Escape From Rome
Page 49
At the beginning of the second century CE, Rome seized the Nabatean kingdom, a client state centered on what is now Jordan but extending far along the Arabian west coast. The reins were tightened: inscriptions document the presence of a Roman prefecture and a garrison on islands of the Farasan archipelago north of the Mandeb Strait that connects the Red Sea with the Gulf of Aden. Merchants from within the empire had long imported ivory, tortoise, and turtle shell from Axum (Ethiopia/Eretria) and myrrh, cassia, and cinnamon from Somalia. Before the second century CE, trade (especially in elephant tusks) with coastal Kenya and Tanzania had been checked by outposts of the Saba-Himyarite kings of Yemen, but now Roman ships would sail directly all the way to Tanzania and deal with local suppliers.91
Effective control over the Red Sea basin and its southern access point also guaranteed unencumbered relations with the Hadramawt kingdom in eastern Yemen and Dhofar, a major source of frankincense, and helped secure the route to India. Direct commercial and diplomatic relations with South Asian rulers date back to the reign of the first Roman emperor, Augustus, who had seized Egypt and was the first to receive embassies from the Indian Ocean region. Roman vessels came to sail both to Gujarat and to Tamil southern India often directly from Africa or Arabia by taking advantage of the seasonal monsoon winds. The city of Muziris in Kerala became the principal hub: it even featured a temple of the deified Augustus. Sizable merchant communities grew up there and at other ports. Black pepper was the top export item alongside indigo and precious stones. As early as 20 BCE, the kingdom of Pandia south of Muziris had dispatched envoys and gifts of pearls and gems to Rome and, later on, local sources point to the presence of Roman mercenaries in the service of local rulers.92
Trade with Sri Lanka had initially been in the hands of Tamil intermediaries, but once direct contact was established in the 50s CE, an embassy from the island arrived in Rome, and high seas crossings began to supplement the ones via Indian ports. Just as in East Africa, middlemen had fallen by the wayside as Rome reached out to more distant polities. Its merchants eventually extended their routes to the east coast of India and crossed the Gulf of Bengal to reach Burma. By the second century CE, they may have sailed around the Malay peninsula to Thailand, and perhaps even all the way to the Han territory in northern Vietnam.93
In monetary terms the volume of this trade was enormous. Literary allusions to massive outflows of specie receive support from a surviving tax document from the second century CE regarding the arrival in an Egyptian port of a ship from Muziris that carried 135 tons of pepper and 84 tons of cinnamon and paid dues in excess of a tenth of a percent of the total annual public revenues of the Roman empire. According to an observer from the Augustan period, when this trade had just taken off, each year 120 ships reached Egypt from the Indian Ocean. The docks at Berenice, a port on the Egyptian east coast, could accommodate vessels that were 120 feet long and displaced 350 tons.94
By comparison, in the early sixteenth century Portugal imported between 1,000 and 2,000 tons of pepper from India, and by 1600 Southeast Asia shipped 4,000 tons a year to Europe—cargo that a small fraction of the 120 ships referenced for the Roman period could have managed. Thus, even though we should be wary of reading too much into flimsy data, there can be little doubt that Roman-era trade flowed on a scale (at least) comparable to that of the early modern period and that Roman merchants were able to move about the Indian Ocean basin unimpeded and often protected by diplomatic relations. With the single exception of Aden, the Romans, unlike the Portuguese, did not normally need to tussle with local potentates and worry about local rivals. The sheer weight of empire and the pull of its market ensured cooperation with exporters that benefited both sides at the expense of intermediaries.95
Under these circumstances, any suggestion to sail around Africa or across the Atlantic to access these goods would have been met with bewilderment. This had not always been the case: near the end of the second century BCE, a Greek explorer in the employ of the Ptolemies of Egypt sailed to India and back and then down the eastern coast of Africa. On both occasions, the authorities confiscated the goods he had brought back with him. Their depredations inspired him to seek out an alternative route around Africa, a bold endeavor that appears to have failed. Once the Mediterranean had been united under Roman rule, even this fantastic option became irrelevant: traders faced a straightforward choice between accepting imperial taxation or staying at home. The decision was not difficult: strong demand and the relative smoothness of long-distance trade sustained high volumes of exchange.96
The scale of state formation was the critical variable. Imperial unity in the Roman period and abiding fragmentation coupled with relatively low levels of state capacity during the next millennium or so generated dramatically different incentive structures. Even if we were to hold demand for imports constant (itself a dubious proposition, given the enormous concentration of wealth in Roman elite circles), post-Roman polycentrism and the resurgence of Middle Eastern polities in the late Middle Ages and early modern period introduced handicaps that were alien to hegemonic empire. In conjunction with the dynamics of interstate competition that locked parties into open-ended commitments, state-society bargaining that mobilized productive exploitation, and unexpected payoffs (especially in the New World), these constraints sustained growing commitments to overseas expansion.
Geography as such mattered less. The Roman experience refutes the notion that it was inherently difficult for Europeans to access the riches of South and East Asia: it was hard for them only when they were not organized as a single imperial entity. It is true that navigating the Atlantic Ocean required sturdier ships and more sophisticated rigging than riding the monsoon winds did in the Indian Ocean and South China Sea: but that fact alone did not compel Europeans to live up to that greater challenge. Otherwise, Japan would have developed similarly advanced shipping instead of orienting itself toward the imperial behemoth to its southwest. Institutions, not winds, turned out to be decisive: imperial hegemony and its gravitational pull on the one hand, inter- and intrastate polycentrism on the other.97
East Is West …
This leaves us with the second objection to an institutionalist perspective: that geography may have lent Europeans a helping hand while stymieing others. Testing this notion calls for a rather extravagant counterfactual: it is only by changing the orientation of entire continents that we can hope to gauge the influence of geography as such. The goal is to make it harder for Europeans and easier for the Chinese to reach the New World. (As there is no obvious way to make this possible for South Asia, which is bound to remain remote from the western hemisphere unless we rearrange the world in an even more fantastic fashion, I focus on the traditional contrast between the eastern and western fringes of Eurasia.)
I can think of three different adjustments that serve our purpose. One is to flip the Old World around its axis, so that Europe faces the Pacific and China the Atlantic. This can be done in a manner that preserves the overall integrity of this cluster of continents and their principal adjacent islands (what we might call “Greater Afroeurasia”): by rotating the entire area between 30 degrees West (to include the Azores and Iceland) and the dateline (to include the northeastern tip of Siberia and New Zealand) east–west around an axis located at 75 degrees East, halfway between these two boundaries (figure 11.1).
FIGURE 11.1 The world with Greater Afroeurasia rotated around its axis at 75°E.
Even though this intervention helps preserve existing connections at the margins of the Old World, it displaces the European coast outward into the Pacific and pushes China and Japan toward the center. This shortens Pacific crossings from Europe to the New World and expands Atlantic crossings for East Asians. Because this conflicts with the main objective of this experiment—to reverse the European advantage—it might make better sense to disregard the integrity of the cluster and focus on relative distance.
In this second scenario, the goal is to approximate the location of
central Japan to that of England, and that of the Yangzi Delta region to that of the southwestern Iberian peninsula, even if that adjustment severs connections between the Afroeurasian mainland (here: “Afroeurasia”) and peripheral island chains. This adjustment, which ensures that the principal East Asian launch points for voyages to the Americas are as close to their destination as the principal Western European launch points are in real life, requires a flip around an axis at 63 degrees East, the midpoint of the averaged distances between London/Lisbon and Tokyo/Suzhou (figure 11.2).
FIGURE 11.2 The world with Afroeurasia rotated around its axis at 63°E.
A third and (even) more radical counterfactual would leave the map unchanged but assume that China equals Europe and vice versa. In this case, Japan plays the role of Britain and east-central China serves as the equivalent of the Iberian peninsula. While this may seem to be the least invasive method of substitution, it does rupture the connections between geography, ecology, and state formation: contrary to what I have argued in chapter 8, it forces us to presume that Europe’s physical environment could have supported Chinese-style imperiogenesis, and that proximity to the steppe would not have affected European institutions. This also completely transforms interactions between “Europe” and neighboring regions: it no longer abuts the Mediterranean or has access to Africa, whereas “China” does.
This final scenario would alter history in a much more arbitrary and unpredictable fashion, and render it unsuitable for controlled conjecture. Perhaps paradoxically, leaving geography untouched would entail the most dramatic changes in starting conditions. I therefore abandon this approach in favor of the two axial rotations that change only oceanic distances without interfering with the dynamics of state formation within the Old World.
Rotated around these axes, Europe faces the enormous expanse of the Pacific Ocean, which covers close to a third of the earth’s surface and contains about half of all its water. Depending on which sea routes we are looking at, this intervention affects logistics in very different ways (table 11.1).
First of all, the distance between northwestern Europe and the extremities of North America does not significantly change: the relocated city of Bergen (employed as a proxy for the origins of Norse expeditions) is roughly as far away from Anchorage (a proxy for the American mainland) as the real Bergen is from L’Anse aux Meadows at the northern tip of Newfoundland, the site of a Norse settlement at the end of the tenth century. While the historical Norse could (and did) use the Faroe Islands, Iceland, and Greenland as stepping-stones, their rotated counterparts would have been able to sail along the Kuril Islands, the Siberian coast, and the Aleutian island chain: at the time, the latter housed an expanding civilization with sizable settlements. By the same token, a rotated Tokyo is not dramatically closer to Newfoundland than the real Tokyo is to Alaska.98
Ian Morris’s claim that it was much easier for the Norse than for the Chinese to reach America disregards these basic facts, and gives short shrift to Japan’s potential role as a forward base and the presence of elongated island chains. Chinese sailors would by no means have been forced to cover ten times as much open sea in a single stretch than the Norse. As far as the northern seas are concerned, there is little to support the view that it was intrinsically much more difficult to cross the Pacific from East Asia than to cross the Atlantic from Europe.99
But neither Newfoundland nor Alaska were particularly profitable destinations. In 1492 Columbus sailed from Palos de la Frontera in Andalusia to San Salvador Island, about 6,500 kilometers as the crow flies. His actual route via San Sebastian de la Gomera in the Canary Islands was a little longer than that, in excess of 7,000 kilometers. After the rotations, departing from Palos’s new location either west or east of Japan, he would have had to cover between one-third and two-thirds as much distance again to reach the American mainland at the same latitude as San Salvador Island, near Todos Santos in Baja California Sur. However, the prevailing winds (the Pacific westerlies) and the North Pacific current would have brought him close to the American mainland well north of that spot. This would have reduced Columbus’s travel without landfall to something not dramatically longer (maybe by one-third or less) than his actual voyage from the Canaries to the Caribbean. The Northeast trades and the Californian and North Equatorial currents would then have allowed him to move down the coast and return to rotated Europe.
TABLE 11.1 Minimum distances (in kilometers) and differences between actual and counterfactual distances (in percent)
Actual
“Greater Afroeurasia” rotation
“Afroeurasia” rotation
Bergen to L’Anse aux Meadows
3,800
Bergen to Anchorage
3,400 (–11%)
4,500 (+18%)
Palos de la Frontera to San Salvador Island
6,500
Palos de la Frontera to Todos Santos
8,700 (+34%)
10,600 (+63%)
London to Jamestown
6,000
London to San Francisco
6,600 (10%)
8,100 (+35%)
Lisbon to Salvador de Bahia
6,500
Lisbon to Panama
7,700
11,900 (+55%)
13,600 (+77%)
Cádiz to Lima
*11,000
14,100 (+28%)
16,000 (+45%)
Cádiz to Veracruz
8,700
Cádiz to Acapulco
10100 (+16%)
11,900 (+37%)
Cádiz to Panama
8,000
12,300 (+54%)
13,900 (+74%)
São Tomé to Salvador de Bahia
9,600
São Tomé to Panama
15,100 (+57%)
17,700 (+84%)
Tokyo to San Francisco
9,000
Suzhou to San Francisco
9,900
Suzhou to Acapulco
13,100
Tokyo to Anchorage
5,600
Tokyo to L’Anse aux Meadows
5,400 (–4%)
3,800 (–32%)
Tokyo to Jamestown
7,500 (–17%)
5,600 (–38%)
Suzhou to San Salvador Island
9,900 (0)
6,700 (–32%)
Suzhou to Salvador de Bahia
8,700
6,100
* Includes land crossing.
Geodesic distances in table 11.1 derived from https://www.movable-type.co.uk/scripts/latlong.html.
In terms of access to much of the New World, northwestern Europe would have been very little affected by either one of the axial rotations. London lay some 6,000 kilometers from Jamestown, only up to one-third less than the distance between flipped London and San Francisco, Jamestown’s latitudinal equivalent on the West Coast. Similarly, it would not have been much more demanding for Spanish ships to reach Mexico: the trip from flipped Cádiz to Acapulco is only between one-sixth and one-third longer than between actual Cádiz and Veracruz.
This need not have impeded conquest of the Aztec empire, which touched both the Atlantic and the Pacific coasts. Instead of the Tlaxcala, the Tarascans near the Pacific coast might have collaborated with the conquistadores. The lack of islands to be used as staging areas (which the Caribbean provided in abundance) was an inconvenience that could have been overcome by the establishment of coastal positions along the Spanish access route north of the Tarascan-Aztec zone.
Much the same applies to the connections between Spain and the Andean region. Cádiz is 11,000 kilometers from Lima including a land crossing at Panama. Although rotated Cádiz would have been between a quarter and half as far away again, an unobstructed all-sea voyage would have helped make up for this disadvantage. Only Brazil would have become much more remote: the distance from Lisbon to the early Portuguese Brazilian capital of Salvador de Bahia would have tripled from 6,500 kilometers to
something closer to 20,000 kilometers plus a land crossing.
Failing as they do to properly account for the varying impact of winds and currents, all these calculations are necessarily exceedingly rough but nevertheless contribute to a more nuanced understanding of geography’s role. Rotating the Old World around its axis and the consequent expansion of the ocean that separates Europe from the Americas would have had no substantive impact on the early stages of exploration and exploitation. The Norse could easily have reached Alaska, Columbus California, Cortés and Pizarro Mexico and Peru. Undermined by Old World pestilence, the Pre-Columbian societies would have fallen just the same and their precious metal would have been shipped off.
Even if this process might have taken a little longer and been a little more costly, there is no compelling reason why it would have been aborted altogether. After all, logistical constraints would for the most part have been only moderately greater. The same intra-European rivalries would have played out, albeit sometimes in different terrain; the same hunger for fame and fortune would most likely have sustained colonial development.
This is not a complacent assessment. Even in the sixteenth century, European seafarers were already perfectly capable of mastering challenges much more daunting than Atlantic crossings. As early as 1520–1521, Magellan’s ships covered more than 15,000 kilometers in 99 days between Chile and Guam without making landfall. From the 1560s, once it had been discovered how to traverse the Pacific in the opposite direction, Spanish treasure ships undertook twice-a-year passages between Acapulco and Manila, a notional 14,300 kilometers as the crow flies and, on the way back, closer to 18,000 kilometers in practice as eastbound vessels had to swerve far into the North Pacific in order to capture the trade winds. Crew attrition was higher than for Atlantic voyages but not enough to deter.100