The colonisation of rainfed land in al-Andalus: an unknown aspect of the eleventh-century economic expansion

ABSTRACT Since the 1980s, there have been significant advancements in the study of the development of rural Andalusi settlement linked to irrigated agriculture, both in relation to large suburban green belts and to the small hydraulic systems of rural communities in mountainous regions. However, recent investigations in several dry areas and, in particular, the intensive archaeological survey that we have been carrying out in the eastern sector of La Mancha, reveal a densely populated rural region that responds to a hitherto unknown pattern. It consists of small villages on the plain, devoid of community defence elements and inhabited by peasants who made the best of the only existing natural resources that have traditionally oriented the region's economy towards cattle ranching and rainfed agriculture. Most of the villages seem to have developed from the beginning of the eleventh century, disappearing at the end of the same century due to the increase in insecurity after the Christian conquest of Toledo in 1085. Their existence would be associated with a general context of population growth that intensified the competition between urban elites and peasants for the privileged agricultural areas and, as a result, the colonisation of less favourable production areas.


Introduction
In this study we deal with a subject that until now has scarcely been present in the historiography: that of the agricultural and cattle exploitation of the dry lands of al-Andalus, that is, the part of the Iberian Peninsula that during the Middle Ages was under Islamic rule. 1 In addition, we demonstrate through archaeological and documentary evidence trees, wheat, and vineyardsmost suitable for the natural conditions of a large part of the Iberian Peninsula.The shift required the construction of hydraulic systems to procure the water that the Mediterranean climate does not naturally provide, which implied intensive use of the land and required a great deal of dedication to cultivation.
It was Barceló's theoretical definition of the rigidity of the irrigation systems and the logic that determines their development which has allowed the study of hydraulic spaces that have survived to the present day as archaeological remains of the society that designed and used them.In this way, the creation of irrigation systems after the Muslim conquest and their relationship with the establishment of a certain social formation and its economic strategies have been studied intensively over the last few decades in different regions of what was once al-Andalus.This work has been based mainly on spatial archaeology and the tools of this discipline: prospecting, toponymy, aerial photography, etc., while it has hardly benefited from the admittedly scanty and scattered written sources.Thus, the idea of Andalusi irrigated agriculture has acquired a hegemonic character in the history of research, which has made it the determining reference point from which to understand other productive options: rain-fed agriculture, livestock farming, and the different ways in which uncultivated areas benefited, all of which have been perceived as subordinate or auxiliary to irrigated production. 9his explains why there is, by contrast, a notable historiographical vacuum regarding the settlement and exploitation of dry lands.Not only is it due to the difficulty involved in analysing medieval productive spaces using spatial archaeology techniques in the absence of physically identifiable elements, such as the hydraulic systems of irrigation, but also, and above all, because it has been assumed that the study of the economic use of dry land would be of less value from a historical point of view for the analysis and characterisation of Andalusi society.This would differ from late antique and medieval societies precisely because of the expansion of irrigated agriculture. 10owever, some recent research projects aimed at studying Andalusi settlement in inland areas of the Iberian Peninsula, where the harsh climate and the geological and hydrographical conditions have traditionally made the development of irrigated agriculture impossible or very limited, are providing information that calls into question what has hitherto been asserted about the history of Andalusi agriculture.This work makes it possible to determine that, contrary to current hypotheses, all these areas that were supposed to be practically unpopulated were, in fact, relatively densely populated and they developed an economy based mainly on rain-fed agriculture and livestock farming.The proliferation of peasant settlements in these regions is evidence of a colonising impulse which, in general terms, began at the beginning of the eleventh century, encouraged by an expansive economic context at a transnational level.In al-Andalus, it ended due to external causes related to the Christian military expansion in the centre of the peninsula at the end of the eleventh century.
Our research project in eastern La Mancha includes an intensive survey of the territory over the past decades.Since 2020, our project also includes the archaeological excavation of one of the most significant sites: the village (qarya) of La Graja. 11 In the present study, we also take into account other recent archaeological projects in geographical areas similar to ours in the centre of the peninsula.The results of these other archaeological projects coincide in general with the results we have obtained, although they provide some interesting nuances.One of those archaeological projects has as its object of study the rural Islamic settlement in the mountain ranges of Gúdar-Javalambre (Teruel), where the existence of numerous hamlets has been detected, mostly located in environments that lack conditions suitable for the development of irrigation systems and whose agriculture was clearly oriented towards the exploitation of dry cereals. 12There are also important analogies with the villages of the Salado valley in Guadalajara, which are being studied with the tools of spatial archaeological analysis, although in this case no excavations have been carried out. 13e territory Eastern La Mancha or La Mancha de Montearagón are generic names for a geographical area located to the east of the southern sub-plateau, bordering the hills that separate it from the eastern part of the Iberian Peninsula and the Segura basin.It is a natural and human landscape that has occasionally coincided more or less with a political-administrative entity such as the Marquisate of Villena in the late Middle Ages, or, in the Andalusi period, with the iqlīm (the castral territory that depended administratively and economically on a madīna, or city) of the madīna of Chinchilla (Figure 1).
The territory has an average altitude of over 750 m, with some areas exceeding 1,000 m.It has a Mediterranean climate with a continental tendency that is manifested, essentially, in low winter temperatures.The absence of rivers and endorheic 14 relief means that much of the rainwater flows into a series of lagoons, some brackish and others freshwater, depending on the geological substratum.These lagoons, as well as the sparse springs and fountains, have conditioned the settlement of the area in all historical periods, as their surroundings have been chosen to establish settlements.All this naturally gives rise to a vegetation cover dominated by holm oak and scrubland (Figure 2).It is not surprising that in this natural environment, agricultural production in the area has traditionally consisted of dry cropscereals, legumes, olives, almonds, and vineyardsand extensive livestock farmingsheep, goats, cattle, and horsesthe latter two associated with draught and tillage work.Irrigated orchards were never viable for various reasons, mainly the absence of fertile plains that could be irrigated with water, and also the harsh winters which prevented the introduction of subtropical crops which, in addition to irrigated cereals, were 11 The project "El poblamiento andalusí en La Mancha oriental (siglos XI-XIII)" is headed by the Instituto de Estudios Albacetenses "Don Juan Manuel," the Spanish National Research Council (CSIC) and the Higueruela City Council.Also participating in its financing are the Junta de Comunidades de Castilla La-Mancha and the Fundación Iberdrola. 12Ortega and Villargordo, "¿Campesinos ricos en al-Ándalus?," 180. 13García-Contreras, "Algunas cuestiones para el debate sobre los asentamientos rurales." 14Endorheic basins are those whose waters do not reach the sea because the rivers flow into inland lagoons or because they are depleted by evaporation, infiltration, or consumption.
specific to Andalusi irrigated agriculture.This was the case at the end of the sixteenth century: There is nothing else in this town [Chinchilla] except farming and raising livestock; there are very few merchants, all live by farming and raising livestock; usually six million bushels of tithe wheat […] sheep and goats are raised in this land and there is a great need for mules in this land because there is so much ploughing work. 15 order to understand the landscape and its ecological potential, data from the statistics on land use in the decade 1980-1990 provided by the Spanish Ministry of Agriculture is very significant. 16It shows that only 0.16% of the region is dedicated to irrigated farming, including areas that make use of water obtained through modern extraction systems.Therefore, the irrigated area must have been even smaller in the Islamic period.In short, these are the physical characteristics of the territory during the Early Modern era and up to the present day, but to what extent can they be extrapolated to the Andalusi period?We have very little evidence to answer this question, although it is worth mentioning that, during the excavation campaigns at La Graja in 2020 and 2021, we carried out the first analyses of the palaeobotanical evidence of the strata located directly on the floors of the rooms.The carpological analyses of seeds show the presence of a series of wild plants that are fairly common in the medieval Mediterranean world, indicating the exploitation of a dry landscape relatively rich in nitrates similar to that of today. 17Among the cultivated species, only grapevines have been found, which would be in line with the hypothesis for the Andalusi period of extensive agriculture based on the Mediterranean triad.The anthracological 18 analysis carried out on the samples show, as preliminary results, a majority presence in all the stratigraphic units of pine wood and leguminous plants, even in the hearth, accompanied by taxa of juniperus (juniper, in Spanish, sabina), monocotyledoneae (monocot plants), quercus ilex and quercus coccifera (holm oak and kermes oak), and rosaceae (rosacea wood). 19This indicates that in general terms the climatic environment of the eleventh century did not differ greatly from that quoted above in the Relaciones of the late sixteenth century.
This territory has been characterised for centuries, from the point of view of human occupation, by a low population density and its concentration in a few small or medium-Figure 2. The site of La Graja and its territory, seen from the northwest.Source: authors. 17Unpublished study in progress by Diego Rivera and Javier Valera, Department of Plant Biology at the University of Murcia. 18Anthracology is concerned with the collection, botanical identification, and conservation of charcoal and wood recovered from archaeological contexts or natural sites. 19Ruiz, "El consumo forestal en el yacimiento de la Graja," 261-68.sized widespread nuclei.Such a low population is in line with harsh physical conditions that only sustain rainfed agricultural and livestock farming activities.For this reason, it has been assumed that these same settlement patterns must have occurred in earlier phases for which there is much less information, such as the Islamic central medieval period of the eighth to thirteenth centuries.This also seems to be supported by the proven Andalusi preference for settlements that would allow the development of hydraulic systems to support irrigated farming areas, conditions that in La Mancha did not exist or are exceptional.However, the Arabic written sources, which are both scarce and circumspect, show sufficient evidence that this stereotype should be subjected to criticism, just as the Arabist María Jesús Rubiera argued, although her assessment has not been sufficiently taken into account: Both this Eastern La Mancha and the rest of the lands known by this name have been labelled as dry and uninhabited since the Muslim period […].In contrast to these hypotheses, we support the opposite: in the Muslim period, La Mancha was not an uninhabited desert, although it undoubtedly did not have large metropolises, which were not the general rule in al-Andalus either. 20biera reached this conclusion after assessing the information provided by the chronicles of two military campaigns that took place more than two centuries apart and in which the Andalusi armies, first Umayyad and then Almohad, crossed these lands for several days without any difficulty in obtaining food and water.
The results of our work confirm the hypothesis put forward by the late Arabist, and even demonstrate that for some time the rural settlement obeyed patterns different from those that have traditionally characterised this region, as we will explain throughout this study.Indeed, the systematic field surveys that we have carried out over the last two decades in this area have shown an intense occupation of the territory made up of villages (ar.qarya, pl.qurā), according to a pattern of settlement and exploitation of the territory in this region that diverges greatly from the traditional pattern for the following reasons: firstly, because it is much more intense and implies an unusual population density; secondly, because of its dispersed nature, based on numerous small inhabited nuclei instead of being concentrated in a few localities.These villages do not correspond to the most common archetype in al-Andalus either, that is, not located on high slopes above the rigid line of hydraulic systems that allowed small intensively cultivated areas to be irrigated.Rather, they were mostly located on the plain and in clearly dry land contexts.The good state of preservation of archaeological remains made it possible in many cases to draw up general plans in which the perimeters of blocks and buildings, and sometimes even the internal distribution of the dwellings, can be identified (Figure 3).This work continued more intensively in 2020, 2021 and 2022 with the excavation campaigns in one of these villages, La Graja, where initial results have provided us with fundamental archaeological information for interpreting the data obtained during the surveys.We have been able to establish the chronology of the foundation and abandonment of the settlement, as well as some particularly relevant aspects of the economy and social organisation of the peasant community that inhabited the village, based on landscape analysis, urban planning, domestic furnishings, and architecture, both residential and religious.The structure of the Andalusi settlement Until the end of the 1990s, there was practically no information about Andalusi settlement in the province of Albacete,21 the southeastern sector of La Mancha.The only data available came almost exclusively from written sources or toponymy, and resulted in a panorama in which the few identified settlements were concentrated around the fertile plains of rivers such as the Júcar, Mundo, and Segura, leaving wide empty spaces in between in which there were only a few localities along certain roads.Yet, this panorama has changed radically because of the systematic surveys we have carried out over the last two decades.
The society that inhabited the territory of what is now the province of Albacete during the Umayyad Emirate period (eighth century to first half of the tenth century) seems to be have been a society in decline and with clearly rural characteristics, structured around the main settlements on high ground, heirs to the elevated sites of late antiquity (Figure 4). 22This model, of which El Tolmo de Minateda would be the finest example, can be detected in other settlements in the mountain ranges or its foothills, such as Peña de San Pedro, El Santo de Alcaraz, La Molata de Letur, Peña Jarota in Nerpio, and Castellar de Meca (Ayora) far to the east.This can also be seen in Chinchilla, in the central plain, and Jorquera, next to the river Júcar, although for these last two cases we do not have conclusive evidence.A few very small hamlets, such as Loma Eugenia (Hellín) or Cerrico de Don Felipe (Montealegre del Castillo), can be found around these larger nuclei, likely suggesting that they depended on them from an administrative point of view.Significantly, most of the settlements mentioned, both the larger and the more modest ones, were abandoned at the end of the Emirate period.
Archaeological surveys show that the settlement patterns of the region underwent intense change around the turn of the eleventh century, with a series of settlements that can be grouped into medium-sized fortified towns (ḥ us ūn), villages (qurā) on high ground or on the plain, and cave shelters (Figure 5).During the Middle Ages, the administrative head of this extensive territory was the town of Chinchilla (madīnat Šantiŷŷāla or Ŷinŷāla),23 called Saltigi in Roman times, which according to al-'Ud  rī,24 was located on the northwestern border of the province (cora) of Tudmīr (see Figure 1).Then, as today, it was located on a rocky spur at the southwestern end of the Montearagón mountain range, in the centre of the present-day province of Albacete, between the Júcar valley and the lands of the Almansa Corridor that allow passage from Vinalopó and La Costera or the highlands of Murcia towards the centre of the Meseta. 25According to al-Idrīsī, Chinchilla was a medium-sized town (madīna), defended by a strong and extensive fortress, which was surrounded by gardens, or almunias, and groves of trees. 26Several figures born in Chinchilla between the second half of the tenth century and the first half of the eleventh were sufficiently important to appear in biographical compilations of the period, such as Ibn Baškuwāl and Ibn al-Farad  ī. 27 Of its medieval urban fabric, we can only identify the general layout of the main spaces today: the citadel located on the highest part of the hill; an albacar, or enclosed sector, although devoid of buildings on the side facing the inhabited area; and the housing sector which extended towards the lower part of the hill, surrounded by a wall that began at the citadel.Inside, the only known remains from the Andalusi period are two bays of a public bath, the hot and warm rooms. 28t the time of the Christian conquest in the mid thirteenth century, several ḥ us ūn (sing.ḥ is n), or medium-sized fortified towns, such as Alpera, Montealegre, Higueruela, Almansa, and Carcelén, depended on Chinchilla.These ḥ us ūn probably emerged and grew in the eleventh century, first due to the demographic boom at that time, and then, as we show below, because of the defensive needs of the inhabitants of the area following the Castilian advance at the end of that century.29These high fortified settlements  remained over time, maybe even experiencing certain development during the twelfth century and until the Christian conquest of the region in the middle of the thirteenth.After this, not all survived and some of them, such as Higueruela and Carcelén, were temporarily depopulated.
The most numerous and significant group of Andalusi settlements in the territory under study are the small villages distributed in the intermediate space between the ḥ us ūn, which we can identify with what the Arabic texts refer to as qurā, as these constituted the basic cell from the administrative and fiscal point of view of rural Andalusi settlement. 30They are only absent in the immediate surroundings of the madīna of Chinchilla and the ḥ is n of Almansa, important localities that generated around them an uninhabited hinterland of some ten to fifteen km, i.e., a two-or three-hour walk, likely because the farmers who worked these lands lived in the urban centre and its suburbs.This group includes La Toyosa and El Boquerón (Chinchilla); La Carrasquilla, El Vallejo de la Casa de la Vega, and El Cerrico de los Conejos (Corral Rubio); La Casa del Sol I and II, La Casa de la Zorra, and Malefatón (Alpera); Hoya Matea (Almansa); Dolonche (Carcelén); El Villar de Hoya Honda, Los Pocicos, and La Graja (Higueruela), the latter being the subject of archaeological excavations since 2020.
We do not have sufficient data to say whether the territory was articulated according to the model described in 1988 by Bazzana, Cressier, and Guichard, in which a series of ḥ us ūn centralised the administration and defence of a rural population distributed in dependent villages; or whether it followed a different pattern, such as that of the Lower Alentejo, where "the dense distributions of village sites found in the survey lack associated fortifications, and there is little evidence of a ḥ is n-qarya pattern." 31In the present case, the villages are, in general, firmly dated to the eleventh century thanks to archaeological evidence, while we have less information about the ḥ us ūn because most of them have remained populated up to the present day.We know that in the eleventh century Almansa and Meca were active, so it is possible that they functioned as administrative and defensive heads at the time when the villages were inhabited, while the enclave of Mompichel may have served as a fortress in case of need for the surrounding villages.For the rest of the ḥ us ūn, we only have information from the time of the Christian conquest in the thirteenth century, so we will have to wait for more data to be able to describe precisely how such settlement was organised.

General characteristics of the villages
Of variable size, some of the villages seem to have had barely a dozen houses, such as La Toyosa, while others must have had forty or fifty, as was the case in Los Villares del Bachiller (Chinchilla); in most cases, the number of houses would oscillate between the two figures (see Figure 3).Even if some villages may have had a tower, most of them lacked defensive elements and not even the dwellings were arranged by way of casa-muro. 32In fact, almost all of the villages were located on the plain, in places protected from the prevailing north-westerly and northerly winds.The few that were situated on high ground were on the slopes of small hills with hardly any natural or artificial defences, apart from the slight slope of the elevation and, occasionally, a summit.This version was usually small and with no signs of occupation, suggesting that they may have served as a refuge, although there is no clear evidence, such as walls, towers, or cisterns, to confirm this.Not even the strategic dominance that could be derived from its high vantage point justifies the location, as in some cases its visual fields are hidden by nearby elevations or are open to spaces without strategic interest.
Most of the villages avoided occupying the most usable farmland and the areas of ponds and lakes.Exceptional are those located next to small meadows, such as Tobillos in Alpera,33 which could exploit irrigated agricultural areas of small size and with no possibility of extension.Many were placed on lands that even today, despite the mechanisation of the countryside, are wastelands, either because the outcropping of the rocky substratum and the scarcity of soil have made them unusable for agriculture, or because they are areas of scrubland, traditionally used for grazing, on the boundaries of large estates that have been used to accumulate the stones from the cultivation terraces.
In general, these villages seem to have had a relatively short life, which is reflected in the rudimentary appearance of their urban planning, characterised by the dispersion of the buildings and the existence of wide and irregular passageways which do not yet have the façade alignments typical of a formed street.In other words, urban planning is characterised by the absence of the phenomena of densification and saturation of the plots and streets typical of advanced phases in the evolution of inhabited places.In some of the villages, long stretches of wall can be detected, similar in width to those of the exterior façades of the houses, which appear to be individual large enclosed spaces which we believe could be communal sheepfolds.
Due to the fact that these villages were based in areas that are not currently cultivated and because during the centuries following their abandonment no population centres had developed in their vicinity to use their ruins as a quarry, these archaeological sites are generally in good condition.Thus it is possible to map them, including, in many cases, the groundplans of the houses and their internal layout (Figure 6; see also Figure 3).
From an urban planning point of view, the settlement layouts demonstrate the absence of intensive planning that could have been imposed by an external authority.The dwellings are laid out according to the preferred orientation and are arranged together in proto-blocks which extend along the contour lines, thus forming built-up strips through which passageways are located.It seems clear that the urban landscape was shaped by decisions reached by consensus among the peasant community, respecting roads and rights-of-way and creating collective spaces, such as the large sheepfolds on the periphery and the central squarewhere the congregational mosque was locatedwhich likely functioned as a temporary marketplace when the inhabitants of the village and the surrounding hamlets gathered for the compulsory collective prayer (Figure 7).However, there are no distinguishable features to suggest social stratification in the village (e.g., landowners vs labourers), although differences in domestic architecture between houses could be argued to reflect various ranks of wealth.
In La Graja, the centre of the village is traversed by two neat, although not totally straight, parallel streets which adapt to the terrain contour lines: the upper one runs through the top of the small elevation and the lower outlines the foot of the hill.Both streets are linked by shorter ones, perpendicular to the slope.In its central section, the upper street widens out to form an open space or square, bounded to the east by the only house with four bays (see Figure 7, building 14).A curved street that runs from the north between what appear to be two large pens enters this square, in the centre of which stands a single-bay mosque (see Figure 7, building 12).This structure presents the canonical features of rural Andalusi oratories: it consists of a simple oblong prayer hall (or ḥ arām), approximately 30 m 2 in size (3×10 m), which could hold up to thirtyfour worshippers.One of the long walls is that of the qibla as it faces southeast, the direction of Mecca towards which believers should prostrate during prayer; in it there is a niche or miḥ rāb, whose interior dimensions are approximately 1.10 m wide × 0.9 m deep (Figure 8).To the west of the miḥ rāb is the doorway that gives access to the mosque from the outside, outlined by jambs made with vertical orthostats, with a span of 1.65 m.In a later phase of construction, the mosque was endowed with a courtyard, approximately 30 m 2 in size, located to the south, facing the qibla wall and the entrance to the prayer hall; a small service annexe to the west side of the prayer room was also built during this phase (Figure 9).This freestanding mosque was probably of the congregational type, in which the heads of household gathered on Fridays for communal prayer.The layout of the surrounding square seems to have been determined by the rights of way generated by the existence of this mosque.Therefore, the urban layout supports the idea that this rural mosque was built at a very early stage in the history of the village and that it was constructed on the initiative of a group of families, one of which (see Figure 7, house 14), probably had a certain pre-eminence over the rest.
The identification and analysis of the mosque of La Graja is important for understanding village planning and the social history of the settlers who colonised this region in the eleventh century.The mosque was built in a central square of the village, next to house 14, which is the most outstanding dwelling in the hamlet from an architectural point of view.The characteristics of this square confirm that this is not a residual area, but that the central layout of the mosque and the surrounding houses was planned from the outset.In accordance with the organisation of the initial nucleus, it can be deduced that the village did not grow by aggregation around a single house, but was the result of the settlement of a group of families, who distributed and divided the space, reserving the central place for the mosque, the construction of which must therefore be attributed to this seminal community.This group of settlers may have numbered ten or twelve families, judging by the  number of houses around the mosque square that seem to form the foundational nucleus.This is consistent with the information provided by the Code of Yūsuf I, according to which the authorisation of a congregational mosque requires at least a community of twelve families, although this source should be handled with caution, as it is much later in date. 34he first settlement was made up of a group of families among whom there were few differences in terms of wealth and social position besides house 14, something which, as far as we know, seems to extend to the other similar villages as well.The need for small groups to colonise these lands, rather than by isolated families, could be due to the fact that the organisation of peasant work in these settlements included the coordination of some agricultural activities.This is comparable to how the expansion of low-input agriculture in northern Europe led to the emergence of villages because peasants shared costly resources such as ox teams and mouldboard ploughs, 35 as well as a communal agreement for cultivating open fields.36It is possible, therefore, that the inhabitants of La Graja coordinated certain agricultural and livestock tasks, although this does not mean that the production system was communal, as there is evidence to the contrary.For example, we know that grain was not stored in communal granaries or in collective silos, but that each house had its own private silo.Likewise, although it is true that there are remains of large enclosures on the periphery of the settlement that may have delimited collective sheepfolds, what we can say with certainty is that a good number of the houses had private sheepfolds, as we will see below.In any case, the organisation of work in this socio-economic model of occupation of the territory is undoubtedly one of the aspects on which future research should focus.

Domestic architecture
Thanks to the well-preserved state of the sites, which in some cases even allow us to draw up groundplans of the dwellings, together with the excavation of La Graja, we know the domestic architecture with an unusual degree of precision (Figure 10).These houses are completely rural, both in terms of their floor plan and the building materials used, mainly masonry and earth, which can be obtained at the foot of the site, as must have been the case with the wooden logs used to support the roofs (Figure 11).Materials that would have required specialised craftsmen to prepare them are absent: no bricks, ashlars, or decorative plasterwork at all, while lime and plaster were hardly used.The fragments of roof tiles found are also scarce, although we believe that this may have been because they were reused after the dwellings were abandoned.
Most of the houses are quadrangular in plan, with a large courtyard occupying twothirds of the total area, onto which between one and four aisles of rectangular rooms open (Figure 12).The preferred side for the built-up area is the north, so that the doors of the rooms that open onto the courtyard face south while the rear of the rooms face north or northwest to protect them from the prevailing winds.
Dimensions of these dwellings are relatively homogeneous: the aisles are between sixteen and twenty m long and three m wide, while the courtyards tend to average around fifteen m to a side, so that the total area is between 200 and 300 m 2 .The  courtyard, therefore, occupied a proportionally very large part of the total surface area of the dwelling, much larger than is usual in Andalusi houses with a central courtyard that are known from urban environments.37This is due to the fact that these open-air spaces in the examples we have been analysing in La Mancha not only served the purposes typical of courtyards, as a place to carry out domestic tasks and a source of ventilation and lighting, but also played an essential role as corrals.Such a use for livestock also explains the width of the access opening to the courtyard from the street, which would be excessive if it served only to facilitate the passage of people and horses.There are also ethnographic arguments to defend this hypothesis, since the traditional farmhouses that were built in the area until the beginning of the twentieth century had courtyards similar to the ones in question, in terms of organisation and proportions, because they were used as temporary shelters for the animals.Perhaps the most decisive evidence in this respect can be derived from a comparison with the known examples of rural Andalusi houses from the same period but located in different productive contexts, such as those inhabited by peasants whose main resources came from irrigated agriculture.Thus, for example, the eleventh-century dwellings of the Foietes village in Finestrat (Alicante) were organised around proportionally smaller open spaces than those we have been analysing.However, the dwellings' architecture is still rural, given that there are no specialised rooms typical of urban houses, such as halls with alcoves or latrines. 38he same characteristics as the houses at Foietes are found at Alcaria Longa in Mértola (Portugal), located next to a fertile plain whose intensive cultivation must have been its economic base in the Andalusi period.39Similar characteristics are also found at Odeleite, a nearby village also linked to irrigated agriculture. 40In contrast, the La Mancha houses resemble those of the ḥ is n and village of Jolopos in the mountains of Granada, a territory in which "irrigated agriculture obviously did not form the basis of the economy here" which was "dominated by stockbreeding and forestry."41Consequently, we hold that in the rural Andalusi architecture of the territory we have been studying, the central space served as a typical courtyard but must also have been used to keep livestock, so it would be appropriate to call it a courtyard-corral.
The only house to have been excavated to date in La Graja consisted of two nuclei articulated around two courtyards (no.15-16), almost equal in size, totalling about 725 m 2 .The house seems to have been built gradually from the northern bay, which was erected before the two courtyards were closed.This house formed a large rectangular block, located south of the mosque and surrounded by four streets.Both courtyards had independent entrances from the outside and were connected to each other through a corridor located in the shared northern corner.Each cluster had four rooms distributed by two L-shaped corridors running along the north and east sides.The organisation and final size of the dwelling suggests that it could have housed more than one family cell, which does not necessarily mean that each of the rooms should be associated with a family, as there is sufficient evidence to establish functional differences between the spaces.This is especially clear in relation to the two rooms in the north corridor of cluster 16, which we can clearly identify as the kitchen (the eastern room) and the main living room with the bedroom (the western room).In fact, although the dimensions of the two nuclei and their organisation may seem very similar at first glance, there are reasons to argue that they were functionally different.The above-noted kitchen and living room are situated in core 16, while some spaces in core 15 appear to have a productive nature, for instance, a segregated corral in the courtyard or a storeroom in the northwest corner.However, there are also traces of domestic use on the short sides of at least four other rooms.This suggests that, at least in its latest phase, this polynuclear dwelling hosted several families that belonged to an extended family group sharing the same domestic unit.
A third of the houses at La Graja, that is, nine of them, also have a second fence of variable size and plan, usually polygonal, always adjacent to the north side of the domestic nucleus.It can be assumed that these pens were used to complement the farmyard to which these spaces opened, for instance sheltering livestock not in the breeding phase, a common practice, even today, to guarantee the survival of the sucklings.No material evidence, however, exists to clarify the uses to which these spaces were put, apart from ethnographic parallels from the area reporting that, traditionally, enclosed areas such as these were used to store livestock and as domestic rubbish dumps.In La Graja, it looks as if the domestic nucleus was built first and the pen was added later, enclosing a space that was previously unoccupied.If our interpretation is correct, this could be understood as a symptom of the growth of each family's herds and the need to expand the stabling spaces beyond the corral-courtyards.
Although few in number, there is also a type of small building with a simple rectangular floor plan, mostly located in peripheral areas or built in isolation from other constructions in the open spaces between the houses.Even if some of them may have had uses other than domestic, most are probably dwellings in their simplest stage, lacking a courtyard, pen, or other bays. 42It seems appropriate to argue that these mono-cellular houses are potential courtyard dwellings, as the evidence suggests that most of the latter began as single-bay houses.The plan of similar sites in in eastern La Mancha confirms that, in general, these simple dwellings are generally located in peripheral areas of the village.In La Graja they appear mainly in the eastern and western sectors and are absent from the centre of the village.In other words, in the final phase of La Graja, mononuclear houses coexist with others equipped with "proto-courtyards," which seem to have gradually developed from the simplest model.Moreover, the urban evolution of La Graja shows that the houses in which the courtyard is already in place, those at the site's core, are earlier than the single-cell houses on the periphery.Therefore, we challenge the hypothesis that the transition from single-nave houses to central courtyard houses was a linear process, directly correlated with the progressive Islamisation of rural society. 43The evolution of the domestic model cannot be regarded as a general process happening simultaneously across the whole of rural al-Andalus; rather, it appears to be linked to the economic and organic growth of individual households and to the urban development of individual settlements.

The economy
Taking into account the characteristics of the productive activities in the region throughout history, the location of the villages next to traditional livestock trails, the characteristics of the courtyard-corrals associated with dwellings, and the presence of what appear to be communal sheepfolds in some of the settlements, we argue that, as in more recent times, the medieval communities must have based their livelihood on rain-fed agriculture and livestock, especially sheep and goats.
One of the main roles of the natural communication routes in these lands has been their traditional use of local and regional transhumance, the essential basis of the area's economy, as indicated in late medieval and early modern texts.Most of the Andalusi villages are located next to these centuries-old cattle routes that cross the region.A large number of villages in the eastern sector of the province sit near the road that later became known as the Cañada de los Serranos, while those in the western area were situated near the branch of the Cañada that leads to the Cañada Real de Andalucía.There is evidence that shows the existence of transterminant and even transhumant pastoralism in al-Andalus.Thus, for example, among the legal rulings of Cadi ʿIyād there is a very interesting one which refers to a group of shepherds who lived in a village (qarya) together with their families and others who must have been farmers.The important point of this proceeding is that these shepherds claimed that they did not participate in the payment of the salary of the imam of the mosque, like the rest of the inhabitants of the census, because they claimed that they only stayed in the village for a few days and the rest of the time they spent outside with their livestock. 44It is not specified how long these shepherds were away from their village, but, given that the imām led the prayer every Friday, the complaint would not be understood if they were absent for just a few days; it could only be sustained if these periods spanned several weeks or even months.
Livestock activity in this region before the Castilian conquest is documented by toponymy; for example, the name of the present-day town of La Roda comes from the Arabic word rutba, which refers to the customs related to wool, while the village of Balazote (Balāt  as -S  ūf) means wool road.Sheep may have supplied the workshops for the production of woollen cloths, which are well documented in the region: in the midtwelfth century, al-Idrīsī singled out Chinchilla, along with Cuenca, as a centre for the production of wool fabrics or blankets (wata' al-s ūf) "which could not be imitated, a circumstance which depended on the quality of the air and water." 45A similar report also appears in the work of the Maghrebi geographer and historian al-Ḥ imyarī, 46 who explains that "some weavings called 'from Chinchilla' are named after this locality, as it is here that they are made." 47Later on, he confirms the narration by referring to Iniesta as a "castle of al-Andalus about two days from Chinchilla, the town where weaving is practiced." 48From these references, it is possible to deduce that Chinchilla is precisely the place near Tudmīr where, in the eleventh century, al-'Ud  rī noted the manufacture of marvellous fabrics and blankets. 49These references suffice to argue for the importance of sheep farming in the economy of the local communities, which has been further confirmed by the discovery of a complete sheep skeleton in La Graja during our 2022 campaign. 50n terms of agriculture, the land surrounding these villages is only suitable, even today, for the rotational cultivation of cereals and leguminous plants, while irrigated land is limited to the bottom of some narrow meadows or ravines protected from the prevailing winds, whose production must have been very limited.Nevertheless, the location of many of them next to cultivable depressions or plains, as well as the presence of silos and the discovery of evidence of the existence of draught animals (a horseshoe and the skeletal remains of bovids) in house 16 at La Graja (see Figure 12), show that dry farming was an economic resource for these peasants, along with cattle breeding.
It seems clear that the mixed economy strategy adopted by the villagers aimed at increasing the community's chances of survival: diversification minimises risks and largely guarantees self-sufficiency in basic foodstuffs.However, the fact that they practiced both agriculture and animal husbandry does not mean that they were not involved in commercial activity.On the contrary, there is enough evidence to argue that they not only produced for self-consumption, but that they seem to have practiced a certain specialisation, namely cereal and, especially, sheep.This afforded them milk and meat surpluses to sell or exchange, as well as wool for the textile workshops of nearby medinas such as Chinchilla.These singular features in the productive strategy chosen by the dry-land peasant communities can be regarded as complementary to those followed by other villages; for example, villages located in irrigated areas also practiced a mixed economy, but emphasising the cultivation of fruits and vegetables, which must have been scarce in the fields of La Mancha.In consequence, rural areas could exchange agricultural surpluses with one another.
Cities, however, were the main markets for agricultural surplus, supplying manufactured goods in exchange.Indeed, the discovery of ceramics decorated in "cuerda seca parcial" 51 and green and manganese, as well as other glazed wares such as basins and pots, reveals that the inhabitants of the village could acquire goods made in urban centres, and that some exchange between them must have occurred.This commercial activity seems to have been rather limited, as can be deduced from the low percentage of glazed wares compared to other sites, 52 yet these goods were certainly brought from elsewhere and, therefore, more expensive than unglazed pottery.At La Graja, glazed pottery accounted for only 4.2% of the total ceramic assemblage, compared to 19% in Foietes de Dalt (eleventh century), 53 and 24% in Las Sillas de Marcén (Huesca), 54 a tenth-to eleventh-century village.Other sites also present a higher proportion of glazed wares than La Graja, but these sites are later in date, so their comparative value is limited: 32% in Algeciras (Almoravid period, 1090-1146); 55 10.6% in Isso (Hellín, Albacete), 56 which dates to the twelfth and thirteenth centuries; 13% in Puça (Petrer, Alicante), 57 also from the twelfth and thirteenth centuries; 29% in Castillo del Río de Aspe; 58 35% in Tossal de l'Almisserà; 59 and 29% in Yakka (Yecla,Murcia). 60To the best of our knowledge, only one site presents a lower proportion of glazed wares than La Graja: the rural settlement of Pajaroncillo in Cuenca (eleventh century), 61 where glazed wares barely account for 2% of the total ceramic assemblage, although this includes eleven fragments of lusterware.
At the opposite end, in the village El Quemao, whose chronology and geographical setting are very similar to those of La Graja, significant quantities of green and manganese ceramics, necklace beads, kohl applicators, and other high-end items have been found, including the fragments of a bone chest decorated with carved epigraphic and zoomorphic motifs.However, La Graja and El Quemao have something in common: both have yielded small weights of the kind used in fairly precise scales.The specimen found in La Graja is a rectangular iron weight with a suspension fitting; its exact weight is uncertain, because the lower section of the piece is lost; the one from El Quemao weighs approximately four ounces.These weights must be related to highvalue goods, most likely gold and silver obtained in mercantile transactions, from which it can be inferred that some peasants of La Graja exchanged their surplus, just like the rich aristocratic landowner of a contemporary estate in Liétor (Albacete), where several scales of this type have also been found. 62e formation and crisis of the peasant settlement Thanks to the pottery found on the surface during the archaeological surveys of the sites we have been analysing, as well as the information obtained during the first three excavation campaigns at La Graja, through stratigraphy, ceramics, and carbon-14 analysis, (Figure 13), we now have reliable data on the chronology and history of this settlement boom in eastern La Mancha.As note above, most of these villages seem to have originated at the beginning of the eleventh century; many of them were abandoned at the end of that same century, and only a few survived until the Christian conquest in the mid-thirteenth century or later.
In La Graja, this chronology is supported by the materials corresponding to the final phase of the site found on the ground and from radiocarbon results, while at the rest of the sites, the absence of later ceramics among those recovered on the surface serves as chronological evidence.The sites that seem to present a pottery record similar to that of La Graja are villages with between ten and twenty dwellings, and most, although not all, are located on the plain, in geographical contexts similar to the one we have been excavating.Two of them are even located in the same valley: Malefatón and Tobillos, both in the municipality of Alpera.In addition, Malas Tardes or Los Villares de Horna (Chinchilla), La Carrasquilla (Corral-Rubio), San Antón (Albacete), Hoya Honda (Higueruela), La Toyosa (Chinchilla), and Los Castillicos or La Rambla (Higueruela), are also all geographically similar to La Graja.It is even possible to include the large settlement of Castellar de Meca in this group.There is also a smaller group of sites which seem to have arisen (or at least experienced a significant growth) at the same time as the aforementioned villages, which survived until the Christian conquest: these are the villages of Villar del Bachiller (Chinchilla); Aguaza; Torre de Bogarra; and, above all, high-altitude settlements such as Mompichel; the castle of San Gregorio (Alpera); the tower of Pechín; and even the Andalusi settlements that preceded the present-day towns of Almansa, Montealegre, and Higueruela.
As for the circumstances of the abandonment of these settlements, in addition to the chronological information provided by the ceramics, there are other interesting 62 Navarro and Robles,Liétor,[65][66][67] indications from the archaeological excavations.The evidence shows that the depopulation the village was relatively calm, as no traces of violent destruction (fires, abandoned household goods) have been found.On the contrary, hardly any complete vessels were left behind and some doorways were even bricked up the hopes of protecting the building, suggesting the intention was to return.63 There is another significant fact: the roof trusses and beams as well as the tiles seem to have been removed to be reused, as can be inferred from the few fragments of tile found and the collapse of the walls towards the interior of the roomsproof that when the ruin of the works took place, the beams,64 which would have generated thrusts towards the exterior, had already been dismantled.This evidence of the reuse of building materials would imply that the inhabitants of the village of La Graja, or at least some of them, did not emigrate to distant parts of al-Andalus, since in that case the cost of transporting relatively low-value materials such as wooden logs and roof tiles would not have compensated.On the contrary, these materials would have moved to nearby sites, so that reuse was profitable.We believe that this phenomenon did not occur in an exceptional way in the excavated village, but that it could have happened throughout the villages of the region that we have surveyed and that show a material record similar to that of La Graja.However, only through future archaeological interventions will it be possible to confirm this hypothesis.
Consequently, the archaeological evidence discussed above, as well as the information derived from the study of the pottery from La Graja and the surveys of the sites in the region, point to a generalised, but not urgent, abandonment towards the end of the eleventh century.This is the same date as the abandonment of the sites in Gúdar-Javalambre (Teruel) and also, approximately, those in the Salado valley (Guadalajara).In the latter case, the phenomenon happened somewhat earlier, although its more nothern location should be borne in mind, as it was conquered by the Castilians as early as the end of the eleventh century.In El Salado it has been possible to verify that the abandonment of some villages was accompanied by the concentration of peasants in high settlements, which should be related to "the Castilian advance and the conquest of the northern part of the valley, around Atienza and its land." 65The union of the dispersed groups of peasants provided them with greater security, even if the sites where they were concentrated did not present better natural conditions for their defence than the previous villages.This could be the reason for the phenomenon that has been recorded by Bertrand and Sánchez for Jolopos (Granada): During the Taifa period, in any case, settlement patterns changed significantly, and hamlets began to merge to form larger, compact villages, such as Jolopos, set below the former emirate-period ḥ is n, Syllar, Diezma, and perhaps Darro. 66milarly, the neighbouring villages of Foietes and Tossal-l'Alfarella (Alicante) apparently coexisted until the end of the eleventh century, which would explain the existence of a rural mosque halfway between the two.Therefore, the abandonment of the former and the permanence of the latter could be due to same phenomenon of population clustering.
All these examples illustrate a historical process that took place at a specific time and in socio-economic contexts similar to those of the eastern sector of La Mancha, where it also seems that the population was eventually able to gather in some settlements on the plain that actually survived, such as Villar del Bachiller, Aguaza, or Torre de Bogarra.We do not know the reasons behind the choice of these places and the depopulation of the others, although it is worth noting that the village of Bachiller, which covered an area of six hectares, was located in the middle of the most fertile lands in the region.In addition to the specific process of concentration in some places on the plain, it seems that what occurred more generally in the territory being studied was the displacement of the population towards settlements better protected by mountains such as San Gregorio de Alpera, Almansa, Caudete, Chinchilla, or Higueruela.Each of these sites seems to have arisen in the eleventh century or, at least, to have undergone considerable demographic development at that time, perhaps coinciding first with the extensive colonisation of this region at the beginning of the eleventh century, and later with the general abandonment of the villages on the plain at the end of the century.It is likely that due to this growth, some of these settlements went from being villages (qurā) to ḥ us ūnthat is to say, nuclei of a higher category from an administrative and demographic point of view.
The reason for this concentration of settlement in the best-defended places can only be the increase of insecurity.It seems logical to suppose that the absence of the minimum defensive conditions of those enclaves on the plain, at a time when the conquest of Toledo by the Christians in 1085 created a serious situation of instability south of the Tagus.The raids by Alfonso VI of León-Castile (r.1065/72-1109) and by El Cid in the southeast, and the establishment of a real Leonese-Castilian bridgehead in Aledo (1086-1092), must have pushed the population of the area to concentrate in sites whose characteristics made them more suitable for protection.The fact that this phenomenon did not occur in those regions of al-Andalus that were not exposed to the Christian advance at the end of the eleventh century is indirect proof that the reason for this transformation of the settlement pattern was due to external causes and not to dynamics related to the evolution of Andalusi society itself.Similarly, this was the case with the Balearic Islands where, according to Kirchner: The ceramics found in archaeological sites indicate that there was an abandonment at the time of the feudal conquest at the beginning of the thirteenth century.With a few exceptions, the lack of abandonment prior to this date suggests that there was not a high percentage of failed initial settlements. 67 is worth remembering that, like the ones we are dealing with here, most of these island villages date back to the early eleventh century and, in some cases on Ibiza, to the tenth century.Therefore, under similar socio-economic conditions, the factor that would explain the abandonment of the villages of Guadalajara, Teruel, and La Mancha, which did not occur in the Balearic Islands, would be the increase in Christian pressure on the lands of the Peninsula. 68owever, the record attests to the abandonment, also in the late eleventh or early twelfth century, of some settlements that met good defensive conditions, beginning with their position on hilltops.One of them is Los Castillicos (Higueruela), which was similar in size to La Graja.The village was located on summit of a high hill, at the centre of a mountainous area and in relative isolation from valleys and plains.Another is Castellar de Meca (Ayora), which occupies the site of an ancient Iberian oppidum, the most important population centre in the region during the eleventh century, before being abandoned during the first half of the twelfth.It is clear that none of these sites was abandoned for defensive reasons, so we must take other factors into account for the abandonment of villages in twelfth-and thirteenth-century Europe, including the inability of the settlers to procure sufficient agricultural land, a phenomenon known in Central Europe as Fehldsiedlundgen. 69According to Pierre Toubert, the regional studies carried out in Germany and Italy, and to a lesser extent in the south of France and Spain, far from challenging this reason for abandonment, support it while contributing to better defining its characteristics.Objectively, the two sites may have experienced this phenomenon: the situation of the village of Los Castillicos must not have allowed for a minimum of agricultural development; this remains the case even today despite the use of modern technology.As for Castellar de Meca (Ayora), its location was not the most favourable from an agricultural point of view, since it was partly surrounded by marshes and had only limited access to water resources, unlike other towns in the area, which succeeded and entered a lasting period of growth, such as Almansa, Alpera, and Ayora, located in the centre of relatively fertile plains.

The historical significance of the colonisation of the dry lands
La Graja, alongside other sites in Teruel, Guadalajara, and occasional finds in central La Mancha, are beginning to prove that the extensive dry lands of the interior of the Iberian Peninsula were not a desert, as argued in the literature to date.On the contrary, these regions were intensively populated and exploited by peasant communities living in villages (alquerías), whenever the economic circumstances were favourable and the volatile political situation allowed.This situation basically ended with the relentless Christian advance in the final quarter of the eleventh century, which culminated with the conquest of Toledo in 1086.Irrigation was the preferred agricultural strategy for Andalusi peasants, with the colonisation and exploitation of rainfed lands only a second-best option that was adopted in this region under very specific historical and economic conditions, as rainfed farming entailed greater risks and lower profits than irrigated agriculture.Despite technical advances and agronomic innovations, Andalusi agriculture was 68 In none of these cases was the economy based on irrigated terrace cultivation, so it is unlikely that the settlement crisis was due to soil degradation and erosion through overexploitation, as has been proposed for the Baixo Alentejo by Boone and Worman, "Rural Settlement and Soil Erosion." 69"Une autre cause majeure de désertions rurales, à côté des crises agraires et des surconcentrations des XIIe-XIIIe siècles, a été désignée par les chercheurs allemands du nom Fehlsiedlungen.Indépendamment de ces recherches, j'avais eu moi-même, dans le Latium, l'occasion de souligner l'importance de cette cause de désertion par ce que je caractériserais comme une incapacité avérée d'habitats castraux nouvellement fondés (Xe-XIIe siècles) à structurer autour d'eux une assiette suffisante de terroirs agraires," Toubert, "Histoire de l'occupation du sol," 35.
exposed to crises due to weather-dependent bad harvests.The same was true for livestock farming, which in pre-industrial period always suffered high mortality rates due to the constant threat of epidemics.Consequently, the colonisation of poor lands that did not allow for the development of the usual productive strategies in the Andalusi rural environment cannot be explained by the general logic associated with natural growth and the consequent segmentation of family groups.However, it is necessary to examine the economic and social context in which this process took place.
Between the second half of the tenth century and the mid-fourteenth century (that is, until the crisis caused by the Black Death epidemics), a demographic and agricultural development took place in western Europe known in historiography as "the great expansion," or the "economic revolution," 70 because it was also accompanied by a notable increase in mercantile exchange and market production.This process, which dates to approximately the ninth century in the Latin West, was characterised from the tenth century onwards by the concentration of the peasant population according to the requirements of defence and social control by the lords who promoted the occupation of the land and the development of space. 71he Islamic territory of the Iberian Peninsula presented notable organisational differences in relation to the societies of the Christian kingdoms.However, al-Andalus seems to have been at the forefront of these transformations during the tenth and eleventh centuries, although Andalusi economic progress would later be cut short by the conquering advances of the Christian kingdoms.While no documents have been preserved that allow us to calculate the population of al-Andalus, there is archaeological evidence of demographic growth, mainly due to the development of population centres, as well as the productive boom that Lucie Bolens called the "Andalusi agricultural revolution of the eleventh century." 72According to Bolens, this important development was motivated by the combination of Hispanic tradition, Islamic knowledge inherited from antiquity, and a historical moment in which princes, governors, and peasants were interested in producing more, driven by demand in the cities and the necessary labour force in the countryside.All this led to the rationalisation of the management of numerous rural properties that belonged to the state, alleviating the centralised taxation of the Umayyad caliphate, which had disappeared exactly at the beginning of the eleventh century and was replaced, after the civil war, or fitna, by small regional states known as the taifa kingdoms.
Guichard has already pointed out that the few textual references we have about the agrarian milieu in the eleventh century force us to revise Bolens's functionalist interpretation. 73Ibn Ḥ ayyān, 74 in particular, explains in relation to Valencia, that the two Slavic co-princes, Mubārak and Muzaffar, 75 levied heavy taxes (ḥ arāğ) which they "collected with the utmost rigour from all categories of the population, to the point that the situation of their subjects deteriorated.People emigrated one after the other from the regions they occupied, which, in the end, were ruined." 76This unbearable fiscal pressure forced the peasants to leave their places origin: "the people could not cope with it except by emigrating from their homes and abandoning their villages [qurā]." 77The impossibility of coping with tax burdens and debts led to the expropriation of farms, which passed into the hands of the landowning elites, who transformed them into private domains or villages and put them under cultivation through labourers, sharecroppers, or tenants who were sometimes the former owners of those same lands.According to Ibn Ḥ ayyān, this not only happened in the region of Valencia, but: It was what most of the rebels who took over the regions of Andalus or revolted in its confines did, after the disintegration of the power of the Community [sultān al-ŷamā'a] in Córdoba at the end of the dynasty of the Banū-Āmir. 78n of religion would have the obligation to oppose the state's fiscal pressure, disallowing illegal taxes in accordance with Islam.Thus they defended the legitimate owners of the land as recognised by some ulema who on one occasion, according to al-T  urt ūšī, 79 accused the viziers: You are the ones who eat up the estates of the people, without profit to them, and consider it lawful to oppress them without reason; who deprive them, little by little, of their means of livelihood, taking it from them as the price of your bribery and your prevarication; who want the land without right. 80wever, the political power had sufficient resources to make the faqīh (pl.fuqahā), or Islamic jurist, accommodate their sentences and rulings according to the whims of the sovereign and his cronies, as the protagonists of the aforementioned story admitted.
The pressure experienced by peasants from both the state and urban elites in the form of taxes and land alienation would lead them to work as labourers; to swell the urban proletariat; and even, uprooted from their family and neighbourhood environment, they could fall into begging; banditry; or resistance and open rebellion, as happened in the Christian societies of the north.There is documentary evidence of the rise of banditry and robbery at this time which hindered the free movement of people. 81Insecurity also extended to the interior of the cities: Córdoba itself seems to have been notorious in the twelfth century for the number of robberies and murders which took place there, while the population was said to be not only unscrupulous but also perpetually 76 "Los recaudaban con el mayor rigor de todas las categorías de la población, hasta el punto que la situación de sus súbditos se degradó.Las gentes emigraron unas tras otras de las regiones que ocupaban, las cuales, a fin de cuentas, se arruinaron," Guichard, "Crecimiento urbano y sociedad rural," 156. 77"La situación llegó a ser insostenible, hasta tal extremo que, a menudo, la gente no pudo hacerle frente sino emigrando de sus casas y abandonando sus pueblos (qurā)," Guichard, "Crecimiento urbano y sociedad rural," 160. 78"eso fue lo que hicieron la mayor parte de los rebeldes que se adueñaron de las regiones del Andalus o se sublevaron en sus confines, tras la disgregación del poder de la Comunidad [sultān al-ŷamā'a] en Córdoba al final de la dinastía de los Banū-Āmir," Guichard, "Crecimiento urbano y sociedad rural," 160.The Banū-Āmir or 'āmirīes were the descendants and clients of the powerful Almanzor, hayib or chamberlain of the caliph Hišām II and de facto ruler of al-Andalus from 976 until his death in 1002. 79Mālikid jurist, ascetic, and essayist (Tortosa, 1059 -Alexandria, 1126).His most famous work is the Sirāj al-Mulūk (Lamp of the Princes), a political text intended to serve as a personality guide for princes, including their relationship with both God and their subjects. 80"Vosotros sois los que os coméis las haciendas de las gentes, sin provecho para ellos, y consideráis lícito el oprimirlos sin razón; los que priváis, poco a poco, de sus medios de vida, arrebatándoselos como precio de vuestro soborno y de vuestra prevaricación; los que queréis la tierra sin derecho," Al-T  urt  ūšī, Lámpara de los príncipes, 1:111. 81Ibn 'Iyād  , Mad̲ āhib al-h  ukkām, 70, 132, 208-09, 487.complaining against authority. 82Although there is no record of peasant rebellions in al-Andalus analogous to those that took place in the Christian kingdoms, it should noted that these conflicts were not always resolved by open revolts but could take the form of legal claims and procedures.Even "secondary" forms of resistance could take place such as non-cooperation and flight, 83 thus formulating their demands under the shelter of widely accepted concepts including the rejection of extra-canonical taxes in Islamic society.
Thus 'Abd Allāh (r.1075-1090), the king of the taifa of Granada, complained of the "senseless hatred of his vassals [the local rulers of al-Andalus] and their reluctance to pay the magarim a-aqta' (extra-Quranic taxes) to which they obliged them." 84For the same reason, Alfonso VI and El Cid each in turn promised the Muslims whom they claimed in vassalage that "they would not do to them as their 'Moorish' kings did, who took more from their subjects than was legally their due." 85Such a tactic was used precisely to win popular support, and the slogan of the contemporary Almoravid movement was "to spread the truth, to repress injustice, to abolish illegal taxes." 86he urban classes interested in the source of wealth that mercantile agriculture offered, in the context of the expansive period that we have been examining, not only directed their efforts towards the acquisition or appropriation of the lands of peasants who were overwhelmed or directly ruined by fiscal pressure, but also used other means to obtain the benefits derived from market agriculture.These benefits included the creation of new cultivated spaces in previously barren areas, in other words, the vitalisation of land, both irrigated and unirrigated.One of the symptoms of the growing interest by urban elites in the benefits of agriculture was the proliferation in the eleventh century of geoponic treatises written by Andalusi authors, including Ibn Abī l-Ŷawād, Ibn ʽĀs im, Ibn Wāfid, Ibn Haŷŷāŷ, Ibn Bas s āl, Abū l-Jayr al-Išbīlī, and al-T  ignarī, on methods and techniques of cultivation and agricultural production. 87The treatises covered topics such as fertilisers and plants, work processes and agricultural techniques, water and hydraulics, and even the preservation of products.The zones preferred by the urban elites would be those that could produce the highest yields, i.e., irrigated land, as is demonstrated by the attention given to this activity in Andalusi agricultural books.The conquest of new irrigated areas by landowners was made possible thanks to economic resources that allowed them to afford the initial investment of work required for the construction of large-scale hydraulic infrastructure.This meant that landowners could take advantage of both the irregular contributions of wadis (dams and reservoirs) and the constant flow of rivers through the digging of irrigation ditches.Land that could only be irrigated by rainfall aroused less interest because it had a significantly lower value than that which was irrigated or was susceptible to artificial irrigation.However, in this general context of agricultural expansion that led to competition for privileged 82 Urvoy, Pensers d'al-Andalus, 72. 83Freedman, "La resistencia campesina." 84"el insensato odio que les tenían sus vasallos [a los sultanes de al-Andalus] y la resistencia que éstos mostraban a pagar los magarim a-aqta' a que les obligaban," Guichard and Soravia, Los reinos de taifas, 135. 85Guichard and Soravia, Los reinos de taifas, 136. 86"propagar la verdad, reprimir la injusticia, abolir los impuestos ilegales," Molina, "Economía, propiedad," 245. 87Jiménez and Camarero, "Los tratados de agricultura." productive spaces, less favourable areas were also developed by urban elites, as can be seen in the almunias (Arabic, munya), or agricultural exploitations with an urban of Šarq al-Andalus 88 and Lleida.Despite their aridity, the dry lands were populated and exploited in the Andalusi period, according to both Arabic and Latin sources.Spaces destined for pasture for cattle predominated and were used from the eleventh century.
Because of the process of the urban elites' pressure on peasant land ownership, demographic shifts took place, yet have hardly left any trace in the written sources.One exception is a mention in Ibn Ḥ ayyān's text that nobles "appropriated the villages whose people had emigrated in order to make them into private farms." 89An obvious destination of the dispossessed peasants were the cities, indicating that this emigration would be one of the causes of urban development during the eleventh century.Muslim authors record some evidence of rural-urban migration in the context of the commercialisation process, as in an eloquent fable told by Ibn al-Jat īb 90 which was likely set in Murcia in the mid-twelfth century.He tells the story of a subject of Ibn Mardanīš, 91 who had in the vicinity of Játiva: A small estate on which he lived, but the taxes exceeded his earnings and he fled to Murcia, although Ibn Mardanīš had established that whoever fled before the enemy, his goods would be confiscated for the treasury.The man from Játiva recounted: When I arrived in Murcia, having fled from my homeland, I took up construction work … 92 The text goes on to relate the tribulations of the character in the city, pursued by all kinds of tax collectors, although for us, the beginning of the story is enough, since it was told precisely for an audience who would be familiar with the situationhow a peasant could lose his land because of tax pressure and be forced to emigrate to the city to become a low-wage worker.Moving to the cities was a risky option because it meant abandoning the relative security provided by the countryside for the vagaries of working for others.In pre-industrial societies, pottery making, and other crafts were a secondary choice for people with insufficient or poor quality land, or no land at all.While agriculture provided food directly for a family, craft production did not, and instead required additional work and higher risks. 93In fact, according to the above fable, the decision to emigrate did not improve the protagonist's quality of life; on the contrary, it entailed a series of hardships until he was imprisoned and subjected to forced labour.Therefore, in a historical context of expansion of commercial agriculture and pressure on the peasant class such as the one in question, it seems logical to expect not only the growth of cities, but also processes of colonisation of new lands. 9488 Navarro and Trillo, Almunias; Brufal, "La Lleida de secano," 261. 89Guichard, "Crecimiento urbano y sociedad rural," 160. 90Ibn al-Jat īb was a famous writer, poet, historian, physician, and politician from Granada.He was born in Loja in 1313 and died in Fez (Morocco) in 1374.He belonged to a wealthy family and was appointed vizier by Sultan Muh  ammad V in 1348. 91Ibn Mardanīš, the so-called Wolf King of the Christians (Peñíscola, 1124 -Murcia, 1172) was a political and military leader that ruled the taifa of Murcia and Valencia, in eastern al-Andalus, as king 1147-1172.During the second taifa period, he opposed the invasion of the Almohads, who did not manage to fully conquer his domains until his death. 92"una pequeña finca de la que vivía, pero los impuestos superaron sus ganancias y huyó a Murcia, aunque Ibn Mardanīš tenía establecido que quien huyese ante el enemigo, se le confiscarían los bienes para el tesoro.El hombre de Játiva contaba: cuando llegué a Murcia, huído de mi patria, me coloqué en la construcción … " Epalza and Rubiera, "La sofra (sujra)," 34. 93Arnold, Ceramic Theory, 193. 94Hilton, Bond Men.

Conclusion
For centuries, and even today, the territory of La Mancha has been characterised by low population densities and concentrated settlement in a few medium-sized small centres, which stand fairly distant from one another.This is a reflection of the region's harsh physical conditions, which only allow for the practice of extensive agriculture and stock-keeping, undoubtedly contributing to the assumption that such a settlement pattern also applied to the Islamic period.This is further supported by the proven Andalusian preference for areas suited to the construction of hydraulic systems and irrigated agriculture, which in La Mancha are very rare.However, our study of Andalusi settlement patterns, through the systematic survey of the southeastern sector of the region as well as the archaeological excavation of the qarya of La Graja, has shown that the area was intensively colonised by peasant communities during the eleventh century.The settlement pattern involved a scattering of numerous villages and hamlets in relative proximity to each other, a very different picture from that presented by the traditional concentrated model that has dominated the region in other periods.The villages differ from typical Andalusi models, as they lacked defensive features and were not located on high slopes, presiding over irrigation channel networks (acequias); rather, they were mostly to be found on the plain and next to rainfed agricultural fields.
The main economic activities of the peasants who inhabited these villages were rainfed agriculture and stock-keeping.The former is proven by the existence of underground grain silos inside the houses, while stock-keeping is amply demonstrated by the architecture of the houses, which include courtyards-pens, by the presence of large sheepfolds, and by the zooarchaeological assemblage.This mixed economic strategy clearly aimed to increase survival chances, as diversification reduces risks.In addition, they not only produced for self-sufficiency, but also specialised in the exploitation of sheep, which afforded them milk, meat and, especially, wool surpluses for the textile workshops of nearby medinas such as Chinchilla.These marketable products, and perhaps also some domestic craft activity, allowed the inhabitants to acquire consumer goods manufactured and sold in the urban markets, as also proven by the archaeological evidence.
No sites from the Roman or late antique periods have been found in the hinterland of La Graja, that is, the territory under the exploitation and control of the village's inhabitants.Therefore, it seems certain that when the first Muslims settled here in the early eleventh century, they were forced to break up what may have always been barren land.The clearing of wild vegetation and stones and the first furrows must have meant a substantial investment of labour.If we add to this the need to build the basic domestic buildings, the task appears to be too large for a single nuclear family.It is, therefore, more plausible to think of a small group of families sharing the burden.This hypothesis is also supported by the analysis of the urban evolution of these villages, whose origin seems to have been the arrival of between six and ten families, which were to become the nucleus of each hamlet.We do not know if these pioneers were bonded by tribal links or by some other type of looser neighbourly ties, for the sources that could be of assistance in this regard (such as toponymy) have disappeared; together with other cues mentioned above, there are indications, such as the absence of collective granaries and the proliferation of individual domestic silos, which suggest that these kin cells were fairly autonomous.
The colonisation of the rainfed lands of southeastern La Mancha was not a local phenomenon.Other research projects undertaken in regions with similar conditionsthe Javalambre-Gúdar mountain in Teruel, La Mancha central, or the Salado valley in Guadalajaraare demonstrating the existence of contemporary settlement patterns that greatly resemble those described in the present paper.Current research thus leads us to conclude that the absence of Andalusi settlements in the dry lands is a fact more historiographical than historical.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by Consejo Superior de Investigaciones Científicas.

Notes on contributors
Pedro Jiménez Castillo holds a degree in Ancient History and Archaeology from the University of Murcia and a PhD from the University of Granada, obtaining the extraordinary doctoral award in 2017.Since 2006, he has been a member of the School of Arabic Studies (CSIC, Granada).A specialist in Islamic archaeology, he has directed more than forty archaeological projects in Spain and Jordan, where he headed the Spanish Archaeological Mission in the Umayyad palace of Amman between 1997 and 2001.He has participated in other international missions in Italy, Algeria, and Morocco.His extensive research publications include Islamic and medieval history, art and archaeology.
José Luis Simón García holds a PhD in History from the University of Alicante.He is currently a member of the Institute of Albacete Studies and a technician for the Generalitat Valenciana, which has led him to direct research projects and work on the conservation of Islamic fortresses and sites such as the Rábita de Guardamar (2021).He has been doing archaeological research in Albacete for forty years, first in the field of Prehistory and Protohistory and then in medieval archaeology, both in the Islamic and Christian periods, including castellology and settlement studies.He is the author of many articles and books on Islamic and medieval sites.
José María Moreno Narganes holds a degree in History from the University of Alicante and an international Master's degree in Medieval History and Archaeology from the University of Lyon Lumière 2 (France).He has participated in archaeology projects in Spain, Morocco, Italy, Portugal, and Jordan.Currently a doctoral researcher at the University of Alicante, his thesis investigates domestic spaces and textile production in al-Andalus, with publications on working tools, ethnography, and gender studies.He participates in research projects on Islamic Almeria and codirects archaeological excavations in the villages of Puça (Petrer), Torre de Haches, and La Graja.

Figure 4 .
Figure 4. Map of the emirate-and caliphate-period settlement, modern province of Albacete and the neighbouring regions.See also the central research area framed in Figure 5. Source: authors.

Figure 5 .
Figure 5. Map of the territory of Chinchilla showing the evolution of Islamic settlement from the eleventh century (above) to the thirteenth century (below).Source: authors.

Figure 6 .
Figure 6.Orthophoto and plan of the La Graja site.Source: authors.

Figure 7 .
Figure 7. Plan of the Andalusi site of La Graja.Source: authors.

Figure 9 .
Figure 9.View of La Graja mosque from the northeast.Source: authors.

Figure 10 .
Figure10.The site of La Graja and its territory, seen from the northeast.Source: authors.

Figure 11 .
Figure 11.General view of house 15-16 from the northeast.Source: authors.

Figure 13 .
Figure 13.Typological table of the pottery found in the archaeological excavations at La Graja.Source: authors.