Permutation City

chemistry -- if such a thing would even be possible. I just want you to construct a system with . . . the same potential.'

Maria was now thoroughly confused. 'When you mentioned a planetary surface, I thought you meant a full- scale virtual landscape -- a few dozen square kilometers. But if you're talking about the Autoverse . . . you mean a fissure in a rock on a seabed, something like that? Something vaguely analogous to a microenvironment on the early Earth? Something a bit more 'natural' than a culture dish full of two different sugars?'

Durham said, 'I'm sorry, I'm not making myself very clear. Of course you'll want to try out the seed organism in a number of microenvironments; that's the only way you'll be able to predict with any confidence that it would actually survive, mutate, adapt . . . flourish. But once that's established, I'll want you to describe the complete picture. Specify an entire planetary environment which the Autoverse could support -- and in which the seed would be likely to evolve into higher lifeforms.'

Maria hesitated. She was beginning to wonder if Durham had any idea of the scale on which things were done in the Autoverse. 'What exactly do you mean by a 'planetary environment'?'

'Whatever you think is reasonable. Say -- thirty million square kilometers?' He laughed. 'Don't have a heart attack; I don't expect you to model the whole thing, atom by atom. I do realize that all the computers on Earth couldn't handle much more than a tide pool. I just want you to describe the essential features. You could do that in a couple of terabytes -- probably less. It wouldn't take much to sum up the topography; it doesn't matter what the specific shape of every mountain and valley and beach is -- all you need is a statistical description, a few relevant fractal dimensions. The meteorology and the geochemistry -- for want of a better word -- will be a little more complex. But I think you know what I'm getting at. You could summarize everything that matters about a pre-biotic planet with a relatively small amount of data. I don't expect you to hand over a giant Autoverse grid which contains every atom in every grain of sand.'

Maria said, 'No, of course not.' This was getting stranger by the minute. 'But . . . why specify a whole 'planet' -- in any form?'

'The size of the environment, and the variation in climate and terrain, are important factors. Details like that will affect the number of different species which arise in isolation and later migrate and interact. They certainly made a difference to the Earth's evolutionary history. So they may or may not be crucial, but they're hardly irrelevant.'

Maria said carefully, 'That's true -- but nobody will ever be able to run a system that big in the Autoverse, so what's the point of describing it? On Earth, the system is that big, we're stuck with it. The only way to explain the entire fossil record, and the current distribution of species, is to look at things on a planetary scale. Migration has happened, it has to be taken into account. But . . . in the Autoverse, it hasn't happened, and it never will. Effects like that will always be completely hypothetical.'

Durham said, 'Hypothetical? Absolutely. But that doesn't mean the results can't be considered, can't be imagined, can't be argued about. Think of this whole project as . . . an aid to a thought experiment. A sketch of a proof.'

'A proof of what?'

'That Autoverse life could -- in theory -- be as rich and complex as life on Earth.'

Maria shook her head. 'I can't prove that. Modeling a few thousand generations of bacterial evolution in a few microenvironment. . . .'

Durham waved a hand reassuringly. 'Don't worry; I don't have unrealistic expectations. I said 'a sketch of a proof,' but maybe even that's putting it too strongly. I just want . . . suggestive evidence. I want the best blueprint, the best recipe you can come up with for a world, embedded in the Autoverse, which might eventually develop complex life. A set of results on the short-term evolutionary genetics of the seed organism, plus an outline of an environment in which that organism could, plausibly, evolve into higher forms. All right, it's impossible to run a planet-sized world. But that's no reason not to contemplate what such a world would be like -- to answer as many questions as can be answered, and to make the whole scenario as concrete as possible. I want you to create a package so thorough, so detailed, that if someone handed it to you out of the blue, it would be enough -- not to prove anything -- but to persuade you that true biological diversity could arise in the Autoverse.'

Maria laughed. 'I'm already persuaded of that, myself. I just doubt that there could ever be a watertight proof.'

'Then imagine persuading someone a little more skeptical.'

'Who exactly did you have in mind? Calvin and his mob?'

'If you like.'

Maria suddenly wondered if Durham was someone she should have known, after all -- someone who'd published in other areas of the artificial life scene. Why else would he be concerned with that debate? She should have done a much wider literature search.

She said, 'So what it comes down to is . . . you want to present the strongest possible case that deterministic systems like the Autoverse can generate a biology as complex as real-world biology -- that all the subtleties of real-world physics and quantum indeterminacy aren't essential. And to deal with the objection that a complex biology might only arise in a complex environment, you want a description of a suitable 'planet' that could exist in the Autoverse -- if not for the minor inconvenience that the hardware that could run it will almost certainly never be built.'

'That's right.'

Maria hesitated; she didn't want to argue this bizarre project out of existence, but she could hardly take it on if she wasn't clear about its goals. 'But when it's all said and done, how much will this really add to the results with A. lamberti?'

'In one sense, not a lot,' Durham conceded. 'As you said, there can never be a proof. Natural selection is natural selection, and you've shown that it can happen in the Autoverse; maybe that should be enough. But don't you think a -- carefully designed -- thought experiment with an entire planet is a bit more . . . evocative . . . than any number of real experiments with Petri dishes? Don't underestimate the need to appeal to people's imaginations. Maybe you can see all the consequences of your work, already. Other people might need to have them spelled out explicitly.'

Maria couldn't argue with any of that -- but who handed out research grants on the basis of what was evocative! 'So . . . which university -- ?'

Durham cut her off. 'I'm not an academic. This is just an interest of mine. A hobby, like it is with you. I'm an insurance salesman, in real life.'

'But how could you get funding without -- ?'

'I'm paying for this myself.' He laughed. 'Don't worry, I can afford it; if you take me up on this, you're not going to be shortchanged, I can promise you that. And I know it's unusual for an amateur to . . . subcontract. But like I said, I don't work in the Autoverse. It would take me five years to learn to do, myself, what I'm asking of you. You'll be free to publish all of this under your own name, of course -- all I ask is a footnote acknowledging financial support.'

Maria didn't know what to say. Lorenzo the insurance salesman? A private citizen -- not even an Autoverse junkie -- was offering to pay her to carry out the most abstract piece of programming imaginable: not simulating a nonexistent world, but 'preparing' a simulation that would never be performed. She could hardly be disdainful of anyone for throwing their hard-earned money away on 'pointless' Autoverse research -- but everything that had driven her to do that, herself, revolved around firsthand experience. However much intellectual pleasure it had given her, the real obsession, the real addiction, was a matter of putting on the gloves and reaching into that artificial space.

Durham handed her a ROM chip. 'There are some detailed notes here -- including a few ideas of mine, but don't feel obliged to follow any of them. What I want is whatever you think is most likely to work, not what's closest to my preconceptions. And there's a contract, of course. Have your legal expert system look it over; if you're not happy with anything, I'm pretty flexible.'

'Thank you.'

Durham stood. 'I'm sorry to cut this short, but I'm afraid I have another appointment. Please -- read the notes, think it all through. Call me when you've made a decision.'

After he'd left, Maria sat at the table, staring at the black epoxy rectangle in her palm, trying to make sense