Complex societies and 'civilisations' seem to develop in areas where scarcity, opportunity, population density, and diversity can be found simultaneously. By that I mean, 'civilisations'* like Sumer and Egypt develop in places where there are obstacles to healthy living and natural geographical boundaries to human development, like deserts; opportunities for growth within the bounded region, like the presence of fertile alluvial floodplains or marine resources; a rising population, possibly as a result of improved agriculture or some other cause; and a diversity of environments (floodplain and steppe, for instance) and people (different language families). I don't want this to be simplistic, of course, because the development of complex civilisation is an incredibly complicated topic - but most of the complex civilisations of which I am aware developed in such environments. It seems as if the development of civilisation is inevitable in areas where population density creates pressure on scarce resources, but where different resources can be exploited by a division of labour allowed for by the stratification of society, all of it enriched by ideas and technologies provided by diverse populations living in diverse environments.
In ancient Mesopotamia, there was a strip of fertile agricultural land extending through Mesopotamia and into the Levant, the product of rainfall and the rivers. Irrigation, practiced in northern Mesopotamia by 7000 BCE, added more land to the crescent. Steppe-like plains were found outside of this agricultural belt, allowing sheep and other animals to be raised on the grass. But beyond this sub-desert steppe was desert, in which agriculture could not develop, and the only people living there were small populations of nomads. Egyptian civilisation, and that of Nubia to the south, also developed in a desert rendered fertile by a river's floods; agriculture in this region created a relatively large population living on a relatively small patch of land, which in turn made land and food scarce. Anyone capable of controlling access to this land and food was in a powerful position over the rest of the population.
Civilisation in the Sahel - Jenne, Timbuktu, etc - arose in a region similarly bounded by desert. The desert created a boundary to agriculture, meaning that it could only be employed in certain areas, compounding the power of people controlling access to its products. People in the Ethiopian highlands also lived in a marginal environment, where mountain peaks and nearby desert limited agricultural potential. These populations lived in environments that were productive enough to generate a large population, but this population then created pressure on the very resources that had allowed the population to grow in the first place, exactly the same pressure felt in Mesopotamia and Egypt.
Similarly, in Amazonia, varzea, or alluvial floodplain, could be found adjacent to terra firme, land that didn't experience flooding but which could be rendered fertile through human action. Add in the rivers and the dense rainforest (in many areas desert-like in its lack of useful resources) and Amazonia actually has a huge diversity of environments, some productive and others hostile, limiting human habitation to areas where the greatest volume and diversity of resources can be had. The rivers and floodplain could generate resources to feed a large population, but the dense forest placed a limit on population growth, leading to stratification through control of scarce resources. Spreading horizontally across the land in order to access a diversity of resources, a process controlled by a ruling elite, would then have followed as a matter of course, something found throughout the world. The world's most successful civilisations all developed the knack of exploiting multiple resources - something that further compounded the power of the elite that controlled the distribution of produce.
Complex societies seem to develop in marginal areas with features that allow for population growth while ensuring competition for whatever those features happen to be. But complex societies prosper best when they use the functional specialisations in their society to access resources spread over several areas. Jenne and Timbuktu were in an area of extreme environmental diversity. From north to south there were bands of 1) desert, 2) sub-desert steppe, 3) wooded steppe, 4) savanna, 5) moist woodland/savanna mosaic, and 6) tropical rainforest. Uruk developed through reliance on resources from both the irrigated plains and the pastoral steppe. Civilisations in western South America developed a penchant for the exploitation of resources spread over a range of habitats, from the dry Pacific coast through the Andean foothills up into the high puna and the tropical forest on the other side. Egyptian polities consolidated their power through the control of precious resources - especially gold, for which Egypt was famous in ancient times, and also natron, salt, and so on - in addition to control over the agricultural surplus. As well as controlling these products, these states all engaged in trade over long distances, another feature which enhanced the prestige of those who enabled and benefited from it. Trade also allowed for contacts of a less purely-economic nature - for ideas to spread, clash, and generate new ones.
In Afro-Eurasia, human cultural diversity allowed for the spread of technologies that altered these circumstances. Horses and the wheel were probably introduced to Mesopotamia, Egypt, China, Nubia, the Sahel, and Europe from the Eurasian steppes, traded south through the Caucasus and east through the Tarim Basin. Both technologies had major consequences, in opening up new lands for cultivation and in providing new tools for agriculture, trade, and warfare - the activities of states. Domestication of the horse and the introduction of iron allowed people in the Sahel to move about over the different environmental zones, allowing them to exploit them all, and iron allowed for easier agriculture. The camel then allowed trade across the Sahara. Introduction of the wheel to Nubia allowed for water-wheels to irrigate new land. And so on. All of these tools allowed populations to increase, and to maintain contacts with one another over large distances.
Without people on the steppe domesticating horses, Eurasian civilisation would have looked very different. The introduction of the horse throughout the continent changed a lot of things - and it was allowed for by a diversity of environments and a diversity of people coming into contact with one another, including settled stratified groups and nomadic egalitarian groups. What you need at the beginning is a small group of people controlling access to needed resources in a bounded environment, but what sustains the whole thing and creates the crazily globalised world we see today is human cultural diversity - people spread over wide areas with different beliefs, practices, cultigens, domesticated animals, and experiences coming together and bouncing off one another, creating new ideas, introducing technologies.
In the Americas, there was little direct contact between groups separated by continental boundaries. Evidence of contact between Mesoamerica and the Andes is controversial, but it seems that if it occurred, it was at least on a very small scale. It seems probable that bronze-working made its way north from Colombia, and maize, beans, and other cultigens could be found from what is now the northern USA all the way south to Argentina - but other things, like llamas, the only large domesticated animals in the Americas before Columbus, failed to make it to the Yucatan. The geography of the American continents made sustained contact between all of these groups very difficult, so that innovations made their way gradually between them. It isn't that the Americas weren't diverse - they were, and incredibly so, and this diversity had a clear role in generating and sustaining indigenous American civilisations. But as Jared Diamond pointed out, the geography of the Americas meant that technologies developed in Amazonia took a long time to reach Florida, and vice versa, unlike in Europe where innovations could be shared and transformed and shared again within a matter of a few generations.
It seems like an outright abundance fails to produce social complexity. It doesn't seem to occur where people are living in an incredibly fertile wonderland or a place where the population is small enough that even meagre resources are enough to go around. In pre-Columbian Guiana, there were large villages of thousands of people controlled by chiefs and elites with considerable power over the other villagers. After the introduction of Eurasian diseases, the population of Guiana was decimated, and where before there was pressure on resources as a result of the large population, afterwards there was no such pressure, and most native communities in Guiana produced enough food for their own use. This led to chiefs having almost entirely nominal roles - they had and have no leverage over the rest of the village, and they serve merely to arbitrate in disputes and to conduct rituals. Villages in twentieth century Guiana had little functional specialisation - there were no specialist basketmakers or goldsmiths - and no stratification. This is because even if someone did want power over everyone else, they would never have the leverage to achieve this.
Living on the edge of a desert with a rising population seems to create sufficient pressure on resources to allow for certain groups of people to act as gatekeepers to them. This control over resources - over the things people want - gives them power. Having power over other people means being able to make them do things that they wouldn't otherwise do, and having control over food distribution generates power over others in a very visceral, obvious way.
Civilisations seem to develop in these weird, marginal areas. But that isn't where civilisation is confined. If anything, the sites where complex societies first began are almost uniformly poor and backward today compared to contemporary societies in less geographically-marginal locations. Egypt is no great power these days. The Sahel and Ethiopia are known today primarily as recipients of aid. The Atacama desert is so dry, even by the coast, that today it is useful to astronomers but to few others. It seems like social complexity and the development of functional specialisation are ideas that can be spread outside of these marginal areas, where they prosper and develop into new institutions with greater power than their marginal-area-dependent predecessors. Furthermore, the power created by controlling resources in these marginal environments allows the group to transcend the marginal environment - to transform it into a less marginal environment, and to increase yields through the development of new techniques, which can also come from and be shared with other human groups.
The British government today doesn't exercise power over the population by visibly controlling the food supply. It affects your judgements, and has power over you, because of a different set of mechanisms and a different set of resources. Power obviously doesn't have to depend on simple economics and the struggle for life in a marginal environment, and among factors I have avoided discussing here are: trade in exotic goods; the use of precious items as status-markers and creators and the imposition of sumptuary laws; the role of religious belief in bolstering the power of an elite; and the use of sheer military force in subduing a population and forcing them to do your bidding. It's obviously a very complicated topic - which is part of the reason
why archaeologists have been debating it since the very formation of the
discipline - but some generalisations can be made, and in the case of sumptuary laws, military might, and so on, it seems like the traction provided by control over economic life is, initially, necessary. It seems, at least, as if civilisations develop out of necessity in marginal, densely-packed areas of human and environmental diversity. And that, of course, has implications for our understanding of just what social structure is and what social facts are. John Searle's view, for instance, that social facts are 'desire-independent' cannot make sense of the rise of states - but that's a topic for another day.
*I'm going to use a folk definition of 'civilisation' here, and won't define it precisely. This is just a general sketch.
Obviously none of this will be new to archaeologists or graduate students.