Saturday, March 26, 2022

Christians minting Muslim coinage (and vice versa)

An Islamic gold dinar minted by Offa, a Christian king of Mercia [source]

If you follow me on Twitter, you may have noticed that I've been more excited about ancient coins than I usually am. The gold coin pictured above, for instance, was minted by Offa, an Anglo-Saxon king of Mercia, in the late 700s. Offa was a Christian, but his coin contains the words "there is no God but Allah alone." In this blog post I'm going to bring together a bunch of my tweets and explore why rulers sometimes produced coins that contained icons and text dramatically at odds with the culture to which they belonged.

Money is like language. Both are characterized by network effects, or lock-in. People living in a geographical location grow up speaking a certain language, and since all their friends, family, neighbours, and colleagues also speak it, it's almost impossible for a new language to get any sort of traction.

That's why the artificial language Esperanto has never taken off. Designed in the 1800s by L.L. Zamenhof to be easy to learn, only 2 million or so people speak Esperanto. Why bother learning a new language, even an easy one, if everyone you know is already speaking English, or Chinese, or Swahili?

The same goes for money. People grow up 'speaking' a certain monetary language: dollars, yen, pounds, hryvnia, whatever. Every Canadian, for instance, internalizes a full array of Canadian dollar prices in their minds: it costs $50 to fill up with a tank of gas, $1 to buy a chocolate bar, $150 to renew a driver's license, etc. And that's why it's difficult for new monetary units to intrude where an old one already dominates. Why adopt a monetary version of Esperanto when speaking in Canadian dollars comes so naturally?

Ancient coins are a good way to illustrate the idea of money as a locked-in language.

Islam emerged in the 7th century A.D., quickly moving from the Arabian peninsula into Africa and Spain. The heretofore dominant Byzantine Empire, with its base in Constantinople, lost Syria and Egypt to the emerging Islamic caliphate.

The Byzantines and the Romans before them had a long history of issuing the solidus, a high-quality gold coin. The solidus was a bit like the U.S. dollar of its day, circulating widely across Europe and the Middle East. Below is a Byzantine solidus from 638 AD, issued by Heraclius, who is pictured with his two sons. Their crowns and staffs have crosses on them, and the reverse side displays a large cross on top of a set of steps.

Byzantine solidus issued by Heraclius, 638 AD [link]

Having conquered such a huge territory in just a few years, the Muslims may have tinkered with the idea of issuing their own entirely new coins replete with Arab text and imagery. But they chose not to, at least not at first. Bowing to pragmatism, their first gold coins were replicas of Heraclius' solidus (below) dated to around 680 AD. That is, a Muslim ruler took the strange step of minting coins that portrayed three Christian emperors. The major difference between the 638 AD Byzantine solidus and the 680 AD Islamic imitation is that it has been "de-christianized" the crosses had been removed.

Islamic imitation solidi, est 680 AD [link]

You can imagine the tradeoffs that the Muslim invaders may have been making when they designed the coin. By creating a near replica of the Byzantine solidus, they were probably trying to harness its network effects and get the first Islamic gold coin into circulation; and by making only a few small changes removing the crosses they wouldn't be compromising too much on their religious values. According to numismatist Clive Foss, these changes may have been too much. Opposition to the crossless coins among the local population was strong enough that the authorities stopped minting them soon after. Foss speculates that's why archaeologists have only found a handful of crossless dinars.

It wasn't until 697 AD that the first fully Islamized gold coins were finally issued, pictured below:

Umayyad gold dinar minted in  Syria in 697 AD [source]

Like the Byzantine solidus that preceded it, the gold dinar circulated outside the Islamic world. Examples have been found as far as Scandinavia, Russia, and England. And like the Byzantine coinage before it, Islamic coinage was copied. Caitlin Green has a very good blog post showing how Anglo-Saxon Mercia (see image at top) and the Carolingian Empire issued their own imitation gold dinars.

The caliphate's silver dirhams were copied, too. According to Green, the first coins issued by Kievan Rus (below) were probably imitation silver dirhams to which a cross and bird were added. "The fact that the first Rus' coinage imitated an Islamic dirham is interesting, although perhaps unsurprising, given that many millions of dirhams appear to have been imported into northern and eastern Europe by the Vikings/Rus' as a result of their trade with the Islamic world (perhaps primarily involving slaves)," writes Green. 

Put differently, don't fool around with a good thing. Copy it with just a few modifications.

A 'Christian falcon' imitation dirham issued by the Kievan Rus' in c. 950 [source 1][source 2]

One reason that European kings may have been particularly eager to copy Islamic gold coins is that while silver coins were common the continent didn't have a long history of minting their own gold coinage. By imitating already-dominant gold dinars, Europeans could coast on the network effect of those coins. In the case of Offa's gold dinar, it may have been designed as a trade coin, similar enough to the original that it would be accepted in southern Europe.

Another example comes from Spain. By the 13th century the Almohads the last Muslim empire to rule Spain had been driven back to North Africa. The monetary influence of Islamic coins was such that the Christian conquerors continued to issue dinar-style gold coins, or maravedis. Around 1180 AD, Alfonso VIII of Castile, a Christian, minted maravedis with Arabic script rather than Latin, "Christianizing" them by inserting a small cross and changing the Arabic message to a Christian one:

Notice that in the previous examples, the new coin issuer used the incumbent coin's style while dropping its religious motifs. But the desire to latch on to the strong network effects of existing coinage sometimes trumped even the demands of religion.

As an example of this, when the Crusaders took over Jerusalem they issued dinar-style gold coins without even bothering to add Christian embellishments (see below). Initially minted in the late 1100s, the Crusader dinars only gained crosses in 1250 when an angry pope pressured the local princes to eliminate coinage which was "not well-formed enough to conceal the scandalous fact that its inscriptions praised Mohammed and bore dates of the Muslim era." Against the pope’s objections, the text on the new Christian design was still written in Arabic. So the idea that money is sticky and locked-in like language may explain some of the strange coins pictured above. It may also explain some of the phenomenon around us today, like:

1) why stablecoins have come to dominate the cryptocurrency space, and why the dominant stablecoins are U.S. dollar stablecoins and not, say, Turkish lira stablecoins. (Bitcoin and other volcoins are awful candidates as unit of account, whereas stablecoins are not only stable, but also benefit from the network effect of existing currencies, none of which is stronger than the U.S. dollar.)

2) why Facebook's original idea to create a new Libra unit of account was crazy overambitious.

3) why Canadians still use the Canadian dollar despite being right next door to the world's mightiest monetary power. We grow up speaking in loonies and toonies, and that's not something that is easily changed. It also explains why Quebecers sometimes refer to the dollar as a 'piasse,' which derives from the piastre or peso, and old coin that once dominated North American trade. The network effects of the long departed Spanish peso were so strong that shades of it still exist.


  1. great post. We think of money (currency) as a protocol (information unit) $, yen, pound, Thaler etc. which is expressed in various media (paper, metal, electronic). Using a network model of protocol acceptance gets interesting valuation metrics. Protocols are positive economic externalities the more use/ acceptance, the more utility.

  2. Why does the influence of the Norman conquerors' French on the English language spring to mind? How quickly would Canadians switch to US dollars, if their rulers told them to?

    1. The euro is a good example of a switch. It required years of planning and careful orchestration by governments, but in the end it worked very smoothly. The euro somewhat mimicked each domestic unit of account via its fixed peg to the lira, markka, franc etc, which may have helped it coast in on the incumbent's network effects.

  3. Wouldn't the Islamic solidus with it's depiction of human figures violate the Islamic prohibition on depicting human figures (because of idolatry).

    1. Yes, I think so. See:

      "A remarkable fact about the coins used in the Islamic Caliphate until about 651 CE is that they bore no Islamic text or symbolism at all. Instead, they had the portraits of the Byzantine and Sassanian emperors and explicitly religious Zoroastrian and Christian symbols – the fire altar and the cross! Thus, paradoxically, the coins used during Islam's most sacred period violate what came to be regarded as one its strictest edicts – the proscription of figural images and un-Islamic symbols. These coins depicting non-Muslim emperors and symbols must have filled the bayt-al-māl (treasury) of the most doctrinally iconoclastic state in history, and the earliest believers whose faith taught them to abhor such images must freely have used these coins in ordinary commerce. Or perhaps these were pragmatic people, and not quite as rigid and doctrinaire in these matters as later generations would make them appear."