Born Anne-Marie Selve in 1976, to a provincial couple already in middle age, Anna had formed herself early. Academically focused at eight, she had been obsessive by fourteen. It was a familar story. Arriving at Girton a year ahead of her cohort she allowed a further year to pass before succumbing to anorexia. Self-harm and her first suicide attempt followed. By then, the parents – never much more than pleasantly surprised to find themselves parents in the first place – were too old to offer emotional help; in addition there remained, according to psychiatric reports, some unidentified tension between father and daughter. Girton patched Anna together. For a time she was, as she put it, everyone’s favourite suicide. ‘They knocked on my door if they thought things were too quiet.’ But soon the Selves’ place in her life was taken by a visiting professor of mathematical physics. This man, Michael Kearney, uncommunicative, narcissistic and easily depressed, turned out to have his own problems. They were married fast and divorced even faster; yet, sustained perhaps by its fiercely mutual structures of manipulation, the relationship proved more durable than either of them believed, lurching along in its disordered way until Kearney ended it on the eve of the Millennium by walking into the Atlantic just north of Scituate, Massachusetts.

At this remove, the mathematician became unidentifiable. He had no family Helen Alpert could trace; while Anna claimed to have ‘forgotten everything’, and wouldn’t be certain about his age or even the colour of his eyes. When she could be persuaded to speak about Kearney, he was transformed into careful fiction. Vague one day, meaninglessly particular the next, Anna’s revision of Kearney presented him as a gap in her life even as he had filled it.

Publicly there was a little more. He had written, probably as a joke, a pamphlet on randomness and the Tarot. Some topological speculations – stimulated by exchanges with the reclusive mathematician Grigori Perelman – had been published a year or two before his death, to cautious peer approval. Otherwise Michael Kearney’s contribution to science lay in an unfinished quantum computing project, most of the work on which had been done by an unassuming experimental physicist called Brian Tate. Tate – newly divorced, unequal even to the brief publicity surrounding Kearney’s suicide, and wrong-footed by a minor funding scandal involving the venture capital firm MVC-Kaplan – went down with the ship. His results proved unrepeatable. With his collaborator dead, and his claim to have coaxed massive parallel processing from a train of cheaply modified desktop PCs dismissed as junk science, he faded from view in a month. All this was a matter of record.