"His questions aren't academic after all, Deet."

"I told you they weren't."

"Hari was right. For all that he seemed to be a dilettante, with his money and his rejection of the universities, he's a man of substance."

"Will the Second Foundation benefit, then, if he comes up with an answer to his question?"

"I don't know, Deet. Hari was the fortune-teller. Presumably mankind is already human, so it isn't as if we have to start the process over." "Do you think not?"

"What, should we find some uninhabited planet and put some newborns on it and let them grow up feral, and then come back in a thousand years and try to turn them human?"

"I have a better idea. Let's take ten thousand worlds filled with people who live their lives like animals, always hungry, always quick with their teeth and their claws, and let's strip away the veneer of civilization to expose to them what they really are. And then, when they see themselves clearly, let's come back and teach them how to be really human this time, instead of only having bits and flashes of humanity."

"All right. Let's do that."

"I knew you'd see it my way."

"Just make sure your husband finds out how the trick is done. Then we have all the time in the world to set it up and pull it off."

When the index was done, Deet brought Leyel with her to the library when she went to work in the morning. She did not take him to Indexing but rather installed him in a private research room lined with vids-- only instead of giving the illusion of windows looking out onto an outside scene, the screens filled all the walls from floor to ceiling, so it seemed that he was on a pinnacle high above the scene, without walls or even a railing to keep him from falling off. It gave him flashes of vertigo when he looked around-- only the door broke the illusion. For a moment he thought of asking for a different room. But then he remembered Indexing, and realized that maybe he'd do better work if he too felt a bit off balance all the time.

At first the indexing seemed obvious. He brought the first page of his questions to the lector display and began to read. The lector would track his pupils, so that whenever he paused to gaze at a word, other references would begin to pop up in the space beside the page he was reading. Then he'd glance at one of the referenees. When it was uninteresting or obvious, he'd skip to the next reference, and the first one would slide back on the display, out of the way, but still there if he changed his mind and wanted it.

If a reference engaged him, then when he reached the last line of the part of it on display, it would expand to full-page size and slide over to stand in front of the main text. Then, if this new material had been indexed, it would trigger new references, and so on, leading him farther and farther away from the original document until he finally decided to go back and pick up where he left off. So far, this was what any index could be expected to do. It was only as he moved farther into reading his own questions that he began to realize the quirkiness of this index. Usually, index references were tied to important words, so that if you just wanted to stop and think without bringing up a bunch of references you didn't want, all you had to do was keep your gaze focused in an area of placeholder words, empty phrases like "If this were all that could be..." Anyone who made it a habit to read indexed works soon learned this trick and used it till it became reflex.

But when Leyel stopped on such empty phrases, references came up anyway. And instead of having a clear relationship to the text, sometimes the references were perverse or comic or argumentative. For instance, he paused in the middle of reading his argument that archaeological searches for "primitiveness" were useless in the search for origins because all "primitive" cultures represented a decline from a star-going culture. He had written the phrase "All this primitivism is useful only because it predicts what we might become if we're careless and don't preserve our fragile links with civilization." By habit his eyes focused on the empty words "what we might become if." Nobody could index a phrase like that.

Yet they had. Several references appeared. And so instead of staying within his reverie, he was distracted, drawn to what the indexers had tied to such an absurd phrase.

One of the references was a nursery rhyme that he had forgotten he knew:

Wrinkly Grandma Posey

Rockets all are rosy.

Lift off, drift off,

All fall down.

Why in the world had the indexer put that in? The first thought that came to Leyel's mind was himself and some of the servants' children, holding hands and walking in a circle, round and round till they came to the last words, whereupon they threw themselves to the ground and laughed insanely. The sort of game that only little children could possibly think was fun.

Since his eyes lingered on the poem, it moved to the main document display and new references appeared. One was a scholarly article on the evolution of the poem, speculating that it might have arisen during the early days of starflight on the planet of origin, when rockets may have been used to escape from a planet's gravity well. Was that why this poem had been indexed to his article? Because it was tied to the planet of origin?

No, that was too obvious. Another article about the poem was more helpful. It rejected the early-days-of-rockets idea, because the earliest versions of the poem never used the word "rocket." The oldest extant version went like this: Wrinkle down a rosy,

Pock-a fock-a posy,

Lash us, dash us,

All fall down.

Obviously, said the commentator, these were mostly nonsense words-- the later versions had arisen because children had insisted on trying to make sense of them.

And it occurred to Leyel that perhaps this was why the indexer had linked this poem to his phrase-- because the poem had once been nonsense, but we insisted on making sense out of it.

Was this a comment on Leyel's whole search for origins? Did the indexer think it was useless?

No-- the poem had been tied to the empty phrase "what we might become if." Maybe the indexer was saying that human beings are like this poem-- our lives make no sense, but we insist on making sense out of them. Didn't Deet say something like that once, when she was talking about the role of storytelling in community formation? The universe resists causality, she said. But human intelligence demands it. So we tell stories to impose causal relationships among the unconnected events of the world around us.

That includes ourselves, doesn't it? Our own lives are nonsense, but we impose a story on them, we sort our memories into cause-and-effect chains, forcing them to make sense even though they don't. Then we take the sum of our stories and call it our "self." This poem shows us the process-- from randomness to meaning

- and then we think our meanings are "true." But somehow all the children had come to agree on the new version of the poem. By the year 2000 G.E., only the final and current version existed in all the worlds, and it had remained constant ever since. How was it that all the children on every world came to agree on the same version? How did the change spread? Did ten thousand kids on ten thousand worlds happen to make up the same changes?

It had to be word of mouth. Some kid somewhere made a few changes, and his version spread. A few years, and all the children in his neighborhood use the new version, and then all the kids in his city, on his planet. It could happen very quickly, in fact, because each generation of children lasts only a few years-- seven-year-olds might take the new version as a joke, but repeat it often enough that five-year-olds think it's the true version of the poem, and within a few years there's nobody left among the children who remembers the old way. A thousand years is long enough for the new version of the poem to spread. Or for five or a dozen new versions to collide and get absorbed into each other and then spread back, changed, to worlds that had revised the poem once or twice already.