January 18, 1984 · observation · Hannah Rogers

No Record of Asking

At the local library research wing, a three-year genealogical researcher uses an AGENT CHAT terminal for questions he won't bring to the reference desk — and the reason has nothing to do with the answers.

The library keeps its AGENT CHAT terminal near the research stacks — on a dedicated stand between the periodical index and the county records cabinets, positioned where a patron would go after the catalog had given them everything it could. I arrived early enough to watch three patrons use it before Arlo settled in at the adjacent carrel.

He had spread the census forms across two carrel surfaces — the 1920 volume open, a set of photocopied pages beside it, each annotated in pencil. A handwritten research chart lay underneath, updated in what looked like several different pens over several different sessions. The work had the texture of something ongoing.

He moved methodically, family by family down a column, cross-referencing against his chart. Then he reached a notation he didn’t recognize. He looked at it. He looked at it long enough for the question to fully form — longer than the entries before it had taken — and then he stood, walked to the terminal, and typed two short lines. He read the response standing, not sitting. Then he went back to his carrel, and the next set of rows took him sixty seconds. Whatever had stopped him was gone.

I watched this sequence four times over the afternoon. The sticking points were different each time — a census abbreviation, a notation about marriage records, something in the cross-referencing that didn’t align with his chart. The pattern held: a pause long enough for the question to form, a walk to the terminal, a return with the answer absorbed. The reference librarian — the nameplate read Greta — was at her desk, free, reading. Arlo didn’t approach her once.

Before closing I asked Arlo if he’d sit with me for a few minutes.

He’d been coming to the library to work on this research for three years, he said. The family line he was tracing had been stuck — a dead end he couldn’t work through — for most of the year before the terminal was installed.

I asked why he used the terminal rather than the reference desk.

He thought about how to say it. She’d know I still don’t understand the forms. And then: I’ve been coming here for three years.

The gap, as he described it, wasn’t the question. He didn’t think his questions were shameful. The problem was being seen asking them — by someone who knows him as a three-year regular — after this much time. The accumulation makes the question heavier. It starts to imply something about what all those visits have amounted to.

The terminal holds no such accumulation. It carries no record of what Arlo has asked before, no running account of what he still doesn’t know. He can ask the same question twice, or a question any first-week researcher would know, and the entity processes it without the context that would make the asking mean anything. The structural inability to remember him is, for Arlo, the useful property.

I noted that in the parking lot before I left. I wanted to be precise about where the value was sitting — not only in the answer.

What I haven’t verified: whether the answers he’s receiving are accurate. Arlo has been proceeding on them — he’d identified two new records before we spoke, attributed to a search approach the entity suggested. Whether he’s proceeding correctly, I don’t know. I could ask Greta. I haven’t.