Ye Wenjia

Li Kaichen's university classmate, working at a history museum, where she had done artifact authentication for 8 years. She was cautious, slow to draw conclusions—when the samples were insufficient, she'd rather say 'I can't be categorical.'

To Li Kaichen, she was one of the few people he could still trust, and still get hold of.

Her office sits at the far end of the corridor behind the exhibition floor, window facing north, the desk permanently covered with three or four half-finished comparisons.

Her work, stripped down, is comparison: how the same symbol is written across different periods and regions, which stroke was there originally and which was added by a later hand. It cannot be rushed, and should not be.

The habit does not make her easy company at the museum—people want a line they can put in a press release, and she gives them three qualifications. The trade-off is that what she does say carries weight.

She and Li Kaichen were university classmates. He was on the crime beat then and she lived in the library, two very different roads, though one thing about them was alike: neither was willing to let "probably" stand in for "is."

They have been out of touch for years. Her number, though, has never once changed.

Li Kaichen sent her screenshots of 4 square-block characters. She studied them for a long time, and her conclusion was cautious: these characters didn't look like Chinese, nor like Japanese, Korean, Vietnamese, Thai, Arabic, or any writing system she knew; their density, arrangement, and strokes were unlike all the scripts she'd ever seen.

Their center of gravity was mechanically equidistant, but to confirm the opening strokes, the closing strokes, and the stroke order, she'd need dynamic frames. She urged him to get some sleep first.

Opinion of her at the museum ran both ways. For a conclusion fit to print on a panel she was the slowest person to ask. To confirm whether something was wrong, everyone came to her.

Her standard was fixed: if the sample was insufficient she said so, evidence carried to the point it reached and stopped there, and nothing was filled in between. In 8 years the standard had never produced a correction, and it made her the hardest person in every planning meeting.

A colleague once said, out of her hearing, that she was too slow. She had heard the remark and let it stand—get one character wrong in authentication and everyone copying it is wrong for decades.

Three or four half-collated documents stayed open on her desk and were not cleared at the end of the day. Clearing them, she said, meant working out again tomorrow where her thinking had reached.

She and Li Kaichen were in different departments and took one general-education course together. They became actual friends after graduating—in those years he was still at the paper and would call her whenever he needed something checked.

After he left the paper to freelance the calls thinned out. Thinned, but never stopped; a few times a year, always.

She knew the kind of thing he was chasing, and knew he would only ever tell her half of it. She did not press. What he sent her was usually a screenshot, a clip, a few symbols, with one line attached: can you tell what this is?

Her answers came back more cautious than he wanted. He complained every time, and every time sent the next one through.