The Record

She had decided on the heading before she began: Como foi sabido: três fios, três pessoas, antes de se encontrarem. Her father would have approved of the formulation. He had taught her that a document's heading was not a summary of its contents but a statement of its purpose — what question it was answering, for whom, on what authority. The heading answered all three: how this came to be known, by whom, in what sequence. The question mark beside the Távora node was on the previous page, where it belonged. This page had no question marks.

She began with the first thread.

The survey thread was the most precise and the most impersonal. She did not have the survey notebook — Diogo had reclaimed it for the deposit he had mentioned, though he had not told her the date — but she had the analysis sheet, copied in his hand, and she had his account of the six plots as he had given it to her in the room in June. She wrote from those:

Fio I — Registo de levantamento (D.M., engenheiro de obras)

Julho–Setembro de 1755: seis lotes adquiridos pela Companhia de Gestão e Terrenos Lisbonenses, Baixa central. Datas precisas: 3 de Julho (Rua dos Fanqueiros, parcela sul), 14 de Julho (Rua do Ouro, lado oeste), 22 de Agosto (Rua dos Ourives, lado oeste, quarta a partir do Terreiro — N.B.: a mesma parcela cujo nicho foi encontrado forçado por D.M. em Janeiro de 1756, ver infra), 9 de Setembro (Rua da Prata, dois lotes adjacentes), 22 de Setembro (Rua dos Correeiros, canto noroeste). Todas as escrituras corretas na forma. Nenhuma com escritura complementar sobrevivente.

She paused. The note about the niche was new — she had not written it in the diagram, because she had not known in August exactly how to characterise it. Now she had thought it through and knew: the forced niche was not additional evidence for the holding company's culpability; it was evidence that someone had removed the evidence before the earthquake could account for its absence. The strongbox, the deed, the Bristow acknowledgement — three removals, one window, one logic. She added a second note:

N.B. adicional: a escritura complementar (referência cruzada ao registo notarial de Outubro de 1755, entrada de 14 de Outubro) foi registada e depois removida antes de 1 de Novembro. Janela: 12–31 de Outubro. O espólio Bristow: nota de débito pessoal, mesma janela, ausente dos registos da firma após o terramoto. Dois instrumentos, um padrão de remoção, antes do terramoto fornecer cobertura.

That was accurate. She read it twice. She did not change it.

The financial thread was harder to write because most of it was not hers. She knew Henrique's payment — that was the instrument she held directly, the entry in the notebook, the four hundred and thirty mil-réis, the instruction not to enter it in the ledger. She knew the holding company registration through her own research at the commercial registry. The rest — the Factory ledger gap, the Bristow debt acknowledgement, the Jesuit network as conduit — she knew only through the merchant widow network, through Beatriz and Senhora Cardoso and Fonseca, at one or two removes from primary knowledge.

Her father's principle applied here with force: she could record what she had been told, sourced, as report rather than as instrument. She could not claim it as her own evidence. She wrote accordingly:

Fio II — Inteligência comercial (fontes: rede de mercadores, via B. Soares, A. Cardoso, advogado Fonseca)

Agosto de 1755: H.V. recebeu 430$000 réis da CGTEL — instrumento primário, entrada no caderno privado de H.V., com instrução: não lançar no livro. (Original retido.)

Registo comercial da CGTEL: Junho de 1755. Antes das aquisições dos lotes. (Verificado por T.V. no registo comercial municipal.)

Registo da Feitoria Britânica: entrada marginal, Julho de 1755, referência cruzada à escritura complementar, riscada a tinta diferente após Novembro de 1755. (Relatado por Fonseca via Cardoso; não verificado independentemente por T.V.)

Dívida pessoal, Secretário de Estado a J. Bristow (período de Londres): nota de reconhecimento retida nos registos da firma; ausente após o terramoto, janela não confirmada. (Relatado por Fonseca via Cardoso; não verificado independentemente por T.V.)

She stopped. The sourcing was careful — perhaps too careful, in the sense that it would frustrate a reader who wanted certainty. But her father had taught her that a notary's job was not to provide certainty where it did not exist, but to specify precisely the degree of certainty that did. She left it as it was.

The testimony thread was the shortest and the most damaged.

Fio III — Testemunho (Padre I. Pereira SJ, falecido ou preso — estado desconhecido)

Janeiro de 1756: P.I.P. procurou um cofre-forte numa casa comercial, Rua dos Ourives, lado oeste, quarta a partir do Terreiro — a mesma parcela identificada no Fio I. O cofre não estava. (Testemunho de P.I.P. a T.V., setembro de 1756; notas destruídas junho de 1758 após a sua prisão.)

Setembro de 1756: P.I.P. declarou que o cofre continha registos de transações financeiras entre um intermediário mercantil e uma casa nobre — transmitidos por via de confessores jesuítas. Não nomeou a casa. Declarou que o intermediário mercantil falecera. (Testemunho destruído; apenas esta paráfrase retida, de memória.)

P.I.P. preso em Junho de 1758 pelos homens do Secretário. Estado atual desconhecido por T.V.

She read this through and felt the thinness of it — the testimony destroyed, the source gone, the only remaining record a paraphrase from memory written in December 1758 of a conversation in September 1756. This was not what her father would have called an instrument. It was what remained when an instrument was taken from you.

She wrote one further line below the thread entries:

Nota: O testemunho foi destruído. O que persiste é a descrição de T.V. do que foi dito. A diferença é material e deve ser lida como tal.

On the second morning she began the copies. She had prepared two sets of materials: the original ledger, which contained the diagram and the timeline as drawn in August; and a new ledger — Henrique's other accounting book, the one she had used for the commercial registry research, largely blank in its second half — into which she copied everything by hand, in order: the timeline, the diagram, the three-thread section.

Copying was slower than composing. She had not expected this; she had expected the act of transcription to be mechanical, requiring less attention than the original writing. Instead she found that every entry required re-reading the original before she wrote it — not to check for errors, but because each entry had been placed where it was for a reason she had to hold in mind as she reproduced it. The sourcing notations in the financial thread required care; a misplaced attribution would change the epistemic weight of an entry, which was precisely the thing the notation was designed to prevent.

She finished the timeline and diagram on the second afternoon and the three-thread section that evening, by lamplight. She read the copy through against the original. It was accurate. She closed both ledgers.

On the third morning she wrapped the copy. She had oilcloth and cord from the kitchen, left over from a parcel she had sent to her sister-in-law in October. She wrapped the ledger in two layers, tied it with cord at both ends and across the middle, and cut a strip of paper from her father's notebook — the last blank strip at the back — on which she wrote an inventory in his notarial shorthand:

Conteúdo: (1) cronologia, Jun.1755–Dez.1755; (2) diagrama de relações, CGTEL e partes; (3) três fios: levantamento / financeiro / testemunho; fontes anotadas. Cópias de: linha de pagamento HV, registo comercial CGTEL. Original retido. T.V., Dezembro de 1758.

She folded the strip and tucked it inside the cord at the top of the package, where it would be found before the oilcloth was opened.

She put the package on the shelf beside the original ledger. Two ledgers now, where one had been, the gap between them narrower than the gap she kept between the ledger and her father's notebooks.

The original she would carry. The copy she would give to Diogo.

On the third evening she took the original from the shelf and held it for a moment without opening it. It was not a heavy book — Henrique's old ledger, the kind that traders used for medium-term accounts, bound in dark board, the spine unmarked. She had not put a title on it. She had not put her name on it. It was legible without either.

She put it back on the shelf between the copy and her father's notebooks. She would not take it down again tonight. She was not finished — there was the matter of the payment receipts, the commercial registration copy, the court filing, all of which needed to be sorted between what she would carry and what she would leave — but the ledger itself was done. The work she had announced to Diogo in August was complete.

She turned the lamp down. Through the window the upper city was dark, the Távora palace unlit for the third night running. She did not look at it for long. She was thinking about what she still needed to do before Thursday, and the list was not short, and the evening was not yet spent.

She sat back down at the table and opened her father's notebook to a fresh page.