A basic fact of pre-industrial life is that it is easier to take wealth away from others than to produce it oneself. This also applies collectively. It is easier to take wealth from other societies than to extract a sufficient amount from your own. For this reason, agrarian societies turn to expansion, sending military expeditions beyond their boundaries to strip wealth from other populations. Armies have to be paid with the spoils of conquest. Further expansion and conquest is thus necessary. It is a positive feedback cycle. Obvious examples are the Roman Empire and the Arab Muslim Empire.
A key element is the taking of slaves. The wealth gained is long-term labor that requires only the most minimal compensation. A society that can produce little needs to import labor that does not need to be compensated, or to be compensated only at very low levels. This transfers the wealth of its production to the elite and its apparatus. In ancient Athens and Rome, slaves counted for more than half the population. Indian civilization solved the production problem slightly differently, with uncompensated labor performed by the so-called untouchables.
Agrarian states were thus hierarchical, centralized, and authoritarian, and the means of coercion were limited to the elite and its army as much as possible. But while the reach of the elite was strong, its scope was narrow. They wanted only two things from their subjects: crops and manpower. They controlled little else, and cared about little else. The welfare of their subjects was of no interest, except that they must be protected from predation by other states or tribes. And, to be sustainable, their own predation of their subjects had to be limited.
These pre-industrial, agrarian states were not large stable blocks of territory with effective state control and sharp boundaries. They were centers of power claiming control and authority over surrounding regions and populations. Over time, these states could vary in economic, political, and military power. Partly in response to the strength of their leadership, they waxed and waned, increasing their effective reach or seeing their control contract.
On the margins of their effective power, these states might make alliances with quasi-independent or independent populations, in most cases tribes. The priority of tribes would have been to remain independent and predatory. Failing that, they would have striven to remain independent, perhaps entering into some lucrative alliance with the state. In the case of an expanding state, the tribes in the path of that expansion would retreat, something fairly easy for pastoral nomads with mobile housing and capital. When a state was weak and began to contract, the tribes would reclaim their independence and return to predation.
This picture of states is accurate up to the eighteenth century, even in Britain. Until then, Britain and the states of Western Europe were ruled by autocrats or absolute monarchs, and their polities experienced constant attempted coups, civil wars, and invasions. It was only in the eighteenth century, not coincidentally the period in Western Europe of the modern agricultural and industrial revolutions, that the state changed. Its structure moved from top-down rule toward more general participation in decision making, and from tyrants toward governments based on law.
Philip Carl Salzman, Tribes and States, Inference, Vol. 2, No. 1.