Chapter 1

Background

If an ethnic group is evaluated solely by the criterion of military success, the Normans were unquestionably the most successful people in eleventhcentury Europe. While best known for the conquest of the Kingdom of England in the late 1060s, both professional and popular historians have, particularly in the last few decades, increasingly turned their attention to the lesser known, but by no means less significant, Norman conquests in Italy, Sicily and Syria. While the reasons for this heightened interest are many, some of the appeal is no doubt attributable to the ‘rags-to-riches’ or ‘against-allodds’ nature of these territorial acquisitions. The conquest of England was a ‘state-sanctioned’ operation led by the Duke of Normandy himself and his most preeminent subordinates and allies, whereas the leaders responsible for conquests in southern Italy, Sicily and Syria often came from the middling ranks of Norman society; they were variously political exiles or disenfranchised sons hailing from relatively meagre patrimonies. Moreover, while the leading families participated in the duke’s invasion in 1066 according variously to tenurial obligations and the promise of increased wealth and prestige, their counterparts in the Mediterranean, despite also being clearly driven by the desire to improve their station in life, did not do so at the request of their duke. Rather, drawn initially by the offers of military service in Italy, the ‘help’, over the course of several decades, gradually became the rulers of their former employers.

Those responsible for the majority of the conquests in the Mediterranean hailed from the Hauteville family, whose military campaigns in the Mediterranean world will be the focus of this book. On the progenitor of these remarkable mercenaries, conquerors, and statesmen, Tancred of Hauteville (c. 980–1041), unfortunately little is known. Despite the concerted efforts of the (almost certainly) Norman chronicler Geoffrey Malaterra to indicate otherwise, at the time of his death Tancred of Hauteville was, at best, a relatively insignificant subordinate of the duke of Normandy, better known to posterity as ‘William the Conqueror’ (r. 1035–87). The siring of sons was clearly of paramount importance, for a total of twelve were borne to Tancred from two wives, Murielle and Frédesende. Four sons would later become counts in Apulia, and one of them, from Tancred’s second marriage, Robert ‘the Cunning’ (Guiscard), became duke of Apulia, Calabria, and Sicily (r. 1059–85). Robert’s son Marc, better known by his nickname ‘Bohemond’, proved to be the greatest of the commanders of the First Crusade (1096–99), during which he established a formidable principality centred on Antioch in northern Syria. But Bohemond would not be the only famous grandson of Tancred of Hauteville. Roger II, the son of Guiscard’s younger brother from the same mother, Count Roger I, became the first king of Sicily, founding a political institution that survived until Giuseppe Garibaldi’s conquest of Sicily in 1860. Another Tancred, the last of those whom his biographer Raoul of Caen called the ‘Guiscardians’, and whose name evidently commemorated his great-grandfather’s, lived much of his life under the shadow of his maternal uncle Bohemond. He was, however, more successful than his uncle when it came to extending the frontiers of the principality of Antioch, doing so at the expense of the neighbouring Byzantine, Armenian, and Selçuk powers. Upon his death, Tancred, as regent of Antioch (r. 1100–03, 1104–12), was in equal measure feared and admired by both his allies and enemies.¹

Southern Italy

The first region of the Mediterranean that the Normans would make their own at the beginning of the eleventh century was a cultural, political, religious and ethnic melting-pot. The inhabitants of modern-day Puglia (Apulia), Basilicata, and Calabria were subjects of the Byzantine emperor. The solider emperor Nikephoros II Phokas (r. 963–9) reorganized the administrative structure of these regions, which had previously been divided into two ‘themes’ governed by generals invested with both civil and military authority. Apulia, formerly the theme of Longobardia, became the ‘katepanate of Italy’. A katepanate was a larger, more important administrative province of the empire, consisting of themes. The other province, the theme of Calabria, was retained, and a new one was created in Loukania (Basilicata). While now under the direct authority of the katepano stationed at Bari in Apulia/Longobardia, the other governors nevertheless exercised a degree of administrative autonomy. Campania, to the west of Apulia, was ruled by three Lombard princes based at Salerno, Capua, and Benevento. However, in the same region, the maritime cities of Amalfi, Naples, and Gaeta were governed by dukes, who despite wielding independent rule, all maintained strong economic and political links with the Byzantine empire. Lastly, although the final imperial possession of Taormina resisted capture until 902, Sicily had effectively changed from Byzantine to Muslim hands by the middle of the ninth century.²

Such was the intricate, overlapping political landscape, but even more intricate was the ethnic one. Apulia/Longobardia, aside from the Greekspeaking south, was predominantly peopled by Lombards. Originally based in Pannonia (Hungary), the Lombards established a kingdom in northern Italy in 568, but their rule was restricted to southern Italy after Charlemagne’s capture of the capital Pavia in 774. A tenth-century Salernitan chronicler referred to the ‘German tongue, long ago spoken by the Lombards’, reflecting that they had gradually integrated with the local population, who greatly outnumbered them. They were therefore essentially Italians, although many, most notably the ruling class, continued to identify strongly with the Lombard heritage even though they spoke an Italian, rather than Germanic, dialect. This continuing ethnic pride is evidenced in one of the local annals written by the Lombards of Bari, which concludes the entry with: ‘and it has been 400 years since the Lombards entered Italy’. Similarly outnumbered by Lombards in much of northern and central Apulia, the Byzantines were content to let the locals be ruled in accordance with ancestral laws – first codified by King Rothair in 643 and substantially revised by King Liudprand (r. 712–44) – under their own officials. Positions within the imperial administration were also open to Lombards, although the office of katepano was restricted to nobles from Constantinople. While some historians have tended to style Lombard leaders who rebelled against the katepano as ‘patriots’ wanting to free themselves from servitude, to use the words of Chris Wickham, they were generally loyal and the exceptions to this general rule were ‘ad hoc hostile responses to individual administrators’.³

Calabria, like southern Apulia, was almost entirely populated by speakers of the official Byzantine language (Greek). In the Campanian region, again the Lombards formed the dominant ethnic contingent, although Greek was still known among some of the élite in the maritime cities. Naturally, since it had been under Muslim rule for over a century by the time the Normans arrived in southern Italy, Sicily possessed large numbers of Arabic-speakers, perhaps comprising up to two-thirds of the island’s population. The non-Muslim inhabitants generally spoke Greek, and were mainly concentrated in the north-eastern area of the island, the Val Dèmone.⁴

The complex political and ethnic makeup of the south was reflected in religion too. First of all, Christians were split between those who recognised the ecclesiastical authority of the Roman patriarch, and those who instead acknowledged his Constantinopolitan counterpart. Both the inhabitants of Campania and the Lombards of Apulia followed the Latin liturgy, and those of southern Apulia, Calabria, and Sicily observed the Greek one. Prior to the rule of the Byzantine emperor Leo III (r. 717–41), the papacy possessed ecclesiastical jurisdiction in southern Italy. However, on account of the Roman Church’s opposition to Leo’s policy of iconoclasm – i.e. prohibition against the worship of icons – the incensed emperor transferred this jurisdiction to the patriarch of Constantinople. Such a move was greatly resented by the papacy, and it was in the eleventh century that the papal curia began to reassert its rights more forcefully in the region. But another important power bloc in western Europe also lay claim to southern Italy: the Germanic emperors of the Salian dynasty, who much like their Ottonian predecessors shared Charlemagne’s belief that the south belonged to the western, not eastern, emperor (as had generally been the case until the deposition of the last western emperor Romulus in 476).⁵

Determining the precise nature of military administration in the regions ruled by Lombards and Byzantines is a difficult task. The former, originally referred to collectively as an ‘army’ rather than a people, fielded a force consisting of ‘freemen’ serving under the leaders of prominent families (farae). Freemen, more frequently identified as ‘soldiers’ (arimanni), were expected not only to provide military service, but also to equip themselves. Those freemen who were wealthier than others were required by the laws of King Aistulf (r. 749–56) to serve as heavy cavalry (i.e. to equip themselves with horses, armour, shields and lances). The men of middling means (minores homines) fought as light cavalrymen (horses, shields and lances, but no armour), and the less financially fortunate as infantry (shields, bows and arrows). These troop types and their equipment were determined by the amount of land held, and while Lombard society had theoretically consisted of a single ‘class’ – there are no explicit references to an aristocracy in the various laws – there were effectively three tiers based on prestige and wealth (farae, arimanni and minores homines). As is evident in a charter of Prince Radelchis I of Benevento (r. 839–51), a century later these property-based groups had become more precisely defined as ‘nobles’, ‘the middling’ and ‘peasants’ (reiterated in a 992 charter with similar terminology). Soldiers were levied from these classes by regional officials hailing from the upper tier of society, defined in the tenth century as consisting of the nobles and their princes, who were entrusted with taxation in addition to military matters. This system was employed throughout Campania, in addition to northern and central Apulia, which, until the Byzantine reconquest in the last quarter of the ninth century, was subject to Lombard rule.⁶

The troops deployed in pitched engagements, however, seem to have been predominantly levied from urban areas, whereas the ones tasked with defence of fortified settlements (castra) – not ‘castles’ or ‘fortresses’ as the Latin term meant in the contemporary French context – were recruited from among the rural population. The mention of ‘the cavalrymen of Salerno’, led in 1052 by Prince Guaimar IV and including among them his brother-in-law Landulf, is suggestive of an élite corps drawn exclusively from the upper echelons. Substantial lay estates are known to have furnished mounted troops, as did those in the ecclesiastical sphere. As for light cavalrymen, in former times recruited from the middling ranks, little is known. However, they almost certainly continued to exist and, as will be argued in the next chapter, can be identified with those known in Apulia as conterati. Lastly, it seems evident that in the eleventh century Lombard infantry continued to consist mainly of missile troops: in addition to the traditional archers, there are various references to the use of slingers (e.g. Prince Gisulf II of Salerno deployed men armed with ‘sling and bow’ in the 1050s). Supplementing these various troop types from the ninth century onwards, when needed, were mercenaries. Arabic-speakers augmented Lombard armies during the civil war of 839–49 and, much like their eleventh-century Norman counterparts, while militarily effective they were rapacious and difficult to control.⁷

While there were detachments of both provincial and élite troops known to have been stationed at Bari in the early eleventh century, their average numerical strength remains unknown. Yet since reinforcements were regularly sent to Italy in times of crisis, it follows that there were never very many. Indeed, the élite troops mentioned were more often officers than divisions or regiments, whose role was presumably to oversee defence, military administration, training and recruitment. The imperial administration in northern and central Longobardia therefore relied predominantly on Lombards – some of whom were ‘Byzantinized’ – recruited from the abovementioned classes, and hence could theoretically field an army of heavy and light cavalry in addition to infantry. The bulk of these men were recruited in accordance with their obligation to provide military service as a form of rent, a duty first referred to in a charter witnessed at Conversano in 980. The local population was also expected to provide ships and pay a naval tax, the proceeds of which presumably paid for the maintenance of the fleet and its marines. Much like the Lombards, in times of need the Byzantine administration hired mercenaries. In 1041, for example, the katepano was said to have been ordered by Emperor Michael IV to hire experienced cavalry from his ‘land’ (i.e. Longobardia/Apulia) in order to help counter the incursions of Lombardo-Norman rebels.⁸

Recruitment and attendant obligations in the theme of Calabria were similar, although the most obvious difference is that the troops, like those in southern Longobardia, were Greek-speakers rather than Lombards. The naval tax is attested in the tenth century, and that it was particularly onerous is apparent given that the citizens of Rossano revolted when it was levied in c. 965 by the governor Nikephoros Hexakionites. The number of ships maintained in Calabria seems to have been considerable, for the fleet led by the Normans in 1061 – i.e. after the region and its capital had been capitulated – was said to have outnumbered the twenty-four Muslim ships sent to oppose its crossing to Sicily. Admittedly this fleet did include ships from Apulia, but Calabria evidently contributed the lion’s share, as there were clearly enough battle-ready ships stationed at Reggio to engage a Muslim fleet in the Strait of Messina prior to the launch of the campaign. Although the last explicit mention of the building of galleys with two banks of oars (chelandia) – used for combat and the transport of cavalry – in Calabria dates to c. 965, clearly a region afflicted by seaborne raids launched from Muslim Sicily in the ninth to eleventh centuries required the maintenance of a fleet to guard against them. That chelandia continued to be produced is confirmed by the knowledge that the Byzantines procured ships from the region in order to transport troops to Sicily (1038), as did their Norman successors in 1061, who had received the surrender of the capital Reggio in the previous year. Lastly, a Calabrian ship identified specifically as a chelandion (sing.), was captured and burnt during a naval engagement in the Adriatic (c. 1064).⁹

The Calabrian theme, originally part of the theme of Sicily, was probably divided into three tactical divisions known as tourmai; at least this was the standard subdivision recommended in various military manuals. The one that anything is specifically known about is the Saline tourma (sing.) based at Oppido in the hinterland of the south-western coast (36km north-east of Reggio), and so called owing to the town’s location in the Saline Valley (the plain of Gioia Tauro). The fortified hilltop town was founded in c. 1044, and the tourma stationed there was subdivided into three vanda/droungoi of cavalry, each under the command of a count. Where the other two tourmai were located is uncertain, but presumably one was at the capital on the south coast (Reggio), and the other at Squillace on the east coast (where Reggio’s commanders took refuge when their city fell to the Normans in 1060). The size of Byzantine units was never static, but two military manuals, the first from the 960s and the second from c. 1000, state that a vandon (sing.) comprised ‘fifty men’. Hence the number of cavalry available to the Byzantine administration in Calabria was around 450 (assuming those in the other six vanda were equivalent in size to the three known to have been stationed at Oppido). Other than that, a signatory to a 1053/4 monastic charter in the same vicinity identified himself as Cyril, commander of the elsewhere unattested unit of ‘Hungarians’, whose numerical strength is therefore unknown. Any other troops – i.e. infantry and light cavalry – were presumably levied when required from the militia residing in the villages and fortified settlements. How such troops were recruited and trained is difficult to ascertain, but no doubt the method employed in Longobardia was also used in Calabria (i.e. by officers from the professional regiments). Indeed, men called Scrivones, presumably from the guards regiment associated with the regiment of ‘the Sentinels’ (Exkouvitoi), were noted by two sources to have been present at Crotone in the 1050s.¹⁰

Apulia and Calabria had developed in line with a tenth-century Italian phenomenon known to specialists as incastellamento: the fortification of either pre-existing or newly founded ‘nucleated settlements’ (castra), as Barbara Kreutz puts it. Such settlements were the opposite of the equivalent in northern France: to rephrase Chris Wickham’s words, those who worked the land resided within the fortified areas, rather than outside them. Apulia/Longobardia was much more urbanised than Calabria; accordingly many people lived in the heavily fortified towns on the east coast (e.g. Trani in the North, Bari in the centre, and Brindisi in the south), although considerable numbers resided in settlements located in the hinterlands of these urban centres. Calabria was altogether different. With Squillace and the capital Reggio being notable exceptions, coastal regions were abandoned in the tenth century in the wake of frequent raids by Arab pirates. Henceforth the majority of the population lived in inland settlements, many of which were fortified. There were, however, a few strongholds established that would equate with the contemporary French meaning of ‘castles’ or ‘fortresses’: for example, the Rocca Niceforo (Angitola) at Maierato in the modern province of Vibo Valentia, named after the abovementioned emperor Nikephoros II Phokas, and built during his reign (960s). There was nothing novel about the building of such a lofty fortress in Italy during Phokas’ reign: Cyprus was seized in 964, after which construction of the impressive fortifications in the Pentadaktylos mountains of the island’s north began. Kantara fortress (Fig. 7), which surely dates from this period, is situated 630m above sea level and, like its equivalents in Italy, was later reinforced by French-speakers. Cyprus, like southern Italy, was a frontier region vulnerable to seaborne invasions and raids; fortified lookout posts were therefore of paramount importance.¹¹

The Byzantine Empire

The second region of the Mediterranean in which the Normans would both make a name for themselves and even conquer portions of – mainly in the short term (Balkans), although three Ionian islands were permanently seized in the 1180s – was the (Eastern) Roman or Byzantine empire. If dated to the inauguration of Constantinople (Istanbul) or ‘New Rome’ by the Roman emperor Constantine in 330, the empire had been in existence for seven centuries by the time the Normans came into direct contact with it, and rather more than that if extended to the principate of the empire’s first ruler, Augustus (r. 27 BC–14 AD). From foundation until the date traditionally assigned to the collapse of the western Roman empire (476), Constantinopolitan emperors had either been the sole or eastern rulers of the Roman empire. Despite the empire’s Latin-speaking roots, by the reign of Herakleios I (r. 610–41), Greek, the lingua franca of the eastern Mediterranean since the age of Alexander the Great’s successors, was declared as the official imperial language. Accordingly, despite referring to themselves as Romans (Romaioi) – and being called that (al-Rūm) by other cultures in the Near East, Africa and Sicily – in western Europe the Byzantines were identified in Latin as ‘Greeks’. Given the empire’s multiethnic, polyglot composition, the term could not have been more inaccurate. Moreover, unbeknown to western European contemporaries, the term Ellenes (‘Greeks’) had since the fourth century transformed from an ethnonym into a pejorative term: as the Roman Pantheon had long been dominated by gods and cults of Grecian and Hellenistic origin, its adherents were called ‘pagans’ – that is, Ellenes. Accordingly, the subjects of the Eastern Roman emperors, including the former ones in Apulia, Calabria, Sicily and Syria, will be called ‘Byzantines’ or ‘Greek-speakers’ throughout this book.¹²

Despite losing Syria, Palestine, and Egypt to the Arabs during the seventh century, the empire remained both wealthy and powerful. At the heart of Byzantine martial effectiveness were the abovementioned administrative and military districts known as ‘themes’. Developed in response to the seventh-century Arab invasions, Byzantine provinces were equipped with semiprofessional, part-time troops – locally recruited, trained, and equipped – who were supplemented when needed by professional, permanent forces stationed at the capital: the tagmata. The latter were normally heavy cavalrymen (kataphraktoi), whereas thematic troops were generally foot soldiers and light cavalrymen, although some were kataphraktoi. Lightly armoured cavalrymen, sometimes collectively referred to as ‘cavalry themes’, formed the bulk of forces in the frontier provinces, where warfare more regularly consisted of skirmishing and ambuscades rather than pitched battles. Ascertaining the tactics and equipment of eleventh-century heavy cavalry is difficult to determine precisely, let alone summarize in brief, and, as is the case with the Normans, imposing a uniform model detracts from their evident fluidity and adaptability. The heaviest form of imperial cavalrymen since the sixth century were covered in armour (head to toe), as were their horses. Armed variously with maces, lances, swords, axes and bows (Fig. 4), they advanced at a trot, and their tactical purpose was to exploit openings created by infantry and light cavalry. John Haldon theorizes that these expensive to maintain ‘oven-bearers’ (klivanophoroi) might have been phased out during the reign of Constantine X (r. 1059–67). If so, their empty ranks were presumably filled by the other form of heavy cavalrymen who were similarly armoured (yet not as heavily), but whose mounts were unencumbered by a layer of protection. Such men seem to have been a feature of field armies since at least the second half of the tenth century, and were presumably the ones who launched an irresistible charge during a battle in 971 (the heavier panoply of the klivanophoroi and their mounts was probably too weighty for the type of charge described). Evidently John I Tzimiskes’ ‘Immortals’ surged forth at full gallop with their lances couched (locked under their right arms) in this battle, for a contemporary source related that ‘violently pricking the spurs to their horses’, they charged with their ‘lances held before them’. However, imperial cavalry adapted to whatever tactical situation faced them – and they may well have been grouped into specialist divisions – for Jonathan Shepard notes that while some contingents in the eleventh century wielded throwing spears, other battalions such as the Immortals (reformed in the 1070s) charged at full tilt with lances. Importantly, the use of the cavalry spear both as a thrusting and throwing weapon is supported by contemporary pictorial evidence – that is, the Theodore Psalter dated to 1066 (Fig. 3).¹³

The medieval incarnation of the Roman empire reached the zenith of both its military power and territorial extent during the reign of Basil II ‘the Bulgar Slayer’ (r. 976–1025). As his nickname suggests, the emperor was hewn from the rock of bellicosity. In addition to the great ability of the emperor and his generals, the military successes of Basil’s reign were in no small part due to the highly effective forces left to him by his equally warlike and expansionist predecessors: Nikephoros II Phokas (r. 963–9), and John I Tzimiskes (r. 969–76), two of the greatest military commanders of the entire medieval period. At Basil’s death, the empire spanned from southern Italy in the west to Crete in the south, to the Danube in the north and to the Euphrates in the east. However, only two militarily-minded emperors reigned briefly in the half century following Basil’s death – Isaac I Komnenos (r. 1057–9) and Romanos IV Diogenes (r. 1068–71) – resulting in an incremental decline in both the numerical strength and overall effectiveness of the imperial army. By the reign of Constantine IX Monomachos (r. 1042–55), the well-endowed treasury left by Basil II had been emptied. A fiscal crisis accordingly developed, resulting in the first-ever debasement of the gold currency established by Emperor Constantine I in 312. But this controversial precedent was not enough to rectify the empire’s financial problems and, as a result, the cost-cutting officials of the imperial bureaucracy turned to the army. The war-hardened troops from the theme of Armenia, estimated by Warren Treadgold to have formed around a fifth of the entire imperial army, were demobilized.¹⁴

Mercenaries were increasingly recruited and, with southern Italy acting as the conduit, French-speakers became prominent among them for more than a century. Substantial Norman and French recruitment had technically begun before the period of financial crisis – when around 300 of them served during the Sicilian campaign of 1038–40 – yet they were to be found increasingly among Byzantine forces from Constantine IX’s reign onwards. Included among these early mercenaries were two sons from Tancred of Hauteville’s first marriage, William and Drogo, who were later to become the first two counts of the Normans in Apulia. As will become increasingly apparent in the pertinent chapters of this book, over the next four decades Normans would both protect and undermine the Byzantine empire on its western and eastern frontiers.¹⁵

Although the empire was weakened by mismanagement, civil wars and a series of incursions on its frontiers, it continued to be a great power until the end of the twelfth century. Moreover, while mercenaries became more prominent, Byzantine troops still featured prominently in the ranks of the imperial army. The word ‘mercenary’, much like the use of it in regard to Hauteville troops, is something of a misnomer. French-speaking battalions, for example, were classed as tagmata; they were permanent soldiers paid with a wage or stipend. As for the power of the empire under the Komnenian emperors Alexios I (r. 1081–1118), John II, and Manuel I (d. 1180), even some of the Latin sources, which were increasingly hostile in the twelfth century, commented on its military technology and combat effectiveness. For example, John II Komnenos (r. 1118–43) was described by William of Tyre in the 1180s as ‘the most powerful prince of the entire world’, who led an army ‘which none of the kings of the earth seemed able to hold back’. William also described Manuel’s great armada of 150 ships (1169). The learned historian was clearly impressed by the dual-masted galleys fitted with sophisticated cavalry transport technology, in addition to larger ships capable of ferrying formidable siege engines.¹⁶

Syria

The third region of the Mediterranean in which Normans would found a long-lasting polity was Syria. Aside from the Byzantine reconquest of northern Syria (modern southern Turkey) in the 960s, including ‘Antioch the Great’ as it was known, much of the region remained under Arab rule until the middle of the eleventh century. By the second half of the same century, however, a new Muslim power had established itself in the Middle East. Originating from Central Asia, the Selçuk Turks were in control of Persia (Iran) and Mesopotamia (eastern Syria, Iraq, and Kuwait) by 1055. Having received formal recognition from the Sunni caliph at Baghdad, the Selçuk dynasty turned its attention towards the west. Prior to the Selçuk incursions, much of Syria was either subject to the rival Fatimid dynasty based in Egypt, or to independent rulers of local origin. However, by the 1070s Selçuk armies had seized authority in eastern Anatolia (Turkey), southern Syria and Palestine. While two of the greatest cities in northern Syria remained in the hands of the semi-independent Byzantine duke Philaretos Vrachamios – a rebel before coming to terms with Emperor Michael VII in 1078 – they had nonetheless been captured by the end of the following decade: Antioch in 1084 and Edessa in 1087. Selçuk authority in Syria, however, was soon greatly undermined by the rivalry between Tutuş, governor of Damascus, and his brother, Sultan Melikşah (r. 1072–92), and between Tutuş and Süleyman, the Selçuk ruler of eastern Anatolia (r. 1077–86). Süleyman, having wrested Antioch from the Byzantines in December 1084, naturally came into conflict with Tutuş when attempting to seize Aleppo in 1086, for the latter had long wanted to add the city to his Syrian dominions. The prize he sought was soon gained, for Tutuş seized control of Aleppo after defeating and killing Süleyman. Relations between Tutuş and his brother Melikşah remained tense, and after the latter’s death in 1092, the former attempted to secure the sultanate for himself. Despite a great victory in the sultanate’s heartland of Khorasan in May 1094, in June of the following year Tutuş was killed in battle against Melikşah’s nephew and successor, Berkyaruk (r. 1092–1105). His sons, Dukak and Rıdvan, now fought each other for the right to inherit their father’s lordship in Syria, respectively establishing themselves in Damascus and Aleppo. During this chaotic period, various towns and cities both on the coast and in the Taurus Mountains to the north became politically independent, variously being ruled by Armenians and Arabs of both the Sunni and Shia forms of Islam. The latter, in the coastal towns, looked favourably towards Fatimid Egypt, from where they hoped to receive military assistance. As the twelfth-century Arab historian Ibn al-Qalānisī lamented, ‘the peoples of Khurusan, Iraq, and Syria were in a state of constant bickering and hatred, wars and disorder’. Naturally, as had earlier been the case in Sicily, the Hauteville scions would benefit from, and even exploit, the political disintegration and rivalries in Syria, and in a wider sense it has often been noted that had the Muslims united against a common foe, the crusaders may well have been annihilated well before they reached southern Anatolia, let alone Syria and Palestine.¹⁷

Like southern Italy, Syria was populated by peoples of various ethnicities and faiths. Though predominantly featuring adherents of the Sunni branch, devotees of the Shia form of Islam formed a substantial minority, particularly in cities such as Tripoli (Lebanon), which had been under Fatimid suzerainty. In addition to Jews, there was a substantial number of Christians – Melchites, Maronites, Jacobites, Armenians and Nestorians – themselves divided like the Muslims when it came to interpretations of their faith. In addition to the various faiths many languages were spoken: Greek, Syriac, Armenian, Arabic and Turkish. By June 1098 another tongue had been added to the mix: northern French. The speakers of this language, it is interesting to note, were thought to have been Byzantine mercenaries by the Muslims when they captured Antioch: a misconception commemorated in an anonymous eyewitness chronicle, which had Gürboğa (Kerbogha), commander of Mosul (N Iraq), querying whether these men were related to the ones ‘who drove all of our ancestors from Romania [Byz. empire] and from the royal city of Antioch, which is the honourable capital of all Syria?’¹⁸

Scholars have in the past questioned whether both the duchy in Italy and the principality of Antioch can accurately be called ‘Norman’, but Graham Loud and Alan Murray have shown respectively that around two-thirds of lordships in both polities were held by Normans (Loud notes the same fraction generally applied to those active in the conquest period in Italy). As had long been the case in southern Italy, it is clear that the core of the Antiochene prince or regent’s cavalry was drawn from his military household, and from those maintained by his vassals or ‘barons’ (e.g. Alan of al-Atārib), a practice well attested during the regency of Tancred’s successor, Roger of Salerno (r. 1113–19). Unsurprisingly, some of the most preeminent vassals were members of the Antiochene rulers’ family (e.g. Mauger of Hauteville). In an interesting passage contained within Raoul of Caen’s biography of Bohemond’s nephew Tancred, those who did not provide military service in return for land not only had to be paid – making them ‘stipendiaries’ or ‘mercenaries’, but not necessarily in the negative sense of the latter term – but also equipped with arms and armour when necessary (although Raoul may have been alluding to household troops in regard to equipment). Again, much like Italy (discussed below), money and supplies seized in raids and battles were the mortar that held the Antiochene martial edifice together, for when Tancred secured access to a substantial amount of gold coins in 1104 he was able ‘to enlarge the military’. In regard to spoils, after a victory in 1115 Roger of Salerno took the portion due to him according to curial law, distributing the rest among his followers as ‘the custom of the same court demanded’.¹⁹

Naturally, the rulers and their soldiers of the same ethnicity or language were greatly outnumbered by the local inhabitants, requiring recruitment from among the latter in order to bolster both defensive and offensive fighting capabilities. While pertinent information is unfortunately far from detailed, a few references in various sources help to provide a basic understanding of the ethnic composition of Antiochene forces during the first quarter of the twelfth century. While perhaps two thirds of the cavalry tasked with both offensive and garrison duties were probably Normans aided by men from northern France, the others hailed, in order of numerical importance, from the Armenian, Byzantine and Muslim populations. An illustrative example of the ethnic diversity of Antiochene troops is to be found in various sources relating to 1119, the year in which a Muslim chronicler described a force ‘of Franks, Armenians, and other troops’. However, since the Antiochene army had been decimated at Harran in 1104, it is likely that, over time, the diminished numbers of French-speakers were increasingly supplemented, primarily by Armenian cavalry. At least, that is what an Armenian source implied, noting that the 1119 force consisted of ‘one hundred Frankish cavalry, five hundred Armenian cavalry and four hundred infantry’. This numerical disparity should not be pressed too far, however, for another Frankish/Norman contingent consisting of ‘three cohorts’ was present at the same battle, but broke away from the main force to either encircle, or mislead, the enemy. Indeed, the 100 Frankish cavalry mentioned by the Armenian source were almost certainly drawn from Roger of Salerno’s household, and the other squadrons that departed from the main body were formed of his vassals and their retinues. As for the ‘other troops’ specifically in this army, but also in others fielded during the period, they seem to have been mainly Greek-speakers: one of the cavalrymen of 1119 was identified by a name originally deriving from ancient Greek mythology: Euterpios, ‘a most esteemed knight’ according to the twelfth-century Norman/French chancellor of Antioch, Walter. Moreover, Usama ibn-Munqidh, a Muslim cavalryman and scholar from Shaizar (beyond the southern frontier of the Antiochene principality) noted that two brothers of Theophilos (Greek for ‘friend of God’) – whom he elsewhere identified as lord of the important southern fortress Kafartab in 1106 – commanded the same stronghold during a siege in 1115. Evidently their brother had passed away by this time, strongly suggesting a hereditary lordship. Importantly Usama was himself present at the siege, so his testimony is likely to have been correct. Lastly, while those serving as garrison troops throughout the principality were generally concerned with defence, it is nonetheless evident that they were called to serve in the Antiochene field army in times of need.²⁰

The reliance on speakers of Greek and Armenian was not merely a necessary expedient: both were warlike peoples with a strong cavalry heritage upon which the Fatimid dynasty had similarly relied in the tenth and eleventh centuries. While the Armenians were not strictly speaking ‘Byzantines’, many in Armenia, Anatolia and Cilicia (southern Turkey) had long been subjects of the Byzantine emperors, supplying them with numerous contingents of infantry and cavalry; they even provided the empire with some of its most warlike and militarily successful leaders (e.g. John I Tzimiskes, r. 969–76). Although some of the Armenians and (former) Byzantines undoubtedly served as heavy cavalrymen, others seem to have fought as Turcopoli, a term originally pertaining to imperial horse archers who were the issue of Byzantine mothers and Turkic fathers. As Raoul of Caen related, ‘a band of Turcopoli bearing bows’ evidently supported Tancred’s heavy cavalry in Cilicia in 1097, and they seemed to have been a detachment from the Byzantine force that travelled with the southern Italian contingent on the First Crusade. Debate has raged as to whether this term later came to denote cavalry or infantry, but Turcopoli generally seem to have been horsemen armed with bows (vitally important against Turkic forces favouring ranged combat), although perhaps some were also equipped with lances and javelins (as was common among light Byzantine cavalry). However, to return to the original meaning of the term, it is likely that included in these ranks of light cavalrymen were men of Turkic and Arabic origin. In 1106, the critically important southern stronghold of Apamea was captured by Tancred, who granted its local Muslim princes command of the fortress and its surrounding districts. Moreover, in the same year, Usama ibn-Munqidh noted that an Arab leader, presumably with his military retinue, ‘passed into the service of Theophilos [of Kafartab]’, with whom he ‘used to go out on raids with the [Antiochene] Franks against the Muslims and plunder them’. To summarize, the Antiochene army commanded by scions of the Hauteville family was much like the one they possessed in southern Italy: a multi-ethnic conglomerate of various troop types capable of employing the combined-arms approach to warfare.²¹

Given the involvement of two Muslim powers of different ethnic origins in eleventh-century Syria, the forces in the region can be presumed to have consisted of a mixture of Fatimid and Selçuk troop types. Both fielded cavalry known as ghilmān: ‘neither lances, arrows nor any other arms do they fear, since they and their horses are completely covered in iron’, observed an anonymous participant on the First Crusade in reference to Selçuk ‘Agulani’.²² While the anonymous observer accurately described the heaviest form of Muslim cavalry who, as Michael Brett puts it, were ‘designed to charge’, it is important to note that ghilmān were also light horsemen equipped with bows. While both types of cavalrymen serving in Selçuk armies were unsurprisingly of Turkic ethnicity, troops of the same origin had long served in Fatimid forces. However, Fatimid armies until the early eleventh century also contained ‘Roman cavalry’ (ghilmān al-Rūm), variously recruited, or seized, from lands formerly (or still) subject to Byzantine authority, in addition to Berber lancers from Egypt and the western province of Ifrīqiya (north-central Africa). Troops from this region had long been of importance to the al-Fāṭimīyūn, who were based in Ifrīqiya prior to moving their capital to Cairo in 969. The numbers of Selçuk cavalry were bolstered by Kurdish mercenaries, who fought deftly with lances and swords, and who had earlier fought in the ranks of tenth-century Fatimid forces. Taking all of this into consideration, clearly there were some similarities between Selçuk and Fatimid armies when it came to heavy cavalry, and the same is generally true in regard to the prevalence of light cavalry armed with bows (although crusader sources indicated that the latter forces no longer featured horse archers by the early twelfth century). That being said, a substantial difference between the two military traditions appears to have been the Fatimid use of javelineers (some of whom threw two-pronged spears), although as will become clear in the chapter concerning the principality of Antioch (and in the Appendix), Latin sources regularly noted the use of javelins by Selçuk cavalry, although those equipped with them must have been substantially outnumbered by their counterparts bearing bows. Another contrast between the two military traditions is the use of heavy infantrymen by Fatimid forces. The latter armies, then, were more like their Byzantine adversaries than those of the Selçuk Turks, who were overwhelmingly mounted and preferred to fight at a distance: their frequently successful tactic was to whittle down foes, to quote a contemporary Frankish observer, ‘by the menace of their arrows’, before advancing to deliver the coup de grâce with their mêlée weapons (generally swords). Accordingly, as R.C. Smail shrewdly observed, ‘Fatimid armies provided, as the Turks never did, a solid target’. Despite Smail’s salient observation, it must be noted that Turkic forces did on occasion feature infantrymen, although the sources are not clear on whether they were of the heavy or light variety.²³

The Normans

The Normanni (‘Northmen’) were the rulers of what was known initially as pagus Rotomagensis, the county of Rouen. Granted to Scandinavian marauders led by Rollo (Hrólfr) of Norway in c. 911 by the Carolingian king of western Francia (France), Charles III (r. 898–922), the early counts extended their borders and, by the reign of the fourth one, Richard II (r. 996–1026), the enlarged county of Rouen had transformed into the duchy of Normandy. The people whom they ruled, however, were speakers of the earliest form of the modern French language, the northern tongue later known as langue d’oïl. Men who spoke this language bolstered the ranks of the county’s Scandinavian troops from the outset: for example, the tenth-century chronicler Flodoard of Reims observed that ‘Rollo their prince sent one-thousand Normans from Rouen as well as inhabitants from that town’ to relieve the besieged garrison at Eu in 925. The bond between the rulers and the local inhabitants was strengthened not only by warfare, but also by familial ties: as would later be the case in Britain, Italy and Syria, intermarriage with the locals was the rule rather than the exception. While leading scholars such as David Bates and Eleanor Searle disagree as to how quickly the Viking warlords transformed into Frankish princes, by the beginning of the eleventh century they, along with most of their subjects, were Frankish/French when it came to language, religion and customs. Historians have in the past been tempted to link the military successes of the Normans with their bellicose Viking heritage. Yet in truth, despite possessing a different past to those of other neighbouring subordinates of the Frankish/French crown, they were similar to them in various ways within a few generations.²⁴

This transformation from Scandinavians to Franks/French is best exemplified when narrowing the focus to warfare. It is generally believed the Normans originally derived from modern-day Denmark, although some did hail from Norway, including their first count Rollo.²⁵ Warriors from both regions of Scandinavia fought in similar manner: that is, they were infantrymen who were ‘covered by a wall of shields’ in pitched battles. As observed in On the Customs and Deeds of the first Norman Dukes (c. 995–1015) by the first historian of the Normanni, Dudo of Saint-Quentin, the weapons Scandinavians used were by no means uniform: javelins, swords and axes, and he also noted that they protected themselves with light shields, helmets and coats of mail. While there were certainly exceptions, generally Scandinavians did not fight as cavalrymen. In contrast, since before the reign of Charlemagne (r. 768–814) the Franks had relied on armoured cavalrymen (milites) as their striking arm, or ‘shock’ force. Indeed, Dudo related that these ‘iron-clad squadrons’ inflicted a heavy defeat on Rollo’s army outside the walls of Chartres (c. 911). His description of the troops led by Rollo prior to this battle is interesting, for both cavalry and infantry were mentioned. However, based on the account of the battle of Chartres, in which he made no such distinction, if mounted soldiers were present they were probably not ‘true’ cavalrymen; rather this contingent consisted of Rollo and his retinue who dismounted to fight. This was certainly the case in a battle with the Frisians prior to Rollo’s ascension as count: while he was on horseback prior to the battle, he and his men fought on foot, forming ‘a tight formation of flashing swords’. Similarly, Flodoard of Reims (d. 966) related that Frankish cavalry ‘proceeded to fight a battle on foot near the [Northmen] camp’ in 925. Evidently the Franks dismounted to engage what was clearly an infantry formation of Vikings, led by another Scandinavian leader, Rögnvaldr (Ragenold), whose army had been assisted by ‘a great number from Rouen’ two years earlier. While those Scandinavians led by Rögnvaldr and Rollo were separate entities, their fighting methods are unlikely to have been different.²⁶

By the reign of Rollo’s successor William I ‘Longsword’ (r. 927–42), Dudo’s itemisation of arms available to the second count of the Normans is still suggestive of Scandinavian-style forces. Swords and axes are mentioned rather than lances or spears, which were rendered in Latin as lanceae and hastae (see Appendix). That being said, the ‘three-hundred iron-covered’ men armed with ‘swords and lances’ used to defeat a rebel army in c. 934, although it is not clearly stated, were probably mounted. In reference to the period prior to the establishment of the county of Rouen, Dudo referred to the hurling of javelins (iacula) by Rollo’s troops. Given that he was writing between 995 and 1015, doubts could be cast on the veracity of the erudite chronicler’s terminology. However, Flodoard, writing much closer to the period in question, related that while some of the Normans escaped the swords of the Franks at Eu in 925, ‘others killed themselves with their own javelins [tela]’. As observed in the Appendix on Latin terminology, spicula, iacula, missilia and tela were the Latin words used to denote javelins although, depending on the context, the latter two could also mean missiles in a more generalized sense of projectiles shot, slung or thrown.²⁷ However, that tela meant javelins in the Norman context is confirmed by consulting Dudo’s narrative of the battle between the cavalry forces led by Count Theobald I of Blois on the one hand, and William I’s successor Richard I (r. 942–96), on the other. The battle, fought in 962 at Saint-Sever – a suburb of Rouen – is the first explicit, unambiguous reference to the transformation of Scandinavian-style infantry into a Frankish-style mounted force. Dudo’s account, presaging the tactics later used by Normans in both England (Hastings) and Italy (e.g. Civitate), accordingly requires a full translation:

In the first clash of the struggle, they were doing battle with shortened [lit. ‘cut short’] javelins [tela] and lances [lanceae]; in the second, with gleaming swords. Then the robust band of Normans joined and interlocked shields, advancing in a formation of flashing swords, and attacking the opposing armoured Franks. And cutting to the front, right, and left, they struck down and shattered the well-formed enemy battalions, riding over corpses of the slain and the opposing tightly packed ranks. After wheeling around, battle was given to the remaining forces.²⁸

Based on the description, the decidedly non-Scandinavian approach to warfare is better termed Franco-Breton than Frankish: while the Franks fought with lances, the Bretons hurled javelins. In the tenth century, Regino of Prüm likened Breton cavalry tactics to those used by the Magyars – a people originally from the Kazakh steppe – noting that the only difference was that the former threw javelins (spicula) instead of firing arrows. Quite unlike the Franks, the Bretons did not advance in a large body; instead they attacked in mobile, smaller units that wheeled around rather than engaging directly at close quarters.²⁹ Regino drew attention to this tactic in reference to the battle of Jengland-Beslé (851):

The Bretons, by habit accustomed to darting here and there with horses drilled in this manner, attacked the tightly packed battle array of the Franks, throwing with great strength javelins into its midst. After simulating flight, they likewise pierced the breastplates of their pursuers with javelins.³⁰

Clearly the first attack at Saint-Sever conformed to the Breton style. Not only were javelins featured, but their purpose was to exploit weaknesses and create openings. Once the latter goal had been achieved, the Breton tactic was succeeded by the Frankish one: advancing in a massed, close formation with brandished swords, the cavalrymen rode through the breaches created in the enemy battle line. This two-tiered order of tactical deployment was still in use a century later, and while various examples will be assessed at length elsewhere, the most famous example will suffice. At Hastings (1066), the infantry opened proceedings with arrows and crossbow bolts, followed by javelins [iacula] thrown by cavalrymen. Then, to use the words of William of Poitiers: ‘Ashamed to fight from a distance, they [now] dare to bear their swords’. That is, having thrown their spears, the cavalrymen launched a sword-based, mêlée assault.³¹

Given their hybridized adaptation of Frankish and Breton tactics, the Normans were therefore relatively unique among northern French cavalry forces. The two influences on their combat methods should not be surprising given the political and cultural ties as well as geography. Bordering Rouen/Normandy to the east was the western Frankish kingdom which, after the accession of the long-lasting Capetian royal dynasty (r. 987–1328), became the kingdom of France. While Dudo noted William I was resented by some Scandinavian subjects for his Carolingian leanings and the favouritism shown towards ‘Frankish friends’, it is surely no coincidence that the non-Scandinavian style of warfare employed by Richard I at Saint-Sever occurred during a reign which, on account of increased involvement in Frankish political and military affairs, demonstrates that by the early 960s the county of Rouen, in the words of Bates, ‘had achieved a sufficient measure of acceptance and integration into Francia’. On the southwestern border was Brittany, whose inhabitants and their customs were rather different to the Franks since their homeland had never been fully integrated into the empire founded by Charlemagne. While Dudo and Hugh of Fleury might have been in error, it remains possible that William had been militarily involved in Brittany in 931. This could be one reason for the decision of King Raoul (r. 923–36), in 933, to grant William control of the counties of Cotentin and Avranchin, which Charles II (r. 843–77) had earlier ceded to the Bretons in 867. Flodoard recorded this grant in ambiguous terms, providing the core for Dudo’s myth that all of Brittany became a Norman dependency. However, William’s claim to be the region’s duke on coinage discovered at Mont-Saint-Michel possibly suggests that his authority, or at least his influence, extended a little further west of the later Normanno-Breton frontier fixed at the River Cousenon (from Mont-Saint-Michel in the north, to Saint-Ouen-la-Rouërie in the south). Contact and conflict with the warlike Bretons, a people who ‘greatly devoted themselves to arms and horses’, remained constant throughout the tenth and eleventh centuries. In addition to Norman involvement in Breton affairs, contact consisted of marriages between the ruling dynasties, the first being the union between Richard I’s daughter and Count Geoffrey of Rennes in 996.³²

The cavalry precedent evidently established during the reign of William I, but perfected by Richard I, irrevocably changed the nature of Norman warfare. Indeed, when the Normans first arrived in Italy in the early eleventh century, they were predominantly a cavalry-based people, and were accordingly identified in Latin sources as milites or equites: heavy cavalrymen. While the terms are often translated as ‘knights’ by modern historians, this word is better applied to the twelfth century, when armoured horsemen were more rigidly associated with high birth and chivalry. Yet, as was the case in other northern French counties and duchies, heavy cavalrymen, other than when raiding and reconnoitring, were usually supported by pedites – infantrymen – who acted as a screen for their mounted comrades. Increasingly Norman forces became more sophisticated as the eleventh century progressed, employing the combined-arms method of warfare. In addition to heavy infantry and cavalry units, archers began to be utilised more frequently, and in Italy their ranks would be bolstered by locally recruited javelineers and slingers (see below).³³

The precise manner in which Norman soldiers were recruited has long been a difficult, even controversial, topic. One reason for this is that scholars are forced to get to grips with the contentious modern-day term ‘feudalism’.³⁴ While the merits of the term will not be discussed here, it is important to note that ‘pure’ or ‘mid-classical’ feudalism, if generally defined as an individual holding an allotment of land (Lat. feudum; Fr. fief) in return for military service, did not exist in pre-1066 Normandy. Some land grants (‘benefices’ or ‘honours’) certainly required military service, while others did not (allodia, ‘freeholds’). The benefice or honour was held for life, whereas the allodium was inalienable, a patrimony. However, some allodia were held by vassals, meaning that despite their hereditary claim to the land, they were nonetheless required to provide troops to their lords when requested. Moreover, since some soldiers did not possess land, they were accordingly paid for their services. Increasingly the benefice/honour and allodium gave way to the fief – that is, in common with the allodium a fief was hereditary, but like the benefice military service was required. The Norman method of land tenure transformed during the eleventh century, slowly but surely moving from a custom based on diverse service agreements between parties, to a more rigidly defined system regulated by law.³⁵

Georgios Theotokis has recently asserted that it is ‘almost impossible to draw a map of comital (military) responsibilities in southern Italy until the beginning of the twelfth century’. Similarly, but somewhat more pessimistically, Donald Matthew stated in the 1990s: ‘Nothing is known about how such [pre-twelfth-century] armies were composed, led or even maintained’. Indeed, historians of the kingdom of Sicily founded by Roger II of Hauteville (Fig. 16) in 1130 are blessed by the survival of The Catalogue of the Barons, a detailed, mid-twelfth-century itemisation of military service and obligations. Without an eleventh-century equivalent, historians are faced with various challenges, primarily because the extant sources were generally, but not always, disinterested in matters pertaining to recruitment and service obligations. That being said, in 1940 Claude Cahen wrote an important work which shed some light on the subject, but unfortunately it remains inaccessible to those without French (which also applies to the more recent analysis by the erudite Jean-Marie Martin [1993]). With the work of Cahen and Martin providing inspiration, what follows is a reassessment of the pertinent terminology used by Geoffrey Malaterra, the observer who paid closer attention to such matters than his counterparts. Despite the at times less-than-precise nature of his observations in regard to recruitment and service obligations, Malaterra’s work nonetheless yields data which provides a glimpse into ‘feudal’ agreements in the region: namely, ‘service’, ‘aid’, ‘counsel’ (or ‘court’), fides/fidelis (fealty, faithful/vassal), homo (man/vassal), in addition to the mention of a practice which became incrementally more common in late eleventh-century Normandy and elsewhere in the subsequent century – a forty-day period, normally pertaining to military service.³⁶

Firstly, the fact that Roger entered Robert Guiscard’s ‘service’ as a young cavalryman (juvenis) in 1057 indicates the latter’s maintenance of household troops in Apulia, and that Roger joined them as his brother’s vassal. Indeed, that other lords in the same region possessed household troops is suggested by Malaterra’s earlier mention of ‘all the young cavalrymen of Apulia’. Secondly, when Roger subsequently violated his ‘contract/agreement’ by rebelling against Guiscard in 1058, ‘he nevertheless was ready to protect his legal status by avoiding injury to his brother for forty days’. So concerned was Roger with observing the rules of ‘feudal’ obligations, that he had earlier ‘placed his case before all of Apulia’s nobles’. As lords in Normandy were known to convene courts in order to settle tenurial disputes from the 1050s onwards, Malaterra’s language strongly suggests that Roger’s appeal was heard before such a court. Guiscard thrice promised control of southern Calabria to Roger (once in 1058, twice in 1062), but failed to honour the agreement until the third occasion. All that Roger had in his firm possession during this period was Mileto, which he received in 1058 with hereditary rights. It seems likely that he established his own military household there, since Malaterra later mentioned Roger commanding ‘three-hundred young men’ in 1061. Indeed, there is no reason to doubt Roger’s maintenance of mounted household troops at his most important bases, for Malaterra was thankfully specific on one occasion: in 1075 Roger’s son-in-law Hugh was recorded as making use of ‘the count’s military household’ based at Troina, Sicily.³⁷

Limited by his single hereditary possession, burdened by having to pay his troops with cash and plunder, and noted as requesting money from his lord to assist with the remuneration of his increasingly disgruntled troops, Roger needed to hold land which could be parcelled out among his lieutenants (who in return could provide him with troops bound by service rather than serving for plunder or payment); a process known to medieval historians as subinfeudation. That this was his aim is evident when Guiscard finally honoured the agreement: Roger distributed plots of land ‘to his faithful [fideles]’, a term Martin notes is regularly found in Apulian charters from the 1060s onwards (e.g. fideles Normanni). While in ‘pure’ feudalism a ‘real’ vassal was a man who performed homage (i.e. became the lord’s man [homo]) before swearing an oath of featly, in pre-1066 Normandy David Bates notes the ‘persistence of the old Carolingian term fidelis [sing.] to describe a vassal’. It accordingly seems probable that the ‘faithful’ granted land by Roger received it from him as vassals, and were bound to provide their lord with military service. Interestingly, some of the Hauteville fideles may have started to perform the act of homage prior to taking oaths of fealty, if the following three examples are anything to go by. In the 1080s, Hugh Falloc’s son ‘was made homo’ during a ceremony in which he was ‘joined to Bohemond by oaths’. As the pact between lord and vassal entailed mutual protection, Count Roger, ‘ready to maintain his legal rights’, was careful to disavow the oath made to ‘his homo’ Ingelmar in the same decade, prior to launching a campaign to suppress the latter’s rebellion. Lastly, in order to secure much-needed military assistance from Roger Borsa in the 1090s, Prince Jordan of Capua ‘was made the duke’s homo’. Although the last two examples only imply homage owing to the use of homo instead of the usual fidelis, the first is reminiscent of the acts of homage followed by oaths at Caiazzo in the 1130s recorded by Abbot Alexander of Telese (who used the same term in plural form).³⁸

Roger’s pool of vassals/faithful created in 1062 was supplemented by others when he was in an even stronger bargaining position: in 1072 he granted sections of land held from his brother in Sicily to two men, one of whom was his nephew Serlo. Perhaps tellingly, not only was the land granted to Serlo and Arisgot as a reward for their military abilities, but also for the value of their ‘counsel’ (consilium). The latter denoted the right of lords to summon their vassals to an assembly (curia, but also called ‘counsel’) and receive their advice, and the fact that Guiscard was said to have approved of the subdivision of his vassal’s lands accords with the general rule that subinfeudation required a lord’s sanction (although this was often ignored in the eleventh century). The importance of counsel cannot be overemphasized: a two-year-long dispute between Guiscard’s sons following his death was only ‘reconciled by vassalic counsel on both sides’. There can be little doubt that counsel was a service required by lords, for Martin cites a charter (1072), in which ‘Count Peter requests counsel from his fideles’. In addition to household troops, and vassals serving in return for land, were the landless who, like their counterparts in England, were at times tasked with garrison duties in return for a stipend (hence the term ‘stipendiaries’). Lastly, there seem to have been a considerable number of troops paid purely from the spoils of war. Taking all of the above terms into consideration, since Malaterra was writing at the beginning of the twelfth century the possibility remains that such terminology could be anachronistic. However, as he claimed on a few occasions to have sourced his information directly from both Roger and the men who served him, it is feasible to suggest that the terminology, while perhaps more reflective of a period when it was starting to be used more precisely, nonetheless accurately illustrates the character of ‘feudal’ obligations between Normans in Apulia, Calabria, and Sicily in the second half of the eleventh century.³⁹

As recently observed, the south Italian kingdom founded by Hauteville scion Roger II (r. 1130–54) was relatively more structured in regard to military recruitment and service. The Catalogue of the Barons attests to widespread cavalry service in return for tenure, although not all knights held land. Nor were the majority of those on horseback knights, for the 8,620 mentioned were considerably outnumbered by more than 11,000 servientes (sergeants): a term first used in the early twelfth century, but which was potentially a standardization of the various names used in the eleventh century for landless, mounted assistants (‘arms-bearers/squires’ and ‘attendants’). As was made clear in the Catalogue, vassals were to provide the king with two quotas: the first, a standard military levy, the second, called an augmentum (‘increase’), was due in times of necessity (i.e. for repulsing an enemy invasion). Effectively the augmentum doubled the standard, less onerous quota. For example, Count Roger of Buonalbergo was required to supply thirteen cavalry ordinarily, but in times of crisis pledged an additional eighteen knights, fifty mounted servientes, and two crossbows. Yet in common with the soldiers fielded earlier in Italy and later in Syria, stipendiary forces continued to be used, in addition to those remunerated solely from plunder seized in campaigns. As Alexander of Telese wrote in regard to King Roger II: ‘Indeed military wages, or whatever according to either agreement or promise was required to be given, he paid without hesitation’. This diverse system of remuneration appears to have still been in use during the reign of Roger II’s grandson, William II (r. 1166–89). Archbishop Eustathios of Thessaloniki, who gathered his information from the king’s lieutenants in the 1180s, noted that while some troops were paid a ‘daily share’ or had received ‘promise’ of payment from the king, the ‘risk-takers’ relied entirely on the spoils of war. Hence, despite the better-organized (or better-documented?) method of cavalry recruitment, vestiges of the age of Robert Guiscard and Count Roger of Sicily, who similarly fielded mounted and unmounted soldiers variously fighting in return for land, wages and plunder, remained.⁴⁰

Military service in Italy was not restricted to those who spoke northern French. Byzantines in Calabria, in addition to paying tribute, were expected to provide troops. The term used, servitium, literally means ‘service’, but like auxilium (‘aid’), in late eleventh-century France it was normally associated with military service (and both terms were used specifically in this sense by Malaterra in reference to a 1068 dispute between Guiscard and his nephew Geoffrey). In the following century, when feudal terminology became more rigidly defined, the noun was appended in Norman England with the adjective militare. Malaterra never qualified ‘service’ with ‘military’, but there is no reason to doubt Ferdinand Chalandon’s belief that the monk specifically meant le service militaire in reference to obligations owed by the local inhabitants. The Greek-speakers, however, were generally not fideles in the Norman sense; rather, their service seems to have been a component of their overall tribute. This was presumably the case given that most of the soldiers recruited in Calabria were infantrymen, including specialist missile troops, and hence were incompatible with a recruitment system based on land and cavalry service. Conversely, Muslims in parts of Sicily were required to provide a quota of cavalrymen in return for the lands they held. While this obligation might seem to be remarkably ‘feudal’, it merely continued the system already in place (iqtā). Based on various descriptions of the Kalbid and Zirid armies encountered by the Normans in Sicily, the Muslim troop types appear to have been similar to those found in contemporary Fatimid armies, which is unsurprising given the heritage of these two dynasties. Firstly, that the Fatimid penchant for javelins was shared in Sicily is suggested by references to the use of this weapon. Secondly, since three attempts by Norman cavalry to rout Ibn al-Ward’s infantry line at Catania were unsuccessful (c. 1080), Fatimid-style heavy infantrymen were undoubtedly present at this battle. As for other troop types, many were probably archers. Although they are specifically identified as such in the sources for the twelfth-century kingdom, the earliest explicit reference before then is 1113: Sicilian ships were observed to have been protected by ‘Saracen men’ who were ‘powerful archers … and whose skill in the shooting of arrows was not held to be inferior in the region of Jerusalem’. Moreover, while written in reference to the siege of Thessaloniki (Greece) in 1185, the abovementioned Eustathios recorded the presence of Sicilian ‘horse archers’, who were presumably the ‘Saracens’ he had earlier mentioned. As mounted archers had been a common feature of Muslim armies since the seventh century, they must have been present in Hauteville armies from the outset. The precise manner in which the non-tenured Muslim troop types were recruited is unknown, but they were likely procured in similar ways to their Byzantine and Lombard counterparts: that is to say, levied in a manner commensurate with service or tribute agreements, paid with stipends or plunder, and, when deemed necessary, conscripted (discussed below).⁴¹

While it is possible to obtain a general idea of the way in which speakers of Greek and Arabic were recruited, the evidence for the Lombards of Apulia and Campania is more difficult, especially when it comes to infantry. Cavalry were evidently incorporated into Norman contingents and, on occasion, these battalions were even led by Campanian Lombards such as Guido of Salerno. The evidence for the Abruzzi (north of Apulia), however, is more precise: in return for cavalry service, Robert Hauteville of Loritello granted lands and fortresses to notable local Lombards in the 1070s. As for infantrymen, the continuation of pre-existing Byzantine and Lombard practices of recruitment was likely the norm. However, when more troops were needed for major campaigns, evidently towns capable of providing additional ones were burdened with conscription, and those unable to comply were required by their Norman lords to furnish them with enough money to enrol either mercenaries or men serving in return for a stipend. For example, in reference to Robert Guiscard’s invasion of the Byzantine empire (1081), Raoul of Caen noted the recruitment of a large contingent of Longobardi who, when reconciled with other sources, comprised Greek-speakers as well as Lombards. While some of these soldiers fought for money, others were ‘forced by an edict’ to participate. On account of their involuntary recruitment, observed Raoul, the Longobardi were ‘not desirous of the glory of serving in the army’. Raoul’s later testimony is confirmed by an earlier one: William of Apulia noted that many troops were unwilling to provide military service, but with ‘threatening words in addition to coaxing requests’ the duke ensured their compliance. Presumably ‘coaxing requests’ was another way of referring to the ‘promises’ attested in various Latin and Greek sources of the eleventh and twelfth centuries.⁴²

According to Malaterra, upon reaching puberty Norman youths destined to become cavalrymen ‘began to be involved in military training and to engage in the study of horses and arms, learning to defend themselves and to fight against the enemy’. This was their formal introduction to cavalry training, for Huguette Taviani-Carozzi has plausibly suggested that many received instruction earlier. Unlike the knights of the twelfth century, who increasingly belonged to a distinct social class (‘order’ or ‘position’), Norman youths did not need to be of noble birth in order to receive such training. The Norman monk Orderic Vitalis noted that Hugh of Avranches’ military household consisted ‘of noble and ignoble boys’; William of Poitiers mentioned cavalrymen of both middling and common origin; and serving Roger of Hauteville in Sicily was Ingelmar, a ‘common cavalryman’. Indeed, the Hauteville progenitor Tancred may have been one of the lower-born sons attached to the household of William the Conqueror’s grandfather, Duke Richard II. With such attention paid to training from an early age, coupled with the various eleventh-century conquests, it should come as no surprise that Norman cavalry acquired something of an invincible reputation internationally. In Italy, the eleventh-century epic poet William of Apulia wrote of the ‘Norman people, distinguished by the fierceness of their cavalry’. Byzantine cavalrymen who charged at full gallop with their lances couched, were figuratively styled by the twelfth-century princess Anna Komnene ‘to have come from Normandy’. She elsewhere maintained that a Norman cavalry charge ‘cannot be checked’; indeed it could ‘bore through the walls of Babylon’. Caution, however, should be exercised when considering such depictions. Clearly the Normans were a people preeminent in arms, but they were certainly not militarily superior to other northern French forces of the age.⁴³

Unsurprisingly the Normans devoted much attention to the breeding of expensive warhorses, known as destriers because squires led them by their right hand (from Lat. dextra). R.H.C. Davis noted that equines were both of local and foreign extraction, the latter almost certainly deriving entirely from Spain. He also suggested that since the abbeys of Saint-Evroul, Fécamp, and Jumièges were well known for the buying and selling of horses, they must surely have bred many of them. The same applies to the duke and his more eminent subordinates, who presumably devoted appropriate sections of their estates to horse breeding. Equestrian expert Ann Hyland has estimated – based on numerous surviving Norman horseshoes – that the average destrier was of medium size, measuring about fourteen to fifteen hands. While this size may well have been the rule, there were no doubt exceptions based on, among other things, personal preference. For example, Richard Quarrel, later to become the count of Aversa and prince of Capua, was reported by his Lombard contemporary, Amatus of Montecassino, to have ridden by choice a horse so small that his feet were close to the ground. In reference to this passage, John France has concluded that Richard’s horse was likely to have measured roughly twelve hands, adding that this could have been the default size of an eleventh-century destrier. Not only does Amatus’ description of Richard’s horse indicate that Norman warhorses were smaller than the heavier, larger equines favoured in later centuries, but it also demonstrates that their riders preferred long stirrups. Longer stirrups gave the rider the ability to stand while mounted, enabling the delivery of more powerful and precise attacks with lance and sword. As the Bayeux Tapestry demonstrates, adding to the horseman’s stability in both charges and mêlée combat was the use of a saddle with a symmetrically raised pommel and cantle (Fig. 2).⁴⁴

The larger body of Norman cavalrymen (bataille, a ‘battle’) consisted of smaller divisions, known in French as conrois, consisting of ten men on average based on a passage in the eleventh-century work by William of Poitiers. Presumably the average conroi (sing.) was drawn from a military household, whose soldiers evidently fought under their lord’s banner fixed to a lance (Figs. 1–2). The banner was of great importance on the battlefield: not only did it allow the commander of a conroi – also known as bannière – to direct the movement of his men, but it also served both as a rallying point and source of morale. When the situation called for it, the cavalrymen would secure lances under their right arms and charge at full gallop. The couched-lance technique, however, was primarily suited to charging against cavalrymen (Fig. 5). William of Poitiers, a Norman with a military background, noted that mounted troops were unable to overcome heavily armoured, disciplined infantrymen arrayed in a densely packed formation ‘without excessive harm’; a fact graphically represented in the Bayeux Tapestry. This the Normans experienced at the battles of Civitate (1053), and Hastings (1066). When engaging an enemy of this sort, other tactics such as flanking and skirmishing were called for. As is evident when examining the Bayeux Tapestry, John France notes that Norman cavalrymen ‘sometimes couched their lances, but at other times stabbed, jabbed and even threw them’ (Fig. 2). Importantly, while not mentioned by France, there is eleventh-century southern Italian evidence that supports his observation: the door frieze at the Basilica San Nicolò di Bari depicts Normans gripping their spears in the overhand, underhand, and couched positions (Fig. 1).⁴⁵

Reacting to ‘Richard Glover’s tendentious remarks’ – i.e. that a Norman cavalryman was ‘for the most part only a mounted javelineer’ (substantiated only with reference to certain scenes in the Bayeux Tapestry) – esteemed scholar of the Normans Reginald Allen Brown dismissed the use of ranged tactics, arguing that those holding spears overhand in the Bayeux Tapestry ‘are in fact about to strike over-arm in the manner most likely against infantry’ (Fig. 2). Yet in addition to the fact that the Tapestry does feature spears in flight, both the overhand and couched-lance grips were most suited to horsemen aiming at those level with them – other cavalry.⁴⁶ Conversely, underhand thrusts were practical against troops situated below cavalrymen (Fig. 3): indeed, in addition to the greater power provided by an underhand grip, the ability to hold a spear towards the bottom of its shaft – as opposed to the overhand grip which required the shaft to be held closer to the middle for greater balance – also provided the horseman with the advantage that the spear extended further from his body. Moreover, despite Allen Brown’s insistence to the contrary, it has already been noted above that the throwing of spears by cavalrymen can be dated as far back as 962 (Saint-Sever). Furthermore, there are also numerous examples of this tactic being used in the eleventh century: William of Poitiers had the young Duke William ‘throwing his lance’ during a mounted engagement in Anjou (c. 1049), dismounted cavalrymen threw lances from the battlements of Ambrières (1054/5), Amatus of Montecassino stated that ‘the valiant Normans … hurled their spears against the Greeks’ in Italy (1041), and in reference to two battles in the same region William of Apulia recorded the use of javelins (iacula [1041], tela [1053]; see Appendix). Additionally, William of Poitiers observed that, at Hastings, the Anglo-Saxons wounded ‘those who flung javelins [iacula] at them from afar’. While the unspecified soldiers who flung these throwing spears at Hastings could be argued to have been infantrymen, William made it clear that the unmounted Norman missile troops were equipped only with bows and crossbows (see below). Precisely which weapons the heavy infantrymen were equipped with was not related, but rather than javelins they likely wielded thrusting spears, as indicated in Guy of Amiens’ account of Hastings. The importance of the throwing spear to the Normans cannot be overstated, for Duke William II was said to have been a threatening sight when equipped with sword, shield, helmet, and javelin (telum) at his coming-of-age ceremony. The hurling of spears certainly made sense when attacking a well-ordered formation of armoured infantrymen; openings could be created, into which charges could be launched. Indeed, creating breaches in a line of heavy infantrymen was of the utmost importance, for horses will instinctively ride around or jump over obstacles, not charge through them.⁴⁷

Taking the various instances of lance and javelin hurling into consideration, it seems that the Normans were notably different to other northern French cavalrymen described in sources of the tenth to twelfth centuries, who uniformly wielded lances and rarely, if at all, threw them. Rather, Norman cavalrymen until the advent of the twelfth century were something of a hybrid force, able to adapt tactically, using the Breton missile approach or the Frankish lancer method (or a combination of the two at Saint-Sever) depending on the tactical situation that faced them. In Italy, both styles of cavalry warfare are well attested, even in the same battle. At Civitate, Apulia in 1053, while the troops led by Richard Quarrel and Humphrey of Hauteville were noted as hurling javelins – respectively iacula and tela – the reserve contingent led by Robert Guiscard was apparently equipped with thrusting spears as he was said to have inflicted mortal wounds with his lance. The usage of the latter weapon by heavy cavalrymen, however, seems to have become increasingly preferred, for throwing spears are not explicitly mentioned after 1082, the year in which a Byzantine source referred to ‘missiles thrown’ by two cavalry contingents. Indeed, evidently the full-gallop, couched-lance deployment described by Walter the Chancellor became the norm over time: in 1119 Antiochene cavalry, comprising men from both southern Italy and Normandy (e.g. Robert of Vieux-Pont) ‘fixed shields to their sides, readied lances, pressed spurs’ and charged in unison. Similarly, in regard to an engagement in February 1097, Raoul of Caen wrote: ‘They loosen reins, they make use of spurs, they brandish lances’. However, there is a danger inherent in pressing this point too far: in both clashes the Normans engaged horse archers, and hence the use of ranged tactics against them would have been futile; such enemies needed to be fought at close quarters as quickly as possible. Despite the great tactical ability of cavalrymen of the northern French type, as already noted they certainly had their limitations when faced with a determined phalanx of armoured infantrymen. Yet it should be emphasized that the latter nonetheless greatly feared them, even resorting to stratagems in order to counter their forceful charge. For example, in the second decade of the eleventh century, Scandinavian forces deriving from Denmark and Ireland clandestinely prepared ditches and stakes to great effect in Brittany and Aquitaine. Moreover, in two successive battles of 1082, Byzantine emperor Alexios I Komnenos placed wagons and iron caltrops in front of his infantry phalanx in order to hamper a charge by Bohemond of Hauteville’s cavalry.⁴⁸

Having covered the basics of the Norman striking arm, it would be unwise to follow the general trend of medieval chroniclers – that is, to pay little or no attention to the infantry. All field armies, to use the words of the twelfth-century Norman chronicler Raoul of Caen, consisted of two important contingents: those who could ‘charge into the enemy’, and those with the ability ‘to repulse a charge’.⁴⁹ The latter group were the heavy infantrymen, variously denoted as ‘armour-clad men’, ‘infantry wearing chainmail’, or ‘armour-clad infantry’.⁵⁰ Protected by a chainmail coat – hauberc in French – the heavy infantryman seems to have generally wielded a thrusting spear (Fig. 1). When arrayed in a densely packed formation, armoured infantrymen were very effective against both infantry and cavalry charges. Provided they held their ground and maintained formation, mounted charges against them were not only fruitless, but could prove costly to both the riders and their mounts. That destriers were regularly killed in battle is evident from the fact that William the Conqueror lost three at Hastings, and Robert Guiscard the same number at Civitate. Moreover, as related by his daughter, Emperor Alexios I Komnenos ordered his infantry archers to target Norman mounts in the 1080s.⁵¹

Despite their general effectiveness, heavy infantrymen were nonetheless vulnerable to assaults from the left and right. Cognisant of this tactical limitation, when the Lombardo-Norman forces engaged the Byzantines near the River Olivento in Apulia (1041), cavalrymen were posted on the infantry flanks. Other infantry such as skirmishers and missile troops were regularly grouped behind the protective screen of the heavy infantrymen, primarily because they were lightly armed and wore little to no armour. That they were generally without protection is evident in the terminology used: infantry (pedites), without the qualification of ‘wearing a hauberc’ or ‘armed’, ‘common infantrymen’, ‘defenceless/unarmoured commoners’, and simply ‘commoners’. Such terms have been interpreted in the past to show that medieval infantry forces often consisted of little more than untrained peasants armed with agricultural implements. This may have been the case in times of dire need, but the economic sense of these terms should also be considered: some of these men might have been trained or part-time soldiers who simply did not have the means to possess coats of mail. That haubercs were particularly expensive is attested in a Norman charter from the 1040s, in which a coat of mail is priced at seven pounds; a considerable amount given that a destrier is listed at thirty.⁵²

Moreover, the unarmoured aspect does not necessarily mean that those without haubercs were unskilled. After all, since classical antiquity the battle effectiveness of skirmishers such as javelineers had depended on their greater freedom of movement. Suffering greatly from missile fire at Sphakteria, Greece (425 BC), élite Spartan heavy infantrymen were unable to engage javelineers and other missile troops who, given their greater mobility, simply retreated with ease when targeted. While javelineers were never directly identified as a troop type by chroniclers, it is nonetheless clear that they featured in Norman ranks. The province of Cosenza (Calabria), ‘mighty in arms’ as an eleventh-century poet termed it, seems to have been the primary recruitment area: Byzantines from Aiello armed with javelins killed some of Robert Guiscard’s cavalrymen in battle (1065), and ‘excellent infantry’ from nearby Cosenza fought alongside the duke’s ‘élite cavalrymen’ in the 1080s. That they were equipped with javelins like the ‘warlike Aiellans’ is evident given that they were valued by Guiscard for their speed. Indeed, as John Birkenmeier has noted, javelineers were esteemed by the Byzantines for their ability ‘to react quickly and nimbly to developing situations’, an assertion supported in a Roman military manual of late antiquity by Vegetius Renatus.⁵³ In addition, other skirmishing troops such as slingers (Fig. 6) were identified explicitly, and many of the best ones, like other missile troops, must have been recruited from the subjected, but often rebellious, province of Cosenza (both the city itself and nearby Aiello), although they were clearly a common troop type in Apulia as well: the Byzantine commander John Doukas recruited ‘a countless mass of slingers’ as late as 1155.⁵⁴ It is important to note here that slingers were only mentioned in the Italian context, reflecting the impact of native traditions and the use of local troops. Archers (Fig. 2), in England, France and Italy, were identified specifically as such, or as ‘infantry armed with arrows’, and crossbowmen variously as arbalists, or ‘infantry armed with crossbows’.⁵⁵

With all the above focus on armies fielded by Normans, it is important to make one thing clear: pitched battles were the exception, not the rule, in eleventh-century northern French warfare. In addition to being determined by the logistical and tactical realities of the age, this general military trend may have been influenced by Vegetius’ fourth-/fifth-century military manual, a work that ranks with the works of Isidore of Seville, Bede, and Paul the Deacon as one of the most copied Latin texts in the post-classical world (see Appendix). In the ‘General Rules of War’ section, Vegetius stressed that alternative methods were to be preferred to the pitched battle, as the latter’s outcome owed more to fortuna (chance, luck) than virtus (courage, manliness). Effectively, in northern France warfare normally consisted of sieging, skirmishing and ravaging. The strategy was to conquer a region by taking its strongholds one-by-one, rather than risking everything in a single engagement. Yet castles were often difficult to take, so it was necessary to devastate the surrounding countryside in order to bring about the capitulation of their increasingly enervated inhabitants. As John Gillingham has observed, William the Conqueror employed this approach regularly in the 1040s–60s, as did those who fought against him: Count Geoffrey Martel of Anjou, and King Henry I of France. In addition to ravaging the countryside, on occasion siege forts were built in close proximity to the besieged stronghold. Their purpose was to blockade and intimidate the enemy or, as William of Poitiers put it, ‘conquer by famine’. This tactic was the primary one used by the Normans in southern Italy; indeed, the slow but inexorable conquest of the region owed more to sieges than victories in pitched battles. While they will be assessed at greater length later in this book, a few examples will suffice here. In the 1050s, Richard Quarrel surrounded the walls of Capua on two separate occasions with siege forts in order to deny its citizens access to their grain and grapes, in addition to destroying the fields outside Aquino. Scorched-earth tactics were used so frequently in Italy that some of the inhabitants knew what to expect if they were lucky enough to be forewarned. Accustomed to the depredations of Robert Guiscard and his brother Roger throughout Calabria in the 1050s, the citizens of Reggio hastily gathered all produce from their fields in an effort to thwart the brothers’ impending investment. While equally assisted by the coming winter, denying the Normans the ability to forage in the vicinity meant that the people of Reggio, for the time being, were able to withstand Norman subjection.⁵⁶

Relative to their Byzantine counterparts, eleventh-century northern French methods of siege warfare, with a few notable exceptions, were less advanced. For while the Byzantines on occasion used the same ‘conquer by famine’ method employed by Normans in France and Italy, given their Roman heritage, itself inherited from the classical Greek-speaking pioneers of siege technology (see Appendix), they possessed siege engineers and engines in towns and fortresses (e.g. in Italy: Melfi, Troia, and Cosenza),⁵⁷ and on campaign. Not only were such men able to maintain siege trains that dwarfed in quantity those infrequently employed in western Europe (and rarely in France), but they were also adept at demolishing walls by undermining them with sapping techniques. Sappers did not become a relatively common contingent of western European armies until the second half of the twelfth century, and John France notes the reason for this: the method ‘demanded skilled labour, which might not available’. Contrast this reality, for example, with the resources available to Nikephoros Phokas in the 960s: his ‘picked engineers’ built a variety of siege engines on demand, the subsequent bombardment of which allowed a team of specialists to gain access to the walls of Chandax (Iraklion, Crete). Equipped with masonry tools they removed the lowest strata of rock, dug a large tunnel underneath the foundations, propped it up with wooden beams, subsequently burnt them, and watched with great satisfaction as the targeted section of the wall, including two towers, collapsed. This strategy, later successfully employed at Mopsouestia (Adana, southern Turkey) in the same decade, was noted as the most consistently successful siege technique in a Byzantine military manual of the early eleventh century. The author of the military treatise, Nikephoros Ouranos, was one of Emperor Basil II’s most experienced generals, and fittingly had conducted a full-scale siege in Syria. While he advised that an investment should begin with missiles hurled by both men and machines, the idea was to clear battlements in order to provide sappers with the opportunity to access, and subsequently undermine, the foundations of the enemy’s walls. His advice was no doubt informed by a technological limitation of the traction-powered (rope-pulled) and hybrid (rope and gravity) catapults: while more powerful than the tension- and torsion-powered catapults used in antiquity – the hybrid could hurl shot up to six times the weight of the projectiles fired by classical engines – they were nonetheless unable to create breaches in the stoutest of medieval fortifications. In the first half of the twelfth century, however, the Byzantines were using the latest development in stone-throwing technology: counterweight trebuchets. Equipped with a sling, windlass, counterweight and beam, stones fired from these gravity-powered engines were not only capable of crushing large houses, but could also demolish walls with concentrated, consistent bombardment. By the last quarter of the twelfth century Hauteville siege engineers had acquired the technology, if not earlier, using these formidable stone-throwers at Alexandria (1174) and Thessaloniki (1185). At the latter siege, one of the trebuchets was so powerful it was known as ‘the earthquake’s daughter’.⁵⁸

Narrowing the focus to the Norman homeland, while Duke William did consider the use of devices such as the catapult (tormentum) when besieging a castle in Maine (1063), his chaplain noted ‘the site was altogether unsuited to siege engines’. Accordingly there were clearly siege specialists available to the duke in Normandy, yet evidently not to his former subjects in Italy until around four decades after they first arrived there, a considerable disadvantage in a region where stone fortifications greatly outnumbered those in France. It is surely no accident, then, that Norman siege techniques in Italy became increasingly sophisticated as they incrementally captured Byzantine strongholds. Capturing, or in many cases receiving the surrender of such strongholds provided the Hauteville commanders with greater access to engineers, technology and techniques necessary to improve their siege operations. A fitting example of the exponential growth in Norman siege abilities can be summarized by consideration of two investments of the Byzantine capital of Apulia, Bari. Given the city’s location on a triangular peninsula, only a third of its walls could be assailed by land. When the Normans under the command of Prince Guaimar IV of Salerno attempted to capture Bari in January 1043, what can only be called an exercise in futility, given their lack of siege engines and a fleet, was abandoned after only five days. Contrast this effort with the one in 1068, when Robert Guiscard decided to make an assault on the city’s impressive fortifications. Using Lombardo-Byzantine engineers hailing perhaps from the stronghold of northern Apulia, Troia, but conceivably from any of the other former imperial towns of Apulia and Calabria now in his possession, Guiscard deployed ‘every variety of catapult’, battering rams, mantlets and forts, in addition to a large siege tower equal in height to Bari’s precipitous walls.⁵⁹

Evidently the only thing missing from the duke’s Byzantine-style siege train was a team of sappers; indeed, there is not a single reference to the use of them by the Normans in eleventh-century Italy. Sappers were certainly used later in Syria – specifically during the Italo-Provençal siege of Ma‘arrat al-Numan (1098) – but such men must have belonged to Count Raymond IV of Toulouse’s contingent (which had already demonstrated its skill in sapping techniques at Nicaea in 1097), not the southern Italian one led by Bohemond. Particularly interesting is William of Apulia’s mention of mantlets or siege shelters (crates) being placed near the gates of Bari, ‘beneath which’ Guiscard placed ‘heavy infantry’ (as Tancred seems to have later done at Tarsus in 1097). Both the description and the term used – ‘wickerworks’ or ‘latticeworks’ – strongly suggests the assimilation of Byzantine laisai, which Ouranos advised in his manual be constructed out of tightly woven vine stalks or tree branches, and that the completed structure should resemble a house (roofed, with four walls and two doors). The use of light, wooden materials was considered essential, so the fifteen to twenty troops within could move the structure without too much exertion while advancing on the enemy fortress. In the twelfth century crusaders were noted to have used a siege shelter called a cercleia, which Paul Chevedden notes is a Latin compound featuring a Greek suffix: Lat. cerc- (‘hooped’) and leia (= Gk. laisa); a hooped siege shelter. Since Guiscard and Tancred may have used them in the previous century, it is possible that Hauteville engineers introduced laisai to other western Europeans in the Levant. Either that, or it was one of the ‘city-takers of every kind’ Alexios I Komnenos introduced to the impressed crusaders at Nicaea (Turkey) in 1097.⁶⁰

Despite this formidable collection of siege equipment, the Bariots did not capitulate for almost three years, an outcome all the more remarkable given Guiscard’s deployment of another Byzantine inheritance: a fleet large enough to blockade the city by sea. As William of Apulia commented in regard to a minor naval victory during the siege: ‘The Norman people were ignorant in naval warfare up this point’. Accordingly he saw the victory as something of a turning point, and quite rightly so. However, he seemed to be unaware that during the course of this victory, when rushing eagerly to one side of their ship to board a Byzantine vessel, 150 armoured Normans were drowned owing to the unsteadiness caused by their inexperienced onrush. Amends were well and truly made in 1072, when the capture of the greatest city in Sicily, Palermo, was made possible by a decisive naval victory. Yet, it is on occasions like this that the term ‘Norman’ seems to be something of a misnomer: indubitably the fleet contained some French-speaking infantry and was under Norman authority, but many of the marines and crew were Byzantines from Apulia and (especially) Calabria, and it is interesting to note that from the time of Stephan Pateranos to George of Antioch (i.e. 1072–1151), Greek-speakers were often entrusted with naval commands. But the same could be said for the land forces which, as more territorial acquisitions were made, increasingly became less Norman (as was also the case in Britain and Syria). Yet observations of this nature should not be taken to diminish what D.C. Douglas in 1969 labelled the ‘Norman Achievement’. After all, the ancient Romans, among other preceding and subsequent military powers, relied on both the manpower and martial talents of those they subjected (e.g. Greek naval skills and technology). Since the Normans were not only able to harness the martial abilities of various peoples, but also to establish powerful, enduring states in southern Italy, England and Syria as a result, they clearly possessed considerable ability, acumen and foresight in the military, logistical, and political spheres.⁶¹