The Lower Bound of Civilization

The View from Oregon – 335: Friday 04 April 2025

In newsletter 333 I discussed the historical process of aggregation as complementary to the process of dissolution. It has since occurred to me that the process of aggregation allows for a significant generalization if we hold that cities are a given threshold of aggregation. I suppose I implied that in this newsletter, since I said that aggregation goes from villages to towns to cities to geographical regions. Each of these named entities constitutes a given threshold of aggregation. Cities as a threshold of aggregation are a limiting case of a general process, and since cities aren’t the only aggregations that form, they aren’t the only threshold, and they aren’t the only limiting case. If we think of villages, towns, cities, and civilizations as natural kinds, then the lower bound of each natural kind is its threshold, and its upper bound is the threshold of the next natural kind. Each of these natural kinds, so defined as falling between two thresholds, is a limiting case from above and from below. (I’m framing this in terms of limiting cases in light of the Einstein quote I used in newsletter 331; effecting a generalization of this kind is all to the good, theoretically speaking.)

I’ve argued that early medieval civilization is distinctive as a civilization in which cities played a minimal role. In the case of early medieval civilization, the aggregation threshold that proved to be durable was the manorial estate and the monastery. Both were like shopping malls in the sense that they attempted (and succeeded for several centuries) to place all functions under a single roof. Even without cities, geographical regions aggregated into kingdoms. Thus in early medieval civilization the aggregation process bypassed cities and still managed to knit itself together into a civilization.

The aggregation of cities into geographically regional networks constitutes a step above and beyond the aggregation of the cities themselves—it is, in fact, a later and higher threshold of aggregation, therefore distinct from the earlier and lower threshold of urban formation—which implies that the view I am suggesting does not equate civilization with cities per se, which is how V. Gordon Childe’s “urban revolution” has frequently been interpreted. It’s not so much that an explicit case is made for the equation of cities and civilization, as that Childe’s ten criteria for an urban revolution have been widely adopted by archaeologists as a diagnostic for identifying a civilization. What is implied by my argument is that even if all ten of Childe’s criteria are present, but there has been no aggregation at a geographically regional level, then there is no civilization, only a city or cities.

I didn’t consciously set out to define a position orthogonal to Childe, and certainly what I’m saying isn’t mutually exclusive of Childe’s conception of an urban revolution and the later tendency to equate the urban revolution with the existence of civilization. However, it can be so taken. There are several ways to approach this. One could argue that a city in the Childean sense can only come into being once cities have aggregated into a geographically regional entity; that is to say, something like a city or cities might exist—we could call them proto-cities—but they don’t fully possess Childe’s ten criteria for an urban revolution until they have further aggregated beyond the mere existence of cities. The idea of a proto-city is already implicitly recognized, after a fashion. Jericho was among the earliest of cities, and from the archaeological evidence seems to have functioned like a city, but another well-known example that I have mentioned many times, Çatalhöyük, endured for thousands of years and may have had a population of around 10,000, but it doesn’t seemed to have functioned like a city in the conventional sense. Göbekli Tepe presents an even more peripheral example (peripheral in terms of urbanization, but not in terms of importance), hence it is rarely called a “city,” though it was clearly a locus of activity. In a continuum of aggregation we can imagine a sequence of Göbekli Tepe, Çatalhöyük, and Jericho, each in turn more closely approximating the structure and function of a paradigmatic city. I’m not suggesting that there was a linear development in this way—these three sites have significant discontinuities; each represents a unique aggregation—but seeing these sites in this context we can see how several proto-cities might define a geographical region without yet having reached the threshold of civilization. One could imagine, for example, a geographical region with a dozen or more settlements like Çatalhöyük but still no Çatalhöyükean civilization.

For some years now I’ve been taking occasional notes on a thought experiment that I call the “single-city civilization thought experiment.” I’ve been thinking of including this among my Thought Experiments in Civilization series of episodes. The question is simple: can there be a civilization that consists of a single city and its hinterland, or does civilization (necessarily) entail more than one city? I tend to think that, in terrestrial history at least, several cities in a given geographical region are a necessary condition of civilization, though in one of the scenarios mentioned above, we could imagine several proto-cities coming into being in a geographical region, but not cresting over the true urban threshold until they mutually develop into a civilization, which is consistent with equating Childe’s urban revolution with the existence of a civilization. The geographical region need not be large. We can imagine a single, fertile river valley along which multiple cities develop at different places along the river, with the cities connected by trade long the length of the river. The obvious example of this is Egypt, although the Nile is a very large river that could easily support many large cities along its length. We can imagine something more modest, with a handful of cities of about 10,000 or less population each along a dozen miles or so of a river. This could be a civilization, as I am presenting it, even while a single city, or a single proto-city, of larger size than the population of the handful of cities previously mentioned might not constitute a civilization. This suggests an interesting question: how large can be proto-city become without passing the threshold of urbanization?

I wrote above as a qualification, “in terrestrial history.” In my thought experiment I imagine different kinds of cities on different worlds, developing under distinctive circumstances. Thus I’m not excluding the possibility of a single-city civilization a priori. There may even have been civilizations of this kind in the earliest history of civilization on Earth, but mostly in our own history we find networks of civilization within a geographical region constituting a civilization. A planet with a very different history might have mostly had single-city civilizations in its history, with networks of cities forming civilizations being the exception to the rule. We can imagine a process in which single-city civilizations develop in distinctive geographical regions of a world, only gradually rising to the level of complexity and integration of true urbanization in the absence of neighboring cities as rivals. However, once these single-city civilizations grow large enough, their borders would eventually come to meet the borders of other growing single-civilization civilizations (the agricultural hinterlands to two cities will come into contact), and on such a world with such a history, a new kind of history would be inaugurated at this point because heretofore interactions between cities had not played a role in the history of such a world. These thought experiments and the problem with which I began—if an urban revolution eo ipso constitutes a civilization—are special cases of the more general problem of the lower bound of civilization. I’ve been thinking about the problem of the lower bound of civilization for many years, and I don’t yet have a definitive formulation of the problem (much less a solution), but I think I’m getting closer.

Consider this approach to the lower bound of a civilization: because a civilization involves social hierarchy and craft specialization, this means that everyone higher up in the social hierarchy or practicing a specialized craft no longer produces food. This in turn means that there needs to be a sufficient population engaged in agriculture of a sufficient level of efficiency, that they are able to feed those who do not produce food. This does not occur in subsistence societies (village agriculture), which only produce food commensurate to the needs of the farmers. Agriculture must produce enough food that a community can afford to support a blacksmith, a wheelwright, a miller, a records clerk, and a headman, and so on. The blacksmith, in turn, needs someone to produce charcoal for him and to dig iron ore for him. Probably the blacksmith starts out making his own charcoal and digging his own ore, but the economic logic of comparative advantage will eventually lead to greater specialization, so that the charcoal burner and the miner will be distinct from the blacksmith. How many people there will be in leadership positions and how many people there will be in craft specialization will vary according to the artifacts created by the culture in question. It’s also relative to the cultivars available in the geographical region in question and the productivity of the soil (which is in turn a function of either rainfall or irrigation water from a river). This calculation would be pretty complex and would require that we fill in several variables to get to a number for the minimum population or geographical area necessary to a civilization. I’m not saying it can’t be done, but I am saying it would require a significant research program to converge on good numbers.

Let’s attack this from another angle. Suppose we hold, with Child’s urban revolution hypothesis, that civilizations are coincident with the appearance of cities. How big does an agricultural village need to be to cross the threshold of being a city? Earlier I mentioned that Çatalhöyük is estimated to have had a peak population of around 10,000, but we don’t need to insist on this number. Even today, in the midst of industrialized civilization, we have many viable smaller cities of a few thousand, so we could set the threshold of a city at a population of 3,000 to 4,000, or even lower. How much lower? That’s a question I will set aside, preferring to focus on our ability to confidently identify a settlement of, say, 4,000 people as a city if that population embodies social hierarchy and craft specialization. Again, we can pursue a sorites paradox here, and ask further questions like how large an agricultural village can become. Could there be an agricultural village of 4,000 persons with little or no social hierarchy or craft specialization? I’m not going to try to answer this, but these are intrinsically interesting questions.

I will employ what I call a pragmatic definition of civilization as a network of cities engaged in relations of cooperation, competition, and conflict. A network of cities that crests over this threshold constitutes what I call a basal civilization. Neighboring cities are more than a stimulus to mutual growth, as I implied above; it seems likely that the origins of civilization are tied to events such as disgruntled minorities leaving an established city and going elsewhere to found their own city, like Muhammad and his followers making the Hegira to Medina from Mecca in AD 622, only return to Mecca under the Treaty of al-Hudaybiya, eventually taking the city in AD 630. In such a scenario, the natural growth of civilization involves rival cities engaged in a range of activities in relation to each other.

Given my pragmatic definition of civilization, a minimal civilization (i.e., the lower bound of a civilization) needs to consist of at least three cities, since with cities A, B. and C, A and B can cooperate, while B and C are in competition, and A and C could be in conflict. Two cities would be insufficient to maintain these dynamics. Note that there could be many more cities, but three is the minimum number, and I’m looking for the minimum number in order to get at the lower bound of civilization. I don’t like to cite hard numbers for something as under-theorized as civilization, but if you held a gun to my head and demanded an answer, I would say that the lower bound of population for a civilization is approximately 12,000 persons, which would be three cities of about 4,000 persons each. The lower bound of the geographical area of such a civilization would be the arable lands (or other sources of nutrition, like fisheries) necessary to feed 12,000 persons, and this would vary, as I said above, based on climate and cultivars. What I have done here, in this approach (in contradistinction to the previous approach outlined), is to fill in the variables with my own definition, whereas in the previous approach I left the variables undefined and consequently came up with no number. In a more detailed study, I would want to see if these two overlapping approaches to defining the lower bound of a civilization converged.

---

---