"...I perhaps should then include a term very strongly discouraging killing people and explaining I am very individualistic, and explain I'm not going to punish a whole country for maneuvers by any single agents."

" - well, then everyone'll just make very sure everything they're doing is pinnable on an individual agent."

"Even knowing I can scry and past-watch?"

"If I wanted Anitam to have the space-to-maneuver to start a war even given the terms, I'd just tell a couple essential people precisely what actions will get Anitam in trouble and that nothing will get them in worse trouble than house arrest and then trust them to leave me plausible deniability on everything. You'll end up doing a lot of parsing of conversations in which absolutely no one says anything about planning a war or a violation of terms."

"And so your suggestion is siding with Anitam and letting people figure out how to convince your country to get what they want?"

"No, no, it's past the time for that, and it does trade off the moral high ground against people having more reasonable expectations about what they can get away with. But I do think people'll have unreasonable expectations along various axes about what they can get away with."

"This is not the first nor the last world we visit, although it is... the trickiest... so far. Is what I described what you would have suggested in advance?"

"Probably - or Voa, to seem less self-interested, but Voa's so conservative I'm not sure they'd even want that -"

"And what meta-rule would you advise me to generalise from that?"

"Predictability is really important? If you don't want - high-variance strategies - it's good to do the things already being done to reduce variance."

"But on the other hand whatever's already being done is typically an approximation to a locally optimal strategy given the current constraints, and the thing we do when we arrive is exactly lift them. And there's the robustness against someone really good at argumentation part."

"I will confess I do not make my plans for robustness against people really good at argumentation."

"Robustness in general is an extremely desirable quality when we're talking about so many worlds and species and social organisations and psychologies."

"I would be very interested in hearing examples."

"Of other species?"

"Yes. Of the - sociological spread of assumptions one has on landing in a new world."

"The world I started existing in was one of the ones I mentioned which had a species with an in-built caste system. They were arthropods that somehow worked around the square-cube law, and were organised in small colonies with one queen per who could reproduce, as well as workers and drones. They had no larger forms of social organisation than these colonies even though they existed all over the world, and little contact even with neighbouring colonies. Wars were very uncommon and fought mostly for food or other basic resources like that, and when they happened were extremely bloody—the worker caste barely had a concept of individual welfare and it was mostly in an abstract way that applied to drones and queens, to the point that it's perhaps more accurate to say a colony's entire worker caste is a single individual. But there was basically zero within-colony variance in strategies, each colony was small enough that the Queen was their entire blue caste, presiding over everything and making all decisions affecting more people."

"Huh."

"Our original species is in fact very similar to yours—we even look very alike, the only changes Sadai made to their appearance was making themself shorter and changing hair colour from black to green—except our years are a quarter of yours, we're fertile all the time once we're past puberty except females stop being fertile when they're older, we don't particularly want as many children as you—most developed countries were below replacement rates, actually—we live half as long as you do, we have as has been mentioned no castes, although gender has historically had some of the same social roles..."

"That's interesting. How common is social mobility of the kind we'd think of as intercaste - people becoming great intellectuals whose parents were farmers -"

"The distribution of wealth is approximately similar, and nutrition and early education play enough of a role that farmer to great intellectual is not extremely common, but what you'd consider mobility from city purple to yellow happens somewhat often, and the yellow, green, and blue category divisions are very much not natural—programming's picked by the same sort of person who studies theoretical physics, and politicians often have superior education degrees from universities. Red jobs aren't seen as a cluster at all, except for the part where they don't typically pay extremely well."

Nod. "Programming could've been green, yellows just grabbed it very hard."

"It's also the norm for people working any kinds of jobs which are part of a hierarchy to, if they're good, eventually be promoted to more managerial positions that would be typically occupied by blues here."

"Appointing good investigators into judges might work sometimes. Blues aren't in a lot of other management roles - you could get a lot done as some hands-off politician's scheduler, I suppose -"

"Lawyer to judge happens a lot, and hands-on activism tends to funnel into politics."