The currencies of late antiquity were dominated by two substantial if quite distinct and partly competing coinages, the late Roman solidus, the leading gold denomination of the period, and the Sasanian drahm/drachm, the dominant silver one. The solidus and the drahm/drachm were each the exclusive standard of their respective monetary systems, so that, viewed as a whole, the currencies of late antiquity rested on a sort of ‘double monometallism’ where the subsidiary coinages were simply ‘representatives’ of the leading metal. This system, if we can call it that, survived for over three centuries till the Arab conquests, when it was subsumed into a new and more expansive empire whose monetary traditions were largely influenced by those of late antiquity.Footnote 1
In this chapter I shall mainly focus on the solidus but argue that its evolution was characteristic of trends in late antiquity as a whole, that is, including the Sasanian Near East, and therefore conclude by looking at the drahm and its wider circulation along the eastern trade routes. The basis of the late Roman monetary system was a gold denomination struck at 72 to the pound. Assuming the conventional figure for the weight of the Roman pound, 327.45 g, is correct, then each full-weight piece would have weighed 4.5479 g.Footnote 2 Three features are remarkable about this coinage. First, its prevalence as the exclusive monetary standard broke with centuries of monetary tradition that had sought to run a viable bimetallic system, on the whole without success. Second, it was struck to an exceedingly high standard of fineness, especially after the monetary reform of the late 360s when new regulations were issued for the form in which taxes should reach the Treasury. As the late John Kent noted, ‘late Roman gold coins were as fine as it was possible to make them’.Footnote 3 Finally, it was struck in substantial quantities and circulated very widely indeed.
For the circulation of the solidus we have not just the testimony of the gold hoards, of which a considerable number survive, though not always in their integral quantities, but also the more local day-to-day evidence contained in the numerous contracts, accounts and other kinds of documentation that survive in the papyri of the fourth to seventh centuries. There is also, of course, the evidence of the historical sources, including numerous references to the very substantial payments made in various contexts – donations to the church, payments to tribal federates, consular largesses and so on. What all of this evidence demonstrates is that there was plenty of gold in the empire, a massive stock of currency, so that if it can be shown that the reserves in the state’s hands showed a declining trend over time (and I think this is certainly suggested by the Roman/Byzantine evidence), the explanation cannot be found in an external flight of goldFootnote 4 but has to be sought elsewhere. My own argument will be that much of the gold ended up in private hands, where it was both actively used and occasionally hoarded.
Most scholars today would agree that late antiquity was a period of considerable monetary wealth,Footnote 5 yet the wider implications of this fact for economic history remain subdued. A major reason for this is that the gold coinage is viewed primarily in fiscal terms, so that even though ‘the monetary wealth of the empire … is conceptually a much broader topic than the imperial finances’,Footnote 6 the two are ultimately conflated in discussions of the late antique Mediterranean. One extreme formulation of this view appears in John Haldon’s book on the seventh century where he says, ‘the presence of gold coins has very little to do with the existence of a market economy, being a reflection rather of the needs of the state and its military, administrative and fiscal machinery’.Footnote 7 By contrast, I want to argue that late antiquity saw a net expansion in the stock of currency (gold in the Mediterranean, silver in the Near East) and that this was largely fuelled by an expansion of trade. Money was needed in circulation as the volume of business expanded. (The theoretical assumption here is that money circulates to ‘realise’ prices and the quantity of money in circulation is therefore determined by the level of prices, once the volume of transactions and velocity of circulation are given, or by the scale of business – number of transactions × price – if the velocity is taken as fixed. Since late antique prices were generally stable in terms of gold, this equation reduces to the volume of transactions and velocity of circulation.)
Monetary expansion in the fourth to seventh centuries
How would one go about constructing such an argument? Die studies of the gold coinage are few and far between, so we have almost no precise estimates of mint output (one of the variables in the ‘stock of currency’) and are forced to rely on some combination of a less coherent body of numismatic data with documentary sources (such as the papyri) and the few scraps of textual evidence that can be drawn directly into the argument. Among late Roman texts one of the most valuable is a fascinating passage of the mid-fourth century tract called De rebus bellicis (‘A Tract on Defence’). The author, perhaps a disillusioned bureaucrat, tells us that Constantine’s move to a new monetary standard (the solidus was first issued at Trier c.310 and not produced in the eastern mints till 324, after which it was struck at ‘almost all the mints throughout the empire’)Footnote 8 was bound up with a massive dishoarding of gold that led to the (renewed) accumulation of monetary wealth in private hands and sparked a veritable ‘passion for spending gold’. The author also links the expanding circulation of gold to powerful social changes. ‘This store of gold meant that the houses of the powerful were crammed full and their splendour enhanced to the destruction of the poor …’.Footnote 9 The reference here is almost certainly to Constantine’s expansion of the governing class of the fourth century, a process continued by his successors that would lead in time, several decades later, to the emergence of a new, more thoroughly Christianised aristocracy in the east, distinct from the older and more traditionalist senatorial families who continued to dominate the central Mediterranean basin.
In the modern historiography of the late empire, the perspectives of the ‘Anonymous’ have had a profound influence in and through the work of Santo Mazzarino, for Mazzarino saw the history of the monetary system as central to the evolution of the late empire.Footnote 10 One clue to how this process worked is that if we look at the first batch of Constantine’s new gold denomination, it turns out that ‘The reverses mainly have a military flavour, such as GLORIA EXERCITVS GALL …’.Footnote 11 The late third century established a tradition of substantial monetary payments to the higher echelons of the late Roman army and within this general movement a strong preference for gold, consolidated by the chaos of third-century bimetallism and the collapse of the silver coinage.Footnote 12 By the first decade of the fourth century, high-ranking officials such as the owner of the Beaurains hoard were in regular receipt of substantial sums in gold in the form of military bonuses (donativa) and unlikely to be willing to forego these payments.Footnote 13 Diocletian’s reform of c. 294 was the last significant attempt to salvage bimetallism but its whole tendency conflicted with the aspirations of the officer corps (militares) and its failure must have been decisive in Constantine’s decision to move to a gold standard.
The relative scarcity of Constantine’s own solidi in the hoard material coupled with their rarity in the collections and sales catalogues has prompted the view that solidi were not struck on any substantial scale till around the middle of the fourth century, well after his reign.Footnote 14 But this is a matter of perspective. It is only true in the limited sense that gold was struck on a truly gigantic scale from the middle decades of the fourth century, dwarfing previous levels of circulation. Constantine produced vastly more gold than the Tetrarchs had ever done,Footnote 15 and his donations to the church imply both substantial hoards of gold and silver and an extensive circulation of solidi in his reign.Footnote 16 The stocks that third-century emperors had hoarded in the TreasuryFootnote 17 were rapidly released into circulation, supplemented by other sources such as temple treasuresFootnote 18 and private hoards.Footnote 19 The forces driving the expansion of the gold coinage were first dealt with by Mickwitz in his book on the monetary economy of the fourth century. From this it was clear that commutation (adaeratio) was a central issue. But Mickwitz saw the mass of landowners (taxpayers) wanting commutation (of taxes) and the bureaucracy resisting commutation (of salaries).Footnote 20 For him inflation was the key factor that explained this pattern of behaviour. There was always the possibility that the gold coinage would be subject to debasement, as the silver had, repeatedly, in the third century.Footnote 21 Mazzarino’s critique of this position developed a more realistic understanding of the late empire. The government was firmly committed to the stability of the solidus precisely because a large and increasing portion of its own revenues and of the emoluments of the bureaucracy were drawn in gold. In fact, officials systematically speculated on the stocks of foodgrains, using commutation and the kinds of arbitrage it allowed.Footnote 22 The general implication of Mazzarino’s critique is of course that the late Roman bureaucracy was a major standard-bearer of the revitalised monetary economy of the fourth century. Commutation was widespread by the early fifth, despite government resistance, and more and more gold had to be churned out to sustain this system.
Analyses of the trace elements in a sample of solidi suggest that the middle decades of the fourth century saw a huge expansion in the volume of coinage struck, sustained by the discovery of fresh mineral sources in the Balkans.Footnote 23 In a pathbreaking dissertation, Jacques Poirier suggested that the ‘monetary mass’ (meaning gold) multiplied by a factor of twenty in the years between 346 and 388.Footnote 24 Whatever the precise forces it saw itself responding to, it is clear that government saw the provision of liquidity as a major monetary objective.Footnote 25 All the other evidence suggests that these were decades of significant economic expansion, with a great deal of money flowing into both agriculture and trade. If the stability of the gold coinage mattered for its fiscal objectives and to the stability of contracts, liquidity was crucial to the level of monetary activity in the economy as a whole. The solidus was a ‘hard’ currency, in the sense that government made sure that it would tend to rise against other currencies.Footnote 26 Gold coins were allowed to find their value in the market in the sense that the exchange rate beween solidi and silver or the base metal currency constantly fluctuated.Footnote 27 People spoke of ‘buying’ and ‘selling’ solidi, much as we do of dollars, pounds and euros. The point to note is that the authorities wanted both consistency and depth in this market. Thus one of the earliest regulations stated that ‘All solidi on which Our face and venerability is to be found are to be valued and sold at one price (uno pretio aestimandi sunt atque vendendi), however diverse the extent of the image.’Footnote 28 The solidus was money, not a piece of metal, and the rejection/discounting of coin on grounds of module or age was strenuously opposed, in part because it undermined the distinction between money and metal.Footnote 29 The workings of the economy required a properly functioning market in gold coins of various denominations, and the regulation of such a market was a consistent objective of government policy.
So, how much liquidity? For lack of the kind of numismatic work that alone can yield reliable estimates of the scale on which individual mints, and Constantinople in particular, struck gold at various times in the later fourth to seventh centuries, our next best option is to scour the documentary sources for reasonably reliable estimates of the amount of gold in circulation. Monetary historians are agreed that the stock of currency was ‘very large indeed’,Footnote 30 there was ‘plenty’ of gold in circulationFootnote 31 and that there was a ‘vast’ coinage in gold.Footnote 32 What the documentary sources can do is establish an order of magnitude to give these descriptions a less intuitive feel.
The solidus was used for a very wide range of transactions, in fact for all but the most trivial exchanges.Footnote 33 The strongest evidence of this is its use in the East Mediterranean countryside, where, in Egypt for example, the papyri show a regular use of gold in all kinds of transactions from loans, rents and taxes to wage payments and rural fines. By the sixth century, commutation of taxes was widespread, undermining any rigid distinction between the fiscal and the economic functions of the coinage, and most producers would have watched the movement of prices carefully.Footnote 34 Anastasius’ ability to generate reserves of 23 million solidiFootnote 35 was inseparable from the reforms in taxation that encouraged further commutation in his reign. Justinian consolidated this trend, as Khusro I would do in relation to Kavād’s seminal reform of Sasanian taxes. In both Byzantium and Persia the main part of the sixth century was the highwater mark of monetary economy, and what is remarkable is how far this was sustained into the seventh century. All of this implies substantial liquidity in both monetary systems, the ability to collect cash revenues in terms of an expanding volume of commercial exchanges. One family of the elite aristocracy, the Apions, generated cash revenues of eighteen to twenty thousand solidi from their estates in one district alone.Footnote 36 Egypt paid 2.6 million solidi a year under Justinian, against an aggregate (cash only?) assessment, for all regions, of over 7 million solidi (100,000 lb of gold).Footnote 37 It is particularly remarkable that the papyri show no signs of monetary recession in the early part of the seventh century. It is this that explains why both the Sasanians and the Arabs could still extract huge sums in gold at various times from 620 to 650. Documents from the Oxyrhynchite show the Sasanian official nicknamed Shahrālānyozān (probably Shahrvaraz, Khusro II’s commander for the western region)Footnote 38 extracting substantial amounts of gold – probably a total of 250 lb from Oxyrhynchus and Cynopolis for the single indiction of 623/4.Footnote 39 Again, on the most conservative estimate – that cited by al-Balādhurī – ‛Amr ibn al-‛Āṣ who governed Egypt in the early 640s was able to impose a tax of 2 million ‘dinars’ on its people.Footnote 40 And in 647, the aristocracy of Byzacena paid the invading forces of ‛Abd Allāh b. Sa’d b.Abī Sarḥ an amount that different traditions variously put at 300 ‘qintars’ of gold (2.16 million solidi) or 2½ million solidi.Footnote 41 In this case, the link to large-scale commerce is explicit. In one of the least embellished accounts to survive, ‛Ubayd Allāh tells us that following the defeat of Gregory’s forces, ‘The Muslims saw that their [the Byzantines’] wealth consisted for the most part of gold. They asked them about this gold and how they had come by it. “By selling olive-oil” was the answer.’Footnote 42
The least one can say on the basis of such figures is that the richest parts of the late antique world had very substantial stocks of gold (and silver) currency running into millions of solidi (and drachms). I want to argue, first, that this was equally true of the Sasanian Near East where the seventh-century trend was in some sense even stronger, and second, that the ageing of the currency probably accounted for much of this accumulating monetary mass, in both Byzantium and Persia. In what follows I shall assume that the theoretical weight of the drachm was 4.2 g of pure silver (or silver of very high fineness, over 95%).Footnote 43
In the early twelfth-century Persian monograph called the Fārs-nāme there is a valuable section on the revenues of Fars (south-west Iran) where the author (conventionally called Ibn al-Balkhi) tells us, ‘When … Anushirvan established his land-tax (kharāj) in all his kingdoms, the land-tax of Fars amounted to 36 million [silver] dirhams equivalent to 3 million [gold] dinars.’Footnote 44 ‘Dirham’ in this context should be understood to mean dirham al-wāfī (‘full-weight’), i.e. the Sasanian drachm whose theoretical weight was that of the later Arabic dinar, that is, one ‘mithqāl’.Footnote 45 The veracity of this figure (36 million drachms) is supported by Ibn Khurradādhbih’s corresponding numbers for the Sasanian assessment of Kerman, Khuzistan and Fars. Here we are told that in ‘imperial times’ (a vague expression but generally used of Khusro I), Fars was assessed at 40 million ‘dirhams mithqāl’, Kerman to its east at 60 million and Khuzistan to its west at 50 million.Footnote 46 Since Kerman, Fars and Khuzistan formed the ‘southern quarter’ of the empire in Khusro I’s reign (according to the scheme preserved in al-Tha‛ālibī),Footnote 47 it seems likely that in the sixth century the region as a whole was assessed for a total of 150 million drachms. As it turns out, the Sawād (Lower Mesopotamia/southern Iraq) was assessed for precisely this amount, 150 million drachms, in (probably) the second reign of Kavād I (498–531), the earlier part of the sixth century.Footnote 48 If the Sawād was Khusro’s kust ī xwarvarān or western quarter, then between them the densely populated south and west, the most developed regions of the empire, would have accounted for a gross revenue of 300 million drachms (assuming Khusro retained Kavād’s assessment of the west). Ibn Khurradādhbih also cites a figure for the assessment of the northern provinces. This was substantially less than the figure for the south, only 30 million drachms.Footnote 49 Suppose Khurasan, the remaining quarter, contributed at least as much as this, viz. 30 million drachms, the aggregate claim may well have been in the region of 360 million drachms, divided, presumably, into three instalments of 120 million each.Footnote 50 This of course is hypothetical but it does indicate a broad order of magnitude for the main part of the sixth century that is not totally out of line with the one exact estimate we do have from the late Sasanian period for the amount of taxes actually collected some decades later, namely, the 420 million drachms (mithqāls) that Khusro II collected in the eighteenth year of his reign, i.e. 607/8, before the huge expansion of reserves that characterised the latter part of his rule.Footnote 51 According to the figures reported by al-Ṭabarī, the level of reserves stood at 800 million drachms in 602 and had escalated to 1.6 billion drachms by 620, at the end of Khusro II’s thirtieth year.Footnote 52 The vast quantities of plunder that Khusro II collected in the occupied Byzantine territories of the Near East (120,000 lb of silver from Edessa alone, in 622)Footnote 53 is one obvious explanation of this staggering level of reserves, but he is also credited with periodic and sharp increases in taxation which led to a breakdown in his relations with the aristocracy. Even if we suppose that annual tax revenues ran into 400 to 500 million drachms and did not exceed that level, as they probably did, the general conclusion that seems to follow from all this is that both mint output levels (production of currency) and the turnover in agrarian and commercial markets would have had to be enormous to support these levels of taxation.Footnote 54
The Sasanians struck their silver in a far greater multiplicity of mints than the late Romans their gold, most situated in provincial capitals.Footnote 55 What is remarkable about Sasanian monetary history is the clearly discernible expansion in mint output that appears to have occurred in the late sixth and early seventh centuries. As Sears states, ‘Hoards document a staggering increase in the output of drachms during the latter part of the sixth and the beginning of the seventh centuries A.D.’Footnote 56 Thus the coins of Hormazd IV (579–90) are ‘among the commonest in the whole Sasanian series’,Footnote 57 and Khusro II’s coinage in silver was massive.Footnote 58 So the numismatic evidence reinforces the impression culled from the textual sources that the late sixth/early seventh centuries saw considerable monetary expansion, at least in the Sasanian empire.
Justinian is said to have depleted the reserves built up by his predecessors.Footnote 59 When Justin II came to the throne, he seems to have used his personal resources to clear the debts that Justinian had run up with a whole throng of private bankers. In the verse panegyric written to celebrate Justin II’s accession, there is a dramatic description of this scene, in which Corippus makes the bankers exclaim: ‘You know, greatest of rulers, how much benefit your treasury has from business. If our strength fails, whence will come the annual tributes for your resources?’Footnote 60 If we believe the bankers, then ‘business’ was clearly central to the monetary wealth of the late empire. However, before coming to this I would like to suggest that the continued circulation of older coin was a major factor in the steady expansion of the stock of currency that took place between the later fourth and seventh centuries.
Now this conflicts radically with the view that the late Roman monetary economy involved an ‘incessant recoining of the gold’. According to Grierson, ‘It is probable that the gold brought in through taxes was melted and re-struck before being reissued … This incessant recoining of the gold was in large measure responsible for the success of the government in maintaining the standard uniform over so many centuries.’Footnote 61 So too Kent who argued in a famous paper, ‘The chance of any individual piece circulating for long was slight. If it escaped recoining, it must have gone quickly to ground.’Footnote 62 These views are now largely outdated. Michael Metcalf has argued, ‘Kent’s emphasis that gold coins enjoyed a short life in circulation, before being clawed back into the imperial coffers by the ruler’s device of insisting that taxes should be paid in gold, may possibly be true of other parts of the Empire, but it would not be an obvious conclusion to draw from the Illyrican finds.’Footnote 63 Harl notes that ‘Annual reminting would have been wasteful … Solidi of good weight, regardless of age or the effigy they bore, were the equivalent of new coins.’ Moreover, ‘the wear on numerous specimens indicates that they had a long circulation life’.Footnote 64
In fact, there is a great deal of evidence to suggest that older coins continued to circulate, and that a considerable if uncertain, and probably fluctuating, portion of the currency comprised such pieces. The emperors repeatedly sought to regulate public responses to the currency in circulation, seeking to combat a widespread public prejudice against older solidi. This is best reflected in Valentinian III’s Novel 16, dated 445, from which it is clear that much older coinage was in circulation at that time and that ‘buyers’ were refusing to accept these pieces except at a discount, obviously because they were perceived to have lost weight through circulation.Footnote 65 The same kind of regulation had to be passed by Valentinian and his brother in the mid-fourth century, on the eve of their reform of the coinage,Footnote 66 and a famous edict of Justinian I, dated 559, shows that, again in the sixth century, in regions with a large and active circulation of money such as Egypt, the problem persisted, and was clearly disruptive of business activity. Justinian accused the coin-weighers and chrysones, probably bankers rather than tax officials, of discounting worn solidi, called ἀπόλυτον χάραγμα, i.e. ‘loose coin’, by as much as 12½ per cent and described this commission as exorbitant.Footnote 67 Papyri survive to show that what the edict did succeed in doing was reducing, not eliminating, the discount to 6¼ per cent.Footnote 68 Finally, the stipulation that solidi should be melted down and re-struckFootnote 69 only applied to the gold that was actually pulled in through taxation, and was in any case no longer being followed by the early fifth century, when, as Delmaire points out, recoinage was no longer automatic or compulsory.Footnote 70
The coinage itself supports this general counter-argument. The so-called ‘circulation’ or ‘emergency’ hoards, that is, those hoards that can plausibly be viewed as random samples of the currency in circulation,Footnote 71 tend to show extended age structures (significant numbers of older solidi). At least four of the hoards in Table 6.1 date from the period after the Arab conquests, when the currency must have aged rapidly, but the general situation is sufficiently clear even from the other hoards listed in the table.Footnote 72
Gravisca Italy | Xanten Germany | Ain Meddah Algeria | Djemila Algeria | Castellana Sicily | Chatby Alexandria | Es-Sermita Tunisia | Rougga Tunisia | Nikertai Syria | South Jordan | Bet She’an Israel | Yildiz Palace 1 Turkey | |
---|---|---|---|---|---|---|---|---|---|---|---|---|
c.408 | c.428 | c.496 | c.496 | c.538 | c.611 | c.615 | c.647 | c.681 | 680s | 680s | 685/95 | |
% solidi in circulation | ||||||||||||
for over 33 years | 14.9 | 46.7 | 33.3 | 29.6 | 66.6 | 17.4 | 54.5 | 31.3 | 65.5 | 53.5 | 63.5 | 51.9 |
Constantius | 1 | |||||||||||
Valentinian I | 26 | 47 | 3 | |||||||||
Valens | 33 | 3 | ||||||||||
Gratian | 1 | |||||||||||
Valentinian II | 29 | 7 | ||||||||||
Theodosius I | 29 | 11 | ||||||||||
Arcadius | 39 | 38 | ||||||||||
Honorius | 51 | 55 | 3 | |||||||||
Theodosius II | 1 | 19 | 34 | 2 | ||||||||
Eudoxia | 1 | |||||||||||
Pulcheria | 1 | |||||||||||
Valentinian III | 18 | 1 | ||||||||||
Marcian | 8 | 16 | ||||||||||
Majorian | 1 | |||||||||||
Libius Severus | 1 | |||||||||||
Anthemius | 1 | |||||||||||
Leo I | 26 | 51 | 2 | |||||||||
Julius Nepos | 1 | |||||||||||
Leo II & Zeno | 2 | |||||||||||
Zeno | 29 | 58 | 4 | |||||||||
Basiliscus | 3 | 11 | ||||||||||
Basiliscus & Marcus | 1 | |||||||||||
Anastasius I | 2 | 2 | 2 | |||||||||
Justin I | ||||||||||||
Justinian I | 2 | 4 | 12 | 1 | ||||||||
Justin II | 9 | 2 | ||||||||||
Tiberius II Const. | 2 | 4 | ||||||||||
Maurice | 43 | 8 | 1 | 3 | 1 | 3 | ||||||
Phocas | 21 | 2 | 83 | 57 | 32 | 95 | 41 | |||||
Heraclius | 5 | 5 | 121 | 276 | 90 | 382 | 142 | |||||
Constans II | 63 | 155 | 80 | 219 | 96 | |||||||
Constantine IV | 22 | 27 | 55 | 61 | ||||||||
Justinian II | 16 |
The late empire’s vast coinage in gold coupled with the tendency for older issues to remain in circulation (subject, of course, to natural wastage) is, I believe, the key to explaining a feature of the monetary history of the seventh century which has not been sufficiently emphasised – the extraordinary fact that an Arab gold coinage did not evolve for some five decades after the conquest of Syria. As Michael Bates points out, ‘The remarkable aspect of the matter is not that ‛Abd al-Malik instituted minting, but rather that a half century elapsed after the Arab conquest before a mint was set up in the capital of the caliphate.’Footnote 73 Balādhurī’s statement that ‛Umar imposed a poll tax on non-Muslims of 4 ‘dinars’ ‘on those who had gold’ implies that these taxes alone would have involved substantial aggregate payments in gold. Yet we know that a pre-reform gold coinage was rare.Footnote 74 It follows that the Arabs relied on the existing stock of gold currency, that is, the stock of solidi in circulation in Syria and Egypt, and that this must have been very substantial indeed to allow the early administrations to function for over fifty years without fresh stocks of currency, i.e. a regular coinage of their own.Footnote 75 For example, under the three-year peace treaty of 659, Mu‛āwiya agreed to pay Constantinople 1000 solidi a day, according to the chronicler Theophanes.Footnote 76 This payment alone would have come to over a million solidi! In July 685, ‛Abd al-Malik sought a treaty on the same terms,Footnote 77 agreeing to an annual payment of 365,000 solidi. This would have lasted till 691 when Justinian II broke the treaty – ‘foolishly’, says Theophanes – because ‛Abd al-Malik now proposed to make the payment in the new gold coinage that began to be struck from the end of 691, and this apparently was unacceptable to the Byzantines.Footnote 78
The solidus vs the drachm?
The most striking feature of the gold coinage of the late empire was its ability to combine a considerable density of value with the functions of a mass currency. The power of the late Roman state was best encapsulated in both the quality as well as the scale of circulation of its coinage, features paralleled only by the Sasanians. Unlike the Sasanians, however, the Roman state was much less dependent on external international trade, and, in any case, unable fully to exploit the potential of that trade, due, in large measure, to a conscious Sasanian policy of controlling access to the east. Whereas the purity of the Sasanian silver coinage was linked to some considerable degree to a far-flung commercial network that used the drachm as a widely accepted means of exchange, the equivalent pressure in the late Roman state was the need to provide a stable monetary basis for the vastly more affluent society that emerged in the Mediterranean in the main part of the fourth and early fifth centuries. This was a society where the aristocracy accumulated enormous sums of money, as Olympiodorus tells us, but also one where the urban economy was largely driven by the monied classes of the empire, that is, those social groups outside the aristocratic elite who drew their incomes from banking, moneylending and trade.
The revival of banking in the fourth century was integrally related to the expanding circulation of gold, as bankers diversified into money changing and began to manage the assets of ‘high net worth’ individuals through an active deployment of funds.Footnote 79 Money changing must have been a hugely lucrative profession,Footnote 80 and its integration with banking would mean that vast sums of money were available to finance the expansion of international trade from the later fourth century. Ambrose depicts the typical merchant of the late fourth century as ‘always on the go, day and night, amassing piles of treasure, or accumulating great new stocks of merchandise’.Footnote 81 This obviously refers to the more substantial merchants, such as Constantinople’s large-scale importers of raw materials such as iron and copper (to hekastou eidous parektikon, hoion sideroteleis, khalkoteleis) mentioned in the mid-sixth-century treatise called Peri stratēgias,Footnote 82 or those who financed and organised trade with the east.Footnote 83 There is good textual evidence of the flourishing of this trade in the fourth century – the repeated reference to wealthy traders in the urban centres of the Near East carrying on a substantial business (negotia maxima) in goods imported from India and China, some at the great annual fair of Batnae, near the headwaters of the Balikh,Footnote 84 others, probably the majority, via Persia,Footnote 85 and the interesting reference in the Hou Han shu, written c.430, to the inhabitants of Ta-Ch’in ‘trading with Persia and India and gaining great profit from their dealings’.Footnote 86
What is abundantly clear from the late antique evidence, however, is that by the early sixth century the Indian Ocean trade was largely in the hands of other groups. The formal proof of this is Cosmas’ own assertion that because of her location Taprobanē was ‘much frequented by ships from all parts of India and from Persia and Ethiopia, and it likewise sends many of its own’,Footnote 87 and the explicit testimony of Procopius that Justinian had to rely on the rulers of Aksum to try and circumvent the Sasanian monopoly of silk.Footnote 88 The clear implication of these texts is that Roman traders were no longer a significant force in the Indian Ocean.Footnote 89 Egeria’s fascinating description of Clysma in the later fourth century distinctly implies that Indian merchants controlled a substantial part of the trade directed to the Red Sea.Footnote 90 There is a resonance of this in the early fifth-century Pelagian diatribe De divitiis, where the author tells his largely aristocratic readership, ‘No one coming from India or Arabia or Egypt on business and intending to spend a short time at Rome or in any other place acquires houses or possessions there for himself or other property which seems immovable, but only gold or silver or merchandise of different kinds or anything which he is able to take back with him to his own country.’Footnote 91 Here the implication is even stronger, namely, that unlike the Sasanians the late empire did not shut its gates to merchants from the east, and thus, by this stage at least, they could carry gold back with them, despite the formal restrictions imposed by Valens some decades earlier. The Arabs too were active traders in the Indian Ocean. From Fa-hsien’s account of his travels in India and Sri Lanka, it emerges that in the early fifth century there was a colony of Sabaean merchants settled in the chief city of the island.Footnote 92 This reference should be read with Ammianus’ description of the flourishing state of the southern coast of Arabia in the latter part of the fourth century, with its ‘dense clustering of commercial ports’ and ‘numerous safe harbours’.Footnote 93 The argument is not that Byzantine vessels did not ply these waters but that they must have been a minor part of the traffic.
The decisive factor, of course, was the growth of a powerful merchant class in Sasanian Persia, which drew its wealth from the luxury trades (Khusro II’s treasures included a ‘vast quantity of jewels, gold and silver plate, furs, precious fabrics’)Footnote 94 and had the strong backing of the state.Footnote 95 Sasanian military expansion secured the conditions for the expansion of trade, but less through conquest than a looser, and less stable, system of control. From eastern Arabia to Sind and the Kushan domains to the north, the Sasanians constructed a series of spheres of influence which laid the basis, from the end of the fourth century, for a crushing Sasanian monopoly of the maritime tradeFootnote 96 and a large-scale circulation of the silver drachm along the terrestrial silk routes. Of 368 drachms in the remarkable Tépé Maranjan hoard near Kabul (date c. 385), 326 are of Shāpūr II.Footnote 97 Coins of Shāpūr II were abundant in the area south of the Hindu Kush,Footnote 98 and they are also the earliest Sasanian coins to appear in China or eastern Turkestan.Footnote 99 Daibul, Makran and the adjacent parts of Sind were acquired by the Sasanians under Bahrām Gōr (Bahrām V, 421–38) as part of a dowry conferred on him by the local ruler.Footnote 100 It was his reign above all that was crucial in consolidating a Sasanian commercial network that led from Sind to Balkh, and from Merv to Bukhara and the desert oases leading to Turfan.Footnote 101 Merv churned out huge quantities of Bahrām V’s drachms,Footnote 102 and these circulated north in sufficient numbers to become the prototype of the Bukhara silver drachms of the sixth century.Footnote 103 Merv was the interface of Persian and Sogdian commercial dominance,Footnote 104 and a major station in the diffusion of Nestorian Christianity into Central Asia.Footnote 105 Even the disastrous closing decades of the fifth century (the losses inflicted by the Hephthalites) did nothing to stop the active trade with Transoxiana, and it is clear that for most of the sixth and early seventh centuries Sogdian merchants used the drachm as an international currency.Footnote 106 So entrenched was the circulation of Sasanian silver that, according to Skaff, ‘silver coins were still being imported around a decade after the T’ang conquest’ of the Kao-ch’ang kingdom in 640.Footnote 107
This is the background against which we have to view the ‘intense local competition between Roman and Persian traders’Footnote 108 reflected in the story Sopatros told Cosmas. To dismiss the narrative as fictionFootnote 109 misses the point that for well over a century the Sasanians had subjected Rome’s external trade to what Callu calls a ‘tactical blockade’.Footnote 110 The Byzantine sources of the sixth century contain some passages of extraordinary value to any reconstruction of events in that period. Justinian’s attempt to use the Abyssinians to open a new line of access to the Indian Ocean was a failure, says Procopius, ‘for it was impossible for the Ethiopians to buy silk from the Indians, for the Persian merchants always locate themselves at the very harbours where the Indian ships first put in (since they inhabit the adjoining country), and are accustomed to buy the whole cargoes’.Footnote 111This was in 531, following the last of a series of Abyssinian occupations of Yemen under the nagaši Ĕlla Asbĕha but before the coup in which Abrĕha overthrew Sumu-yafa’ Ašwa’ (Esimphaios).Footnote 112 When the Sasanians recovered control of Yemen in the last years of Khusro I’s reign (in 575), the real reason, as Theophanes of Byzantium pointed out, was the formidable threat of the emerging Byzantine alliance with the Türks. It was the Türks who approached the Byzantines. The passage in Theophanes documenting this is one of the most coherent in all of Byzantine historiography, for he notes that the Türks were now (in the 560s) in control of the trade with China, having destroyed the monopoly which the Hephthalites had wrested from the Sasanians in the later fifth century.Footnote 113 Zemarchus’ embassy to the Kaghan was warmly received the following year, in 569. ‘It was this’, Theophanes says, ‘that led Khusro to launch his expedition against’ Abyssinian control of the Yemen.Footnote 114 What Photius’ summary of these parts of this late sixth-century history lacks or leaves out is any reference to the information conveyed by another historian, Menander the Guardsman, that the whole Türk mission to Constantinople arose in the first place as a backlash to the Sasanian refusal to allow the Sogdians access to the Persian and Near Eastern markets for raw silk. It was the Sogdians, says Menander, who instigated the embassy to Byzantium, advising ‘Sizabul’ (Ishtemi?) ‘that it would be better for the Turks to cultivate the friendship of the Romans and send their raw silk for sale to them because they made more use of it than other people’.Footnote 115 From this it is clear that in the late sixth century Byzantium was the world’s largest market for silk, and since Sasanian merchants charged probably well over 15 solidi per pound of this material,Footnote 116 it is easy to see why so much energy was invested in constructing and retaining their ‘monopoly’. Moreover, it is clear from Menander’s description of the kind of questions Justin II asked the Türk envoys that the Byzantines knew little about Central Asia c. 570.
If the alliance was designed to contain Sasanian competition, it was clearly a failure. The late sixth/early seventh centuries saw a veritable explosion of Sasanian trade into Inner Asia.Footnote 117 Cosmas had explained the Persian advantage in silk in terms of the differential turnaround times linked to the length of voyages by land and sea,Footnote 118 and De la Vaissière in his excellent book on the Sogdian merchants has argued that the terrestrial routes were still dominant in the late sixth century.Footnote 119 The ‘Turkish consolidation of Inner Asia in the middle of the sixth century’ was itself a major factor stimulating revival or further expansion of the silk trade.Footnote 120 Secondly, as Skaff argues, the demand for international trade was largely driven by political conditions in China, and the latter part of the sixth century saw the pendulum swinging back to unification.Footnote 121 With the T’ang reunification of China in 618, this impact would have been even stronger. At any rate, for the Sasanians the late sixth century was a period of renewed commercial expansion, related perhaps to Hormazd IV’s victory over the Türks. Khusro II’s coinage of the 610s and 620s represents almost half the surviving coins from Turfan’s tombs that have a mint date (eleven out of twenty-five),Footnote 122 and his coins were plentiful in various parts of Afghanistan.Footnote 123 The hoard from Wuqia in the Xinjiang Uygur Autonomous Region museum, which was never properly published since its discovery in 1959, contains at least 546 drachms of his (96 per cent of the Sasanian coins).Footnote 124
The peak of Sasanian silver circulation in Turfan seems to have occurred later, in the second quarter of the seventh century. It was the purity of the silver coinage that won general acceptance for it from Bukhara to Turfan,Footnote 125 and this must have been a major factor in the state’s decision to maintain high levels of purity. On the other hand, late Sasanian silver circulated in other forms equally amenable to trade. The late sixth century also saw an expansion of trade with the regions north of the Caspian.Footnote 126 Here, in north-eastern Russia, the chief means of exchange (in the Kama–Perm regions) were the large numbers of silver bowls and other silver objects that have turned up there. A lot of Sasanian silver was recycled to the north by Sogdian merchants,Footnote 127 showing that beneath the incessant political fluctuations in this vast landscape, the caravans continued to circulate because traders from different nationalities depended on each other. The late Sasanians had an abundance of silver which they used for both coinage and silverware.
To return to the solidus, Cosmas’ story, told to him by Sopatros, about the Sri Lankan king scrutinising the solidus and the drachm is formal proof that both solidi and drahms circulated in the Indian Ocean trade. Since early Roman gold and silver coins had always circulated in the Indian Ocean as bullion, that is, valued for their bullion value,Footnote 128 there is no compelling reason to believe that the solidus would have been treated any differently in late antiquity. This contrasts sharply with the export of low-value copper coins to South India and Sri Lanka. Their import into these regions was probably linked to the undervaluation of the Late Roman bronzesFootnote 129 and designed to meet a need for local currency.Footnote 130 The solidi, by contrast, would have been melted down or holed and used for jewellery. The same was almost certainly true of the few stray solidi that have turned up along the silk routes to China.Footnote 131 The scarcity of Byzantine gold in this sector of the international trade is not so much an indication of the scale of its circulation as reflective of the fact that these coins had a limited monetary use in this part of the world.Footnote 132 But gold did circulate in the Central Asian trade, and this could only have come in, initially, as solidi. The evidence here is partly textual. When Qutaiba ibn Muslim conquered Baikand in 706, the Arabs found ‘innumerable gold and silver vessels’,Footnote 133 total assets valued at 150,000 mithqāls,Footnote 134 so this is some indication that the Sogdians had access to gold from the west. The really crucial pieces of evidence, however, both come from the remarkable early seventh-century history of Theophylact Simocatta. His sources on China were of an exceedingly high quality.Footnote 135 Describing the ‘realm of Taugast’, as he called China, Theophylact says, ‘They have a custom, which resembles law, that males should never embellish themselves with gold adornment, even though they have become owners of a great abundance of silver and gold as a result of their large and advantageous trading.’Footnote 136 A second passage from the same author states (in the context of Bahrām Čōbīn’s campaign against the Türks):
[T]he Huns [Türks] were subjected to tribute by the Babylonians [Persians], although formerly the Huns [Türks] levied from the Medes [Persians] forty thousand gold coins as cause for inactivity. The Turkish realm, then, had been made very rich by the Persians, and this particular nation had turned to great extravagance; for they hammered out gold couches, tables, goblets, thrones, pedestals, horse-trappings, suits of armour, and everything which has been devised by the inebriation of wealth. Subsequently when the Turks broke the treaty and demanded that they be given more than the customary money and that there be a very heavy supplement, the Persians, intolerant of the burden of the imposed tribute, elected to make war. When the Persians won a splendid victory, affairs together with fortune reversed their flow, and the Turks were subjected to tribute by the Persians and were also deprived in addition of the wealth which they had previously accumulated.Footnote 137
So the Persians were paying the Türks in gold but conflict erupted and the Türks were defeated, the flows reversed and gold poured into Sasanian coffers. In what form? Theophylact refers to ‘gold coins’.Footnote 138 These are crucially important passages, for they seem to be about the only textual references we have to the actual circulation of solidi in the geographical area covered by Türko-Sasanian interactions, both political (tribute payments) and presumably commercial as well, in the late sixth century.
The coins Theophylact refers to could only have been solidi, since they are not likely to have been Sasanian dinars.Footnote 139 Since no solidus hoards have turned up in the former Sasanian territories (at least none I am aware of), we must suppose that these payments, like the plunder the Persians collected in the Levant, were melted down.Footnote 140 We can thus visualise two situations for the late Roman/Byzantine gold circulating beyond the eastern boundaries of the empire: one where solidi retained their form of coin but circulated by weight, and did so with other coins such as the drachm, which also circulated by weight. This situation is formally attested by a passage in Balādhurī which says, ‘The dînârs of Heraclius used to be current among the people of [Mecca] before Moslem times, but also the baghlîyah [full-weight] dirhams of Persia; but it was not customary to buy and sell with them except by considering the coins as bullion.’Footnote 141 Both solidi and drachms were used in the Hijaz but they circulated by weight. The second situation, by contrast, was one where the solidi were basically treated as a form of bullion, valued for the purity of their metal content, hence readily melted down, holed, hoarded and so on. The best examples of this are surely China, India and Sri Lanka, and of course it is also possible that in these markets the coins were brought in by traders of other nationalities, such as Sogdians, Arabs from the peninsula, Indians and so on. When Cosmas writes that solidi were carefully selected (eklekta) for export to Sri Lanka,Footnote 142 this must refer, in the first instance, to the fineness of the coins, and to their weight only insofar as there was a demonstrable connection between the weight and fineness of gold coins.Footnote 143 Finally, the two situations together would also account for his other assertion that ‘all countries use the solidus in their commercial dealing’, and that the coin was found from ‘one end of the world to the other’.Footnote 144
In the western Mediterranean the Roman monetary system had of course disintegrated by this stage, but the use of Byzantine prototypes shows the profound influence of the solidus in this sector.Footnote 145 Like the Arab administrations in Syria and Egypt, and a whole century before them, the Merovingians based their gold coinage largely on existing stocks of solidi.Footnote 146 Gold was almost certainly being imported into Italy under the Ostrogoths,Footnote 147 and a great deal of Byzantine gold made its way north via the Ostrogoths and transpadana.Footnote 148 In the east, one of the most interesting hoards to have turned up is that from al-Madhāriba in Yemen.Footnote 149 This contained Aksumite (= 868) and late Roman (= 326) gold coins in a single hoard, but the baffling feature is the considerable gap between the solidi and the Aksumite coins. The bulk of the solidi date between 350 and 367, whereas the hoard stems from the second quarter of the sixth century!Footnote 150 At any rate, it is always interesting to be able to match a numismatic find with textual or other evidence. In the early sixth century, the wealthier Christian families of Najrān in southern Arabia used solidi in their transactions. The new ‘Letter’ of Simeon of Bēth-Arshām reports that, following her martyrdom, the princes of Najrān interceded with the king Yuśuf to grant a proper burial to Ruhayma, daughter of Azma’, arguing:
This (woman) has done many good deeds to everyone, to the king, to the notables, and to the poor; and in like manner she formerly treated M’DWKRM <Ma’dī-karib>, who before you has been king in this land. For he was in difficulty and borrowed from her twelve thousand denarii, and finally, when she saw that he was in difficulty, she renounced claim to the debt.Footnote 151
Ruhayma’s contract can be dated to within a few years. Ma’dī Karib was the last of the Tubba’ kings (the Tubba’s were the ‘ruling house of the kingdom of Himyar’)Footnote 152 and he seems to have been placed on the throne of Himyar by the Abyssinians in late 518 or early 519.Footnote 153 Since he is still attested as ruling in 521–2,Footnote 154 Ruhayma’s loan can be dated to 518–22. She herself was one of the victims of the great carnage at Najrān, in the autumn of 523,Footnote 155 and a superb example of the (largely Monophysite) aristocracy of that town.Footnote 156 So we can conclude that solidi were almost certainly legal tender in this highly commercialised sector of the Arabian peninsula, from where, moreover, much of the trade with India and Sri Lanka was conducted.
To conclude, the seventh century was far more vibrant than it has been made out to be from the narrow perspectives of a fragmented western Mediterranean, or even those of Haldon’s Byzantium, and much of this was driven clearly by a volume of commercial exchanges that was as substantial as ever. Trade with the Far East saw a new period of expansion coinciding with the T’ang unification of China, and culminating, of course, in the great upsurge of the ninth century. Ibn Ḥabīb describes the international fairs at Sohar and Dibba on the Batinah coast in the last years of Sasanian rule in terms that recall Ammianus’ description of Batnae. ‘Merchants from Sind, India, China, people of the East and West came to it,’ he says about Dibba.Footnote 157 Here we have a clear reference to the Chinese sailing to the Gulf, but is the testimony credible? The answer is ‘yes’. The Armenian geography now ascribed to Ananias of Širak completely bears it out. This survives in two versions, a long and a short one. The shorter recension, which dates from the late 630s or 640s, says of Yemen (Fortunate Arabia) that ‘It produces everything and everything is found there either in the country itself or brought by merchants from China or India.’Footnote 158 Baṣra, which was undoubtedly a relatively new foundation at this time, is described as ‘filled with merchants and ships coming from India and all parts of the orient’.Footnote 159 ‘The Sogdians’, says Ananias, ‘are wealthy and industrious merchants who live between the lands of Turkestan and Ariana.’Footnote 160 In short, behind the devastating conflict between Byzantium and Persia, the commercial world of the seventh century lay intact, supported by an abundant circulation of the precious metal coinages.
In terms of external trade, the Byzantines were effectively restricted to the ‘northern’ route (through the Astrakhan steppe or across the Caspian), from where in fact the bulk of solidi moved in other directions – to Scandinavia to finance the trade in Swedish furs,Footnote 161 to the Black Sea steppe in political payments.Footnote 162 Lombard’s famous paper on the three monetary zones saw the late Romans and then the Byzantines progressively exhausting their reserves of gold in a multiple flight that was beyond control.Footnote 163 And it is true that measured by the standards of the late RepublicFootnote 164 or even the second century,Footnote 165 the reserves of gold in the state’s hands had dwindled substantially by the late empire and even more so by the ninth century.Footnote 166 But we can certainly rule out an external flight as the chief cause of this. What does seem to have happened is a progressive dispersal of the hoard into private hands, through largesse (profusa largitio), the vast endowments of the church and, above all, the expansion of private business. In the Sasanian realm, by contrast, the hoard remained largely concentrated in the various and numerous treasuries of the shah till it, too, was largely dispersed through the lack of a strong central administration in the Arab conquests.Footnote 167