authorial statements on April 12, not a final extent or schedule.2626Grothendieck and Brown, Correspondance Grothendieck–Brown, 24–29.

The April state also fixes the medium. Grothendieck described the pages as typewritten, in English, and growing out of correspondence with Brown and Quillen. Those facts distinguish this dossier from the handwritten Long March and explain why copies could be proposed to readers while composition continued. “Typewritten” does not mean printed, edited, or stable; the same April letter anticipates further writing and reader comments.2727Grothendieck and Brown, Correspondance Grothendieck–Brown, 24–29.

The April forecast was already wrong by September in a documentable sense: the first batch had not become the finished publication anticipated within one or two months, and Grothendieck now described the first large part as still expanding. The comparison does not establish failure or motive. It establishes revision of an authorial schedule between two dated letters and shows why later volume plans cannot be treated as a completion calendar.2828Grothendieck and Brown, Correspondance Grothendieck–Brown, 65–69.

Correspondence continued across the September interruption. Grothendieck wrote on September 8, Brown replied on September 15, and Grothendieck wrote again on September 21 and 28. Those dates document an active exchange around the same period in which the manuscript records the September 7 failed hypothesis and September 26 resumption. They do not show that every letter concerned that hypothesis or that correspondence filled the manuscript gap. The two chronologies overlap without becoming identical.2929Grothendieck and Brown, Correspondance Grothendieck–Brown, 69–82; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 490.

Packets, replies, corrections

Sending and receipt are separate events. Grothendieck’s May 10 letter anticipates arrival of a first bunch of notes; it is evidence of sending intention and expectation. His May 23 letter accompanies four omitted pages. Brown’s May 25 letter acknowledges 191 pages. Only the acknowledgment fixes receipt at Bangor. The chapter does not infer that every page Grothendieck believed sent arrived in the same state, or that the 191-page set was identical to a later bound or scanned witness. Packet identity must be established through dates, page ranges, and surviving copies.3030Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 42–45.

The May 23 correction belongs to composition because Grothendieck names an assertion in the notes and withdraws it while sending missing pages. Brown’s May 25 response supplies counterexamples and references. The chapter can report that feedback reached the author and that a claim was corrected. It does not replace the withdrawn assertion with a modern tutorial, decide the final formalism, or assign a completed proof to the exchange. A correction is a dated change in claim status; it is not automatically a theorem.3131Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 43–45.

June shows what a reply could change without proving coauthorship of the manuscript. Grothendieck explicitly credits Brown’s patient explanations and correction of misconceptions. That acknowledgment belongs in the history of learning. Brown’s authored results and examples remain Brown’s, while Grothendieck’s manuscript remains Grothendieck’s. The correspondence documents influence at the level of identified comments and revised understanding; broader claims require matching a letter to the manuscript state it affected.3232Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 45–49.

The September correction has a different source path. It is recorded on a dated manuscript page and described by the archive exhibition, not established through Brown’s reply. Grouping it with the May exchange is useful only at the level of composition practice: both show an active claim changed after another mathematician’s intervention. Their evidence chains, people, and mathematical content remain separate. This avoids creating a generic story in which “correspondents corrected the manuscript” and preserves Brown’s and Illusie’s distinct documented roles.3333Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 43–49; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 134, 490; Maronne and Toën, Le temps des réflexions, 21.

Brown’s role is visible because the correspondence preserves response as well as gratitude. The May 25 letter does more than acknowledge a parcel: it supplies objections and references to the withdrawn assertion. The June 14 letter then attributes corrections of misconceptions and patient explanation to Brown and records a changed understanding after months of reflection. That chain supports influence at named points. It does not make Brown coauthor of the manuscript or imply that all intervening revisions came from him.3434Grothendieck and Brown, Correspondance Grothendieck–Brown, 44–49.

Illusie enters through another object and another date. The exhibition identifies the September 7 manuscript page and its report of the telephone conversation; page 490 anchors the entry, and September 26 marks the return to the interrupted line. Unlike Brown’s May–June exchange, no packet acknowledgment or written counterexample is the inspected witness. The common feature is correction during composition. The documentary routes remain separate.3535Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 490; Maronne and Toën, Le temps des réflexions, 21.

Brown and Tim Porter accepted the proposed reader role in an April 21 reply and asked for the notes. Brown also explained why Quillen had not yet received the original letter: Quillen was then in Oxford, and the mail had not followed him. Brown showed him the correspondence, supplied copies, and reported Quillen’s promise to write. This account establishes a route by which Quillen saw copies; it does not establish a surviving reply or the exact state of the copy he read.3636Grothendieck and Brown, Correspondance Grothendieck–Brown, 30–31; Centre for Grothendieckian Studies, “Correspondence”.

On May 2 Grothendieck confirmed that he would send the first batch and explicitly sought comments about readability. Brown’s May 7 letter then discussed copying logistics and proposed Jean-Marc Cordier as another careful reader, while offering to make the copy at Bangor if Grothendieck wished. These letters show prospective readership being negotiated before Brown’s May 25 acknowledgment. They do not prove that Cordier received the material or define a general circulation list.3737Grothendieck and Brown, Correspondance Grothendieck–Brown, 31–33.

The first readers were asked to perform a bounded task: test whether the developing English typescript communicated its point and offer mathematical corrections. That differs from peer review of a submitted article and from later editorial preparation. Brown’s replies preserve both kinds of response—questions about readability and substantive objections—but no source says that he accepted responsibility for the manuscript as a whole.3838Grothendieck and Brown, Correspondance Grothendieck–Brown, 24–49.

The April and May numbers are not contradictory totals. On April 12 Grothendieck said he had written nearly two hundred pages; on May 25 Brown acknowledged receiving 191. The first is an authorial approximation of pages then written, while the second is an exact recipient report for one packet after separate omissions and supplements. Some written pages need not have entered that packet, and no inspected consultation proves identity. Keeping the predicates attached to their witnesses is more accurate than choosing one number as the manuscript’s length.3939Grothendieck and Brown, Correspondance Grothendieck–Brown, 24–45.

An architecture still open

The manuscript’s changing size is evidence of expansion, not of completed design. Brown acknowledged 191 pages on May 25. In September Grothendieck still projected chapters, an introduction, appendices, indexes, and eventual volumes, and he said that a part expected to take a month or two had consumed much longer. A received packet and a projected architecture answer different questions: the first fixes one transmitted state; the second records a plan still under construction.4040Grothendieck and Brown, Correspondance Grothendieck–Brown, 65–69.

The surviving result can be described without teaching the manuscript’s formal subjects. The work began from one foundational question, expanded into a long first part, accumulated digressions and corrections, and did not reach all the later parts announced by its title and plans. The archive exhibition characterizes the first part as incomplete and notes that projected continuations were not written in the intended form. That is a documentary claim about table of contents and manuscript state, not a judgment that the pages achieved nothing or that later mathematics simply completed them.4141Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 19–21.

For a general reader, the object resembles a laboratory notebook that grew into the draft of a treatise while experiments were still changing the plan. The analogy explains why dates, errors, and proposed chapters coexist. It stops before genre identity: Pursuing Stacks was intentionally composed prose, not raw notes, and its mathematical arguments demand proof-level assessment. A notebook analogy cannot decide which claims are correct, which sections form a book, or whether an announced continuation counts as part of the work.4242Grothendieck, Récoltes et Semailles, ed. Carmona (1985), 17; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 17–21.

The last dated page accordingly marks a manuscript state, not a finish line. Later correspondence can clarify an earlier page; a planned volume can remain unwritten; an editor decades later can publish selected sections without changing what existed in November 1983. The source-locked composition claim is that the architecture remained open.4343Grothendieck and Brown, Correspondance Grothendieck–Brown, 17–97; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 20–21, 33–34.

The January 8, 1984, letter supplies the only page-six-hundred statement used here. Grothendieck wrote that he had produced no further notes for nearly two months and joked about the boldness of assuming a reader remained “up to page 600.” The number is therefore authorial internal pagination visible at that date. It is not a facsimile leaf count, the extent of a transcription, the extent of the 2022 edition, or proof that the projected manuscript was complete.4444Grothendieck and Brown, Correspondance Grothendieck–Brown, 95–97.

Architecture may be described by documentary labels—part, chapter, introduction, appendix, index, projected volume—without treating the mathematical theories named there as completed. The labels show how Grothendieck was organizing a book and which components remained prospective. A table of contents is evidence of plan and sequence; it is not evidence that every listed section exists or that every existing section reached the same editorial state.4545Grothendieck and Brown, Correspondance Grothendieck–Brown, 1983, 65–69; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 19–21.

Page count and architecture likewise answer different questions. A packet total can grow while a plan contracts, and a projected volume can expand while remaining incomplete. The dossier therefore reports numbers only with a date and witness, then states whether they describe pages received, a conventional overall estimate, or a forecast. No number stands alone as “the length.”4646Grothendieck and Brown, Correspondance Grothendieck–Brown; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 33–34.

The surviving objects occupy three documentary tiers. Cote 134 names the original manuscript family at Montpellier. The edited Brown correspondence gives letters that date dispatch, receipt, correction, and authorial forecasts; it is not a transcription of every manuscript page. The Société Mathématique de France’s 2022 printed Volume I is a later edition prepared by Georges Maltsiniotis. An original leaf, a correspondence transcription, and a printed editorial volume therefore carry different pagination and cannot be counted as one witness.4747Grothendieck and Brown, Correspondance Grothendieck–Brown; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 134; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.

The official SMF notice defines the printed selection precisely: Volume I contains the first four chapters, sections 1–91 and 95–98. It assigns sections 92–94 and 99–140, the Breen letters, and related correspondence to a projected second volume. That division is evidence about the 2022 editorial state, not about what a reader held in May 1983 or about completion in November. It also explains why “Volume I” cannot be used as a synonym for either the 191-page packet or the entire archive object.4848Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.4949Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.

The same official notice describes the realized part of the project as The Modelizing Story or Histoire de Modèles and says that the manuscript comprises 140 sections. Its four-chapter Volume I crosses a nonconsecutive section boundary because sections 92–94 are reserved with later material. This editorial arrangement is direct evidence that section number, printed sequence, and projected volume are not equivalent measures. It also confirms that the 2022 editors treated the 1983 object as a structured but incomplete project rather than silently declaring the original plan fulfilled.5050Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.

Pagination shifts at every tier. “Page 490” is an internal manuscript locator used for the September entry; “page 600” is Grothendieck’s own January 8 reference to where a hypothetical reader might have reached; the Roman-numbered preface and new Arabic pagination in the SMF sample belong to the printed edition. None of these numbers converts mechanically into archive leaves. Their coexistence is not an error to average away but evidence of an original, an authorial working sequence, and a later editorial object.5151Grothendieck and Brown, Correspondance Grothendieck–Brown, 95–97; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 490; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.

The SMF sample also exposes editorial layers before the manuscript text: a preface, an introduction, section-by-section summaries, “Réflexions Mathématiques,” and a table of contents precede The Modelizing Story. Those materials belong to the 2022 edition and its editors. They help readers navigate the dossier but are not pages composed in the February–November 1983 logbook. The distinction is visible in the sample’s Roman and Arabic pagination.5252Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.

The title page fixes the printed object as Documents Mathématiques 20, edited by Georges Maltsiniotis and published by the Société Mathématique de France in 2022. Its rights page expressly reserves reproduction rights. The edition is inspected for editorial scope without treating its sample as permission to reproduce pages in this biography.5353Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.

The January 8 remark that no more notes had been written for nearly two months is consistent with a last dated manuscript checkpoint on November 12, but “nearly” prevents conversion into a second exact endpoint. The letter is valuable because it combines an authorial pause statement with the internal page-six-hundred reference. It does not say that all earlier pages belonged to one finalized volume or that scratchwork had stopped.5454Grothendieck and Brown, Correspondance Grothendieck–Brown, 95–97; Grothendieck, À la Poursuite des Champs (Pursuing Stacks).

Composition is not completion

The later edition is one documentary state among several: composition, receipt, correction, and editorial publication require different verbs.5555Grothendieck and Brown, Correspondance Grothendieck–Brown, 44–45; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 134; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.

The open state also blocks motive claims. An unfinished architecture does not prove abandonment, discouragement, withdrawal, or exhaustion. The dated record supports initiation, enlargement, correction, interruption of one line, resumption, and a last surviving date. Explanation beyond those events would require an explicit statement tied to the manuscript history. The chapter therefore refuses to turn a documentary gap into a private ending.5656Grothendieck and Brown, Correspondance Grothendieck–Brown, 17–97; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 19–21.

The title “Long Silence” can now be resolved without rhetoric. Silence describes the absence of ordinary new journal publication under Grothendieck’s name, not an absence of pages, correspondents, or correction. The composition dossier is positive evidence against the stronger claim and does not require speculation about why publication differed. The surviving state combines bibliographic quiet, documentary abundance, and an unfinished architecture.5757Grothendieck and Brown, Correspondance Grothendieck–Brown, 17–97; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 17–21.

This resolution is deliberately nonpsychological. “Silence” is a statement about a publication channel; “open” is a statement about manuscript state. Neither term describes Grothendieck’s feelings.5858Grothendieck and Brown, Correspondance Grothendieck–Brown, 17–97; Grothendieck, À la Poursuite des Champs (Pursuing Stacks); Maronne and Toën, Le temps des réflexions, 17–21.

The resulting endpoint is open in two different senses. In 1983 the author’s architecture still contained projected work; in 2022 the official first volume still selected only stated sections from a larger manuscript and assigned others to a later volume. These are not the same incompletion, and the later editorial one cannot prove the earlier authorial state. Read together with the dated letters, they show why composition, manuscript survival, transcription, and publication require separate chronologies.5959Grothendieck and Brown, Correspondance Grothendieck–Brown; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), 134; Grothendieck, À la Poursuite des Champs (Pursuing Stacks), vol. 1.

Brown’s December 12 letter adds one more recipient checkpoint: he apologized for not acknowledging pages 500–550 earlier. The range proves receipt of a numbered packet state after the November 12 manuscript endpoint; it does not date the composition of those pages or extend the logbook. This late acknowledgment is another reason recipient pagination and authorial chronology remain separate.6060Grothendieck and Brown, Correspondance Grothendieck–Brown, 93–94.

The door into the next decade

In January 1984, a CNRS application named the next project.