“I think subsequent meals would be best,” I said.

“If we could only do it in the middle of the afternoon. But we need a family reaction…” She tapped her pencil thoughtfully on the edge of the stainless-steel sink.

“Yes, well,” I said, “I’d better be getting back.” Deciding for them what they wanted to know wasn’t part of my job.

Sometimes I wonder just which things are part of my job, especially when I find myself calling up garage mechanics to ask them about their pistons and gaskets or handing out pretzels to suspicious old ladies on street corners. I know what Seymour Surveys hired me as – I’m supposed to spend my time revising the questionnaires, turning the convoluted and overly-subtle prose of the psychologists who write them into simple questions which can be understood by the people who ask them as well as the people who answer them. A question like “In what percentile would you place the visual impact value?” is not useful. When I got the job after graduation I considered myself lucky – it was better than many – but after four months its limits are still vaguely defined.

At times I’m certain I’m being groomed for something higher up, but as I have only hazy notions of the organizational structure of Seymour Surveys I can’t imagine what. The company is layered like an ice-cream sandwich, with three floors: the upper crust, the lower crust, and our department, the gooey layer in the middle. On the floor above are the executives and the psychologists – referred to as the men upstairs, since they are all men – who arrange things with the clients; I’ve caught glimpses of their offices, which have carpets and expensive furniture and silk-screen reprints of Group of Seven paintings on the walls. Below us are the machines – mimeo machines, I.B.M.