"Budapest used to be two cities, each with its own enclave, Buda on one side of the Danube, Pest on the other. The Pest wizards got involved in the '48 rising along with a lot of other people and most of them died.

"They didn't want to recruit foreigners, tried to build back by natural increase — there weren't as many mals back then — but it wasn't working, at least not fast enough for them. After a few decades they decided they could grow faster if they spotted mundaneborn wizards, stole them from their parents and brought them up in the enclave.

"That worked, but not very well; there weren't enough mundaneborn. So they decided to make more. One of the Pest wizards — it's only one in the stories — went around Budapest seducing mundane women. If the child turned out to be mundane he got brought up by his mother and her husband as theirs. If he turned out to be wise-born — supposedly they could spot them at two or three — they stole the child and raised it in Pest.

"I expect the women were pretty busy raising the kids, maybe not so eager to have more of their own, so after a few decades the enclave was more than half janissaries. The seniors and their own kids tried to keep control but that got harder and harder. Eventually the janissaries rebelled — according to some of the stories their dad helped them — and when it was all over everyone was dead. In some versions of the story there was one survivor who sealed the gates, gave up on wizardry and joined a monastery. We get told the story as an example of how not to run an enclave — sort of the opposite of yours.

"So that's why there is a lost enclave. Supposedly Pest is still there with sealed gates and full of treasure, including an anti-aging potion the seniors were working on that had something to do with setting off the rebellion, maybe the janissaries figured that if it worked the seniors would be running things forever. Father says you should never trust a historical account that makes a good enough story to have survived on its literary merit but it might be true, or at least the enclave might still be there; what we are interested in isn't treasure but space. If we could find the gates and open them we could link the two enclaves — we already have a tunnel under the river to an exit on the Pest side.

"My affinity is divining and I'm good at finding things."

"Wow. I'd be surprised if a war that left everyone dead didn't do some damage to the infrastructure, enough to take it out eventually if not right then, but it's not my area of expertise and it would definitely be super cool if you found it. Do you know how many other diviners have gone looking?"

"I don't know that any have — most diviners are more interested in the future than the past and it might not have occurred to them that if someone was going to discover it eventually they could divine that now. I think that's how my book finding works. Lots of other people have looked for it, mostly independents. It would be good for us but a lot better for them — we already have an enclave and they don't. A generation ago there was a project to find it by people in Buda but they didn't and nobody since has tried, at least that I know of."

Ellen has been listening attentively.

"If divining is acausal — information from the future affecting the past — it would solve Newcomb's paradox. But it gives a weird world. You find something only because you saw yourself finding it and if you hadn't seen yourself finding it you wouldn't have found it, so either pattern is internally consistent. I have to think about whether there could still be a consistent physics."

She falls silent, eyes closed.

"Most divination in real life isn't perfectly reliable, though, so either it's not acausal or it's just really really hard to do well. If it is acausal I have no clue how the universe decides which self-consistent possiblity is real.

But also you could probably get a good enough predictor for Newcomb's Problem with enough intelligence and maybe mind-reading, which might not work in real life but it wouldn't break physics if an alien that could do it showed up."

"If he is a perfect predictor you have to abandon free will, that's always been the problem. Backwards causation solves that. If he is usually right you don't need backwards causation, just that people's choices are in part predictable.

"The story gets used as an argument for making yourself into someone who will always choose to only take the one box. But in order to choose to make yourself into such a person you have to be able to choose and if you don't have free will you can't, you are just observing what you do and imagining that you chose to do it. That's consistent with physics, at least before quantum mechanics, but I don't think anyone believes it.

"I wish I was back home so I could ask Apa what he thinks about it — whether divining is acausal or just being magically good at extrapolating from current data. We don't actually have an alien who is a perfect predictor but Mari did find that book for Masozi. If we had a true random number generator, using some mechanism that was entirely unpredictable like when an unstable atom would split, we could see if Mari can predict it. If she does better than chance ..."