He did grab one of the best quotes, though (reposted for discussion):
[Edmund Leach] suggested that we look at the precolonial Burmese state not as a physically contiguous territory, as we would in the contest of modern states, but as a complex patchwork that followed an entirely different logic. We should picture the kingdom, he insisted, in terms of horizontal slices through the topography. Following this logic, Burma was, in practice, a collection of all the sedentary, wet-rice producers settled in valleys within the ambit of the court center. These would be...the state spaces. The next horizontal stratum of the landscape from, say, five hundred feet to fifteen hundred feet would, given its different ecology, contain inhabitants who practice shifting cultivation, were more widely scattered and were therefore less promising subjects of appropriation. They were not an integral part of the kingdom, although they might regularly send tribute to the central court. Still higher elevations would constitute yet other ecological, political and cultural zones. What Leach proposed, in effect, is that we consider all relatively dense, wet-rice settlements within range of the capital as "the kingdom" and the rest, even if relatively close to the capital, as "nonstate spaces".
The role of statecraft in this context becomes that of maximising the productive, settled population in such state spaces while at the same time drawing tribute from, or at least neutralizing, the nonstate spaces. These stateless zones have always played a potentially subversive role, both symbolically and practically. From the vantage point of the court, such spaces and their inhabitants were the exemplars of rudeness, disorder and barbarity against which the civility, order and sophistication of the centre could be gauged. Such spaces, it goes without saying, have served as refuges for fleeing peasants, rebels, bandits, and the pretenders who have often threatened kingdoms.
Noisms then uses this to extrapolate a wilderness as dungeon model (of which I am a known fan), with height up from the river valleys corresponding roughly to increasing dungeon level, and with PCs as agents of the river-valley states who go into the wilderness to extract tribute.
I think this model does a particularly good job of making sense of the Wilderlands of High Fantasy. Something that always bothered me about the Wilderlands was Dearthwood. By itself, a forest full of orcs is unremarkable. But when it's within spitting distance of a powerful city-state, it becomes somewhat more straining of disbelief. If you interpret the City-State as a river-valley and estuary power, though, the elevated forests full of non-cooperative, semi-nomadic humanoids fit nicely into the "non-state space" mold proposed by Leach. In the absence of precise cartographic surveying or friendly local guides who know the terrain, the forces of the City-State are largely unable to bring the orcs to battle in the forest itself. A similar dynamic applies to the Troll Fens, another piece of uncivilized land right on the City-State's doorstep (I recall Seeing Like a State mentioning the Marsh Arabs as occupants of a similar non-state space).
Moving to the "PCs as agents of the state in the wilderness" point, a model I've kicked around over the years for this is tax-farming. Throughout history, states have found it annoying to collect taxes and tribute from uncooperative people. Some historical polities (I recall reading of the Mongols doing this) hired independent contractors (private individuals) to go collect tribute in distant provinces and bring it back. This is a wonderful way to play low-domain / high-wilderness levels. You bid against other tax-farmers ("I will collect 20000 gp for the crown this year from the province of Baluchistan!" "Pfah, pathetic! If appointed tax agent to Baluchistan, I would collect 25000 gp for the crown this year!"), take your private army to Baluchistan, and then shake down the locals for everything you can get. Cue feuds with local bandits and tribesmen, previous tax farmers who set themselves up as warlords Heart of Darkness-style, wilderness monsters, peasant rebellions... After a year of this, you bring the crown's share of the tribute (the amount you bid) back, and keep whatever's left as your rightful payment. And if you didn't make your tribute goal... well, maybe don't come home. This also works very well with the City-State, which is nominally a powerful state with many tributaries. In practice though these tributaries are distant, the infrastructure and bureaucracy is insufficient for easy and safe extraction and transport of goods, and waging a military campaign to go take tribute by force is expensive, risky, and messy. Much better to let private adventurers assume the risks and reap some of the profits. It also, like the dungeon, has a nice boundedness property - you're only allowed to collect taxes in the province of Baluchistan, and if you go raid across the border into Afghanistan (to make your quota, say), the tax official for Afghanistan's going to be right pissed if he hears of it. Also, as in the dungeon, the players are very much outsiders, with poor maps and noncooperative locals.
Another passage from Seeing Like a State suggests a related approach:
The demography of precolonial Southeast Asia was such that control of land per se, unless it was a strategically vital estuary, strait, or pass, was seldom decisive in state building. Control of the population — roughly five persons per square kilometer in 1700 — mattered far more. The key to successful statecraft was typically the ability to attract and hold a substantial, productive population within a reasonable radius of the court. Given the relative sparseness of the population and the ease of physical flight, the control of arable land was pointless unless there was a population to work it. The precolonial kingdom thus trod a narrow path between a level of taxes and corvee exactions that would sustain a monarch’s ambitions and a level that would precipitate wholesale flight. Precolonial wars were more often about rounding up captives and settling them near the central court than asserting a territorial claim. A growing, productive population settled in the ambit of a monarch’s capital was a more reliable indicator of a kingdom ’s power than its physical extent.This reminds me somewhat of this post, where PCs build domains out of whatever sort of creatures they can find and coerce in the wilderness. Taking peasant, serf, and slave as a muddied equivalence class, it would not be difficult to conceive of PCs as "labor collection agents" for a state, going into the wilderness to acquire peasants (some of whom might be green and have tusks) and bringing them back by force to settle near the capital, where they can be monitored and taxed (help help I'm being repressed!). Whether any given set of players would veto such a campaign on moral grounds is another question.
The precolonial state was thus vitally interested in the sedentarization of its population — in the creation of permanent, fixed settlements. The greater the concentration of people, providing that they produced an economic surplus, the greater the ease of appropriating grain, labor, and military service.
Anyway, there's a lot of good stuff in Seeing Like a State (I'm particularly interested in his concept of legibility, the limits of a sovereign's ability to measure and understand the people and places that he rules) as it applies to ACKS, and I'm excited to see other parts of the OSR picking up "wilderness as dungeon" as an operative metaphor.