The researcher arrived in Uvira in March. She had come for the report. The report would come out in May. In March it was still a thing to be done, and the thing to be done was this: to speak with people, one at a time, and to write down what they said.

The woman received her at home, in the front room, the one with the door onto the street. The door was wood, with an iron bolt that was drawn from inside. The researcher sat at the table. She opened a notebook. She set the notebook on the table and a pen beside the notebook. She said the woman could stop whenever she wished. She said she could decline to answer a question and move on to the next.

The woman offered something to drink. The researcher accepted. This was the beginning, and the beginning had to be done in this order.

Then the researcher started with the dates. The dates were fixed, she already had them from other interviews. The M23 forces and the Rwandan soldiers had entered Uvira on the tenth of December. They had remained until the seventeenth of January. Thirty-eight days. During those days, in the woman's neighborhood, fighters had gone house to house. They knocked. They asked about men and boys. They said they were looking for anyone with ties to the militias that had sided with the government.

The researcher explained how the report worked. It would be twenty-three pages. Behind the twenty-three pages were one hundred and twenty interviews, and the woman's was one of the one hundred and twenty. The report would count three things: people executed, women raped, people taken away. For each of the three things there would be a number.

The researcher had a method, and the method was always the same. First the large facts, the ones that do not change: the dates of the occupation, the units, the names of the commanders. Then the facts of the neighborhood: who had passed through which street, on which day. Then, only at the end, the facts of the house. You moved from the wide to the narrow, from the city to the room, and you arrived at the door last. The woman recognized that method without having studied it. She understood it from the order of the questions.

Then the researcher asked the woman to tell her about her night. Everyone had a night. The woman's night had been between the sixth and seventh of January.

The woman told it in objects. She said that at that hour the radio was on, at low volume, on a frequency that played only music. She said her husband had gotten out of bed. She said someone had knocked at the door three times. Three blows, a pause, and then nothing more. The husband had gone to the door barefoot. He had drawn the bolt himself. This the woman said precisely: the bolt had been drawn by him, from inside, with his own hand. Then she told about the street, the sound of the engine, the time she had read from a clock. She told everything that surrounded it. She left the center empty.

The researcher was writing. She wrote quickly. She skipped nothing. At a certain point she stopped. She said that for the report she needed one thing. She needed the man's name and the date. Without the name, she said, the man remained inside a number. The number, for the people taken away and never returned, was twelve. Every name written in the report took a man out of the number, placed him among the people with a name.

The woman did not answer at once.

Since January the woman had been cooking for one and a half. Not for two, because her husband was not at the table. Not for one, because saying one was something she had never done. It was a quantity that did not close the door. As long as she cooked for one and a half, her husband was a man who could still come home at night and knock. She would count the blows. She would know them.

Giving the name to the report was another thing. The name in the report stood in the row of the twelve people taken away and never returned. Never returned were two words already written, and the name went beneath them.

The researcher waited. The pen was still on the notebook. She did not press. She only waited, with the pen still, and that was her way of asking. She had done one hundred and nineteen interviews before this one. She knew that the name comes or it does not come, and that pushing does not help.

The woman said her husband's name. She said it in full, the given name and both family names. Then she said the date: the night between the sixth and seventh of January.

The researcher wrote the name. She wrote the date. She read back aloud what she had written, so the woman could confirm, and the woman confirmed. The researcher closed the notebook.

Then she stood. The woman walked her to the door. She drew the bolt, the same bolt, and opened the door. Outside it was March, it was afternoon, there was the full light of the street. The woman stood at the threshold until the researcher had reached the far end of the street. Then she went back inside. The door, that afternoon, she left open.