Why Religions Spread From the Top Down

Constantine. Vladimir of Kyiv. Clovis. Ethelbert of Kent. Trdat of Armenia. The kings of Ghana and Mali. Mindaugas of Lithuania.

The pattern is so consistent it is nearly a rule: a ruler adopts a faith, the court follows, the aristocracy follows, and the countryside follows over a period measured in generations or centuries. Mass conversion from the bottom up is the exception, and it describes early Christianity before Constantine and not a great deal else.

Which means the question to ask about any conversion is not what the population came to believe. It is what the ruler was buying.

The answer is usually a bundle of very concrete things. A scriptural religion arrives with a literate class: clerks who can write, keep accounts, draft laws and maintain archives, trained at somebody else's expense. It arrives with a body of law, and with a chronology, and with diplomatic standing: a Christian king in the Latin alphabet is legible to, and marriageable with, every other Christian king. A pagan one is not.

And it arrives with a legitimacy that does not depend on ancestry, which for a consolidating monarch is the decisive item. In a traditional aristocracy your claim rests on descent, and the families you are trying to subordinate have genealogies as good as yours. A universal god is not anybody's ancestor, which means anointing by his representative outranks everybody's grandfather. This is why the conversions cluster around exactly the moments when rulers are trying to turn a first-among-equals position into a monarchy.

Note what this does not require. It does not require the ruler to be insincere (Vladimir and Constantine may well have believed every word) and it does not require the population to believe anything at all at the point of conversion. Belief arrives afterwards, slowly, and in a form nobody quite intended.

That form is syncretism, and it is the normal mechanism rather than a failure of it. Local deities become saints. Festivals keep their dates and change their names. Holy wells stay holy. Practices with no scriptural basis anywhere become, locally, what the religion is.

Missionaries frequently encouraged this on purpose. Gregory the Great's letter to Augustine of Canterbury is explicit: do not destroy the temples, convert them, because people return to places they are used to. The result is that popular religion almost everywhere is layered, which is also why reform movements keep appearing, every few centuries, to demand a return to something purer, and why they find so much to object to.

The Islamic case is the one that most people get wrong, and the correction is striking. The conquests were extraordinarily fast: within a century of Muhammad's death, from Spain to Central Asia. Conversion was extraordinarily slow. Richard Bulliet's analysis of naming patterns in biographical dictionaries suggests Iran, Iraq, Syria and Egypt did not reach Muslim majorities until something like the tenth century: two to three hundred years after the armies arrived.

Part of the reason is fiscal. Non-Muslim subjects of the early caliphates paid the jizya, and a state whose revenue depends on non-Muslim taxpayers has a direct interest in not converting them all. There are recorded instances of governors discouraging conversion for exactly this reason. It is a clean case of two motives pulling against each other, and the revenue one slowing what the religious one would have accelerated.

The last thing worth understanding is why rulers so reliably ended up fighting the institution that legitimised them.

Because the institution that can legitimise can also delegitimise, and because it is not merely a chaplaincy. A medieval church owned a very large share of the land, ran its own courts with their own jurisdiction, collected its own revenues, appointed its own officers, and answered to an authority outside the kingdom. That is a rival power with a religious costume.

The Investiture Controversy (fundamentally a fight over who appoints bishops, and therefore over who controls a third of the kingdom's wealth and its literate administration) ran for fifty years and produced the spectacle of an emperor standing in the snow at Canossa. The same structural conflict recurs wherever a religious hierarchy has an independent material base, from Byzantine iconoclasm to the Ottoman management of the ulama to the English dissolution of the monasteries.

It looks like a quarrel about God. It is a quarrel about jurisdiction, and it is the natural consequence of the deal the ruler made when he converted.