time under way
sail day and night, and thereby doubled the distances they could routinely covet We will need to consider what historical, and perhaps especially, what economic forces might have encouraged skippers who were not acting under imperial orders to shift toward a different structural pattern of their voyages, and what the consequences of such a growing shift might have been. Were there more cargoes that were more fragile? Was demand increasing on one or both sides of the water providing an incentive to quicker trips? But first we need to scrutinize our early medieval travelers' movements from another angle, that of routes.
"Spaces of sea": Europe's western Mediterranean communications
I n Charlemagne's day, travel to Constantinople meant salt water.
1 "Crossing the spaces of sea," was how a court poet imagined the king's daughter would get to Byzantium, while Sea Verses seemed an apt title for a poem about Anialarius' trip a generation later.1 The Mediterranean Sea has dominated scholars' investigations of early medieval communications no less than royal courtiers' imaginations. Rightly so, since the Roman communications that once coursed along the continental corridors of the Danube or the Via Egnatia had ceased entirely in the seventh century (Ch.3.2).
I n the age of Charlemagne and Harun al Rashid, what routes funneled contacts between eastern and western areas of what had been the Roman empire?2 Routes, we sometimes imagine, are predetermined by geography. Physical geography and climate do in fact define the possibilities. But they do so in complicity with human technology and culture, as historical developments affected which possibilities would be exploited. So far, the history of routes has played little part in the debate over long-distance trade in the early Middle Ages. Yet we have seen that routes changed. A systematic analysis of the spatial patterns of travelers' movements forces a new appraisal of the geographic structure of Carolingian communications. Data of varying and independent origins discloses some clear patterns. Development and growth are among them.
How people got from the western to the eastern areas of the former Roman empire changed dramatically between 700 and 900. The routes increased, from one main route to five or more, depending on how finely one distinguishes them. 1 hey also changed in nature, as what were mainly sea routes came to be flanked by
1 Peter of Pisa, Carmino, no. 12,11, p. 62; ef. Paul Deacon's response, no. 13,10, p. 67. Anialarius himself calls his poem "Versus "larini": Opera litur^ica, 1.231.19.
2 I our earlier studies merit mention: Pertusi ■964, esp. 82-92; Hubert 1964; the
imaginative though poorly documented work of Lombard 1972,73-94; and Micheau 1979. On the concept of routes, as opposed to roads (or, indeed, navigational courses), see Denecke 1979, 439-42.
500
501
ivuu i a r^wm i ULY TO THE AEGEAN
combined land and sea or almost exclusively overland routes. They may have changed, finally, in structure: we shall have to watch for their components whether they be interlocking regional transport networks or long-distance carriers.
Around 700, such contacts as persisted between eastern and western Christendom were essentially seaborne. Just one great sea lane remained open the ancient, northerly trunk route that linked Rome and Byzantium via Sicily and around Greece. We shall begin with that route and its traffic. The rest of this chapter will concentrate on the western Mediterranean. A swift glance will discern the quickening pace of communications along the southern rim, in the new Islamic polity that was rising out of the rubble of Rome. This will help us make sense of striking new developments in southern Italian waters. Further north, we will observe the increasing number of ships along the shores of Tuscany. In the course of the eighth century, communications will grow denser in each zone. And they will begin to spill over into the neighboring region, as new networks took shape along the edges of the empty spaces of sea.
i. The ancient trunk route from Italy to the Aegean
For much of the eighth century, a seaborne axis extending from Rome around Greece to the Aegean dominated Christendom's communications (Map 8.1). It skirted the coast of the Tyrrhenian Sea, struggled through the turbulent strait of Messina and crossed the deep and dangerous Ionian Sea to the Greek coast and the shelter of the Ionian islands. From there, travelers sailed around the Peloponnesus and north into the Aegean. Sicily was the crucial hinge. Whether or not it was smoking, massive Mount Etna guided westbound sailors securely toward the strait of Messina and Sicily's eastern seaboard, along which a branch of the main route ran southward, down to the provincial capital at Syracuse. Naples, Reggio di Calabria, Messina, the great rock of Monemvasia (Figure 5.1) and one or another Aegean island were the main ports along the lifeline linking Constantinople and Rome. And, as we have seen, a host of lesser settlements, coves and islands offered weary sailors respite along the way (Ch. 13.3). At either end, this great trunk route could extend northward. Some late Roman ships had plied past Rome toward Marseilles, while in the east ships worked northward through the Aegean toward Constantinople. A further extender ran east along the Asian coast, past the maritime district of Pamphylia, toward the Byzantine border town of Seleukeia, and across to Cyprus and the Levant.3
3 The route has been described many times over, from the Middle Ages down to
modern Sailing Directions. See, for the Middle Ages, esp. Pryor igQ2, 92-7; for the
502
The central segment was ancient. It had been one of the main courses set by the great grain fleets of Alexandria that fed Rome.4 Activity on the northern leg to Constantinople had surged and declined as the city mushroomed and shrank in late antiquity, but it never ceased. In one form or another, the Sicily-Monemvasia sector continues in use down to the present day. What was new in the seventh century was that the disrupted overland traffic and the withering of other shipping lanes gave this ancientaxis a near monopoly on east-west communications.
Though it is easy to glimpse segments, few witnesses depict the whole trunk route between Constantinople and the west. Martin I's death trip from Rome to Constantinople, down the Tyrrhenian coast, to the layover in Messina and landings in Calabria, on "many islands," at Naxos and Abydos, has provided a clear conspectus (above, p. 486). Two excellent descriptions show that this seventh-century pattern lived on in the eighth. For this was the course Pope Constantine set to Constantinople over the winter of 710/11, and this was how Willibald and his companions sailed, in the spring of 723, toward Jerusalem.
The papal court boarded their ships at Porto Romano (Portus) and sailed to Sicily via Naples.5 From there they reached Reggio and then sailed directly (trons-jretavit) across the Gulf of Taranto from Crotone to Gallipoli.6 As we have seen, the winter weather prevented the papal convoy from crossing the Ionian sea immediately: the ships stalled at Otranto until the winds shifted (above, p. 463^. The next notable stop came in the Cyclades. On Kea, the patrician and commander of the Imperial Fleet welcomed the papal party as "they reached the region of Greece."7 From there they sailed to the harbor of the Hebdomon ("At
experience of i4th-C. Genoese navigators, Balard 1974; cf. e.g. the i5th-C. l'arma-Magliabecchi Portulan. The other was the coastal way along Africa, CassoniQQ5,297; cf. Rouge 1966, 86-99; and Ch, 4. Nonetheless, as Pryor 1992,6-7, 90 and 99, makes abundantly clear, for navigational reasons, the more northerly trunk route normally musthave been the preferred course.
R73; Liber pout, Mommsen, 222.18-223.21. On Porto Romano, see Schmiedt 1978, 159-63; Naples: ibid., 173-6. The pope may have sailed down to Syracuse to see the stra-tfgos, who was suffering from a fever, or the latter may have come to see the pope at Messina: the Latin ("occurrens pontifici," 223.10-12) is ambivalent, since it allows both a ceremonial occursus and actual travel to greet the pope.
6 Cf. the identical course of I4th-C. Genoese navigators, depicted in Balard 1974, map facing 248, and the difficult crossing from Crotone in Pellegro Maraboto's log, 1-5 October 1369, ibid., 261. On the ports of Reggio, Crotone and Gallipoli, Schmiedt 1978,185-6,188-9 and 194.
7 Theophilus was strategos of the Karabisianoi. Mommsen's punctuation of this passage ("Unde [sc. Hydrunto] egressi, partes Graeciae coniungentes, in insula, quae dicitur Cea occurrit Theophilus..." Liberpont, 223.18-20) rightly corrects Duchesne's (Liber pont., 1.390.9-n). Although coniunfjere is often used intransitively when it means "arrive" at this stage in Liber pont (e.g. Mommsen 223.16), a transitive usage was not unthinkable in 8th-C. Italy (cf. e.g. "coni-unxerat.. . Romam" of the "A" class
503
europe's western mediterranean communications
route from italy to the aegean
the Seventh Milestone") complex outside Constantinople.8 Since the papal ships reached Greece only via the Aegean Sea, the convoy manifestly did not transit across the isthmus of Corinth; rather the ships circumnavigated the Peloponnesus. Willibald's voyage makes this explicit twelve years later. After a first, short sea link from Gaeta to Naples, he sailed to Reggio, over to Catania and Syracuse and across the Ionian Sea, and around Greece to Monemvasia. Then they crossed the Cyclades to Kos, Samos and Ephesus. Later, he would catch another ship in the Aegean and take the trunk route's eastern extender along the Asian coast to Syria, via Cyprus (Ch. 5.1).
These trips are exceptional only in that they are so well documented, for they followed the habitual route of their time. Infection is once again a precious informant. Like its fourteenth-century successor, the spread of the last great plague of antiquity maps with deadly precision the main route of contemporary shipping.9 Its path proves that the trunk route was the usual one, some twenty years after Willibald's trip. From Italy to Constantinople, the contagion struck the imperial capital in the summer of 747 (R153). In a contemporary's words, it "started from Sicily and Calabria, spreading like some kind of fire to Monemvasia and Greece and the adjacent islands," before reaching Constantinople.10
The way the plague moved reveals more than the main shipping line between east and west in the 740s. The epidemic actually had begun much closer to Constantinople. It broke out in northern Mesopotamia in the winter of 743-4 and spread that spring. It seems to have hit Syria in 744, and it certainly returned in 745; it reached Egypt in 744. By 745, it had also infected north Africa, where it would recur regularly for seven years.11 In other words, an exceptionally lethal strain of bubonic plague had been raging inside the Caliphate for more than three years before it made its way to the Byzantine capital. What is more, the course of the infection seems to defy geography: it traveled not from the Levantine provinces closest to Byzantium's Aegean heartland, but from the empire's extreme western edge. If we remember that shipboard rats were the prime long-distance vectors of the infected fleas which carry the bacillus, and that their preferred niche was the kind of cargo storage associated with bulk transport, the
Footnote 7 (cont.)
ofMSS, Lrber pont., Duchesne, i.4o8.4\vith the apparatus). The partes Graecioe correspond exactly to the Greek Ta Tfjs 'EXXdSoj ttEpn which in the nearly contemporaneous MiraculaS. Demetrii (BHG 516Z-23), 2,156-7 with nn247~8, designates this region.
8 Liber pont., Mommsen, 223.20-1: "A quo loco navigantes venerunt ad septimo milia-rio Constantinopolim." The Hebdomon
port went back to Valens: C. Mango, ODB 2: 9°7-
9 On these correlations, see in general McNeill 1976,101-2,109-14,128-31, Biraben and Le Goff 1975, Conrad 1981, 299-300, and McCormick 1998a, 52-65.
10 Theophanes, a.m. 6238,1.422.29-423.5. The place of "Greece" in the sequence matches that of Pope Constantine's trip.
11 Conrad 1981, 293-9.
implication is unmistakable. In 744 and 745, rare indeed were the cargo ships entering Byzantine space directly from its immediate Arab-controlled neighborhood. The Middle East, including Alexandria, the hub of Mediterranean shipping in late antiquity, was sending no ships to Constantinople in 745.
Mo source explicitly documents the transmission of the disease from the area of modern-day Tunisia 45 NM across the strait of Sicily. But chronology and circumstance make a powerful case that this was the course of the infection.12 From Sicily, the Byzantine shipping network launched the contagion on its lethal route toward the capital. It was through its most distant margin, over the most circuitous sea route imaginable, that the Byzantine heartland connected to the Islamic economy around 745. A last striking feature is how gradually the plague spread within the Byzantine empire. This feature will tell us something about the structure of early medieval shipping routes.
Westbound ships sailing the main trunk route took a slight variant in the lower Tyrrhenian Sea, where eastbound travelers hugged the Italian peninsula. Navigational conditions imposed it. Late in the seventh century, the pilgrim Arculf had sailed from Constantinople to Rome, via Sicily and the Aeolian island of Vulcano; a generation later, Willibald returned homeward to Naples, via Syracuse, Catania, Reggio, and Vulcano (R43; R125; R214). The stop on the Aeolian islands (Map 16.1) underscores the exactness of these accounts, l'revailingwinds thereabouts vary from southwestto northwesterly. Itwould have been very hard for early medieval ships to beat northward along the Italian coast against them. They had to skirt the islands with whatever winds came from the south, and then head northeastward toward the Calabrian coast.13
Only the old trunk route and its extensions connected Italy and Constantinople for most of the eighth century. Fleets dispatched from Constantinople to the western Mediterranean necessarily set this course. Between a failed reconquest of Africa in 697 and a squadron sent to relieve Messina in 001, warfleets moved along the trunk route at least fifteen times. They sailed more frequently as Byzantine Italy came under growing threats from the 780s.14 But warships were not alone on the ancient route: an eighth-century relic tag seems to trace an unknown pilgrim's voyage along it.15
504
12 Thus already in Amari 1933-g, 1:302; cf. Conrad 1981,298-9.
11 Mediterranean Pilot 1978-88,2:203 (10.1); cf. Diagrams 12-13, "Wind distribution."
J4 In these cases, the nature of the movements implies identity of the ships from beginning to end of the route - whereas travelers might take overland legs - and the practical absence of other routes rules out any ambiguity. See R52 and R56; R139;
15
R198; R220; R310; R410; R439; R450; R604; R623; R653; R681; R701; R720; R747; another fleet was rumored to have been sent from Constantinople against Rome and Francia-which can only mean the Provencal seaboard at this date: R165. AtChelles.ChLAno. 669, 35 may reflect relics collected in Gaul, northern Italy or Rome, and Ephesus: see above, p. 312.
505
europe's western mediterranean communications
The route remained active. At least three more Holy Land travelers turned to it for the long haul to western Europe in the course of the ninth century.16 Around 892, a merchant ship with business in the Aegean was still sailing it from Rome (R729). Glimpses of ships are pretty common. But it is rarely clear whether the segment we can see was the whole voyage, or only one leg in a longer trajectory. Passengers and goods crop up as they board ships, for instance, from Sicily to Rome, or from Reggio to the Peloponnesus, and, of course, from Sicily to Constantinople.17 Historical fiction tells the same story. Ships move routinely along the ancient trunk route in the Italo-Greek novels written around 800: papal convoys travel to Rome from Sicily, bishops and monks naturally sail the same route, and the emperor commands his official to sail to Sicily and return to Constantinople within 60 days. Even magic can chart no other course: this was the only way to Constantinople and the east. A sorcerer's ship flies north from Catania to Reggio, then she streaks through the air to Crotone and zooms across to Otranto en route to the imperial capital, along the same sea lane the pope and pilgrim had sailed a century earlier.18 Another hagiographer's hero hopped into a ship on the beach near Naples and sailed off to Constantinople, but then changed his mind and disembarked on Chios.19
In sum, for the eighth century, there is almost no hint of any other itinerary between Constantinople and Rome, even in the imaginary world of fiction. This makes it more than likely that communications between Rome and Sicily, or between old and new Rome which transited via Sicily, also sailed this sea lane when we do not get conclusive details of the route. Routine obviated mention of infrastructural details.20 Once in a blue moon other routes may have been
J
16 Count Gebahard's deacon sailed home from Jerusalem to Rome: R242. This must have been the route Michael Synkellos had in mind when he set out for Rome from Jerusalem, via Constantinople: R286. Bernard almost certainly traveled this way when he went from Lydda to the vicinity of Salerno in sixty days; the time involved makes it virtually impossible that he could have traveled along the coast of Egypt and north Africa: R580. On the advantages of the northern route over the southern for sailing and oared ships, see e.g., Pryor 1992, 6-7, 902nd 99.
17 A few examples which certainly involved ships: the transport of huge beams from Calabria to Rome must have been over water: R45 and R90; Elias Speliotes sailed from Sicily to Rome: R692; Elias the
Younger fled Arab attacks, from Calabria to Greece: R685 and R719; a mother and son sailed from Reggio to Naples: R742; Peter ofTauriana found a ship bound for Constantinople in Sicily and was able to carry out the mission entrusted to him by the local governor: R323. Leontius, V. Greg. Agrig. e.g., 61,219.23-4; 25,176.29-31,46,202.15-17, respectively; V, LeonisCataniae, 9,89.1-90.6; 12,93.1-17-Passio Niconis, 5, *i6B. 20 A few examples: documents and/or news R84 and R322; delegations and embassies: R132-3; R206; R208-9, where the legates presumably felt they could escape from Sicily to Rome or return to Constantinople as the tumultuous events surrounding II Nicaea dictated; R217 (presumably the sptitharioi were coming, ultimately, from
18
19
route from italy to the aegean
attempted, but they saw little or no regular traffic. Thus, writing c. 731-41, a member of the papal entourage blamed the strategos of Sicily for the repeated failure of Roman emissaries to reach Constantinople: there was no alternative to the main trunk route via that island.21
Early eighth-century western connections with the Levant also prolong this axis' seventh-century pattern in another way. So far as we can tell, western travelers usually transited via the Byzantine empire. This is not so obvious as it may appear. One might think that endemic warfare between Byzantium and the Caliphate might have made the old trunk route less appealing to neutral travelers like Aquitainians or Anglo-Saxons; it would in fact soon be supplanted. Nonetheless, all trips between the west and the Holy Land around 700 whose routes are known transited through the Byzantine empire. They follow the main axis we have described between Rome and Constantinople or Aegean ports like Ephesus. Further connections to or from the Holy Land ran via Cyprus or Crete and Alexandria.22
The great sea lane channeled more than just hagiographical heroes and pilgrims, even in the most obscure years of the early eighth century. It has been observed that all major naval engagements between Arabs and Christians were fought on the Mediterranean's major routes; most ninth-century battles occurred precisely along this one.23 Many of the Greek immigrants to Rome must have boarded ships sailing some segment of this course, and some commodities, ideas, and books traveled with them. Traffic on this route helps explain immigration to Rome from Sicily, and probably also from Athens (e.g. R134 and R147). Along this ancient route the bishop of Monemvasia must have fled to Rome or, in the opposite direction, a cleric from Reggio gained appointment to the episcopal see of Kerlcyra.24 Whether by commerce or imperial command, coins struck at ninth-century Syracuse probably traveled this sea lane eastward and followed its northern extension all the way across the Black Sea to Cherson.25
Some developments had reinforced the main trunk's importance in the seventh and eighth century, while others diminished potential rivals. The African and Italian fiscal exports to Constantinople that swelled briefly around 655-70 must have increased trunk route traffic.26 The short-lived transfer of the imperial headquarters to Sicily in the 660s dictated, however briefly, substantial reinforcement of lines of
Constantinople); R2i4;R237; R243 (possible).
Liber pont., Duchesne, 1.416.16-417.1. Four trips from the 680s to the 720s: ^42-3; Thomas' seven-year pilgrimage had taken him from Rome to Jerusalem by an unspecified route, butc. 720 he returned to Italy via Ephesus: R62; Willibald went in both directions via the Aegean and the Asian coast: R108, R110 and R117.
23 Pryor 1992,106-11, making the pointalso that Islamic colonization followed the same pattern.
24 R339. Another possible case: R314. On contacts between neighboring regions of the empire, see McCormick 1998b, 41-2.
25 Sokolova 1965 offers no details; cf. Morrisson 1998.
26 McCormick 1998a, 78-80.
506
507
communications leading back to Constantinople, as well as toward Africa and continental Italy and may even have left tangible traces in the Sicilian waters facing Africa.27 The withdrawal of the imperial administration of Africa to Sardinia implies a resurgence in administrative contacts between that island and Constantinople-those contacts too traveled via Sicily and the main trunk route. The numismatic evidence bears this out: a hoard of Byzantine coins probably reflects someone's flight from fallen Africa to Sardinia c. 685-95 (Appendix 3: B20); tellingly, Sicilian coins figure alongside a majority of Carthaginian issues. A Sicilian deacon represented the archbishop of Sardinia at the ecumenical council in 787, ecclesiastics traveled at least sporadically between Sardinia, Sicily, and Constantinople in the ninth century, and Sardinian soldiers left their mark on Byzantine ceremonies.28 As far as rival routes are concerned, the Slavs so dominated Greece that even the impregnable offshore fortress of Byzantine Monemvasia appeared to be "in Slavinia land" to Willibald.29 The Arab conquest of the Near East blocked the overland route to Byzantine Africa through Palestine and Egypt. At the same time, expanding Arab control of the southern shoreline and growing naval power intensified for northern ships the inherent navigational disadvantages of southern Mediterranean shipping routes.
For a variety of reasons, the old sea lane from Rome to the east via Sicily and Monemvasia persisted almost alone as the Europeans' eastward route in the first decades of the eighth century. Yet within a generation or two of the nadir, change was in the air. Toward 750, other routes which fed into the main trunk line quickened again.
2. Southern rim: communications between the Maghreb and the Muslim center
While the northern sea lane maintained the precarious links between what was left of the Christian world, new communication patterns were solidifying within the Caliphate. Those that ran along the Mediterranean's southern rim linked the Maghreb, or "west," to the Muslim heardand in the Middle East, and their
27 On the route from Africa to Constans IPs capital of Syracuse, the exceptionally large cargo ship found at Pantano Longarini (Parker 1992, no. 787), is dated by C to a.d.5oo(± 120)andA.d.622( ±48);cf. Table 13.1. Two ships, reportedly of late 7th-C. date and of undetermined size, were discovered and looted at Punta Secca: Parker 1992, no. 968.
28 Pros. "Epiphanius" and Anons. 237-8; acclamations: Constantine VII, De ceremo-niis, 2,43, 650.14-651.12. See in general McCormick 1998b, 3on25; cf. for Sicilian coins on Sardinia, Appendix 3, e.g. B39, B41, B45; and Morrisson 1998.
29 "In Slawinia terrae" (sic), Hugeburc, V. Willib., 93.12-13.
508
ignif>cance reaches beyond the Muslim world, although western historians have so &r Pa'°- t^em Htt'e heed- For now, the movements of immigrants, armies, pil-
rims, scholars, and traders best sketch these communications within the Muslim world, but coin deposits are beginning to fill in the picture.30 Specialists have started to recover their importance in the eighth and ninth centuries.31 Ships certainly followed the coastline of northern Africa between Egypt and the Muslim west. For instance, a Visigothic king's granddaughter sailed from Spain t0 Ascalon en route to Damascus, showing that the Levantine shipping hub of late ■intiquity was still (or once again) functioning in the second quarter of the eighth century.32 But it was a treacherous shore, and many more travelers must have done as our own Elias the Younger, and moved overland.33
In fact, the volume of traffic in the eighth century was big enough to leave tracks in the Christian north (Ch. 17.3). This suggests that the eleventh-century caravans tramping across Ifriqiya, Muslim north Africa, between Morocco and Egypt, began bv the ninth century, or even earlier.34 The annual pilgrimage (hah) to Mecca was obviously capable of stimulating communications, although non-Muslims were not supposed to travel with its caravans. The late eighth century was a time of heightened investment and prestige for the huge pilgrimages to the holy cities of Arabia, and the annual religious caravans may well have had spillover effects on the broader transport system along the southern rim of the Mediterranean.35 We
E.g. Levi-Provencal 1950-3,1:44-7, 81-5, etc. For an overview of 8th-C. immigration to Spain, seeTaha 1989,110-65.,n smaller numbers, itcontinued into thegth C, e.g. the historian Ahmad al Razi's father, a trader who came to Cordova from Persia c. 864: Taha 1989,4. Cf. Levi-Provenqal ^o^, 3:488-93, on gth-C. intellectuals. On Spanish, Idrisi, and Aghlabid African coins found in the east, see Manzano Moreno (in press).
Most notably in an unpublished paper delivered by Olivia R. Constable at the Medieval Academy of America meeting in April 1999.
R.113. On Ascalon and Gaza shipping in the 6th and 7th C, see above, Ch. 4.3. Many of the later voyages listed in Udovitch 1978, 'able 1 seem to follow this course. Pryor 1992,21-4, on the difficulties of navigating along the African shore; it is unclear whether the Alexandrian envoy to Rome went overland from Egypt to Africa: R171.
Isaac and the Carolingian delegation to Baghdad certainly did so: R254; also R674; and from the realm oflegend, Leontius, V. Grcgorii kgng., 9,154.1-2; cf. 18-19,164.18—3 (sic). One suspects that this was true of Frotmund, R521; whether he traveled by land or sea between Egypt and Carthage, his story clearly reflects this route.
34 For instance, Bell 1913 interpreted a Berlin papyrus to mean that sailors supplied by Antinoe in 712 were to travel overland to Cyrenaica, where they would join the fleet being prepared for the raid of7i3; for the caravans, Goitein 1967-93,1:212; 275-81.
35 Ibid., 281; see in general A. J. Wensickrt al., "Hadjdj," EI 3 (1971), 31-8 (esp. B. Lewis, 37- 8); on the early Abbasid and, particularly, Harun al Rashid's enormous financial and personal investment in the Hajj, see Peters 1994,69-71; an edifying parallel comes from the commercial dimensions of the Hajj under the Ottomans, Faroqhi 1990, 200-22.
509
europe's western mediterranean communications
might even suspect that the great pilgrimages affected commerce directly: a few better-documented, centuries later, the pilgrimage caravan transported more than pilgrims. The countless sacks of spices and other oriental wares it carried to Egypt from Mecca could be counted on to flood the Nile's markets.36
But peaceful land and sea communications along the Muslim shore were only part of a bigger picture. As the eighth century opened, sea raids surged across the straits of Sicily. They may well have been coordinated from Damascus. The Arabs had driven the Byzantine navy from Carthage in 698, and the caliphal governor quickly set about establishing a naval infrastructure in Tunis.37 Between 700 and 710, five Arab squadrons are known to have sailed north toward Italy, hitting Pantelleria, Sicily (probably repeatedly), and Sardinia (R.58, R63-4, R66, R76). Then came a lull of nearly two decades broken by only one expedition (R104) until raiders returned to Sicily in force between 727 and 735. After two more major spasms in 740 and 752-3, attacks ceased altogether for almost fifty years. The raids show a north African shipping world that was coming to life again with unpleasant consequences for Byzantine Italy; but it is the lulls that catch our attention.
There is good reason to believe that the clustering of attacks is real, rather than an artifact of source preservation.38 That attacks on Sicily subsided between 711 and 727 is readily explicable. Other colossal enterprises absorbed the African Arabs' energies: the invasion of Spain in 711 and, less successfully, the greatsiege of Constantinople, which entailed such heavy losses of ships and men.39 After
36 Goitein 1967-93,1: 276.
37 R57. On state sponsorship of the early 8th-C. raids, Guichard 1983,57-8.
38 The second surge in attacks might actually have begun in a.h. 107, a.d. 725-6: Amari '933-9) i: 297n2. The later Arab historians of Africa are our main witnesses, but their testimony receives sporadic confirmation elsewhere. The papyri from Aphrodito in Egypt provide contemporary evidence of a coordinated effort to support sea raids against the Byzantine empire from Africa, Egypt, and Syria, chiefly in 708, 709, and 710: e.g. P. Lond. 4.1349, I35I> I353~5,1392, etc. This fits well with the later Arab historians' account ofthe refurbishment of the naval infrastructure at Tunis, including the transfer there of Copts: R57. P. Lond. 4.1350 (a.d. 710), independently corroborates one of the raids on Sicily: R.63. That raids continued
against Sicily after the peace of 728 is known from a papal report of 813: R124. An ill-dated raid of the 8th C. described in Greek hagiography can plausibly be assigned to the first half, and might be identical with one of those recorded by the Arab sources: R59. 39 Amari 1933-9,1:296-7, observed the correlation with the conquest of Spain; on the fleet's possible role in Spain, see Ageil 1985,147-57. The lengthy involvement there of naval forces seems now to be confirmed: R99, although the chronology raises the possibility of confusion with the return ofwhatever survived of the fleet sent against Constantinople; R95. An Arab and Greek papyrus of November 713 documents the levy of Egyptian sailors from Antinoe for a raid to be mounted from Africa in indiction 13, but whether the attack involved Spain, Sicily, some
from africa to the southern tyrrhenian sea
-740, Arab attacks ebbed again, as civil disturbances in Ifriqiya and vigorous Byzantine defensive measures gained strength.40 And of course the last great episode of the plague made the marshaling of ships and men especially difficult.
When they were occurring, there is no mistaking that the raids were war. They caused certain suffering and damage where they struck, and they unfolded jaainst a backdrop of Arab imperial expansion. But from an African perspective, the violence had a positive economic dimension, for the attacks produced slaves uid plunder.41 In this sense, these slave- and booty-gathering expeditions foreshadowed the integration of southernmost Italy into the new north African economy. This is only made clearer by the fact that when peaceful commercial links between Italy and Africa reappeared, they concerned the same wares: Christian slaves.
3. Southern links: from Africa to the southern Tyrrhenian Sea
Africa's exports across the western Mediterranean died with Roman power. Early archaeological returns suggest that even to nearby southern Italy, they plummeted in the later seventh century and then ceased. Coins minted in Africa continued to make their way to eastern Sicily right up to the end.42 At later seventh-century Naples, for instance, lamps manufactured locally (35.7 percent) and in Sicily (12.5 percent) together outpaced the African lamps (39.2 percent) which had once dominated there. African lamps disappear altogether from the early eighth-century strata.43 Still undetected exchanges may have persisted
other place, or even occurred atall is unknown: Bell igi3; cf. Fahmy 1966a, 64-5. For the abatement of military activities after the disastrous siege of Constantinople, see also Ageil 1985, 159-60. Some raids were launched in these years, but those whose targets are known aimed at Sardinia and Corsica: R106, R112; those islands were presumably also the target of the raids of R107 and R115: cf. Ageil 1985,168-9.
Jo R159;Talbi 1966, 36-8; cf. Guichard 1983, 58-9 who argues strongly for an interruption after 752.
+ 1 Slave-hunting seems to be the intention attributed to the Arabs by Peter ofTaurinia, Miracula Pharmni (BHG 1509), 18,70.1031 (leelateo and cognates are used to refer to captive-taking in medieval Greek: e.g.
Theophanes, 2.757, s-v- leelasia). Captives are explicitly attested for the raids of 720, 727,730-1,732 and 752: see Appendix 4 under these dates.
42 Morrisson 1999.
43 Panella 1993, 674, with n259, thinks a few lamps were still landing around 700. The recent results from Naples are less encouraging. Some of the African lamps are probably earlier, "residual" deposits which found their way into the later 7th-C. context, so the change may have been even sharper than the numbers suggest, judging from a relatively small sample (fifty-six useful fragments) in the Carminiello ai Mannesi excavation at Naples: Garcea 1994, 325. In any case, arrivals were approaching the vanishing point.
510
5"
between Africa and southern Italy right after 700, but they can have been little more than a trickle.
The slackening of Arab raids on Sicily appears to have been followed by better relations. We have already seen that the last wave of the bubonic plague to assail early medieval Byzantium entered the empire from Africa. The rats which conveyed the disease certainly did not swim across the straits. In other words, ships were traveling between Sicily and Ifriqiya in the summer of745, probably loaded with cargo and certainly carrying rats.44 The further implication is that the economic integration of Sicily into the Muslim world had begun over three-quarters of a century before it was politically integrated by military conquest.
The earliest explicit witness to renewed commercial links between Italy and Africa connects them with the resurgent Adriatic, about three years after some ship transmitted the plague from Muslim north Africa to Sicily. The last Greek pope - who himself must have traveled from Athens over the old trunk route -discovered to his dismay that many Venetian slave dealers had set up a market in Rome, where they were buying up Christian slaves to sell to Islamic north Africa.45 Slaves were abundant enough in Rome around 748 to supply external markets; the demand for labor in Africa is obvious, given the plague that was still raging there. It was an unusual opening for foreign shippers, an opening seized by sailors from the upper Adriatic, a region not known to have had direct contacts with north Africa or western Italy, for that matter, since late antiquity. According to his biographer, Zachary put an end to the Venetians' enterprise and redeemed the slaves at cost. Whether or not he was fully successful, Rome was surely not the sole place in western Italy where the Venetians, who had come so far, were seeking slaves for Africa.
This incident was only the beginning of renewed communications between Africa and Italy. Around 767 the patriarch of Alexandria communicated with the pope and the Frankish king via the reviving route that connected Egypt to the Muslim west. He sent a monk to "Africa," where he caught a ship for Rome.46 This contrasts, revealingly, with another religious traveler's itinerary half a century before. Back then, the pilgrim had followed the old trunk route in its most northern incarnation: to get to Rome from Alexandria, he had sailed to Crete and then Constantinople. About a decade after the Alexandrian envoy had reached Rome from Ifriqiya, the markets of Africa again look like the most likely destination of the Greek slavers who were collecting their human cargoes along the central Italian coastline and exporting Lombards to the Muslim world; it is no
44 Note that no African raids on Sicily are recorded between 740 and 752/3: Appendix 4.
45 Pros. "Zacharias 1"; R157; for moreon the Roman slave market, Ch. 21.2.
46 R171. Pippin's ambassadors c. 768 myht have traveled via Africa.
coincidence that five years later, Arab gold coins materialize in Rome (R186; above, pp. 335O. A forgotten chapter of church history points in the same direction. Pope Hadrian I, during whose reign that Arab gold first reached Rome, was also credited with restoring Christian bishops to Africa (R181). He was but the first of the Carolingian-era popes to enter into relations with Africa (see below).
The African route was certainly followed by the envoys of Harun al Rashid and Charlemagne when they returned to Europe from Syria by traveling overland to Tunisia. There they found a ship bound for Pisa (R254-7). In the closing months of Charlemagne's reign, we catch another glimpse of shipping between Africa and Sicily - again, aboard Venetian ships - when ambassadors from Tunisia or, possibly, Morocco, visited the military governor of Sicily.47 The same year, a Christian in Africa sent a letter to an official in Sicily.48 All of this fits the picture of routine shipping links between Sicily and Africa drawn around the same time by the Italo-Greek inventor of the Life of Gregory of Agrigento, although he pretends to be depicting late antiquity.49
These travelers show us two very important things. The first is that Christian travelers began to avail themselves of the rising Muslim infrastructures of travel. This was true of Charlemagne's ambassadors. But it was also true of private travelers like the Alexandrian monk of 767 and Bernard. The latter's description of the camel renting service in Egypt (Ch. 5.1) exemplifies how western travelers naturally piggybacked on the locals' travel infrastructure. On a more personal level, in Charlemagne's lifetime it did not seem strange for a Christian and a Muslim to travel together as pilgrims from Egypt to Jerusalem, in order to venerate the shrines of their respective religions and, incidentally, take in some sights, notwithstanding the widespread opinion that early medieval people never traveled for pleasure.50
The second thing is that communications between Muslim Africa and C hristian Italy were increasing before the Muslims invaded Sicily. The gradual conquest of the island and its forcible integration into the African polity and economy naturally expanded the emerging communications network, peaceful and otherwise. In fact slave raids from Africa into Sicily and southern Italy had
47 R321. Amari 1933-9, 1:356-7, claims that an eyewitness account preserved in the Riyad an-nufus documents that the treaty provided for the free passage of merchants between Sicily and Africa, but his own rendering (Biblioteca 1.304-5) refers only more generally to a Sicilian agreement to free any Muslim who went to Sicily and wished to return home. As translated, the subsequent discussion mentions only Byzantine prisoners.
48 R320; their friendship means moreover that one of the men had probably traveled between the two places.
49 Leontius, V. Greg. Agrig., 5,149.1-150.16.
50 Leontius, Vita Stephani Sabaitae (BHG1670), 99-100, AASS Iul. 3.543F-544B; the Muslim accompanied the Christian to Mar Saba for the sake of the view. Dr. Mattina McGrath kindly alerted me to this passage.
512
resumed no later than 820, and probably as early as 812.51 From 826 on, military or corsair squadrons criss-crossed the straits of Sicily. In later years, they expanded their activities to the coastline of the Italian mainland and beyond. But the violent incorporation of Sicily into the vast Muslim economy implied commercial contacts as well. In fact, the nature of slave raiding along the shifting Muslim-Christian frontier makes commercial and bellicose contacts even more difficult to disassociate than had been the case in the eighth century. Take for instance, the story of twelve-year-old Elias the Younger, whom we saw sold to an African Christian slave trader right on the beach near his Sicilian home (see above, p. 245^). For the boy and for the Byzantine warship that intercepted the slaver, this was doubtless an act of war; for the merchant, his crew, the slave hunters and disappointed slave markets of Africa, it was a risky commercial enterprise gone very sour.
The aggregate patterns of Mediterranean communications have already undermined modern scholars' supposition that warfare ended trade. Shipping links between Africa and southern Italy further challenge that supposition. Ships transporting slaves from Italy continued to sail to Africa throughout the ninth century and into the tenth (below, pp. 768ff). Muslim-sponsored depredations in fact involved economic and social interactions with Christians which were more complex than one might first imagine. These included communities in addition to the Neapolitans (and with them, we may think, the Amalfitans), whose collaboration with the Arabs was notorious. For instance they helped conquer Messina, and supplied food, weapons, and other support so readily to the Arabs that Naples looked like Africa itself, according to a Frankish ruler.52 Some Christian pilgrims certainly and safely boarded ships laden with Italian slaves destined for the markets of Africa, raising the question of what sort of shippers transported Christians to the block in Egypt but respected the status of three Christian ecclesiastics and paying passengers (R579).
In fact, some non-belligerants sailed without difficulty between Italy and Africa in the midst of the raids, warfare, and slave shipments. Thus, in the 850s or 860s, a group of Frankish pilgrims returned to Rome from Jerusalem via Egypt and "Carthage" or, more probably, Tunis (R521). About twenty years later, another pilgrim, a manumitted Italo-Greek slave, made his way without observable difficulties from Antioch to Africa, and thence to Palermo, Taormina, and the Peloponnesus, even as he watched an Arab fleet preparing to sail against Byzantine towns, or fled an impending raid himself. A few years later, the same
51 R3 59; R308, where the Arabs are called "Mauri," which may pointto Africa rather than Spain. See, however, Guichard 1983, 62 and 64-5, for the possibility that Mauri
sometimes referred to Berbers settled in Spain.
52 R462 and Louis II, Ep. ad Basilium, 393-!7-*3.
514
Lan, now living near Byzantine Reggio, received visitors from Arab Sicily who informed him of his mother's death in Palermo.53 Communications had become 0 routine that around 892, an ecclesiastical legation traveled to Rome from Africa. They asked the pope to intervene in a schism which was developing in their church, and assumed he would be able to do so. So we are not surprised that an Italo-Byzantine abbot might send a monk to Africa to redeem five associates who had been carried off by Arab raiders. The monk was successful, and sailed [ home on an Arab ship (R502). Cases like these give substance to an African jurist's ruling that Christians sailing to Muslim ports on business were immune from attack (R514).
More than Tunisia was involved at the African end of this renascent shipping circuit. Whether or not envoys from the Atlantic imamate ofMorocco traveled to Sicily, contacts grew with Libya. Early in the ninth century, Roman fiction with unmistakable Sicilian connections depicts late Roman Tripoli as a junction for land and sea routes linking Italy and Africa to the Middle East. There is little evidence of such a route from the final century or two of antiquity, although archaeology suggests that Tripoli was faring better than its sister cities ofLepcis Magna and Sabratha.54 A few decades after the novel, in fact, papal correspondence proves Tripoli's new relations with Italy. In 853, the pope responded to reports reaching him from Tripoli by writing a critical letter to its bishop, Galerius (R512). A Greek monk, pilgrim, and small-time trader from Jerusalem took a ship to Rome from Tripoli a few years later (R519). Around 867, Bernard the Monk saw two slavers loaded with Europeans sail for Tripoli from Taranto just behind two ships, similarly loaded, that shoved off for "Africa" (R577-8).
4. Northern links: Tuscany, the upper Tyrrhenian Sea, and Liguria
A pattern of contraction and expansion also occurs further north, along the coast of Tuscany and in the upper Tyrrhenian. There the old shipping circuits had shrunk to a shadow of their former selves around 700. Early in the eighth century
Ik
Pros. "Elias 1."
1'ilgrims traveled overland from Carthage to Tripoli en route to Jerusalem: Leontius, v- Greg. Agrig. 9-10,154.1-155.1; pilgrims returning from Jerusalem traveled to 1 ripoli, where they found a ship belonging to the bishop of Palermo which had come there on business; they sailed aboard it to Licata and Agrigento, en route to Rome: 19,164.1-9. Hints ofTripoli's
success, founded on its oasis and remarkable harbor, predate the Arabs: Mattingly 1995,122-5; cr- the role Tripoli played in the Arab conquest of Roman Africa: e.g. Taha ig8g, 56-7. The 9th-C. nexus of shipping and the coastal road at Tripoli may have been reinforced by caravans arriving there from subsaharan Africa (with gold?) via Ghudamis, ifLewicki 1980,133-4, 's correct; see however Devisse 1982,173-82.
515
we see an occasional ship or two sailing through the Tuscan archipelago, linking the Ligurian and Tyrrhenian seas. But we know that, compared with a century before, this sea lane had declined.55 Still, the court of King Liutprand (713-44) at Pavia was able to send someone to Sardinia to purchase St. Augustine's relics -they had been moved there from Africa - and translate them to Pavia (R100). Coin hoards again confirm the pattern of contacts implied by the movement of relics-two separate finds on Sardinia (B42 and B44) mixed gold coins struck by the local Byzantine mint with tremisses issued by Liutprand.
By 754, the Lombards had conquered Corsica from the Byzantines, which means that there were enough ships along the coast to get them there from the mainland.50 Although at the closest point it lies some 80 km off the Italian mainland, the island was sufficiently integrated into the continental economy that the eighth-century bishop of Lucca's family owned land there.57 This also implies some local shipping linking Corsica to the coast, precisely via the Tuscan archipelago that centers on Elba. Around the same time, a family from the vicinity of Populonia regularly ran grain and salt up the coast in their ship; they later engaged to do so for the church of Lucca.58 In 776, the pope assumed that the duke of Lucca had access to "many ships"; the duke certainly acquired land on the coast, near the mouth of the Cecina river, which it is tempting to connect with shipping.59 A few decades later, a Corsican bishop joined the local prelate in presiding over an ecclesiastical court in Lucca (R317). Corsica was nonetheless isolated enough to serve as a place of exile (R380).
From 754 or, at the latest, 774, the popes expressed their interest in the island by laying claim to it, which would have been pretty pointless in the absence of sea links.faU In the next century, Gregory IV and Leo IV - the latter repeatedly - intervened across the Tyrrhenian Sea in Sardinia about ecclesiastical discipline (R401 and 496). Leo IV also solicited armed retainers from a Sardinian leader and asked him to buy a sea dye for him there, even as he acquiesced in the leader's request for a canonical favor (R503). John VIII too wrote to Sardinia (R625).
The papal communications imply improving ship transport in the upper Tyrrhenian Sea. This is only underscored by Sardinian and Corsican immigration to the mainland in the mid-ninth century. A "Sardinian quarter" appears at Rome, and the popes eagerly recruited the islanders into their military forces and domestic service, a pattern of intraregional movement not unfamiliar even
55 R.61, by ship, along the coast. On the shifting of sea vs. land routes from Italy to Gaul, seeCh. 3.3.
56 Schneider 1914,70^3.
57 CDL, 1, no. 114, p. 335.12-14; cf. Schneider 1914,79n3 for other mainlanders with property on the island.
58 CDL, 2, no. 223,26 August 768; cf. Schneider 1914, Ii6n2.
59 R.186; on Allo's coastal holdings, acquired in 782, and their connection with naval activities, Schwarzmaier 1972,167-8.
60 Hahn 1975,78-82.
516
europe's western mediterranean communications
today.61 Immigration suggests a disparity of economic opportunity between the Italian mainland and the island, and proves that the ship movements existed to make migration possible.
It is very hard to see beyond the Tuscan archipelago into the Ligurian Sea and along the old coastal arc that had linked Spain to the French and Italian rivieras The ship movements which had brought the "usual wares" as well as the plague from Spain to Gregory of Tours' Marseilles show few descendants for a long time And we now know that around 700, land links replaced the old coastal route between Italy and Gaul. Maritime developments may have followed a different later chronology from the Tyrrhenian littoral. Nonetheless, to the west, in 813 the count of Ampurias (Empuries) had access to ships to get to Majorca (R318). But more sustained signs of coastal merchant shipping set in around the 840s, when tolls were collected at Narbonne on merchant ships that frequented the French coast.62 Then too Charles the Bald granted to the bishop of Gerona - whose diocese included the landings of Ampurias and Rosas (Roses) - a third of the royal tolls on "merchants of land and sea."53 But the first hints that this western zone linked up with points east of Marseille come only in the 860s (Ch. 21.3).
In the ninth century, alongside the burgeoning land links, ship movements between the French and the Italian coasts increase somewhat. Envoys had sailed from Rome to Marseilles in 756 and 773, but in both instances, the sea route was considered exceptional (Ch. 3.3). As late as 820, travel between Italy and Spain was thought of in terms of roads, at least at Rome (R351). Two years later, however, we can detect a shorter, sea circuit. Louis the Pious granted to St. Victor of Marseilles the tolls on ships coming to Marseilles from Italy, a grant confirmed in 834-64 Ceramic manufactured in central Italy was certainly reaching southern France by the tenth century; the flow may actually have started in the ninth.65
61 Pope Leo IV's reign displays particularly close relations with Sardinia and Corsica. One of his earliest gifts was to the "vicus Sardorum": Liber pont., Duchesne,
2.108.12-13; cf- e-g^ I27-2I-3; 128.8-9; he also restored a ruined monastery of the "Corsi," ibid., 112.7-12; later he settled Corsican refugees atPorto: ibid., 125.27-127.7. Nicholas I had "domesticos suosgenere Sardos": ibid., 2.162.12-13.
62 Accepting that the editors are correct, and thatDChB 1, no. 49,20 June844, is reliable at 1.142.21; "deinceps concessum estab omni integritate, de quocunque commertio ex quo teloneus exigitur vel portaticus ac
de navibus circa littora maris discurrenti-bus." Cf. Ganshof 1959,161141.
63 D ChB i, no. 47,11 June 844, St. Sernin of Toulouse; cf. on the phrase "telonei mer-catorurnqueterremarisque," ibid., 3-35°> s.v. "mercator."
64 R372 and R426; it is unclear whether these particular revenues had already been granted by Charlemagne; the revenues in salt and "other condiriones" at "Leonio" may be the gth-C. survival of the old Merovingian customs station at Fos, according to Ganshof i960,133.
65 Bonifayetal. 1991, 44.
518
tuscany, northern tyrrhenian sea, liguria
Further west, Arab raiders who were perhaps based in Spain's rebellious eastern seaboard operated in the northwest Mediterranean from 798 to 813, before they moved to the greener pastures of the eastern Mediterranean. Charlemagne's defensive measures along the Tyrrhenian shore were triggered by these raiders; the defenses involved ships and may explain the raiders' departure from the region.66 Squadrons sent to protect Corsica successfully drove off Arab attacks in 806 and 807. That the count of Genoa was killed in the earlier battle shows that some of the ships came from that port.67 Unless we assume a Lombard war fleet that no one has ever suspected, they must have been cargo vessels requisitioned for the cause. Defensive measures for Corsica continued in the 820s.68 In fact, one Tuscan squadron even took the offensive, and raided the African shore. The ascendancy in Italy of the Bavarian family that commanded that raid owed much to Tuscany's naval development in those decades.69 At the close of the ninth century, Tuscan ships were again preying on African ones (R736)-
The overall picture which emerges suggests intensifying ship movements in the Tyrrhenian Sea. A good number were hostile, but even they entailed the buying and selling of booty, of slaves. Others were certainly mercantile, or at the least, non-commercial transport of goods. Raiding shows a subtle shift which confirms this picture. Earlier Arab forays looted coastal sites almost exclusively; die rare captured merchant ships of which we hear were Arab craft taken by the Byzantine navy (R159). In the ninth century, Arab raiders expanded their targets to include Christian ships as well as coastal settlements. The small crews of early medieval merchant mariners will not have produced large troops of slaves, so the ships themselves and whatever they were carrying must have been the prime attraction.
Like the Byzantine slavers of 776, some of the ships plying the Tyrrhenian waters in the ninth century were certainly merchant vessels. Notwithstanding some attacks, a lull in large-scale raiding in the northwestern Mediterranean may have set in for a decade or so after 813, if the Andalusian pirates who moved to Alexandria, and ultimately, Crete, had previously been operating in that area.70 1 he Prankish court nonetheless claimed that Arab pirates captured and sank a convoy of eight merchantmen returning from Sardinia to Italy in 820.71 A "large"
66 R24iandR275;Guichardi983.
°7 R269 and R276; see in general Schwarzmaier 1972,364-7.
68 No one except Lewis 1951,77-8, who claims that the Lombards gave "naval assistance" to Charles Martel in 738; his source, Paul the Deacon, Hist. Lang. 6,54, 183.22-5, says no such thing: R388.
69 R404 on Boniface's raid on Africa; Schwarzmaier 1972,172-3.
70 Guichard 1983, 66-8.
71 Even if this were an attempt to conceal a naval defeat, it is significant that trading ships could serve as a veil: R362.
Ill
519
Christian merchantman captured by Arab raiders off Sardinia led to an African legal conflict in the first half of the ninth century (R514). The same Islamic jurist who resolved that case declared another time that Christian ships trading with Christian ports were fair game for corsairs, but that Christian merchants who were known to do business with Islamic ports were off limits, whether they were near or far from their Muslim destination. The fact ofthe ruling itself implies that Christian trading ships posed recurring opportunities and problems for Arab raiders, while its terms show that some specialization in routes was occurring.7^ As the century advanced, Sardinia served as more than a hunting ground for slaves. By 873 it was a market where Byzantine sailors were selling off Christian slaves they had recaptured from the Arabs (R625). Once again the economic implications of warfare are unmistakable. A gold hoard (A32) buried on the island about a decade later mixes thirty-seven solidi issued at Constantinople between 829 and 879 and three African dinars dating from 850-1 to 883-4, as well as some pieces of gold ornament. The assembly of these coins on Sardinia points to at least indirect connections with distant places.
Whether they involved intermediaries or long-distance carriers, some ship movements linked the northeastern and southern zones of the western Mediterranean. Sea raiders from Africa struck Sardinia in 710, the 720s, 732-3 and 752, the last time in conjunction with attacks on Sicily.73 The threat had led the Lombard king to seek St. Augustine's relics, but it later subsided for a generation or two. Less bellicose shipping links with Africa are probably implied by the Greek slavers operating off Tuscany in 776; they are certain in the back-and-forth trips of Charlemagne's envoys in 801. When Isaac and the Arab envoys landed at Pisa, the head of Charlemagne's writing office assembled a convoy in Liguria which crossed over to Africa to fetch the famous elephant, and whatever survived of the Frankish delegation (R186; R254-6). Sicilian coins unearthed on the edge of Liguria and Tuscany tell a similar tale of contacts up and down the ninth-century Italian coast.74
About the same time that an Alexandrian ecclesiastical envoy sailed from Africa to Rome, Frankish ambassadors accompanied by envoys from Baghdad arrived at Marseilles; for the return trip, the Arabs returned to the same port (R175). A few Arab and Byzantine coins dribbled toward the Rhone still, or again (Table 12.4). Around 760 a large Byzantine fleet was rumored to be en route from Constantinople, via Sicily, to attack Rome and the Frankish seaboard, so the idea
ui, 1 i\IN   ůCrt,   L1VJ u Rin
72 It is just possible that the traditional port of the Arabs mentioned in 869 near Aries was or had started out as an emporium for Arab merchants: see R596.
73 R76,Rio6,Rn2,Ri2i,Ri36,Ri5g;cf. R107 and R116; Boscolo 1980 rehearses the
74
events, based largely on Amari i933-9! where the authors seem to differ, I have followed Amari.
Appendix 3, B28;six out ofseven date from 821-56.
of such ship movements was plausible (R165). Relics which were brought from Carthage by a Carolingian legation reached Aries by sea around 801 (R257). Spanish slave hunters touched the threshold of the eastern Mediterranean when tney attacked the Byzantine monks on Pantelleria (R272). A pilgrim from Constantinople is supposed to have sailed to Sardinia from Constantinople via Africa a decade or so later (R333). In 828, a Frankish squadron commanded by Count Boniface circumnavigated Corsica. On Sardinia he found pilots who were knowledgeable enough to guide his ships to the Gulf of Tunis, which he raided (R404). This suggests that at this date, Tuscan sailors knew their way to the islands, but not to Africa, whereas Sardinia had stayed in contact with its former provincial capital. A pilgrim from the east reached Rome from Corsica and ultimately traveled on to France, via Genoa (R472). So, even without a rabbinical responsum which has been erroneously interpreted to reveal northern traders in Africa, long-range movements in the western Mediterranean were picking up.75
The renewed traffic within the Tyrrhenian Sea and its occasional extension to Africa is important in its own right. It is even more important with respect to the broader pattern ofwestern European links to the Middle East. For it is difficult to imagine that links across the Tyrrhenian Sea and Sicily to Africa do not reflect, indirectly, communication patterns within the Caliphate, linking the Muslim Maghreb, or "west," to the east.
The traffic along the southern rim of the Mediterranean leads to yet another, more indirect route which came to link the Carolingian west with the Middle East. By the middle of the ninth century, one could reach the Frankish empire from the Levant via Spain. This was the route the monk George intended to improvise once he had reached Cordova from Palestine, and it may well have been the route followed, most ofthe way, by the Spanish pilgrim Hyacinth, traveling in the other direction.76 George's body ultimately did travel all the way to Paris, in the care of
75
Agus 1965,1, no. 1, p. 53 claims thatRabbi Paltai Gaon observed that Jewish traders from Germany traveled to Africa almost continuously in summer. I am grateful to my colleague Bernard Septimus foran enlighteningdiscussionofthis Babylonian wponsum from the 840s. Among the numerous points he clarified forme, there is "othingtoconnectthetextwith Africa; the conjectural emendation of Ashkenaziim to Arviim is groundless; the text and its logical structure indicate that theAshkenaziim in questionare not Jews, but traders dealing with Jews; the meaningofAshkenaziim is uncertain, since Paltai'stextisabouta
century earlier than the first sure use ofthe word to mean "Germans." Anotherpossible interpretation ofthe word might be "Slavs," so the text might point to trading contacts in central or eastern Europe. Babylonian scholars in this period had contacts with Jewish communities in Europe (Ch. 22.3). I wonder if this might warrant locating the addressees in Spain, to which we know that caravans of German traders were traveling inthemid-9thC. (below, p. 677)?Other options would include Italy or, possibly, central Europe, as well as the eastern terminus ofthe northern arc. 76 Pros. "Georgius8"; "Hyacinthus."
520
521
Usuard, whose detailed account illustrates the routes connecting Cordova and France (R528). An Arab observer explicitly notes that some traders traveled from Frankland to the Middle East (and beyond) via Spain and the land route across north Africa.77 However many (or few) travelers or traders followed such a circuitous route from Frankland to the Middle East, it seems clear that interlocking communications and probably also trade zones provided yet another alternative in the ninth century for long-distance exchanges. Certainly Jewish and Christian traders as well as some diplomats traveled between Frankland and Spain. Their movements were probably reinforced by the powerful Frankish bridgehead around Barcelona.78 Since most of this indirect route lies within the House of Islam, it is enough for us to have recognized its implications for our theme, and leave more exhaustive study to those who are better equipped for it.
Around 700, western Europe was joined to the Middle East only by sea, only through Byzantium, and only sporadically. A century or so later, shipping had become considerably less sparse in the Tyrrhenian Sea. Sea links between Italy and Africa, between the Tyrrhenian and Ligurian Seas, between Italy and France had come back to life. The north-south sea route connected western Italy to the major communications axis of the Islamic world that ran along the Mediterranean's southern rim. Within a generation, another almost entirely overland route was available through Spain. The birth or rebirth of these routes was important indeed. But the most dramatic change was occurring elsewhere. It was occurring in the once dormant Adriatic Sea.
77 Ibn Khurradadhbih, trans. Jacobi, p. 253; for the overland and sea routes linking Africa with Spain, see Ya'kubl, trans. Wiet, pp. 210-11 and 217.
78 On diplomats, El-Hajji 1970,119-52; one
suspects that this was the route by which a rabbinical letter traveled from Syria to Barcelona in the middle of the gth C: R526; for traders, see Ch. 23.3.
522
18
Venetian breakthrough: European communications in the central Mediterranean
Th e Venetians' arrival on the western shores of Italy around 748 casts a sudden shaft of light on the central Mediterranean. The Adriatic returns to life in the eighth century amidst a flurry of ship movements. A couple of generations later, another communications corridor funneled into the Ionian Sea and reinforced it: the land and sea routes which converged at the gulf of Corinth. Tracing the return of shipping to these two great corridors will fill out our picture of the origins of medieval European shipping patterns and allow us to address their structure.
1. Venetian breakthrough
In the twilight of antiquity, local shipping in the upper Adriatic sputtered along.1 As empire ended, the Roman population had fled offshore to security among marshy lagoons and islands; necessarily, boats moved them around their watery world. Some administrative traffic up the 740 km-Iong sea had persisted through the late seventh and earlier eighth centuries, funning a lifeline to outposts like Ravenna or those along the Balkan coastline. But the ties that bound were stretching ever thinner.2 No imperial
1 E.g., the duke ofFriuli had sailed from Istria to Ravenna and then up the Po to Pavia in the late 7th C: Paul the Deacon, Hist. Lang. 6,3,165.15-16, cf. Krahwinkler <992, 53-4; the Venetians participated in an attack on Ravenna c. 732-3: R137.
i For the deportation of prisoners and their return to Ravenna c. 709-12, see R70-2; an embassy from Ravenna to Constantinople, Rgi;on trading activity, Claude 1985b,
151-2 and on the administrative ties of Ravenna, Istria, and Constantinople in general, Brown 1984,144-63. On present archaeology, the recently discovered port of Classe ceased functioning in the 7th C: Maioli and Stoppioni 1987,33-55. The chronology of other landings further east and north which may have supplemented this port is unclear: see Schmiedt 1978, 223-6 and pi. xlvi. The Rižana Plea,
523