PROLOGUE
BEFORE THE FIRST MOVE
Professor Elias Aldridge disliked forms that pretended to ask permission after a decision had already been made.
The one on his screen was titled PRELIMINARY CANDIDATE IDENTIFICATION — VOLUNTARY FACULTY REFERRAL.
It arrived six days before the autumn term.
The sender was Sebastian Cross.
Aldridge had not spoken to Cross in nearly a year.
That alone was reason enough to leave the message unanswered.
Instead, he read it twice.
Cross had once been the most promising research assistant in Aldridge’s lab: quick with models, patient with data, impatient with people. The difficulty had never been whether Cross understood ethical rules. He understood them perfectly. He simply believed a sufficiently interesting result should be allowed to negotiate with them.
Aldridge did not.
That disagreement had ended their work together.
Now Cross was attached to the Cairn Institute and proposing a future, fully reviewed study on adversarial reasoning. He wanted names of students who might eventually be invited to volunteer.
No data or contact was emphasized in the message. Names only for preliminary planning.
Aldridge should have deleted it.
Instead, he thought of his new seminar roster.
Twenty-seven names.
Four stood out for reasons that had little to do with grades.
Dev Sharma had submitted an entrance essay arguing that good institutions were not those that eliminated individual discretion, but those that made discretion accountable. The argument was polished enough to sound settled. Aldridge suspected the student behind it was not.
Ryan Vaughn had won a regional negotiation competition after persuading the judges that the problem’s stated objective was inferior to an unstated one. His evaluator described him as “exceptionally persuasive and occasionally intolerable.” Aldridge considered that promising.
Theo Hart’s academic file was less orderly. His test scores were excellent when he bothered to sit the tests. His attendance history looked like weather. Three different instructors used the phrase distractingly perceptive in comments that were not meant to be compliments.
Jun Shen’s mathematics record was extraordinary. His introductory computing professor had written, “He does not accept the question until he has checked whether the question deserves to exist.”
Aldridge typed four names into the reply.
Then stopped.
He removed them.
He wrote instead:
I HAVE SEVERAL STUDENTS WHO MAY FIT. IF THE STUDY RECEIVES APPROVAL, SEND ME THE FINAL RECRUITMENT MATERIALS AND I WILL DECIDE WHETHER TO FORWARD THEM.
He sent the message.
Thirty-one seconds later, Cross replied.
UNDERSTOOD. OUT OF CURIOSITY, ARE SHARMA, VAUGHN, HART, OR SHEN IN YOUR SEMINAR THIS TERM?
Aldridge stared at the screen.
He had not typed the names.
He checked his original message even though he knew he had not.
Then he checked the seminar roster permissions. Faculty course lists were accessible through normal university systems to approved research partners only after term enrollment locked. Enrollment had locked that morning.
There were innocent explanations.
Cross could have heard the names elsewhere. All four students had drawn attention in different departments. Cairn had university access. None of it required conspiracy.
Aldridge replied with one sentence.
WHY THOSE FOUR?
No answer came that day.
The next morning Cross sent a formal recruitment-planning sheet with the four names already entered. His explanation was bland: public academic records, competition results, faculty recommendations. Aldridge still disliked it. He nevertheless confirmed one fact, that all four were enrolled in his seminar and wrote beneath it: NO CONTACT WITHOUT APPROVED MATERIALS.
It seemed, at the time, like a boundary.
On Monday morning, Aldridge entered Wren Hall and found four blue index cards on the lectern.
Blank.
The cards were the same blue stock kept in the department supply cupboard. There was probably nothing remarkable about them. Still, Aldridge slipped them into his briefcase rather than leave them on the lectern.
By the first Thursday of the term, Aldridge had found an inconsistency in the Calder scoring sheets. By Tuesday evening, a second discrepancy had convinced him it was deliberate. He scheduled a formal audit for Friday morning, before the fellowship finalists could be notified.
On the fourth Thursday of term, before his seminar arrived, Aldridge stood alone at the board.
He thought about bad evidence.
He thought about good motives.
He thought about the four students who had not yet met one another in any meaningful sense.
Then he wrote on the board:
THE MOST DANGEROUS LIE IS THE ONE THAT IMPROVES THE TRUTH.
Across the city, in an office Aldridge had never visited, a computer woke from sleep.
Four profile windows were already open.
No names appeared on the screen.
Only labels.
SYSTEM.
LEVERAGE.
BEHAVIOR.
UNCERTAINTY.
A cursor moved to a new field.
COHORT 4 — STATUS.
Someone typed a single word.
READY.
