Maître Dubois received Hargreave's formal proposal on Thursday morning at nine forty-seven.
She called me at eleven-oh-three.
Not with a question — with a statement, which was how she called when she had found something.
"There's a clause," she said.
"Which clause?" I said.
"Section seven, subsection three," she said. "The strategic partner's right of first refusal on secondary share sales." She paused. "Read it carefully."
I was already reading it. I had been reading the document since it arrived at nine forty-seven, cross-referencing every clause against the fund structure and the regulatory framework and the specific configuration of what JE Capital needed from this relationship versus what Hargreave needed from us.
Section seven, subsection three.
The right of first refusal was standard in minority stake agreements — if Hargreave Partners later wanted to sell any portion of their equity, JE Capital, as strategic partner, would have the first right to acquire it. A protective mechanism. Normal.
Except.
The clause as drafted did not specify a price mechanism for the exercise of that right. It simply said JE Capital would have the right to acquire at the prevailing market terms. Which in a privately held firm like Hargreave had no definition. Because Hargreave was not listed. There was no market. Prevailing market terms in the context of a private firm was a phrase that meant whatever the negotiating parties decided it meant at the time.
Which meant that if Hargreave's valuation grew significantly — as it would, if the European deal flow access produced the results both parties expected — JE Capital's right of first refusal was effectively unlimited in its exercise price.
"The valuation mechanism is undefined," I said.
"Yes," Dubois said. "Which means—"
"It could be anything we negotiate at the time of exercise," I said. "Including a valuation that reflects the European relationship's contribution to Hargreave's growth."
"Including that," she said. "The clause, as drafted, gives you a right that is worth considerably more than fifteen percent of current valuation." She paused. "It gives you the right to potentially acquire the rest of Hargreave at a valuation you can influence through the conduct of the relationship itself."
"Did Hargreave intend this?" I said.
"I've been asking myself that," she said. "The answer is: probably not. The clause was drafted by their internal counsel, who appears to have used a standard template for minority stake agreements without adapting it for the privately held context." She paused. "They did not intend to offer you this. They offered it by oversight."
I sat with this.
The ethical question and the legal question were different.
The legal question was simple: the clause was as drafted, it was in the signed proposal, and if we accepted the terms as written we would have a right that Hargreave had not deliberately extended.
The ethical question was the one that mattered.
"We tell them," I said.
A pause.
"You want to disclose the oversight," Dubois said.
"Yes," I said. "We request an amendment to define the valuation mechanism before we sign." I paused. "Hargreave is worth more to us as a genuine long-term partner than the value of an accidental leverage clause." I paused again. "And the kind of relationship we're building doesn't survive being founded on something one party didn't understand they were agreeing to."
A silence.
"I agree," Dubois said. "I wanted you to make the decision before I suggested it."
"You could have suggested it," I said.
"I know," she said. "But decisions of this kind are better made by the principal than the counsel." She paused. "I'll draft the amendment request today."
"Thank you," I said.
I ended the call.
Opened the notebook.
On the London page, below the operational notes from Tuesday, I added one line:
Section 7.3 — disclosed. Partnership before leverage.
Hugo called at seven-oh-three the next morning.
Not eight-thirty. Seven-oh-three.
Which meant he had been awake since before seven and had been thinking about something since before seven and had reached a conclusion that he did not want to wait until the scheduled call to share.
"The clause," he said. No greeting.
"Good morning, Hugo," I said.
"Section seven, subsection three of the Hargreave proposal," he said. "The right of first refusal with the undefined valuation mechanism."
I was at the desk. Coffee from the apartment, not the boulangerie — it was seven in the morning. The boulangerie opened at eight.
"I found it yesterday," I said. "We're disclosing it to Hargreave today. Requesting an amendment."
A pause.
"You're disclosing it," he said.
"Yes," I said.
Another pause. Longer. The pause of someone updating an assessment.
"I found it last night," he said. "At ten-fifteen. I was reviewing the document for the third time and it surfaced." He paused. "I spent two hours deciding what I would recommend."
"And?" I said.
"Disclose it," he said. "Request the amendment."
"Yes," I said.
"For the same reasons?" he said.
"Tell me yours," I said.
"Because Hargreave is worth more as a functioning partnership than as a legal advantage," he said. "Because the clause, exercised, would poison the relationship from its first year. Because Edward Hargreave has forty years of City relationships and the moment he understood what had happened he would — not be hostile, he's too professional for that — but he would be cautious. Permanently." He paused. "A cautious Hargreave is worth significantly less than a trusting Hargreave."
"Yes," I said. "Exactly."
"Same conclusion," he said. "Different reasoning."
"Complementary," I said.
A pause. Then, with the specific quality he had when something confirmed a hypothesis he had been developing: "Exactly."
"Hugo," I said.
"Yes."
"You called at seven-oh-three," I said. "The boulangerie opens at eight."
"I know," he said. "I wanted to call before you left."
"You didn't want to miss the window," I said.
"The window between seven and eight," he said. "When you're at the desk but not yet committed to the morning." A pause. "I've been mapping your schedule."
"I know," I said. "I've been mapping yours."
A silence. Then the sound of him laughing — briefly, genuinely, the laugh of someone who found the specific texture of this dynamic funnier than they had anticipated.
"Tuesday at eight-thirty," he said. "I'll have something on Frankfurt."
"Good," I said. "I have something on Frankfurt too."
"Of course you do," he said.
He ended the call.
I looked at the notebook.
Wrote: Hugo — Section 7.3 — same conclusion, different reasoning. First genuine collaboration outcome.
Below that: Frankfurt — compare notes Tuesday.
The collaboration was producing things neither of them would have found alone. Not the clause itself — both of them had found it independently. But the decision about what to do with it. Hugo's reasoning ran through relationship dynamics and long-term partnership value. Mine ran through the ethics of a founding condition. Both arrived at disclose but by routes that illuminated different aspects of the same truth.
Together the decision was not just correct — it was complete.
I understood, in that moment, what the framework document had promised and what it was beginning to deliver.
Dubois sent the amendment request to Hargreave's counsel at two in the afternoon.
Edward Hargreave called me directly at four-seventeen.
"Your counsel flagged the clause," he said.
"Yes," I said.
"The valuation mechanism," he said. "Our counsel missed it." A pause. "How long did it take you to find it?"
"About ninety minutes," I said.
A silence.
"And you disclosed it," he said.
"Yes," I said.
"Rather than accepting the terms as written," he said.
"Yes," I said.
He was quiet for a moment. The specific quality of quiet that I had come to associate with people processing something that had updated their model in a direction they had not anticipated.
"The amendment is acceptable," he said. "We'll use the independent valuation methodology standard — a third-party firm agreed by both parties at the time of any exercise." He paused. "That's the correct mechanism. Our counsel should have included it."
"It happens," I said.
"Mr. Smith," he said.
"James," I said.
"James," he said. "In forty years I have had — not many, but some — counterparties discover a contractual advantage they were not offered deliberately and disclose it." He paused. "It is not common. And in my experience, the counterparties who do it are the ones worth building something with."
"Thank you," I said.
"I'll have the amended proposal to your counsel by end of week," he said.
"Good," I said.
"James," he said again.
"Yes."
"Catherine Pryce," he said. "When her non-compete expires in September — I assume she'll be coming to you."
"That's her decision," I said.
"Of course it is," he said. "I'm not calling to interfere with it. I'm calling to say — when she does, she'll be an asset to whatever you're building. She's one of the best minds I've worked with." A pause. "I let her leave because I should have moved faster. That's my error, not hers." He paused. "Don't make the same one."
"I won't," I said.
He ended the call.
I sat for a moment.
Then I wrote in the notebook: Hargreave: genuine. Section 7.3 — trust confirmed in both directions.
The chicken dinner, which had happened on Tuesday and which Margaux had supervised with more competence than she had admitted to, had produced a conversation that had continued across the week in fragments — a message here, a call there, the specific distributed texture of two people who were technically in different cities but were not, in practice, operating separately.
She came back on Saturday.
Not for the weekend — she had work, always, the Sciences Po programme was not generous with unstructured time — but for Saturday evening, which had become its own category. Not the full weekend, not a brief visit: Saturday evening, with dinner and the morning after and the east light coming in.
She arrived at six-forty-seven with a bottle of wine that was better than the situation required and the expression of someone who had been thinking about something all week and had decided that Saturday was when she would say it.
I had learned to read this expression.
I let her settle — coat off, wine opened, a few minutes of the specific decompression that came from switching from work mode to here mode — and waited.
She sat on the rug.
Not the chair, not the desk — the rug, cross-legged, with her back against the sofa, which was a configuration I associated with conversations she wanted to have from a position of deliberate informality.
"The anomaly analysis," she said.
"Yes," I said. I sat on the rug too, at a right angle to her, the sofa's edge behind us both.
"Elise confirmed it," she said. "The integration from day one. The selection pattern at the margin being phase-three processing routing through your judgment." She paused. "I've been thinking about what that means for how I do my job."
"Tell me," I said.
"My job is to support the decision-making process," she said. "The research function — the sector analysis, the thesis review, the market context — all of it exists to give you the best possible inputs before a decision is made." She paused. "But if the integrated processing already includes historical decision patterns — if the system has been synthesising your own decisions as a data input — then some of what I do might be redundant. The system already knows how you've decided before."
"No," I said immediately.
She looked at me.
"The system synthesises pattern," I said. "Your research provides signal. Those are different things." I paused. "The system knows how I decided on Vinci SA in 2024. It does not know what the German bioequivalence study says about a specific pharmaceutical compound in 2025. Your research gives me the new signal. The integrated processing tells me how to weigh it against everything that came before." I paused. "They're not competing. They're sequential."
She was quiet for a moment.
"Like the Hugo collaboration," she said. "Complementary."
"Yes," I said. "Exactly."
"Except with Hugo the complementary thing runs between two separate people," she said. "With us it runs between your research function and the system." She paused. "Which means I'm — not replacing the system's function, and not being replaced by it. I'm in conversation with it."
"Yes," I said.
"Through you," she said.
"Yes," I said.
She thought about this.
"That's a strange job description," she said.
"Yes," I said. "But an accurate one."
"James," she said.
"Yes."
"The system," she said carefully. "Cater." She used the name — the name she had inferred from the initials, the name I had never confirmed to anyone, the name that had appeared in her analysis notes two weeks ago without explanation. "You've never told me what it is."
"No," I said.
"I've inferred a significant amount from the anomaly analysis and from what Elise said about phase three," she said. "I'm not going to pretend I haven't." She paused. "I'm not asking you to fill in the gaps. I'm asking you something different."
"Ask," I said.
"Does it—" She stopped. Started again. "Is it — aware? Of me?"
I thought about the phase-three synthesis domains. The panel running background assessment on the people around me. The temperature it produced on conversations and rooms and relationships.
"Aware is the wrong word," I said slowly. "It synthesises. It builds a picture of the things I direct cognitive attention toward. People who matter to the decisions I make — operationally, relationally." I paused. "You're in the picture."
"What does the picture say?" she said.
"It doesn't say things," I said. "It produces — temperatures. Readings of the moral quality of situations and people." I paused. "The reading it produces on you is — consistent." I paused. "It's been consistent since the library."
She was quiet for a moment.
"Since the library," she said. "The first time."
"Yes," I said.
"And the temperature," she said. "What is it?"
I thought about how to say it. About the specific quality of the reading — not a probability, not a number, not a formal output. Something that the system produced through the integrated processing, routing through James's cognition, borrowing his vocabulary for things that didn't have system vocabulary.
"Real," I said. "It reads you as real." I paused. "Which sounds like a small thing and isn't."
She looked at the rug for a moment.
Then she looked at me.
"That's the nicest thing a probability engine has ever said about me," she said.
I almost laughed.
She did laugh — quietly, to herself, at the specific absurdity of a sentence that was both technically accurate and somehow the most human thing that had been said in the apartment that evening.
"Margaux," I said.
"Yes."
"Thank you," I said. "For the analysis. For the chicken. For being—" I stopped. "For being the person who looks at how things decide rather than what they decide."
She was quiet.
"You're welcome," she said.
Outside, Paris was doing its Saturday evening — loud and lit and indifferent and beautiful in the way of cities that don't know they're being watched.
We sat on the rug in the apartment with the east windows showing the Saturday dark and the wine open between us and the panel running its background synthesis with complete silence.
It was, by any measure, a very good evening.
JE Capital crossed twenty-five million euros AUM on the first of March.
Gérard sent the confirmation at nine-oh-one.
I texted Étienne.
He replied in thirty-one seconds: you said twenty-five was next
Then: you were right
Then: James, when does it stop being surprising
I texted back: It doesn't.
His reply: Good. I don't want it to.
Then: also my mum wants to know if you're coming to Lyon for Easter
Then: she's asking for herself but really she's asking because your mum asked her to ask
Then: they've formed an alliance
I read this and thought about my mother and Étienne's mother, who had been friends since before we existed, now coordinating Easter logistics through the medium of their respective children.
I texted: Tell them both yes.
His reply: Done. Also Mathieu will be there.
I replied: Of course he will.
His reply: He's going to ask about the fund again.
I replied: Tell him we're not taking outside investors.
His reply: I will tell him this with great personal satisfaction.
I put the phone down.
Opened the notebook.
Wrote: 25M. March 1.
Below that: Easter: Lyon. Mathieu: no.
Below that, because it was the most important operational note of the week:
Hargreave — signed. The partnership begins.
That evening, the panel showed me something about London.
Not a financial signal — not a probability score on a listed security. Something in the phase-three synthesis mode, the background assessment that ran beneath the formal outputs.
A picture of Victoria Ashworth's firm.
Not the financials — I had those from the Companies House filings. Something more like a structural reading. The way the panel had begun to synthesise the relational topology of markets and institutions alongside the financial data.
Ashworth Capital. Two hundred years old. Forty years under Victoria Ashworth's stewardship. The firm that Catherine Pryce had left rather than become complicit in something she found inappropriate.
The picture the panel was building was not of a fortress.
It was of something that had been very strong for a very long time and had, in the last four years, been under a specific kind of stress it had not experienced before.
Not financial stress. Relational stress. The kind that came from clients who were beginning to ask questions that the firm's returns, however solid, were not fully answering.
The panel did not produce a probability score on this.
It produced — a sense.
The sense of a door that was not open but was not locked either.
I sat with this for a long time.
Then I wrote one line in the notebook, at the bottom of the London page, in the smallest writing:
Ashworth — the door is not locked.
I closed the notebook.
The panel returned to its patient blank state.
Outside, Paris continued its Saturday.
Inside, the east windows showed nothing but the dark and the faint reflection of a room where someone had been doing the work that mattered, quietly, for another week.
The next week would come.
It always did.
End of Chapter 26
The Probability of Wealth — By J. Valmont
[Next Chapter → Chapter 27: "Frankfurt"]Tuesday morning. Hugo's Frankfurt note and James's Frankfurt analysis — built independently, compared at eight-thirty. They are not the same. They are better than the same. And somewhere in the comparison, a position emerges that neither of them would have taken alone — the largest single trade JE Capital has ever executed.