The Crusader Lordship of Transjordan (1100–1189): settlement forms, dynamics and significance

This paper presents the results of a study of the 12th-century Crusader Lordship of Transjordan and discusses the traditional view that the principal role of this region was that of frontier of the Kingdom of Jerusalem. The possibility of applying the concept of frontier to Transjordan is discussed in the context of a debate on the relationship between frontiers and castles, and of the conclusions obtained from the analysis of settlement patterns of this case study. On the basis of the documentary and archaeological data reviewed here, it is argued that the lordship had several roles, including military, political, economic and social, that were of crucial importance for the entire kingdom, to which it was tightly connected. Simply seeing the lordship and its castles as defining a frontier is not only incorrect, but also fails to reflect this region’s complexity and identity. Additionally, it is demonstrated that the common understanding that Transjordan was an isolated and peripheral region needs to be modified; instead, the elements of continuity with the rest of the kingdom were numerous and significant, to the point that several important socio-economic, military and strategic aspects of the kingdom depended largely on the existence of the Lordship of Transjordan. 1 1 The present study is based on a PhD thesis defended at Cardiff University, School of History, Archaeology and Religion, under the supervision of Prof. Denys Pringle, with the title: 'Settlement in Crusader Transjordan (1100–1189). A Historical and Archaeological Study' and published open access by Cardiff University (Sinibaldi 2014). The thesis presents a comprehensive overview of the available historical and archaeological sources on Crusader-period Transjordan; its publication is currently in progress for Archaeopress.


Introduction
In the framework of the long-established tradition of Crusader studies, the Lordship of Montréal, or Lordship of Crac, or Lordship of Montréal and Crac, better known as the Lordship of Transjordan (Mayer 1990: 5-16), and explored and settled by the Franks in the years 1100-1189 AD, has hitherto been, surprisingly, in view of its great significance for the kingdom of Jerusalem, a relatively neglected subject.Crusader Transjordan is characterized by a scarcity of documentary sources; it is worth noting that very few charters are available by comparison with other areas, and that the sort of detailed topographical studies that have been done for other areas of the kingdom (e.g., for the Mount Tabor area in Galilee in Khamisy (2016)) are not, therefore, possible for this region.Despite this, several historians have written on the subject, in particular Hans Eberhard Mayer, who has undertaken a comprehensive study on Crusader Transjordan, based exclusively on historical sources (Mayer 1990).However, when archaeological studies have been carried out, they have been much more limited in scope and very fragmented.
Transjordan has often been commented upon, albeit relatively briefly, by historians and scholars writing about political events in the Latin Kingdom of Jerusalem.Despite the scarcity of available sources, the general assumption appears to be that for the Franks, the main function of Transjordan was to serve as a frontier land located on the south-eastern periphery of the kingdom.Mayer's work on contextualizing the few available documentary sources (Mayer 1990) has highlighted both the complexity and importance of this region, and its significant connections to the rest of the kingdom.Despite this work, there is still a tendency to consider Transjordan a region of lesser importance; functioning merely as a military frontier.Although Transjordan was certainly always on the edge of Frankish controlled territory, scholarship has, in the main, seen this region as a different kind of frontier: a defensive system intended to protect lands further west in the centre of the kingdom.
Deschamps identified military, commercial and agricultural advantages behind the Frankish settlement of Transjordan, but the focus of his discussion was still on the relationship between political history and fortifications (Deschamps 1939: 35-98).He stated that the location of the main fortresses in Transjordan, by forming a line east of the natural borders already present therethe Jordan Valley, the Dead Sea and the Wadi Arabahwas crucial to blocking the potential access of the enemy to the core of the Latin Kingdom (Deschamps 1934: 16-42).Prawer described Transjordan in terms of its military and commercial importance to the Latin Kingdom, mainly because of its position between Cairo and Damascus, and being on the Darb al-Hajj (Prawer 1975, I: 247).He too, observed that the castles of Transjordan could block potential attacks from the enemy, forming a first line of security in addition to the Jordan River, the Dead Sea and the southern desert (Prawer 1972: 285) As Pringle (2013) recently noted, when discussing the relationship between castles and frontiers, these two elements have been commonly associated by researchers, and then these, in turn, have often been linked to the idea that fortifications reinforced natural defences, forming fortified lines to protect the vulnerable parts of frontiers; this view, predominant since 19th-century studies by Rey, was generally accepted until the 1970s (Pringle 2013: 227-31).Smail however, who published in the 1950s, was an exception to this trend, arguing that castles were not particularly useful for defending borders.He also highlighted the fact that castles had a variety of functions and that in areas such as Transjordan, they were used principally to establish control in strategic areas and become centres of colonization, thereby controlling territories, the local population and the revenues generated from the local economy.They were also administrative and residential centres, policing posts and barracks, but above all, they were centres of authority (Smail 1956: 60-61).
Indeed, some 13th-century authors did think that Karak and Shawbak were important in protecting the kingdom, in particular Jerusalem, from Muslim attack; it is possible that the traditional view of Transjordan as a frontier partly originates with these authors.Jacques de Vitry was of the opinion that the campaign of Egypt in 1168 AD pushed the Franks to reinforce these defences.In his opinion, 'because they could not subdue the towns of Cairo, Alexandria and Damietta, and others in the interior of the country, at the limits of the extents of their possessions, in order to defend the borders of the territory they controlled, built some very strong and impregnable castles between them and the enemy: beyond the Jordan, these were Shawbak and Karak' (de Vitry 2008: 216). 2 While, as will be discussed below, it is very clear that this was one of the reasons for building castles in Petra, this observation cannot be related to the main castles (Karak and Shawbak); his lack of correct information is reflected by the fact that Karak and Shawbak were built much earlier than 1168 AD, and is explained by the fact that Jacques de Vitry was not a contemporary of the events he comments upon.His writings however, might help in understanding the opinion of the time with regard to the castles of Transjordan.
The refusal of al-Kamil to return Karak and Shawbak to the Christians in 1219 AD, at the time of the Fifth Crusade, is also meaningful.Regarding the negotiations that occurred that year, Oliver of Paderborn, wrote, in his Historia Damiatina (1219 AD): 'Now these two places located in Arabia which have seven very strong fortresses through which merchants of the Saracens and of the pilgrims, going to Mecca or returning from it, usually cross; and whoever holds them in his power can very seriously injure Jerusalem with her fields and vineyards when he wishes' (Peters 1971: 86).According to James Powell (Powell 1986: 160): 'The main reason to reject the truce was the refusal of the sultan to include the fortresses of Kerak and Krak de Montréal, located east of the Jordan and regarded as essential to the defense of Jerusalem … While the military orders believed these forts to be critical in the defense of Jerusalem, from the viewpoint of the Moslems they posed a serious threat to the continued communications between Damascus and Cairo.… the crusaders were divided on this issue, but could hardly ignore the threat to their control of Jerusalem if the castles remained in Moslem hands'.
Oliver of Paderborn's opinion, that at this particular historical moment Muslim control of those castles could seriously threaten Jerusalem, appears very reasonable since the castles, especially Karak, were not far from Jerusalem, and, as specified below, the presence of castles in a strategic position and a relatively short distance from the enemy was a system shared by both the Franks and the Muslims.This was certainly even more important in a situation where the Franks, after 1187 AD, were in a muchweakened position.This opinion does not contradict the fact that, when the castles were built, the Franks were not minded to build a shield to protect Jerusalem.What might have become true after 1187 AD, had the Franks managed to regain control of those castles, was not necessarily true when those castles were built in the 12th century, a time of military expansion and confidence.Furthermore, Oliver of Paderborn mentions that the value of the castles to the Muslims, in addition to being able to damage Jerusalem, still largely lay in being able to control communication along the Hajj Road.It is very likely that this continued to be the main reason the Muslims wanted to retain control of the castles, and that this remained the situation for much longer than merely the Crusader period, with the castles functioning as administrative and residential sites, as well as ensuring smooth communication along the Hajj Road.This paper argues that the traditional interpretation of this region as an isolated frontier, is largely based on both a scarcity of detailed studies on the subject, and on interpretations of the role of the Lordship of Transjordan that were formed after its actual existence.Based on a detailed examination, critical analysis and combination of the available historical and archaeological sources, including the results from a case study of Petra, this paper challenges the traditional view of defining Transjordan simply in terms of a frontier and proposes an alternative interpretation.It argues that both the region of Transjordan and its castles fulfilled more complex functions than defence and supports Smail's definition of the function of castles.It also demonstrates that for Transjordan the definition of frontier is not only very reductive, but also incorrect.Moreover, it shows that while the lordship had its own, clear identity within the context of the rest of the kingdom, it was also, simultaneously, tightly and deeply connected to it.Finally, it is argued that this region had, overall, a great importance for the entire kingdom, something largely understated until now.The paper discusses four key aspects: settlement forms and dynamics of Crusader-period Transjordan; the structure and function of its castles; the economy of Transjordan; the relationship between the Franks and the local population.

Forms and dynamics of settlement in Crusader Transjordan
The interest of the Franks in the area beyond the Jordan River had already begun by 1100 AD, the year following their conquest of Jerusalem.In addition to needing to secure safe access to Jericho, located on the Jordan River, and which provided agricultural resources as well as being an important pilgrimage site, early exploration of the areas beyond the Jordan River was stimulated by Baldwin I's desire to increase his popularity through military achievements (Mayer 1990: 16-20).Several carefully planned expeditions took place for the selection and organization of areas the Franks wanted to control, and continuous efforts were invested in holding these areas throughout the period of the Frankish presence in Transjordan.The interest in settlement beyond the Jordan River focused on both the Petra region and the area north of the River Zarqa, around the valley of the River Yarmuk, known as the Sawad.In 1100 AD some villages in this area, which was part of Galilee, were granted, by Tancred, to the Benedictine Abbey of Mount Tabor (Röhricht 1893: 5-6, no. 36).Originally it was probably hoped that this more northerly area, which was very fertile and allowed the Franks to control enemy territory, would take the form of permanent settlement under Frankish rule.However, only partial control of this region was ever achieved through a few fortified points, mainly the castles of al-ʿAl and Habis Jaldak, founded in 1105 AD and 1109 AD respectively (Ibn al-Qalanisi, trans.Gibb 1932: 71-72; Sibt ibn al-Jawzi, RHCOr III: 529-30) and by agreements between the Muslims and the Franks on sharing the territory's resources.Settlement developed slightly earlier in this area; however, this area did not become part of the Lordship of Outrejourdain (Fig. 2a).
South of this area and in particular in the area between the Wadi Mujib and Petra (an area that eventually became part of the Lordship of Transjordan) settlement developed in a constant, gradual and carefully planned manner during the 12th century.This was also the area that eventually contained the   despoiling the local population.That the king brought back to the other side of the river about 60 Christians from Wadi Musa (Albert of Aachen 10, 28-31, ed. Edgington 2007: 745-47;trans. Edgington 2013: 136), clearly shows that Frankish settlement in this area did not yet exist, although it was certainly already contemplated for the future.Between 1100 AD and 1115 AD the Franks came to realize that it was this region that offered the possibility of permanent settlement; the areas north of the Wadi Mujib and south of Petra were not considered as suitable, in part because they were both too exposed to potential military attacks from either Damascus or Cairo, and were, therefore, not considered sufficiently safe.Moreover, Petra was the southernmost location where it was possible, given the nature of the territory, to develop large settlements based on agriculture; the kind of settlements that were able to support large castles and their populations.This exploratory phase was followed by a first phase of construction, marked by the important foundation, in 1115 AD, of the first fortress in Transjordan, Montreal (today al-Shawbak), a royal castle, and, presumably, a number of smaller local settlements (Fig. 5).The Frankish interest in settling permanently in the region is clear from the description given by the contemporary Fulcher of Chartres, who outlines specific long-term economic benefits and advantages for Christendom as a whole (Fulcher of Chartres, II.55, ed. Hagenmeyer 1913: 592-93).Shawbak had a safe location that controlled both the surrounding fertile area and the passage of merchants and pilgrims along the King's Highway from whom they could levy taxes.The presence of a Christian population in the areas of Petra and Shawbak was a further encouragement when planning this kind of Frankish settlement.In contrast to what happened north of the River Zarqa, settlement in the south was, apparently, aimed not simply at collecting economic benefits and controlling the enemy from a close distance, but also at settling a Frankish population, and establishing agriculture and a trade network.During this phase, Transjordan formally became a lordship (Mayer 1987: 201); the extent of the territory of the Lordship of Transjordan continued to increase over time.The settlement located in ʿAmman (Ahamant) (Röhricht, 1893: 96-97, no. 366;Strehlke 1975: 3-5), as well as other possible sites in the area between ʿAmman and the Zarqa River, technically within the borders of the Lordship of Transjordan, were not likely to have been particularly significant, since they were clearly not capable of defending a captured Muslim fort located only a short distance north of ʿAmmanthe one in Jarashwhich, in 1121 AD, the Franks chose not to maintain because even to reach it would have involved too much labour and risk (William of Tyre, 12.16, ed. Huygens 1986: 365-66): this suggests that this area of Transjordan was, at that time, considered too dangerous to plan permanent, extensive settlement (Fig. 2b).A first phase of exploration and looting beyond the Jordan River, succeeded by a second phase, from about 1115 AD, when the first permanent settlement was planned, was followed by a third phase that saw the construction of several castles in Petra around 1130/1140 AD, including Wuʿayra (Pringle 1998: 373-77; Sinibaldi 2016b), which was probably constructed around 1130 AD (Fig. 6), and Karak in 1142 AD.The Franks devastated and robbed Wadi Musa in 1127 AD and on this occasion, enslaved and scattered its inhabitants before withdrawing (Ibn al-Qalanisi, trans. Gibb 1932: 182-83; Sibt ibn al-Jawzi, RHCOr III: 566).In 1144 AD, after a siege of a few days, they reconquered al-Wuʿayra Castle, which had been taken by the Seljuqs with the support of the local inhabitants (William of Tyre, 16.6, ed. Huygens 1986: 721-22).
The 1130s and 1140s were meaningful in terms of planning a new phase of settlements in southern Transjordan.This phase relied on a now solid understanding of the territory of Transjordan by the Franks and shows increased confidence in territorial control, culminating in the foundation of Karak Castle (Fig. 7) in 1142 AD by Payen (Paganus), Lord of Transjordan (William of Tyre, 15.21, ed, Huygens 1986: 703-04).Although Qalqashandi (1355-1419) was living much later, his account is detailed in reporting that the Franks decided to build the castle and took the important initiative of making it the administrative centre of Transjordan, after they had been settled for only few months or years in the town, which was surrounded by very fertile land, was abundant in cereals, and after they had attracted Christian communities from the surrounding area to live nearby (cited in Gaudefroy-Demombynes 1923: 131-32).This decision must, therefore, have been taken once they had come to appreciate the many advantages of the area, including the short distance to Jerusalem (which may have even been visible from Karak on a clear day), the fertility of the land, the various opportunities for trade with the local villages of the Gawr and the Jordan Valley, the largely Christian  population, and the position of control over both the King's Highway (for both military and commercial reasons) and the main connecting roads to Frankish territories in the west.The Franks already held possessions in the Balqaʿ, the highlands between the Zarqa River and the Wadi Mujib, as early as 1126 AD (Tibble 1989: 35-36) and it is therefore reasonable to hypothesize that settlement in Karak may have started as early as then, especially if the Franks had already explored the Gawr-as-Safi during the expedition of 1100 AD to southern Transjordan.
A fourth phase of settlement, from about 1160 AD to about 1170 AD, can be identified as one of consolidation of control over the conquered territories.It was during this phase that completion of settlement in the Lordship of Outrejourdain reached its full extent.ʿAqaba was probably not settled by the Franks until after 1160 AD (Pringle 2005) and was already lost in 1170 AD, although as early as the beginning of the 12th century AD it was clearly considered a potential site for settlement.This interest originated not only from a desire to control the southern segment of the Hajj Road, which would have provided additional revenues from pilgrims (Prawer 1975, I: 298), but also for much more important strategic reasons, since control of ʿAqaba would also impede the movements of troops between Cairo and Damascus.In 1170 AD ʿAmman was again lost by the Franks (Mayer 1990: 162).During this phase, in 1168 AD, the seat of the newly revived archbishopric of Petra was transferred to Karak. 3dentifying and clearly distinguishing the phases of settlement outlined above reveals that the Franks acquired, over time, an increasingly clear understanding of the most promising areas for settlement and how best to utilize those areas; this, in turn, led, in the 1160s, to a firmer consolidation of territorial control.The Franks concentrated their main settlements, Karak and Shawbak, south of the Wadi Mujib, because, despite the fact that reconnaissance parties had originally identified the whole of Transjordan as free from direct Muslim control, the exploration and conquest of the areas south of the River Zarqa did not provoke as intense and immediate a reaction from the enemy as had occurred further north.Moreover, the area south of the Wadi Mujb was found by the Franks to already have welldeveloped centres of settlement when they arrived, in particular Karak and Wadi Musa (cf. Walmsley 2001: 518), and the Franks appear to have consciously chosen to settle at these sites.In summary, the choice of location for the main Frankish settlements appears to have been mainly guided by the combined aspects of economic advantage, a strategic but safe location, the presence of a Christian community, and, often, the presence of an existing settlement.It is very clear that through this process, Transjordan became increasingly important and de facto integrated with the rest of the kingdom.
At the same time, as reconstructed by Mayer, the king only very gradually conceded his territories to the Seigneurie; Shawbak only became part of the Lordship in 1130/1136 AD, while al-Wuʿayra wasn't integrated until 1161 AD (Mayer 1990: 159).
The reason for such a gradual process was the clear understanding, on behalf of the crown, of the crucial political and economic importance of this region; hence the caution with which direct control was relinquished.This awareness was present when the king appointed vice-count Pisellus, a very close collaborator, to administer Transjordan for him in 1115-1118 AD (Mayer 1990: 69): Pisellus was the first royal viscount, we know of, in the whole Latin East (Mayer 1990: 81).
It is important, for the present discussion, to summarize the timeline of military confrontations with the Muslims.While in the area north of the Zarqa River, and therefore outside the area of the Lordship, conflicts were continuous until at least the construction of ʿAjlun castle in 1184-1185 AD, the military actions of the Muslims against the Franks in Transjordan can, again, be grouped into several phases.A first phase, before c. 1140 AD, was characterized by raids and the construction of forts, such as the one in Wadi Musa, destroyed by the Franks in 1108 AD.A second phase, from c. 1140-1169 AD, saw, specifically, attacks from Damascus on the new Frankish settlements in the south, as far as Wadi Musa (1144 AD), and from Egypt to Shawbak and Tafila (1156 AD) and to Wadi Musa again (1158 AD).A third phase, 1169-1189 AD, was characterized by continuous attacks on Karak and Shawbak (but especially Karak) by Saladin and Nur al-Din.Saladin reconquered ʿAqaba (1170 AD), aimed more consistent attacks on Karak from 1183 AD and founded ʿAjlun castle 1184-1185 AD.It appears, therefore, that during the first phase the Muslims were reacting to the episodes of exploration and raids by the Franks, while in the second phase they were reacting specifically to the construction of settlements in the south.In the third phase, Saladin, as a wazir in Egypt and then when independent from Nur al-Din, clearly aimed at reconquering the Hajj Road, concentrating on Karak, Shawbak and ʿAqaba, thus reflecting the comparative importance of these sites and of Karak in particular.What is interesting is that military action taken by the Muslims appears to have been a reaction, mirroring the phases of increasing Frankish confidence in their settlements.Moreover, and crucial to this discussion, this military action was never aimed at attacking the castles to conquer what was to the west of them, but rather at damaging the castles themselves, which for most of the time and particularly in the last phase, obstructed their freedom of movement.
The castles of the Lordship of Transjordan: location, function, structure and socioeconomic aspects While identification of the main castles, such as Karak and Shawbak, has always been well known, the location of other castles mentioned in the sources has been the subject of debate.In addition to the castles mentioned in the sources, there were probably other fortifications of lesser importance, which could have been located on the Hajj Road or in other strategic locations.
Oliver of Paderborn, writing in the early 13th century AD, mentions in his history of the Fifth Crusade against Damietta, the existence, at that time, of seven very strong Ayyubid fortresses dependent on Karak and Shawbak, but does not name them (Oliver of Paderborn, Historia Damiatina, ch. 31, ed. Hoogeweg 1894: 222-24). 4Deschamps and Pringle both identified ʿAmman, Tafila, Khirbet al-Hurmuz, al-Wuʿayra, al-Habis (these last three being located in the Petra area) and ʿAqaba, as being among them.Deschamps, however, interpreted the seventh castle as a second structure located in ʿAqaba, one castle being placed on the mainland and one on Jazirat Faraʿun (Deschamps 1939: 39, no.1), while Pringle identified the seventh castle with Khirbat as-Silaʿ (south of Tafile) (Pringle 2001: 678).
Pringle has shown that it is likely that there was only one castle in ʿAqaba, located on Jazirat Faraʿun, an island about 15 km from ʿAqaba (Fig. 8), as is clear from the accounts of ʿImad ad-Din, Abu Shama and Maqrizi: (Pringle 2001: 678;2005: 338-39); the main castle structures currently visible on the island belong to a later, Ayyubidperiod castle, which has been recently heavily reconstructed.
Only three castles apart from Karak and Shawbak have, so far, been securely identified with their material remains: al-Wuʿayra (Li Vaux Moysi: Pringle 1998: 373-76), al-Habis (Al-Aswit: Pringle 1997: 49, no.97; Zayadine 1985: 164-67) (Fig. 9) and Khirbat as-Silaʿ/Sela, north of Buseirah al-Silaʿ : Musil 1907: 318;Pringle 1997: 95, no. 202;Zayadine 1985: 164-67) (Fig. 10).While the first two castles are already known as Crusader castles, my own recent archaeological surveys support the Damiatina , ch. 31, ed. Hoogeweg 1894: 222-24), who makes the same arguments on another occasion; therefore, this information relates to the year 1219 AD.At this time, because the castles of Karak and Shawbak were held by the Muslims, this list might be misleading in terms of identifying fortresses originally held by the Franks, i.e., they might not have been the same fortresses.This should be taken into account regarding the observations reported here by Deschamps and Pringle before this update, when it was thought that the source was Jacques of Vitry.I thank prof.Denys Pringle for sharing this information.
identification of the less known site of Khirbat al-Silaʿ, based on an observation of similar building techniques and ceramics at all three sites.When Karak and Shawbak are added, the material remains of five castles can, therefore, be analysed and commented upon.The military and strategic functions of all of these castles are undeniable, as was true of others in the rest of the kingdom; indeed, these were, for contemporaries, the defining characteristics of a castle.The castle on Jazirat Faraʿun was clearly threatening the safe passage of Muslims through the area, but was probably too small to accommodate more than a small garrison.Most of the castles of Transjordan played a role in guarding the east-west connections in order to block unforseen invasions from Egypt; while Karak controlled one of the main accesses west through the Dead Sea, the castles of Tafile, al-Silaʿ Shawbak and Wadi Musa were all reachable from the Wadi Arabah (Mayer 1990: 206).Given the large concentration of castles in the area of Petra (al-Wuʿayra, al-Habis (al-Aswit) and Hormuz) one of their functions was clearly to check the potential arrival from Egypt, through the Wadi Arabah, of enemy armies of the Frankish territories in the area.Egyptian armies had sieged the castle of al-Wuʿayra  for eight days in 1158 AD (Ibn Muyassar, RHCOr III: 472).Al-Habis (al-Aswit) in particular had a very strategic position, overlooking what, at the time, was one of the main roads to Petra coming from Egypt.The castle of Al-Wuʿayra, on the other hand, was mainly intended to protect south-western accesses to the area, both from the King's Highway and through Petra, and to command the road between Wadi Musa and the northern Jabal Shara, beyond which Shawbak was located (Fig. 3).
The castles of Transjordan were also bases for launching attacks, as is clearly demonstrated, for example, by an episode occurring in 1182/1183 AD, when the Franks, having heard that the sultan had left Egypt followed by a large number of merchants, met in Karak and planned a raid on the caravans (Abu Shama: 217-18).Karak is frequently mentioned as a base for these attacks on caravans organized by the Franks, but a similar function could have been performed by smaller castles located in a strategic position.
As well as being considered the most important Frankish site in Transjordan, and the most threatening castle for the Muslims' freedom of movement in the region, Karak castle had an additional role: its foundation had brought the centre of the lordship considerably closer to Jerusalem.In addition to reinforcing the position of the Franks in Moab and becoming the residence of the lords in Transjordan, the newly founded castle also made it possible to improve safe communications through the region (Prawer 1975, I: 330-31).In 1115 AD, on the occasion of the construction of Montreal (Shawbak), Fulcher of Chartres stated that the castle was located three days' journey from the Red Sea and four from Jerusalem, potentially a convenient location for controlling the area all the way up to the Dead Sea (Fulcher of Chartres, II.55, ed. Hagenmeyer 1986: 592).It is clear, therefore, that both the main castles were, from the moment of their foundation, regarded as having an important role in reinforcing the kingdom's connections between the southern region of Transjordan and Palestine, and were an integral part of the Frankish expansion strategy.
Nevertheless, as was the case for most castles of the kingdom, Karak, Shawbak and the other castles had a far wider role than purely military; they were also centres of lordship (Pringle 2010: 225-26).Five fortresses depending on Karak and Shawbak were located in the region between Tafile and Petra.As outlined above, documentary sources clearly show that the main castles were built in the safer areas of the region, between Karak and Petra, within the core of the lordship, leaving as much less populated the most exposed areas.Despite the constant potential attacks from Egypt, the castles in these areas were relatively far away from Cairo and Damascus.The main castles, as would be expected of a castle in Europe and indeed in other areas of the kingdom, were also primarily centres of settlement and authority, as well as a tool for controlling a territory, its population and its economic activities: archaeology and architectural surveys clearly show this point.
Karak Castle, built on the most obviously defensible point of a mountain, was located within a naturally defended town inhabited by a largely Christian population that the castle defended in times of danger.When Saladin arrived in Karak in October 1183 AD, crowds of people sought refuge in the fortress, including many local Christians from the surrounding countryside.Reynald, Lord of Karak, suggested that people leave the burgus and transport their things into the citadel (William of Tyre 22. 28-29, ed. Huygens 198628-29, ed. Huygens : 1055-57)-57).Shawbak had its own court of burgesses (John of Ibelin, RHCLois I: 420).The sources clearly outline that the purpose behind the construction of the castle was to colonize the area.William of Tyre explicitly mentions that the castle was founded 'because the king wanted to extend the boundaries of the kingdom in that area of the kingdom' and that 'knights, sergeants and villeins' were living in the castle.He also states that 'the town was well fortified by a wall, towers, an outer wall and ditch' (William of Tyre, 11.26, ed. Huygens 1986: 534-35).The description clearly suggests the presence of at least two lines of walls, which has been confirmed by surveys, and within which lay the core of the Crusader settlement.Surveys have confirmed the presence of an upper church, a Latin parish church for the ruling class, built in the most heavily fortified part of the castle, as well as a lower church, in the outer ward, probably built for the local Orthodox Christian population (Pringle 1997: 75-76;1998: 304-14;2001: 678;2004: 35), all of which confirms that a substantial Christian community was controlled and defended by the castle.The community was probably mainly based in the suburb, which was located outside the second line of walls: such an arrangement reflects a similar set up to that found at Karak; the existence of an unfortified space, dedicated to the local community, immediately outside the inner castle.
Similarly, at al-Wuʿayra, the castle included an inner and an external naturally and artificially defended area.The church in the inner part of the Frankish castle is contemporary with the castle's foundation (Pringle 1998: 375-76).A Byzantine-period church, located in the southern part of the broader castle area, might also have been in use during the Crusader period (Brown 1987a;1987b;Leporatti and Vanni Desideri 2018).Byzantineperiod churches in use during the Crusader period were indeed very common in the kingdom (Pringle 1993;1998); this is well documented beyond the River Jordan, for example at the Crusader cave castle of Habis Jaldak, formerly a monastic site (Nicolle 1988;Pringle 1993: 26).Al-Silaʿ Castle was managing an agricultural estate (see below), and, therefore, was likely controlling a population working the land.In contrast, the castle of al-Habis does not appear to have had an important residential function, nor to have managed a large territory; rather, its function appears to have been the control of roads through Petra.That it would not have controlled a large community is supported by the nature of the landscape in the Petra Valley, which was never as fertile as the areas surrounding it, where most of the population naturally concentrated (Sinibaldi 2016b).The castle's main role was therefore, not, as others have stated, to control the majority of the Frankish population based in the Petra area, located in the Petra Valley (Vannini and Vanni Desideri 1995: 513), but rather to control against potential enemy incursions from Egypt.Interpretations from the 1990s, such as those that argued that 'the headquarters of the European newcomers were located […] within the carefully guarded bottom of the valley' (Vannini and Vanni Desideri 1995: 513) have now been revised on the basis of more up-to-date studies (Sinibaldi 2016b).
Archaeology has helped determine the possible identity of castle inhabitants, as well as reconstructing their daily life.At al-Wuʿayra, the small children's rock-cut graveyard, located in a prominent position by the entrance of the church, was clearly for the use of families of commanding rank.Faunal studies, indicating that at least part of the population was living on a varied diet that included products of high quality (Brown and Rielly 2010;Corbino and Mazza 2013), support the theory that a ruling class was resident at the castle; a similar scenario has been reconstructed at Shawbak.The group managing the castle was able to afford imported objects, including fritware from Syria (Vannini and Tonghini 1997: 382).It is unlikely that this ceramic type was commonly available at the local Wadi Musa markets, and probably came from the network of the Hajj Road.
Financial investment in the construction of the main castles' structures was not limited to military considerations, as some parts of these castles reflect high living standards and the involvement of specialized workmanship from outside the area.The southwestern tower of al-Wuʿayra Castle, where a finely built and plastered cross-vault is still standing, and externally slaistered (Sinibaldi 2014), demonstrates a residential function for this part of the fortress, probably as the castle's donjon (Kennedy 1994: 26).At al-Wuʿayra and Shawbak, a sustained level of investment in high-quality construction is evident in some areas, in particular in the innermost churches, where the ashlars have a surface treatment characterized by a very fine tooling and are bonded with a hard lime mortar of very good quality .5 In the upper church at Shawbak, the central nave and aisles had groin vaults (Pringle 1998: 307-11).None of these elements, nor the surface finishing of the wall plaster, are known to be used locally at this time. Prticularly significant from this point of view is the surface treatment of the ashlars of the churches of Shawbak and al-Wuʿayra, which must have been costly to produce.It has been calculated that at this time, it took, at the very least, one full day for a skilled mason to prepare a special stone, such as one for an apse (Shotten-Hallel and Kool 2016).Finelydressed ashlars were used quite extensively in the upper church at Shawbak, where the stones comprising the groin vaults were also very time-consuming to prepare; if the church was indeed completed in only three years, as suggested by an inscription mentioning the date of 1118 AD (Pringle 1998: 308-09), this must have meant a considerable deployment of money and labour in an unusually short time.
The inner church of al-Wuʿayra and the lower church at Shawbak may both have been planned by the same architect (Pringle 1998: 376;2004: 35); all the Frankish churches at the castles of Petra and Shawbak show an architectural style that has much in common with other Frankish churches in the Middle East (Pringle 1998: 307-08).The presence of specialized workmanship was also present in the Karak Castle church, which was originally decorated with frescoes (Pringle 1993: 290).This is in contrast with what can be seen at al-Habis Castle in Petra, which although similar in building technique to Shawbak and al-Wuʿayra does not display the same high level of financial investment.This contrast probably reflects the different purpose of the castle, which was intended less as the centre of an agricultural estate and a centre of power than as a fortification with mainly military functions.In general, it is rarely possible to know with certainty whether the lord of the castle physically resided at the site, since castles were often run by stewards (Pringle 2010: 224).However, it is clear that an important element of representation of local power was present at the three main castles of Transjordan (Karak, Shawbak and al-Wuʿayra) and therefore, that these castles were, at least at times, centres of residence for the representatives of such power.This is apparent in the financial investment.The documentary sources support these conclusions.For example, when Shawbak was still a royal possession, a royal castellan was appointed to reside there (Mayer 1990: 64).Al-Wuʿayra Castle was probably the residence of both a crown vassal and of a royal commissioner for most of its existence (Mayer 1990: 191, 200).
In light of these observations, gathered from recent surveys, on the investment in high-quality and specialized workmanship, earlier conclusions referred to the castles of Transjordan, such as that 'perpetual manpower shortages, the lack of skilled workers and the task's immensity dictated that the fortified desert line be built economically' (Brooker and Knauf 1988: 186) now need to be revised, or at least nuanced.Among the elements mentioned as characterizing an especially noticeable economy of construction at the castles of Karak, Shawbak, al-Habis and al-Wuʿayra, is the intense reuse of earlier materials, which is represented, for example, by the first phase on the east front of Karak Castle, suggesting that this was either inadequate or perhaps temporary (Brooker and Knauf 1988: 186), or influenced by the rushed construction (Deschamps 1939: 80-81).However, the state of conservation of Transjordanian castles attest to their, generally, very high quality of construction, mainly due to the good standard of their building mortar.Moreover, the reuse of earlier building materials does not, in itself, indicate poor quality of construction: reuse is a very common characteristic of Frankish architecture in the rest of the kingdom, where the presence of a former site, affording a supply of construction material, was often one of the main elements guiding the choice of the location for a new settlement; this might have been the case the case, for example, at Jifna, Khirbat al-Marjama, Iribbin, Saffuriya, Qaqun, Ramla, ʿAbud and ʿAmwas (Sinibaldi 2002).
All five castles of Transjordan whose material remains are identified were built on ruins of an earlier period, making it logical to reuse ruined structures and building materials.It is known from the sources that when the Franks built the castle on Jazirat Faraʿun, they found the ruins of a previous castle which they refortified (Pringle 2005: 339).In general, in addition to the advantage of using a supply of materials already at the site, very often a formerly occupied site had been selected for the same characteristics that defined a good location for a new site.William of Tyre, for example, stated that one of the main motivations for King Amalric's choice of location for Darum, was the presence of remains of earlier buildings (William of Tyre 20. 19, ed. Huygens 1986: 936-37).
Besides a careful choice of location, the Franks used the very successful strategy of building almost impregnable castles, which could hold out until help arrived from other locations.A strong reliance on passive defence, meaning that the Franks had to wait until help arrived from other locations, was a rather precarious balance in terms of distribution of forces.The Muslims were eventually successful in their overall defeat of the Franks because they understood that, when defending their strongholds in Galilee, the Franks would have to weaken their defence of the castles of Transjordan and vice-versa (Prawer 1975, I: 596).All three castles of Karak, Shawbak and al-Wuʿayra surrendered between 1188 and 1189 AD, but none were taken by military force.Karak castle withstood several intense attacks and sieges before the town, though not the castle, was taken for the first time in 1183 AD.William of Tyre states that the village present in Karak at the time of the Franks' settlement was in a safe location and that it could be defended by only a few men.The castle of Karak was reinforced, after its construction, by the successors of Payen the Butler, Maurice and Philip of Nablus (William of Tyre, 22.28, ed.Huygens 1986: 1056), reflecting the growing number of attacks on the region by the Muslims.The existence of these different phases and the reinforcement of the fortified system over time are clearly shown in a structural survey of the castle for the Crusader period (Sinibaldi: 2019), and it is likely that similar improvements to the original structures occurred at other castles as well.However, while the planning of and maintenance of defence for these castles appears to have been extremely efficient, these, as well as the more general defensive aspects, do not appear to have been particularly different from those in the rest of the kingdom.All five castles analysed here (Karak, Shawbak, al-Wuʿayra, al-Habis and al-Silaʿ are strongly protected by natural defences.At Karak and at the Petra castles, the Franks adopted the strategy of using the bedrock; isolating a rock spur from the surrounding space by cutting a moat through the connecting side (Fig. 13).This has the advantage of isolating the site artificially and not having to construct walls.In Petra, adopting the strategy of carving the sandstone, a locally well-developed system by the Nabataeans, was particularly efficient.While the fosse excavation at al-Wuʿayra has been interpreted as requiring an immense effort and reflecting urgent strategic needs (Brooker and Knauf 1988: 186), this should rather be seen as a skilful and efficient adaptation of the Franks to the local environment.This system was not, however, limited to the castles of Transjordan, but was also used at several other Frankish sites including Banyas, Beaufort, Qalʿat al-Dubba, Qalʿat Jiddin and Montfort (Sinibaldi 2002).In fact, it has been noticed that the moats of al-Wuʿayra and Karak have similar depths (between 20 and 30 m) to the one at Sahyun in Syria (Biller et al. 1999: 54), and it is likely that at al-Wuʿayra, at least part of the sandstone which was not reused from existing buildings came from the rock-cut moats, as at Sahyun.In summary, based on the evidence of those castles whose remains can be identified at present, the Franks in Transjordan took advantage of local building techniques and made skilful use of the topography of the local territory when fortifying their castles.The building of those castles also involved an element of economy; however, they did so in a way generally similar to that seen in other areas of their kingdom, as confirmed by recent surveys referred to above.Moreover, they also clearly included the use of specialized workmanship, as well as a high technical level of craftsmanship.

The economy of the Lordship of Transjordan
On 31 July 1161 AD, King Baldwin III donated Shawbak, with all its land, to Philip of Nablus and his heirs, as well as the castle of Karak, ʿAmman, the castle of Wadi Musa (al-Wuʿayra) and the lands previously owned by Baldwin, viscount of Nablus (Röhricht 1893: 96-97, no. 366;Strehlke 1975: 3-5).Interestingly, the king also introduced, in this agreement, limitations to the power that he had conceded to the Lord of Transjordan: he kept for himself the revenues collected from the caravans passing through the country, as well as those from the Bedouins, and he also requested that John Gothman, a vassal who held lands in the region, continue to pay homage to the king.All of this was, clearly, both to continue to take advantage of the economic revenues of the region and to limit the independence of the lordship, whose enormous strategic importance the king clearly understood (Barber 2003: 69).The decision of the king, to have exclusive rights to control all caravans passing through the Lordship, was a wise political choice, since the Muslims would not have accepted instability on this routea situation that was eventually challenged by Raynauld de Châtillon's behaviour, with catastrophic consequences (Mayer 1987: 201-02).The demand of the king, to retain entirely for himself the taxes from the traffic of the Hajj Road and those paid by the Bedouins, indicates that these must have generated a considerable income.This policy shows, crucially, how much Transjordan and its resources, not only strategic and political, but also economic, were valuable for the rest of the kingdom.In addition, these aspects show that the economic and political connections between Transjordan and the rest of the kingdom were not only of great importance, but were also very close.Petra's crucial role in both the east-west and northsouth main roads in the region had a long history prior to the Crusader period: besides the good road conditions, the way through Wadi Musa towards Gaza offered, in ancient times, a rare opportunity for abundant water and provisions, and the Via Nova Traiana, in the north-south direction, passed through Ayla ʿAqaba, Wadi Musa, Shawbak, Karak, Rabba, Madaba and ʿAmman (Mayer 1990: 201).According to Mayer, during the Crusader period, in order to try and avoid the control and taxation from the main castles (Karak and Shawbak), the caravan road switched eastwards and passed through Maʿan, which was closer to, and therefore controlled by, Wadi Musa, rather than Shawbak (Mayer 1990: 128) (Fig. 3).Additionally, although during the Crusader period the traffic in the north-south direction had originally been controlled by Shawbak, once no longer directly controlled by the crown, when it became part of the lordship in the 1130s, control of the traffic switched to al-Wuʿayra Castle, which remained under direct control of the king until 1161 AD.
The location of al-Wuʿayra Castle also made it possible to control, in addition to the nearby village of Wadi Musa, the main access from the Wadi Arabah, which was not through Petra, but rather through the Wadi Nemela and Bir Madkur.However, as stated by Musil (1907: 220), given the distance from al-Wuʿayra Castle, control could only be effective with an extra fortified point, Hormuz Castle, in the area of Baydha, overlooking access from the Wadi Nemela (Mayer 1990: 203-04).From Baydha, at the end of Wadi Nemela, travellers coming from Wadi Arabah could continue either north towards Shawbak or south-east through the village of Dibidba, to Udruh and Maʿan, and access one branch of the caravan route.In summary, with the construction of Hormuz, al-Wuʿayra, as the main castle of the region, was able not only to prevent attacks from the south and west, but was also able to control caravans travelling those same roads, in addition to the caravans travelling to Maʿan through Baydha and the Wadi Nemela (Fig. 3).This is why the king only released possession of al-Wuʿayra and the Petra castles to the lordship in 1161 AD, and why he retained control over the trade routes controlled by the castle (Mayer 1990: 207).Raynauld's systematic disregard of the agreement between the king and Saladin was probably aimed at gaining control of the rich revenues from the Hajj Road, which must have been very significant: his expedition to the Sinai in 1177/1178 AD was probably undertaken in order to demonstrate to the Muslims that he could control the most important roads and impose taxation on travellers (Mayer 1990: 60).
Petra and the Jabal Shara present an ideal subject for a case study, both because the area was intensely settled in the 12th century AD and because it has recently provided new evidence from archaeological excavations and surveys (Sinibaldi 2016b).Based on the analysis of ceramics from several stratified assemblages, a local ceramic chronology, based on chronologically diagnostic local handmade pottery, has been created for Petra (Sinibaldi 2013a).Results have been matched with a study of building techniques characteristic of the Crusader period and with data from documentary sources (Sinibaldi 2014).In addition to producing some preliminary archaeological tools for better identifying the 12th century AD in the Petra region, such tools have also made possible the confirmation or rejection of Crusader-period chronology previously assigned to several lesser-known sites.This study (Knodell et al. 2017;Sinibaldi 2009;2013a;2013b;2016a;2021) has led to the rejection, beyond any doubt, of the long-established theory (Hammond 1970, followed by Vannini andVanni Desideri 1995) that Petra was completely abandoned after the Byzantine period and remained so for the entire Islamic period, with the exception of an important, though short-lived, revival during the Crusader period.
Archaeological evidence of continuity of occupation from ceramic assemblages, both inside and outside the valley,6 supports the conclusion that the Petra Valley and region were inhabited continuously, though not intensely, throughout the Islamic period, and that there was no Crusader-period revival between two long periods of abandonment.Part of the Petra Valley population gradually moved, after the Nabataean period, to nearby areas, such as Wadi Musa and Baydha, where water was more plentiful and the conditions for agriculture more favourable.This, however, was never a complete process; the Petra region and valley continued to be inhabited.The Crusader period did not, therefore, impact deeply on the Petra area; this conciliates information from documentary sources, which ascribes increased population in the area, as in many other areas of Jordan, not to the Crusader period, but to the earlier Fatimid period (Walmsley 2001: 518;554-55).Additionally, although some fragments of imported pottery were present at al-Wuʿayra Castle, results have clearly shown that ceramic imports were generally very limited in this period in the Petra region.
In summary, in addition to showing essentially dominant patterns of continuity and adaptation to the local environment in Petra, rather than disruption and innovation as formerly assumed, this research clearly showed that the core of the local economy in the Petra region was agriculture and not trade.While Mayer's opinion, as outlined above, that the castle of al-Wuʿayra was crucial in controlling trade from the south and the west is persuasive, the relatively scarce number of ceramic imports found at sites in Petra suggests that the economy of the Petra area was not particularly influenced by international trade, and that the basis of the local economy was agriculture, which also supported the castles of the Petra region.It will, hopefully, be possible to comment further on this once the Crusader phases at Shawbak and Karak have been excavated and published, and can, therefore, be compared to Petra.
Control of trade was not limited to the Hajj Road.Karak Castle's position had the advantage of controlling traffic on the Dead Sea as well as the region to its south and west, an important part of the economy of the lordship.In a charter of 1152 AD Maurice, a lord of Transjordan, added the right to transport goods on the Dead Sea without charge to several concessions made to the order of the Hospitallers (Delaville le Roulx 1894: I: 160, no.207).These donations were later confirmed to the order by Raynauld de Châtillon's in 1177 AD (Delaville le Roulx 1894, I: 355-56, no. 521;Röhricht 1893: 71, no. 279;146-47, no. 551).Although some of Idrisi's statements are based on earlier reports, he mentions boat traffic on the Dead Sea, some of which transported dates (Marmardji 1951: 15); such traffic may have still existed at the time he was writing in the mid-12th century AD.The construction of Karak Castle also allowed the Franks to control the Karak plateau and most likely the Ghawr, including trading relationships with villages in the area.Zughar (today Khirbat Shaykh ʿIsa), in the Ghawr, overlooked one of the most important crossings to the west and was a site with significant resources (Fig. 1).This is witnessed in a map of c. 1300 AD by an anonymous Italian cartographer,7 depicting the site, Segor, as the main point of arrival to Karak from the other side of the Dead Sea (Brown 2013: 723;Rohricht 1891: 8-11, plate 1).Indigo, bitumen, salt, dates and minerals, including sulphur, are all known to be important products from this area before the 12th century AD, while during the 12th century AD, dates and indigo production are mentioned as important in as-Safi (Al-Muqaddasi, trans. Collins 2001: 151, 154;Le Strange 1965: 31, 65-66).In the 14th century AD, sugar from Cranco (or Cracco) de Montreal is mentioned; this was considered of higher quality than that produced in Alexandria, but of lower quality to that of Cyprus, Rhodes or Syria (Pegolotti 1936). 8he name Cranco di Monreale refers to Karak (Brown 2013;Mayer 1990: 5-16).The production area was not the castle itself, but the area controlled by Karak, which would have probably included the Ghawr as-Safi, where sugar production, important in the early Mamluk period, was probably already underway in the 12th century AD. 9 Once Karak Castle had been founded, the Franks would have been able to exercise control over the villages of the Ghawr, as well as the distribution and sale of sugar, probably through some sort of commercial agreement.The intense connection between this area and the region west of the Jordan River is supported by archaeology.At Khirbat Shaykh ʿIsa, in as-Safi, excavations have shown the use of imported pottery during the 12th century AD, the result of intense contact with other areas of the kingdom and the Latin states.Our inability to comment in detail on the role of Karak and the Kings' Road in connecting Transjordan and the Palestinian region is heavily influenced by the almost complete absence of excavations at Karak and the lack of published, clearly isolated, 12th-century levels at Shawbak.However, the archaeological evidence gathered to date, strongly suggests that the trade in ceramics reflects an intense use of the Hajj Road, through which, between the 12th and the 14th centuries, objects from Palestine travelled all the way to Petra in the south (Sinibaldi: 2013b).In summary, the available evidence from ceramics suggests that the main area of Frankish settlement between Karak and Petra, was mainly characterized by its connections with the Palestinian region, whether directly or indirectly, through the King's Highway.
The 1161 AD document implies another important aspect.If the king was able to grant the land in Transjordan without also granting the tolls as an economic resource to support the costs of the lordship, including the very high costs associated with the building and maintenance of castles and their garrisons, it follows that the lords must have been able to cover these costs with other significant means.These must, necessarily, have been the well-documented agricultural resources, which were flourishing in the area beyond the River Jordan.We are aware, for example, of specific names of fiefs granted in Transjordan, between ʿAmman and Karak, and almost all the sites occupied by the Franks were in the highland area east of the Ghawr, where it is possible to grow barley, wheat and fruit trees (al-Bilbisi 2013: 44-45).The whole region south of the Wadi Mujib is described as very fertile by contemporaries, and Al-Idrisi, who wrote in the mid-12th century AD, described the Sharah and Jibal districts as producing large quantities of pomegranates, figs, almonds, olive trees and grapes (Le Strange 1965: 35).At the time of Thietmar's visit, in 1217 AD, to the 'plains of Moab', he comments that the land was flat, green and abundant in goats and corn, although without trees (Pringle 2012a: 119).Johns' study of the Karak plateau, supported by both archaeological and historical sources, concluded that in the 12th century AD there were a large number of agricultural villages and a lively and varied economy (Johns 1994: 11-14).Ibn Jubayr's claim, in 1184 AD, that Karak supported no less than 400 villages is likely an overstatement in terms of numbers, but it does testify to the prosperity of the area and to its being based on agricultural activities (Ibn Jubayr, trans. Broadhurst 1952: 301).
There is even more direct and specific information on the area of Petra and Shawbak, especially Petra, both of which are described as very fertile in the medieval sources; this is fully supported by the archaeology (Sinibaldi 2014;2016b).Early in the 14th century AD, for example, Abu al-Fida described Shawbak as a small town located in the Shara province, with many gardens and mainly Christian inhabitants.He adds that at the foot of the hill were two springs, whose water ran through the town and irrigated the gardens in the valley to the west, whose fruits, including apricots, were exported even to Egypt (Abu al Fida, cit.Le Strange 1965: 536).It is stated explicitly that Shawbak Castle was at the centre of important agricultural activities, and we have names of villages as its dependencies, including a village called Benisalem (Delaville le Roulx 1894, I: 160, no.207).Excavations in the Nawafla district, in the north-eastern sector of Wadi Musa, have discovered a site with a long history of occupation ('Amr et al. 2000: 241), including the 12th century AD,10 which was mainly agricultural, as is very evident from the series of olive-presses recovered ('Amr et al. 2000: 233).This ties in with correspondence found in the historical sources, which states that in 1144 AD, when the Franks tried to reconquer the castle of al-Wuʿayra, which had been taken by the local population, olive trees were the main source of income for the local community (William of Tyre, 16.6, ed. Huygens 1986: 721-22).These observations corroborate those made by Fulcher of Chartres concerning the fertility of Wadi Musa.A village called Hara, in the area of Wadi Musa, was granted to the Hospitallers around 1160 AD (Mayer 1990: 98-99), suggesting that by that time Frankish settlement was well established, not only in the castles, but also in the town area.Al-Wuʿayra Castle is mentioned in the 1161 AD list of donations as the castle of Wadi Musa, revealing that it was the main fortification at the centre of a system of other castles, including al-Habis and Hurmuz, for controlling the Petra area.This clearly suggests, therefore, that Wadi Musa was controlling several villages and agricultural production in the area.Hormuz Castle, whose specific location is still under discussion, lay in the Baydha area, located a few kilometres north of Petra, in an area that is still largely agricultural today.Excavations at Baydha have revealed that the site was occupied during the Middle Islamic period, including, most likely, the 12th century AD (Sinibaldi In prep a; In prep b); if this was indeed the case, it is almost certain that this village was controlled by the Franks, given its ideal position between Wadi Musa and Shawbak, its potential for agricultural productivity, and, of course, its vicinity to Hormuz Castle which must have required support from an agricultural village (Fig. 14).In addition, 12th century AD ceramics found in another village of the Baydha region, Baʿja, suggest that other areas of this fertile and strategic region were inhabited at the time (Sinibaldi In prep c).Finally, there exists a very interesting document that casts some light on the organization of agricultural territory in southern Transjordan.In 1187/1188 AD, a letter from Saladin to his brother included a list of sites conquered by the Muslims, which are called 'cities': each of them had, as its dependencies, villages, fields, smaller sites and territories (Abu Shama: 303).This suggests that the sites of Hurmuz and al-Silaʿ which are included in this list, were each at the centre of a territory and administrative district of some sort, even if they were not the main fortified centres of power.This is solid and important evidence confirming that, at least in southern Transjordan, settlement and economy were largely based on a well-developed agricultural system, as was the case in most of the rest of the kingdom.
In summary, he Franks controlled resources in Transjordan that were much more important than those controlled to the north of the River Zarqa.The Franks introduced themselves into an area of long-established agricultural prosperity in Karak, Shawbak and Wadi Musa.The main centres of settlement, which were also the main castles of the region, were economically sustained by the management of agricultural resources through the control of the local community.Agricultural potential was, therefore, one of the key aspects in selecting the location of the main settlements.In addition, the Franks collected important resources from the control of pilgrimage and trade, which were an important part of the income for the rest of the kingdom and from which the crown took important benefits.Transjordan was, therefore, not only a largely prosperous and economically independent region, but was also tightly connected to the rest of the kingdom, providing it with important resources.Such connection is reflected and demonstrated by the, albeit limited, ceramic trade, including imports from the Palestinian and coastal region of the Frankish-controlled territories, that arrived in the Petra area.

The Franks and the local population
The subject of the relationship of the Franks with the local population is of crucial importance when defining the identity of Transjordan, in particular the one between the Franks and the Bedouins.11For the purpose of this discussion, however, it is relationships with the Christians that will be analysed in more detail.The first aspect to be analysed is the suggestion that a strong reliance on the local Christian population was a central element of settlement strategy in Transjordan, and that this was especially important in the very first stage of Frankish explorations and settlement.During this first phase, the strategy for dealing with Christians was to move them from Transjordan to west of the river.According to William of Tyre, one of the main concerns at the very beginning of Frankish settlement (1099-1100 AD) was the fact that the few sites under Frankish control were surrounded by enemy territory and, therefore, safety was considered precarious.An additional problem was that local Muslims refused to cultivate the fields (William of Tyre, 9. 19, ed. Huygens 1986: 445-46).King Baldwin I therefore decided to populate Jerusalem with local Christians, and following careful investigation discovered that in Arabia, beyond the River Jordan, there were many Christians.The king relocated them in sections of the city where it was felt they were most needed (William of Tyre, 11. 27 ed. Huygens 1986: 535-36).
While the claim that the Muslims' refusal to cooperate in agricultural production with the Franks clearly referred only to their initial reaction, the passage appears to support the argument that from early on the Franks understood that the agricultural system's best chance of success required the involvement of the local Christian population.Both statements suggest that, at least in the initial phase of settlement, the hope of receiving support from the local Christians was high.The Franks must have realized early on that their own survival as a minority was highly dependent on the relationship with the local population, in particular because they had to rely on the functioning of an already established agricultural system.The second statement is also indicative of the large number of Christians to be found in Transjordan at that time.
Local Christians were regularly employed as scouts by the Franks, including during King Baldwin I's expedition to Transjordan in 1100 AD (Fulcher of Chartres, II.4, ed. Hagenmeyer 1986: 374-75) and during the 1108 AD expedition (Albert of Aachen 10: 28-30, ed. Edgington 2007: 745-47).They were also in charge of running important castles in Transjordan.In 1182 AD the castle of Habis Jaldak surrendered to the Muslims: William of Tyre blamed this on the fact that the garrison in charge was Syrian, an 'effeminate and weak race' (William of Tyre, 22.15, ed. Huygens 1986: 1028-30).This passage suggests that it was probably necessary to hire Syrian Christians due to the scarcity of western men of arms; this is all the more likely for several castles in Transjordan, relatively far away from Palestine, where most of the Europeans were settled.William of Tyre himself reported that the population of the village of Shawbak was Christian, meaning that more reliance could be placed on it (William of Tyre, 20.27, ed. Huygens 1986: 950-51).This statement, which contrasts with the one made about the cave castle of Habis Jaldak, suggests that this relationship was of crucial importance and, despite the inherent risks, was a worthwhile investment on the part of the Franks.
The relationship with the Christian population was, however, far more than simply a useful tool for settlement strategies.It also corresponded to a broader aim, one of the declared core goals at the origin of the Crusades: the protection of Christendom.This is clearly reflected in the fact that, as mentioned above, a garrison was placed in Shawbak in the interest of the Christians, probably meaning both native Christians and Franks, and therefore, the protection of the Christian community in general (Fulcher of Chartres, II.55, ed. Hagenmayer 1986: 592).Additionally, Ibn al-Furat tells us that Frankish settlement started in Karak because the monks of an existing Orthodox monastery asked some Franks to move there to protect them from the Bedouins (Ibn al-Furat, ed. Lyons 1971: 51).
It is clear that Karak, Shawbak and Wadi Musa were largely Christian areas when the Franks first arrived.There are good reasons to think that the presence of an Armenian community in Karak in the 12th century AD was strongly encouraged by the arrival of the Franks, probably with the aim of including them in the benefits arising from commerce along the Hajj Road. 12n Wadi Musa, around 1160 AD, a village called Hara, which included a parish church of St. Moses, was administered by an individual with the Christian name of Saba, who was the son of George.The village was originally given to Saba by King Baldwin II (1118-1131); Saba most likely acted as the king's steward (Mayer 1990: 98-99;Pringle 1998: 377).In the mid-10th century AD Jabal Harun is listed as a Christian site in the possession of the Melchites (Schick 1997: 76).Some of the structures of the Byzantine monastery were still inhabited in some form, although not extensively, by monks in 1100 AD (Fiema 2008).In 1217 AD, Thietmar visited the site and stated that two Greek Christian monks were living in the church on top of the mountain (Pringle 2012a: 121).Archaeological evidence suggests that the monastery may have been in use until the 13th century AD (Fiema 2016).At Udruh, a few kilometres south-east of Wadi Musa, a Christian population was evidently still present until at least 1866 AD, as recorded by Mauss, who was travelling in the region at that time (Brünnow and Domaszewski 2004, I: 462).The village of Dibidba, near Baydha, is known to have been Christian until the end of the 19th century AD.It is therefore also possible that Baydha, where 12th century AD pottery has been recorded, was Christian, and that, like most, or all, of the other villages controlled by the Franks, was not newly founded by the Franks, but was a long-established indigenous settlement that the Franks relied on for agricultural supplies and with which they maintained friendly relations (Sinibaldi In prep a; In prep b).
The archaeozoological and ceramic records suggest that the inhabitants of al-Wuʿayra probably depended on the population from Wadi Musa for food provisions and purchasing ceramic objects.Textual and archaeological sources make it clear that in Transjordan, as they also did in areas west of the Jordan (Ellenblum 1998;Pringle 2001), the Franks chose to settle mainly in Christian areas.It is notable that they did not invest in large settlements in areas such as ʿAqaba, which, among other issues, were probably largely Muslim.
It is also likely that at this early stage, in the full spirit of the First Crusade, the plans of the Franks may have included protecting pilgrims travelling to religious sites.As noted above, archaeology has shown that the religious community at the monastery of Jabal Harun in the 12th century AD was probably quite small and that the visitation of the mountain was probably not intense (Fiema 2008: 434-41;Sinibaldi 2013a;2016a: 206); it can therefore be reasonably assumed that whatever pilgrimage there may have been to the site in the 12th century AD would have been minimal.It should be noted, however, that when writing about the first expedition to Transjordan in 1100 AD and the arrival of Baldwin I's party in Wadi Musa, Fulcher comments on the sanctity of the place because of the proximity of Aaron's tomb (Fulcher of Chartres, II.5 ed. Hagenmeyer 1986: 381).This suggests that when the Franks first arrived in the area, the idea of settling in places of Christian religious significance was at least contemplated.This suggestion becomes more significant with the realization that the depiction of pilgrimage sites in Jordan, on 12th-century pilgrims maps, includes Mount Nebo and Jabal Harun, in addition to Karak and Shawbak (Brown 2013: 718;Rohricht 1895: 176-78, plate 5).
It should also be noted that the main route to the pilgrimage site of St Catherine's Monastery (Fig. 15) in the 12th century AD was through Transjordan, and remained so until the Mamluk conquests in the mid-13th century AD.Thietmar, visiting the site in 1217-1218 AD, passed through Madaba, Karak, Shawbak, Jabal Harun and ʿAqaba (Thietmar 8-22 in Pringle 2012a: 107-28).
Moreover, the mention by several sources that the archbishop of Petra's suffragan was the abbot of St Catherine on Mount Sinai (Edbury 1997: 192; John of Ibelin, RHCLois I: VII; Mas Latrie 1871: 68) indicates that the Franks clearly had ambitions of controlling this important pilgrimage site, even though Latin jurisdiction over St Catherine's was not acknowledged by the Orthodox community living there (Hamilton 1980: 182, 185).Such ambitions were already apparent in King Baldwin's unfulfilled desire to explore St Catherine's Monastery in 1116 AD.The goal of controlling Christian sites for the sake of organizing and managing pilgrimage in the region would certainly have combined very well, financially, with that of controlling pilgrimage and charging Muslim pilgrims and merchants along the Hajj Road.
In summary, the support of the local Christians in Transjordan, as in the rest of the kingdom, was part of a broad agreement according to which, in exchange for the advantage of controlling the territory and its resources, the Franks would protect the Christians and promote their safe settlement and pilgrimage activity in accordance with the principles of the first Crusade.The presence of a Christian population was, therefore, another of the principal reasons for choosing to build new settlements in Karak, Shawbak and Wadi Musa, although this enterprise was not without challenges and required considerable investment.
The Lordship of Transjordan: a frontier of the Latin Kingdom of Jerusalem?
On the basis of the evidence presented above, two conclusions have emerged: that the castles of Transjordan were never intended to form a frontier to protect the Palestinian region; and that Transjordan as a region was not a frontier in and of itself.
A frontier is traditionally understood to comprise a series of forts strategically placed to protect an area against an external enemy.This, however, was not the reason for the geographical location of the main castles of Transjordan along a roughly north-south line.Their alignment was naturally determined by both their position on the Hajj Road, from which the Franks benefited for both military and economic reasons, and by the presence of earlier sites.These two aspects offered the opportunity to take advantage of building materials, an agriculturally promising land and often a pre-existing, organized Christian settlementa very common choice throughout the rest of the kingdom.Therefore, the siting of castles was strategic, not only for military reasonsand these castles in any case were not connected to creating a frontierbut also for economic, practical and social ones.
Since the traditional understanding of a frontier implies that there is a clear-cut borderline connecting castles, the definition of the borders of the lordship will be discussed.ʿImad al-Din wrote, in 1189 AD: 'We are today owners of all the Latin Kingdom, which is bordered on the Hijaz side by Karak and Shawbak' (Abu Shama: 391-92).To the Muslims, Karak and Shawbak were certainly the most important sites threatening their security of movement, so it is natural that these were mentioned in this context.However, the Frankish control of territory did extend, geographically, east of the King's Highway (the important ancient trade route connecting Transjordan to Syria), on which Karak and Shawbak were located.As shown above, Udruh and Maʿan, east of these castles, were at times controlled by the Franks.The main castles were not, in fact, sited at the eastern extremity of the territory controlled by the Franks and there was no borderline connecting them.There would have been no point in setting a borderline to the east of the King's Highway, an area where settlement was very sparse and where, moreover, there was certainly no agreement with the Muslims on any line of demarcation.Aside from this, although there were occasions within the kingdom when borders were set by treaty between Muslims and Franks, there is no evidence for the creation of borderlines connecting Frankish castlesin the sense of clearly defined lines indicating borders drawn between one castle and another.In the charter of 1152 AD, wherein the extent of Frankish control over Transjordan is described, this is done in general terms by elements of geography: on the north by the River Zarqa, on the south by the Red Sea, and on the west by the Jordan Valley, the Dead Sea and the Wadi Araba.Castles were meant to control and protect not a defensive line, but an area; including smaller sites which fell within their sphere of influence.Smail's opinion that the Frankish dominion 'could best be represented on a map not as an area bounded by a frontier line but as a series of points, which were the fortified places' (Smail 1956) is fully confirmed by the case study of Transjordan.
Crucially, the main evidence that the castles of Transjordan were not defending a frontier is the fact that the strongest castles, in particular Karak and Shawbak, were built in the safest areas of the lordship, while the parts of Transjordan most exposed to attacks from the Muslims, like ʿAqaba and ʿAmman were left with much more limited investment in terms of fortified sites.Smail pointed out that the areas most exposed to Muslim attacks from Damascus were left more or less unfortified, despite the fact that the danger coming from Damascus was always, for the Franks, the most immediate.For instance, the ford of the Jordan at Sannabra, south of Lake Tiberias, was used on numerous occasions by the Muslims to enter Frankish territories; it was, nevertheless, left virtually unfortified (Smail 1956: 207-08).This interpretation has been followed by Ellenblum, who observed that in the kingdom before the late 1160s there was almost no relationship between the construction of Frankish castles and military confrontation with the Muslims, and that many of the strongest castles were, in fact, built in the safest areas, in particular, in centres of agricultural production.He concluded that danger from the Muslims was not the reason for the construction and use of castles (Ellenblum 2007: 159-60;172-74).
Prawer observed that between 1182 and 1184 AD, Muslim attacks were directed principally towards Transjordan and Galilee, but that Galilee was given priority since it was easier to attack this region from Damascus than it was to attack the castles of Transjordan, because the castles impeded military co-ordination between the two parts of the Ayyubid empire (Prawer 1975, I: 596).This point is fully consistent with the conclusion of this study that the castles of Transjordan played no role in preventing Muslims attacking areas to the west of them, such as Galilee; in fact, what happened is exactly the opposite; it was the very presence of the castles that at times has led to the targeting of Galilee in the first place.
By building their castles, the Franks inserted sites they controlled into Ayyubid territories, thereby creating an obstacle to the movement of the enemy along the Hajj Road, as well as between the north and the south.The military motive behind the construction of the castles on the Hajj Road is confirmed by Muslim sources.Ibn Shadda ¯d reports the reasons for multiple attacks on Karak and Shawbak; he notes that Saladin decided to start with Karak: because it was the nearest to Egypt and because it was an obstacle on the route of anyone travelling to Egypt.No caravan was able to get through unless he went out in person to convey it through the enemy's lands.He wished to widen and improve the road, so that the regions might be in contact one with another, and it make things easier for travellers.(Ibn Shadda ¯d RHCOr III: 53; trans.Richards 2002: 48.) The fact that ʿAqaba was, at times, not under their control was a considerable burden to the Muslims.ʿAqaba was traditionally located at an important crossing point, including, during the medieval period, the Darb al-Hajj and the Darb al-Shiʿwi, leading through the Sinai, as well as the Darb al-Ghazza.In addition, ʿAqaba is well attested as a meeting point for the Muslim pilgrimage (Zayadine 1994: 500-01).The presence of a Frankish castle on Pharaoh's Island (Jazirat Faraʿu ¯n) would, therefore, have been crucial for the Latin Kingdom in controlling the Damascus-Cairo road, in particular the critical point of Naqb al-ʿAqaba, only 10 km away.This would have served the double purpose of protecting Christian pilgrims travelling to Sinai and controlling the roads through the southern deserts, as part of a wider policy of the Latin Kingdom of Jerusalem (Prawer 1975, I: 298;Pringle 2005: 347).
Moving to the concept of 'frontier area', this could be defined as an area on which there is agreement between two parties for the establishment of a de-militarized area, or shared control through a condominium.This situation never existed for the region of Transjordan, or parts of it.This concept may, however, be applied to the area of the Sawad (Terre de Suète) and north of the River Zarqa.As Prawer noted, an effective solution for obtaining security in the Sawad and Jabal ʿAwf was to leave the area devoid of important castles (Prawer 1980: 471-73).The Terre de Suète, referred to by Frankish chronicles as being located on the fringes of the kingdom (Devais 2010: 72-73), was separated from the lordship of Transjordan by the River Zarqa, while the eastern and northern borders were never fixed, and fluctuated throughout the 12th century AD following the changes in territory controlled by the Franks.This area, like Transjordan, was also at its greatest extent around 1160 AD and is well recorded as fertile (Ernoul, Eracles, RHCOcc I: 1105), but was always formally dependent on the Principality of Galilee.While both the Terre de Suète and Transjordan had an offensive role, since both areas contained sites used to launch attacks on the enemy and both, in very general terms, were also 'buffer zones' since they isolated the heart of the kingdom from its enemy, it was Terre de Suète that, throughout the entire Frankish presence, contained Muslim armies entering Frankish territories from Damascus (Devais 2010: 76-77).In the Sawad, conflicts with the Muslims were constant, in contrast to the areas of Karak and Shawbak, where confrontations did not start to become frequent until the late 1160s.In the area of Sawad, the castle of al-ʿAl13 was never reconstructed following destruction by the Muslims and the castle of the Cave de Suète, the main castle of the Sawad, although important regarding control of its area, had much more limited functions than Karak and Shawbak.
Devais also suggests that in both areas the high number of knights indicates at least an intention of seeing these areas as a border (Devais 2010: 78).The list of the baronies owing knight service set out in John of Ibelin's Livre des Assises, dating from the mid 1180s (Edbury 1997: 129-31), reveals that the lordship of 'Crac et Montreal' (Transjordan) had to provide 40 knights for the king and that of Hebron, to which it was linked at the time, a further 20 (John of Ibelin, ed. Edbury 2003: 607-08;cf. Edbury 1997: 196).Considering that the Principality of Galilee owed 60 knights, the 40 knights owed to the king by Transjordan represents a relatively high number, perhaps revealing the importance or wealth of the lordship compared to the rest of the kingdom.However, this number, when compared to that of the Terre de Suète, also 40, does not appear particularly high given that Transjordan was much larger.In practice, the deployment of more military forces in the Terre de Suète might suggest a military 'frontier' function, a function which appears much more appropriate to this region than to Transjordan.Nevertheless, this was only the case until 1118-1119 AD, after which the Franks were able to organize control of Sawad as they wished, while the Muslims, in 1121 AD, lost the fort in Jarash and, therefore, their last fortification in the area (Devais 2010: 75).
The growing involvement, throughout the 12th century AD, of military orders in Transjordan and territories beyond the Jordan River has been interpreted as acknowledging the importance of defending the borders of the Latin Kingdom of Jerusalem, while the royal policy of concentrating military security on the borders has been interpreted as a growing interest in the control of Egypt, especially during the 1160s.This strategy is revealed, for example, by the important concessions made to the Hospitallers in the charter of 1152 AD, and was promoted by King Amaury (Barber 2003: 60-61) who, in 1166 AD, confirmed the Templars in their possession of ʿAmman and its territory, in addition to half of what Philip of Milly owned in the Balqaʿ (Delaville le Roulx 1905Roulx -1908: 183-84, no. : 183-84, no. 2), and who, the same year, ordered 12 Templars who surrendered an important castle beyond the Jordan River, to be hanged (Barber 2003: 72-74).Furthermore, this policy was also behind the choice of Baldwin III in 1161 AD to grant to Philip of Nablus and his heirs the territories of Transjordan: hence Baldwin was able to charge someone very experienced with the defence of an important area on the borders (Barber 2003: 68-71).
It is interesting that the Cave de Suète in the Sawad had significant numbers of Templar knights in its defence.As already observed, 12 is a substantial number when compared to large fortresses, like Safad, which counted around 50 (Devais 2010: 79) and this appears to support Barber's theory that there was a close connection between military orders and crucially important castles at the edges of the territory under Frankish control.In this case, the role of military orders would not have been to extend territorial control, as much as to defend a key area.The relationship between castles and borders does not necessarily mean, however, that their role had to be purely defensive, as will be discussed below.
In summary, while the definition of a frontier area was perhaps suitable to the Sawad, it was not suitable to any area of Transjordan.As Pringle observed, in some parts of the Latin East 'the area of contactor potential contactbetween Christian-and Muslim-held lands was separated by a frontier 'zone' governed by a treaty or truce, which defined the rights of both sides, usually for a fixed period since such truces were not normally permanent arrangements' (Pringle 2012b: 470).An example of this 'condominium' was the area around Banyas and Jacob's Ford, in northern Galilee in the 1170s and 80s, where Muslim and Christian officials divided the rents from the villagers (Pringle 2012b).Such a situation is comparable to the one in the Sawad, where both parties realized that it was necessary to find an agreement.It seems, however, that the Lordship of Transjordan, as far as is known, never reached such an agreement with the Muslims, and always had full control over its own territory and people.The same is true of the Hajj Road, which, although regulated by a kind of agreement with the Muslims, was not, in itself, a frontier.Ellenblum (2007: 275-86), states that in Transjordan the existence of a 'frontier' began in the late 1160s and early 1170s, with the beginning of a constant offensive by Nur al-Dı ¯n and S ̣ala ¯h ̣ al-Dı ¯n, when most of the attacks were aimed at the kingdom's southern and eastern fringes.During this time, these areas of the Latin Kingdom of Jerusalem 'turned into frontier areas par excellence', where life was less secure than in 'the heart of the kingdom' (Ellenblum 2007: 161-64) and the population of Transjordan, especially during the 1180s, lived in an endangered area, comparable to Ascalon in the first 15 years of Frankish rule (Ellenblum 2007: 149).It is certainly true that this is when Transjordan was confronted, for the first time, with almost incessant attacks by the enemy, but the definition of frontier in the traditional sense is not applicable to the castles of the lordship, not only because the Franks did not built them to contain attacks from the enemy, but also because the attacks on Karak and Shawbak were not aimed at conquering the areas west of them.In fact, well before 1187 AD, the Muslims attacked these castles in the hope of conquering Frankish territories west of them by surprise; their strategy, however, was to pull the Franks towards Transjordan, thereby leaving areas to the west undefended.
It should be noted that the timeline of the Muslim attacks shows continuity with the rest of the kingdom.Ellenblum (2007: 149-51) observed that offensive actions by the Muslims can be separated into different stages.According to his model, the first stage (1099-1114) was characterized by frequent military engagements as the Franks established their hold on the country.The second stage (1115-1167) was a time of continuing offensive campaigning against the Muslims, but also one of relative peace and security inside the kingdom, during which a number of castles were built.From the end of the 1160s, confrontations with the Muslims increased.It is worth noting that this was also, more or less, the pattern recorded for Transjordan.
In summary, while castles closer to the Muslim territories were, of course, more exposed to attacks by the enemy than others, it was never the intention of the Franks, when they built these castles, to create a frontier defending Jerusalem and the lands further west, nor did the Muslims see these castles as a frontier that needed weakening in order to penetrate the territories beyond.
Finally, Transjordan as a region was never a frontier in and of itself.Besides the fact that settlement included a variety of settlement types, not just castles, the extent of the territory controlled was always seen as an area with the potential for greater expansion; its limits were never, necessarily, meant to remain static.The northern border with the Terre de Suète as the River Jarmuk, did not become clear until 1129 AD (Mayer 1990: 26) and the entire region was partially royal until 1161 AD.The fact that both the borders and the extent of the lordship were defined through a process that took decades indicates that having fixed borders was never part of the initial plan, and certainly not along the line of castles.
With regard to the military function of castles, Smail points out that castles on the borders had a specific importance because these were more exposed to sudden and frequent attacks, and that their locations were often consciously chosen by the Franks as being places from which they could either carry out offensive attacks, or sally forth and meet the enemy while having the logistical support of a fortress; however, he also points out that this does not correspond to the function of protecting a frontier (Smail 1956: 208).An interesting example, that illustrates the underestimated offensive aspect of castles, is the fortification history of the area around Ascalon.As pointed out by Smail, while historians have mostly interpreted the establishment, between 1136 and 1149 AD, of Frankish fortified posts around Ascalon (Gibelin, Ibelin, Blanche Garde and Gaza) as one of defencedefending the southern borders of the kingdomthey appear to have overlooked William of Tyre's statement (William of Tyre, 14.22, ed. Huygens 1986: 659-60) that the specific purpose of these castles was to contain the raids from Ascalon, as well as enabling attacks against it and serving as centres of Christian colonization (Smail 1956: 213).William of Tyre's statement that the purpose of these fortifications was also to weaken Ascalon has been interpreted by Ellenblum as not reflecting the reality, and as originating from an exaggerated attribution of danger, on behalf of William, to Ascalon (Ellenblum 1996: 528).Archaeology, however, has been crucial in fully supporting the original military function of the castle of Beit Gibelin highlighted by William of Tyre, before its military strength was weakened in a later phase (Pringle 1993: 99-100) and, therefore, in validating William's information (Sinibaldi 2002: 74-75;352); it also offers an example of how, over time, the dominant function of Frankish castles often changed.
As noted above, the years between c. 1140 AD and the end of the 1160s, were a phase of high military confidence and territorial expansion for the lordship of Transjordan, culminating, around 1160 AD, with the occupation of ʿAqaba which was controlled throughout the 1160s, and the widest territorial control of the lordship, when it was completely relinquished from royal control.That renewed, energic efforts would be invested in defending the conquered territories would seem to be a natural consequence.The focus of the Franks' attention on Egypt during this time explains why the Petra area had several castles controlling a region vulnerable to attacks from that quarter, as well as controlling trade on those same roads.The connection between borders and defence was crucial; however, it seems that the connection between castles and expansion was at least as important.The fact that borders were not seen as fixed, but as dynamic and potentially in continuous expansion, is very clear.Although they eventually failed, several attempts were made to expand the lordship of Transjordan south-east and south of ʿAqaba In 1116 AD King Baldwin left Montreal, with nearly 200 knights, in order to see a possible new territory to conquer (Fulcher of Chartres II.56, ed. Hagenmeyer 1913: 594-95).Albert of Aachen reports that on this occasion, because the king was always eager to explore new things, once he had confirmed that the castle of Montreal was built, he proceeded towards Egypt with 60 knights, 'to see if they could do some distinguished deed, such as capturing Saracens or Beduins, or invading towns'.He arrived at the Red Sea, from where he intended to visit the Monastery of St. Catherine on Mount Sinai; however, he was asked not to do so by the monks and eventually decided not to go.Albert notes that it was possible to travel from St. Catherine's to Cairo in less than four days (Albert of Aachen, 12.21, ed. Edgington 2007: 857-59;trans. Edgington 2013: 196).As observed by Prawer (1975, I: 611-12), the Frankish claim that St. Catherine's monastery should be under the authority of the archbishop of Karak, is indicative of their ambitions to control this area.This desire is further supported by an episode reported by Thietmar, writing about his pilgrimage in 1217-1218 AD; according to Thietmar, a lord of Karak had travelled to Mount Sinai where he had attempted to kidnap the body of St. Catherine (Mayer 1990: 214;Pringle 2012a: 126-27).Such ambition continued throughout Frankish presence in the lordship and is reflected, together with the attempt to control the Hajj Road to the south, by Raynald of Châtillon's expedition of 1182-1183 AD on the Red Sea, which tried to reach the Muslim holy cities, and, early on, by Baldwin I, following the foundation of Shawbak Castle, which was considered a suitable base for further exploration and settlement expansion southwards towards the Red Sea and along the Hajj Road.
The conquest of Egypt was always present in the(?) broader Frankish policy, and in the 13th century AD was considered a possible solution in order to recover lost lands.In the 12th century AD, gaining control of Egypt was a core policy of Amalric; there was, at this time, an awareness that this would be much more difficult should Nur ed-Din conquer it first, a move which would have significantly strengthened the position of the Muslims against the Franks.This was considered so important that Amalric made no less than five attempts, between 1163 and 1169 AD, to conquer Egypt (Phillips 1995: 132).
In summary, the royal policy of attaching great importance to the borders of Transjordan was always present; however, it originated mainly from the desire to keep attacks from Egypt under control and the ambition to conquer Egypt, rather than from the goal of defending areas of the kingdom located to the west of it.

Conclusions
Three main conclusions, challenging the traditional assumption that Transjordan's main role was to be a frontier of the kingdom of Jerusalem, have emerged from the evidence presented above.The first, that the military function of the lordship was as much offensive as defensive, as dynamic as static; the second, the close connections between Transjordan and the rest of the kingdom; the third, the great importance of this region for the kingdom.
Concerning the first of these conclusions, part of the region's role was based on the idea of ambitious expansion, in particular during the maturity of the lordship (c.1130-c.1160AD), and specifically on the conquest of Egypt.Realizing this goal would have expanded the control of Christendom and given the opportunity to include under this control an important pilgrimage site, reached through Transjordan and the Hajj Road.The Franks' aim of conquering Egypt was always vigorously and systematically attempted, not only from Palestine, but also from Transjordan.
In support of these conclusions, this paper has shown that the main castles of Transjordan were not very different from those in the rest of the kingdom with regard to being especially, or better, equipped for passive defence, and that there is nothing in their physical structure to justify their definition as 'frontier castles'.There is no evidence, therefore, that the castles of Transjordan were ever created to defend a frontier.
Moreover, it has been proven that the castles of Transjordan had a variety of functions that went beyond the military ones; Karak and Shawbak were strong castles, but were also administrative centres for the collection of agricultural and caravan taxes, centres of authority and important centres of settlement for local Christians.In addition, they were both bases for offensive attacks on the enemy and on caravans, were able to deter passage in their area of influence and functioned as a base for exploration.Smail stated that 'new castles were built in those areas into which it was desired to bring the Latin dominion and in those in which force was particularly required to support the work of administration or exploitation.It is easy to see that there was a military element in such a use of fortified buildings, but it was fused with economic, administrative and social considerations' (Smail 1956: 215).As an example, the castle of Darum, with its foundation, 'extended the limits of the kingdom by bringing a new area under control'.Another significant example of how a castle could shift the balance of power in an area is the castle of al-ʿAl, which was immediately destroyed by the Muslims because it gave control of the entire area of the Terre de Suète west of it, to the Franks.The quick destruction of this castle sent a message similar to the one sent by Baldwin I when he destroyed the Muslim fort at al-Wuʿayra in 1107: allowing the enemy, on either side, to control a strategic point which led to the control of a much broader area was not acceptable (Mayer 1990: 27-30).Such examples indicate that these castles had little to do with the control of a frontier.In fact, the territories under the control of the Franks would be best represented by a series of fortified places rather than a frontier (Smail 1956: 62).In summary, the general definition of the function offered by Smail for Crusader castles fits, very well, the castles of Transjordan.
The Franks' strategy was not dissimilar to that adopted by the Muslims with their foundation of Qalʿat al-Rabad (ʿAjlun), which had the potential to be used as a base for launching attacks, controlling strategic points and disrupting the enemy's communication between two points.The castle was built north of Jarash, outside the borders of the Lordship of Transjordan, just after the Franks had lost control of Habis Jaldak.According to some scholars, Saladin's decision to construct the castle in 1184/85 AD was part of a strategy to challenge Frankish control of Transjordan, which included intensified attacks on Karak Castle.Its position, between Damascus and Karak, the strongest castle in Jordan, also provided surveillance of Belvoir Castle on the opposite side of the River Jordan.In addition, it served to prevent the Franks from raiding the Sawad and, specifically, was closer to the castle of ʿAin al-Habis, clearly an important place strategically (C.N. Johns 1931: 23-24).This strategy was, therefore, similar to the one adopted by the Franks when using Shawbak as a base for further exploring the south and Egypt, using ʿAqaba to threaten and control the enemy's territory, or using Karak to launch attacks to discourage movements of the enemy.Similarly, the construction of the Ayyubid castle on Jazirat Faraʿun, after the Franks were chased from the area in 1170 AD, originated from the intention to control a key strategic point along an important communication road.This intention, to gain control over important communication roads, was the rationale behind the decision to occupy the main Crusader castles of Karak and Shawbak.The advantages of controlling communications and using the castles for administrative purposes were the same for the Ayyubids, who made the best use of the Crusader castles.
The comments, expressed in the 13th century by Jacques de Vitry and Oliver of Paderborn (see above) appear to support Smail's opinion that 'when the Franks were the invaders, the castle was used as an offensive weapon.When they themselves were invaded, the castles were the final refuge of their authority' (Smail 1956: 215).In summary, the military functions of the fortresses of Transjordan evolved over time: during the 12th century, a time of military conquest and confidence, they had mostly an offensive role, as well as one of controlling the local population; in a time of military weakness, their role of passive defence became more prominent, and this is the role that has, most traditionally, been associated with them.
The second point to emerge from the evidence is the close connection between Transjordan and the rest of the kingdom, demonstrating the great importance of this region for the kingdomthe third point of these conclusions.This is reflected in the continuity seen between the two areas on both sides of the Jordan: from an economic perspective, Transjordan, with its significant agricultural resources and trade, was an important source of support for the entire kingdom; from an ideological perspective, since the region shared the same aims of supporting Christianity and Christian pilgrimage; and from a military perspective, because it was a region deeply involved in the attempt to conquer Egypt, an important goal of the kingdom as a whole.Transjordan was also important for institutional and political reasons: this was the most crucial region for maintaining the balance of diplomatic relations with the Muslims, which were essential to the very existence of the kingdom.Awareness of the crucial importance of this region for the kingdom existed from very early on.Donations in the Moab and Edom regions supported by the king show that, as early as 1110-1112 AD, the incorporation of these areas into the kingdom was firmly on the political agenda (Mayer 1990: 33).
The traditional idea of Transjordan as a frontier assumes that the region was both isolated, or, at least, not well integrated with the rest of the kingdom, and also of less importance.The archbishopric of Karak and Hebron, established in 1167 or 1168 AD, is listed last among the archbishoprics, which are organized by seniority (Edbury 1997: 180); however, the archbishop based in Karak was one of only five suffragans to the patriarch of Jerusalem (John of Ibelin, RHCLois I: II ;Edbury 1997: 191).After 1170 AD, Raynald of Châtillon held Transjordan and the lordship of Hebron at the same time; he was the only lord of Transjordan to hold an additional territory that was not part of Transjordan itself.The separation of the two territories is reflected by the fact that Raynald had two separate seals, one for each territory (Mayer 1987: 202;1990: 15).The lordship was, however, formally linked to Hebron.In its short lifetime the lordship achieved not only a wider extension and greater importance, but also greater political independence; this tendency, however, was always contested by the crown, as shown above.The several episodes of insubordination, in 1181, 1183 and 1186 AD, on behalf of Raynald of Châtillon, and the fact that the king was unable to discipline him, indicate that, at that time, the king was unable to fully exercise his power in these territories (Prawer 1975, I: 595-96); this, however, needs to be considered within the more general framework of gradual loss of control, by the king, of territories within the kingdom of Jerusalem.At the same time, it is also clear that the crown always thought of this region as one tightly connected to the rest of the kingdom and one which it was important not to lose control of.The Hajj Road, which had political, military and commercial value, was the core resource to control and played an important role in the economy of the whole Kingdom, although agriculture was the basisas in most of the rest of the Latin Kingdomfor supporting settlements.
All these aspects represent important elements of continuity with regard to settlement choices made in the rest of the kingdom.The choice to settle in intensely Christian areas and make agriculture the basis of the local economy have been highlighted by Ellenblum as general aspects of settlement throughout the whole kingdom (Ellenblum 1998), and this is fully confirmed for Transjordan.The case study of Transjordan would appear to confirm Prawer's opinion that local Christians were never treated equally to the Franks and were not seen as part of the Frankish community (Prawer 1972: 219), but it also confirms Kedar's opinion that they felt relatively at ease among them (Kedar 2006: 210-13, 221).Transjordan also reflects the special relationship that the Franks had with the Armenians, if, as seems to be the case, they attracted an Armenian community to Karak.The presence of the Franks in the region was heavily reliant on a strategy of adaptation to the local environment; a clear pattern of continuity with former periods, as confirmed by the case study of Petra, where elements of continuity and adaptation were far more important than those of disruption.The only element that radically impacted on the society and landscape of Crusader-period Transjordan was the construction of castles, which remained, through the Ayyubid and Mamluk period, important administrative centres along the Hajj Road, and which were enlarged by Muslim rulers, who considered the existing structures and their positions worth reoccupying.
In summary, neither the image of Transjordan as a 'frontier' in the traditional sense, nor that of it being merely a peripheral region can be maintained anymore.The aims and ambitions of settling in Jordan were much broader than just establishing a south-eastern frontier for the Latin Kingdom of Jerusalem.Both the written historical and archaeological sources complement each other in demonstrating this, and the reconstructed picture is very clear.The importance of this region to the kingdom is proven by the investment involved in conquering and trying to control it, by the early interest in settling in the region and by the insistence in attempting to control some areas, such as Petra.Extensive efforts were also invested in the establishment and modification of major institutional agreements, building new trade networks, advancing large financial resources to build castles, and establishing completely new and diversified relationships with each group in the local community in order to obtain their support.It is clear that the Franks were confident that their time and effort would be repaid by economic gains, longterm stability, the maintenance of an agricultural and trade system, a relationship of mutual support with the local Christian population, and a base for further expanding conquered territories and the pilgrimage network.
The common understanding that this was an isolated and peripheral region requires modification.This image of isolation is largely due to the limited amount of research conducted to date.Other contributing factors are the scarcity of historical sources on the subject, the fragmentary nature of the archaeological information, and the traditional concentration on the study of castleswithout doubt the most impressive monuments of this period, but which, until now, have often been, erroneously and unnaturally, seen in isolation from other types of sites, but without which, as is always the case for castles, they could not exist.The role and significance of this region in relation to the kingdom have, as a consequence, been, automatically, mainly associated to the role of its best-known sites, castles, which were seen mainly for their position on the map and their defensive function towards the more important region to their west.It appears that this traditional viewthe association of frontiers and castles, as well as that of Transjordan as a frontier, still persists.
The traditional acquisition, in reference to the Lordship of Crac and Montreal, of the name Outrejourdain, a translation of the name, Terra trans Iordanem, which was given to the area in the 13th century, may, somehow, have contributed to the focus on the geographical location of the lordship beyond the Jordan River, as well as creating an impression of separation from the rest of the kingdom.Nevertheless, the core of the association of this region to the idea of isolation, is probably to be found in late history.The rediscovery of the Middle East in Europe started at the end of the 18th century, with the Napoleonic campaigns in Syria and Egypt, though travel to the region of Transjordan, in particular to Petra, appears to have been considered dangerous and difficult for the whole of the 19th century (Miettunen 2008: 28-30).The publication of Seetzen's explorations in 1806, one of the first western travellers of that time to the region, is opened with his statement that his friends in Damascus tried to dissuade him from his journey to the east of the Jordan and the Dead Sea; they warned him that it was dangerous and impracticable, and that he would find deserts impossible to traverse.The preface of the book declares that Transjordan was one of the few places across the globe to be still completely unknown to Europeans (Seetzen 1810: iii-vi;7-9).Finally, while it is true that Transjordan, in the Crusader period, was relatively far away from centres of power, such as Jerusalem, this distance perhaps appears greater today than it was in reality because of modern political boundaries.

Figure 1
Figure 1 The Lordship of Transjordan in relation to sites in the broader region (copyright: Google Earth).

Figure 2
Figure 2 a: Regions and features of physical geography in the Lordship of Transjordan and the Sawad.b: Location of sites in the Lordship of Transjordan and the Sawad mentioned in the text (copyright: Google Earth).

Figure 3
Figure 3 Location of sites in the Shawbak and the Petra regions mentioned in the text (copyright: Google Earth).

Figure 4
Figure 4 The pilgrimage site of Jabal Harun, with the sanctuary to Aaron at the top of the mountain (photo by M. Sinibaldi).

Figure 6
Figure 6 al-Wuʿayra castle: the main entrance gate, seen from east (photo by M. Sinibaldi).

Figure 8
Figure 8 Jazirat Faraʿun, Egypt: the Ayyubid castle, constructed on top of the former, Crusader-period one, seen from east (photo by M. Sinibaldi).

Figure 9
Figure 9 Al-Habis castle, seen from its lower ward (from east) (photo by M. Sinibaldi).

Figure 10
Figure 10 Khirbat al-Silaʿ, structures part of the castle remains (photo by M. Sinibaldi).

Figure 12
Figure 12 Traces of tooling, indicating specialized workmanship, on building elements of the upper church at Shawbak castle (photo by M. Sinibaldi).

Figure 14
Figure14The Baydha region (at the center of the picture), seen from the Shara Mountains (in the foreground) and with Petra in the background (photo by M. Sinibaldi).

Figure 15
Figure 15 St. Catherine's Monastery in the Sinai, Egypt (photo by M. Sinibaldi).