Somehow, I whipped the interview into shape. I was terrified of paraphrasing James Rourke; with most people, it was easy enough to judge what was fair and what wasn’t, but here I was on treacherous ground. I wasn’t even sure that the console could convincingly mimic him – when I tried it, the body language looked utterly wrong, as if the software’s default assumptions (normally used to flesh out an almost-complete gestural profile gleaned from the subject) were being pumped out in their entirety to fill the vacuum. I ended up altering nothing – merely extracting the best lines, and setting them up with other material – and resorting to narration, when there was no other way.

I had the console show me a diagram of the segments I’d used in the edited version, slivers scattered throughout the long linear sequence of the raw footage. Each take – each unbroken sequence of filming – was clearly “slated”: labeled with time and place, and a sample frame at the start and end. There were a few takes from which I’d extracted nothing at all; I played them through one last time, to be sure I hadn’t left out anything important.

There was some footage where Rourke was showing me into his “office” – a corner of the two-room flat. I’d noticed a photograph of him – probably in his early twenties – with a woman about the same age.

I asked who she was.

“My ex-wife.”

The couple stood on a crowded beach, somewhere Mediterranean-looking. They were holding hands and trying to face the camera – but they’d been caught out, unable to resist exchanging conspiratorial sideways glances. Sexually charged, but … knowing , too. If this wasn’t a portrait of intimacy, it was a very good imitation.

Sometimes we can even convince ourselves that nothing’s wrong. For a while.

“How long were you married?”

“Almost a year.”

I’d been curious, of course, but I hadn’t pressed him for details. Junk DNA was a science documentary, not some sleazy exposé; his private life was none of my business.

There was also an informal conversation I’d had with Rourke, the day after the interview. We’d been walking through the grounds of the university, just after I’d taken a few minutes’ footage of him at work – helping a computer scour the world’s Hindi-speaking networks in search of vowel shifts (which he usually did from home, but I’d been desperate for a change of backdrop, even if it meant distorting reality). The University of Manchester had eight separate campuses scattered throughout the city; we were in the newest, where the landscape architects had gone wild with engineered vegetation. Even the grass was impossibly lush and verdant; for the first few seconds, even to me, the shot looked like a badly forged composite: sky filmed in England, ground filmed in Brunei.