Start with a number that reframes everything else.
At its height the Roman Empire governed somewhere in the region of fifty to sixty million people across three continents. The number of salaried imperial officials, on the usual modern estimates, was in the low hundreds for the whole empire, rising to a few thousand if you are generous about who counts. Britain, a province of perhaps two to three million, was run by a governor with a staff you could fit in a large room.
That is a ratio no modern state would recognise. It tells you immediately that Rome did not govern in the way we mean by the word. It set the taxes, kept the peace, judged the big cases and let the towns run themselves.
The empire was, administratively, a franchise. Local elites collected the taxes, maintained the roads and buildings, staffed the councils, and settled most disputes. In exchange they got Roman citizenship, status, and protection from their neighbours and from their own populations. It is a cheap way to run a continent and it works as long as the local elites think the bargain is worth it.
The road network eventually ran to something like eighty thousand kilometres of main highway, engineered in layers, drained, cambered, and built straight through obstacles rather than around them where that was cheaper in time.
They were built by the army, and they were built for the army. A road is how you move a legion from a quiet frontier to a noisy one before the trouble spreads, and how a governor's order reaches a town in days instead of weeks. There was a state courier system with staging posts at intervals, and it was reserved for official traffic; using it without a permit was an offence, and the permits were a form of patronage.
Ordinary trade used the roads too, but heavy goods went by water wherever water existed. Moving grain by sea was cheaper by a large factor than moving it overland, which is why the empire's cities cluster on coasts and navigable rivers, and why the road map and the trade map are not the same map.
The other great engineering effort was water. Rome itself was supplied by a set of aqueducts delivering an enormous daily volume, mostly by gravity over long distances, feeding public fountains, baths and a minority of private houses. Cities across the empire copied it at smaller scale. The surviving arcades are the famous part; most of the length ran underground, which is duller and better engineering.
Both systems have the same political logic. They were public works paid for by the state or by rich men buying status, they were visible, and they made the case for the empire to the people living in it. A town that got an aqueduct had a reason to prefer Roman rule to the alternative.
The phrase for the long internal calm is the Roman peace, and it was real. For roughly two centuries most of the Mediterranean world was free of the wars between neighbouring states that had been normal before, and the population and the economy both grew.
It was also a peace maintained by an army of around three hundred thousand men stationed almost entirely on the frontiers, and by the certainty that a province which revolted would be crushed. A historian writing a century after the events puts a speech into the mouth of a British chieftain describing the Roman method: they make a desert and call it peace. The line is a Roman writing his own side's critique, which is interesting in itself, but the destruction it describes is documented.
Underneath the peace, slavery. The empire was a slave society in the strong sense: slaves were a large share of the population, they were legally property, they could be worked, beaten and sexually used at their owner's discretion, and their children were born owned. Manumission was common by ancient standards and a freed slave became a citizen, which softened the institution's edges without changing what it was.
And the tax. The empire was funded by a land tax and a head tax, assessed by periodic census, paid mostly by farmers, and the money went to the army. For most people the empire was an assessor with a list and a soldier standing behind him. When the archaeological record shows a province flourishing, that is compatible with the same province producing the tax revolts and the peasant unrest that also show up in the sources.
If the administration was thin, the army was the one imperial institution most provincials actually met, and it did far more than fight.
A legionary served twenty five years. The auxiliary regiments, recruited from non citizens along the frontiers, served the same term and were granted citizenship on discharge, for themselves and their children. We are not inferring this: hundreds of the bronze certificates recording those grants have been dug up, folded double and sealed, issued to individual men whose names and units and home districts are on them.
So the army took a farmer from the Rhineland or the Balkans, moved him a thousand miles, paid him in imperial coin, taught him enough Latin to take an order and enough of a trade to be useful, and released him at forty five as a citizen with savings and a plot of land, usually near where he had served. Do that for two centuries and you have manufactured a provincial middle class that has a personal stake in the empire continuing.
The forts did the same thing to the landscape. A permanent garrison needs grain, leather, pottery, metalwork and drink, and it pays cash. Civilian settlements grew outside the walls to supply it, and several of them are cities today. A number of European capitals and major towns began as the shop and tavern district outside a Roman camp.
The cost side is equally plain. Army pay was the largest item in the imperial budget by a wide margin, and when emperors needed the soldiers' goodwill they raised it, which is one of the pressures behind the debasement of the coinage in the third century. An institution that both held the empire together and consumed most of its revenue is a difficult thing to reform, and nobody managed to.
The roads decayed and the aqueducts were cut. The thing Rome built that is still standing is the law.
Roman law began as the procedural rules of a small city and grew, over eight centuries, into a body of professional legal reasoning. Its distinctive move was the separation of a case's facts from the legal question underneath, and the development of a class of specialists whose written opinions on those questions accumulated into a literature.
What that produced was a set of concepts still doing daily work: contract as an agreement the law will enforce, and the different kinds of it; ownership as distinct from possession; the trust obligations of someone managing another person's property; the idea that an injury creates a duty to compensate. Modern lawyers in most of Europe, Latin America and beyond are handling categories that were refined by Roman jurists.
It reached us by an odd route. In the sixth century, long after the western empire had gone, an eastern emperor ordered the whole mass of it compiled, edited and made authoritative. That compilation was largely forgotten in the west, and then a manuscript of it turned up in Italy in the eleventh century and became the object of study at the first universities. From there it was taught, argued over, and gradually absorbed into the law of most continental states, and later written into the codes that Napoleon and his imitators exported.
So the transmission is not continuity. It is a rediscovery, six hundred years later, by people who had a use for it.
It is worth asking why the arrangement lasted as long as it did, because the answer is not mainly force.
Citizenship was the mechanism. It began as the privilege of one city, was extended to Italy after a war, and then spread outwards steadily, granted to individuals for service, to communities for loyalty, and automatically to soldiers of the auxiliary regiments on discharge. In the year two hundred and twelve an emperor granted it to virtually every free inhabitant of the empire at a stroke. Ancient writers say he did it to widen the tax base, and that may well be true, but it completed something: by then a man from Spain, Africa or Syria could hold any office, and several emperors did come from the provinces.
The system's weakness was that it had never solved the succession. There was no rule. An emperor might adopt an heir, or father one, or be murdered by the guard that was supposed to protect him. As long as the transfer worked, the machine ran. When it stopped working, everything else was exposed.
It stopped working in the third century. Over about fifty years from the two hundred and thirties there were dozens of claimants, most of them proclaimed by their own armies and killed by others, alongside plague, currency debasement and simultaneous pressure on two frontiers. The empire came back together, but it came back as a different and harder thing, and that is the subject of the next chapter.
Firm: the physical record. Roads, aqueducts, forts, harbours, milestones and the layout of towns are all still there to be surveyed. Roman law survives in a vast edited compilation. Coinage lets us track debasement almost year by year.
Soft: the population. Fifty to sixty million is a modern estimate built on census fragments, city sizes and agricultural capacity, and serious scholars have argued for figures well outside that range, particularly for Italy. Treat any single number as a bracket.
Soft: how good it was to live in. The evidence is skewed towards the people who left buildings and inscriptions, which means towns and elites. Rural life, which is where most people were, is known mainly through archaeology and tax documents. Skeletal evidence suggests that ordinary health did not obviously improve under Roman rule and in some places got worse, which sits awkwardly beside the aqueducts.
Contested: whether the peace was worth the conquest. That is not a question evidence settles, but it is worth noticing that the most famous condemnation of Roman imperialism in ancient literature was written by a Roman senator and put into the mouth of an enemy, which suggests the Romans were capable of asking it themselves.
One correction to a common impression. The empire did not run on engineering genius so much as on organisation and on the ability to spend enormous quantities of labour. The arch, concrete and surveying instruments were mostly inherited or improved rather than invented. What was new was a state that could decide to build the same standard of road for eighty thousand kilometres, and then do it.