I told Justin that too. He said he understood. He said sometimes when you re young you have to think about things, because you're forming your value-sets and you keep coming up with Data Insufficient and finding holes in your programs. So you keep trying to do a fix on your sets. And the more powerful your mind is and the more intense your concentration is, the worse damage you can do to yourself, which is why, Justin says, Alphas always have trouble and some of them go way off and out-there, and why almost all Alphas are eccentric. But he says the best thing you can do if you're too bright for your own good is what the Testers do, be aware where you got which idea, keep a tab on everything know how your ideas link up with each other and with your deep-sets and value-sets, so when you're forty or fifty or a hundred forty and you find something that doesn't work, you can still find all the threads and pull them.

But that's not real easy unless you know what your value-sets are, and most CITs don't. CITs have a trouble with not wanting to know that kind of thing. Because some of them are real eetee once you get to thinking about how they link. Especially about sex and ego-nets.

Justin says inflexibility is a trap and most Alpha types are inward-turned because they process so fast they're gone and linking before a Gamma gets a sentence out. Then they get in the habit of thinking they thought of everything, but they don't remember everything stems from input. You may have a new idea, but it stems from input somebody gave you, and that could be wrong or your senses could have been lying to you. He says it can be an equipment-quality problem or a program-quality problem, but once an Alpha takes a falsehood for true, it's a personal problem. I like that. I wish I'd said it.

And once an Alpha stops re-analyzing his input and starts outputting only, he's gone completely eetee. Which is why, Justin says, Alpha azi can't be tape-trained past a certain point, because they don't learn to analyze and question the flux-level input they get later, and when they socialize too late, they go more and more internal because things actually seem too fast and too random for them, exactly the opposite of the problem the socialized Alphas have—too fast, though, only because they're processing like crazy trying to make more out of the input than's really there, because they don't understand there is no system, at least there's no micro-system, and they keep trying to make one out of the flux they don't understand. Which is why some Alphas go dangerous and why you have trouble getting them to take help-tapes: some start flux-thinking on everything, and some just go schiz, de-structure their deep-sets and reconstruct their own, based on whatever comes intact out of the flux they're getting. And after that you don't know what they are. They become like CITs, only with some real strange logic areas.

Which is why they're so hard to help.

I think Yanni is right about Justin. I think he's awfully smart. I've asked my uncle Denys about him, about whether they don't make him a Special because of politics; and Denys said he didn't know whether Justin qualified, but the politics part was definitely true.

Ari, Denys said, I know you're fond of him. If you are, do him and yourself a favor and don't talk about him with anybody on staff—especially don't mention him in Novgorod.

I said I thought that was rotten, it was just because of what his father did, and it wasn't his fault any more than it was mine what my predecessor and his father were fighting about.

And then uncle Denys said something scary. He said: No. I'm telling you this for his sake. You think about it, Ari. He's very bright. He's possibly everything you say. Give him immunity and you give him power, Ari, and power is something he'd have to use. Think about it. You know Novgorod. You know the situation. And you know that Justin's honest. Think about what power would do to him.

I did think then, just like a flash, like lightning going off at night, and you can see everything, all the buildings you know are there, but you forget—you forget about the details of things until the flash comes—and it's gray and clearer than day in some ways. Like there could be color, but there's not quite enough light. So you can see everything the way you can't see it in daylight.

That's what it is when somebody throws light on what you've got all the pieces of.

His father is at Planys.

That's first.

Then there are all these other things—like him being my teacher. Like—we're best friends. But it's like Amy. Amy's my best friend but Florian and Catlin. And we couldn't get along till Amy knew I could beat her. Like she was going to have to do something about me until she knew she couldn't. Then we're all right.

Power stuff. Ari was right about that too—about us being territorial as hell, only territory is the wrong word. Territory is only something you can attach that idea to because we got used to the root concept when we were working with animals.

That's what Ari called science making itself a semantic problem. Because if you think territoriality you don't realize what you're really anxious about. We're not bettas.

The old Greeks talked about moira. Moira means lot in Greek. Like your share of things. And you can't grab somebody else's—that's stealing; you can't not do your own; that's being a coward. But figuring out what yours is would be a bitch, except other people and other animals help you define what your edges are when they react back: if they don't react back or they don't react so you can understand them you get anxiety reactions and you react with the fight or flight bias your psychset gives you, whether you're a human being or a betta. I got that from Sophocles. And Aristotle. And Amy Carnath and her bettas, because she's the first CIT friend I ever had I worked that out with, and she breeds fighting fish.

It's not territory. It's equilibrium. An equilibrated system has tensions in balance, like girders and trusses in a building.

Rigid systems are vulnerable. Ari said that. Equilibrated systems can flex under stress.

The old Greeks used to put flex in their buildings, moving joints, because they had earthquakes.

I'm leading up to something.

I think it has to do with me and Justin.

I don't trust too much flex any more than too little. Too much and your wall goes down; too little and it breaks.

I'm saying this for the record. If I was talking to Justin or Yanni or uncle Denys I'd say:

Non-looping paths don't necessarily have to be macro-setted on an individual level.

Except then—

—then you have to macro-set the social matrix, which you can do—

—but the variables are real killers to work out—Justin's proved that.

God, talking with yourself has some benefits.

That's what Ari meant by macro-values. That's what she was talking about. That's how she could be so damned careless about the random inputs with her designs. They all feed into a single value: the flux always has to reset off the central sets. That's what Justin was trying to explain to me: flux re-set functions.

Gehennans identify with their world. That's the whole center of what Ari did with that design. And no flux-thinking can get at that.

But where does that go?

Damn! I wish I could tell it to Justin. . . .

Define: world. There's the worm. God! It could be one, if you could guide that semantic mutation.