Early medieval household archaeology in Northwest Iberia ( 6 th-11 th centuries )

Carlos Tejerizo* Universidad del País Vasco

Country, Asturias and Galicia has been done, taking into account more than sixty sites.This paper will be structured in four main parts: in the first one the theoretical background that frames this study will be analysed; the second will be an overview of the sampled sites; the third will encompass the study of domestic architecture; and lastly, the existence of «constructive forms» will be proposed, together with some global considerations regarding the archaeology of domestic units.

THE THEORETICAL FRAMEWORK
Domestic architecture is a very useful material record, as it allows both social and economic analyses on human societies as well as on the relationships of power that frame them (Foucault, 1981;Lull, 2005).However, the analytical potential of domestic architecture is directly linked to household archaeology.This type of studies, which have a long history in other archaeological traditions (Bender, 1967;Deetz, 1982), have hardly been applied to early medieval peninsular societies (Gutiérrez Lloret and Cañavate Castejón, 2010;Vigil-Escalera, 2007b).
Following Fernández-Posse and Sánchez Palencia, a domestic unit can be described as the «elemento articulador de la ordenación y desarrollo del espacio construido, el núcleo básico organizador de la producción y el mecanismo de regulación y control del comportamiento y relaciones sociales de la comunidad» (Fernández-Posse and Sánchez-Palencia, 1998: 130).The domestic unit is in this sense the link between the individual/family unit and the wider socio-economic context (Deetz, 1982).
According to Wilk and Rathje (1982), the domestic unit as an anthropological concept, is based on three elements: a demographic unit (including the number of individuals and their relationships); a material unit (domestic architecture proper, activity areas as well as other properties); and the behavioral unit as relating «to the activities it performs».Similarly, they propose that domestic units in pre-capitalist societies fulfilled four functions: Production, distribution, transmission and reproduction (Stanish, 1989;Wilk and Rathje, 1982), all of which happened in the material unit, that is to say, the domestic architecture.
As the authors mention, archaeologists excavate dwellings and domestic artifacts, which are the material evidence.But the analytical process should not stop there, but rather «we have to infer dwelling units from the material record; then we must infer households from the dwelling units» (Wilk and Rathje, 1982: 618).This epistemological leap from the object to the social relationships which they reflect is nothing but the necessary interpretation of the remains in social terms («social archaeology») beyond a simple artifact or monumental archaeology (Deetz, 1982;Lull, 2005;Mcguire, 2002;Sanahuja Yll, 1988).
As Amos Rapoport mentions, «folk architecture is the direct and unconscious translation of the physical forms of a culture, of its needs and values, and of the wishes, dreams and passions of a people»3 (Rapoport, 1972).Various authors (for instance, Appadurai, 1986;Ayán Vila, 2003;González Ruibal, 2003) warn us about the need to avoid normativist of functionalist points of view, and to value the importance of ideology when it comes to forge a materiality (Demarrais, et al., 1996;Mcguire and Bernbeck, 2011).If materiality is not only a performative frame of production, but also of social reproduction, domestic architecture should reflect both ideology and the relations of power directly linked to it.This should not be taken as an aspect of ethnicity in a determinist way, as has been done (Fronza, 2009;Valenti, 2009), but should be understood rather from a materialist perspective, with different ways of confronting the organization of production and the power relations between different social groups and genders (Quirós Castillo, 2011a;Tejerizo García, 2011).
The domestic unit as a structural concept must be introduced in a given socio-economic context which sets a frame in which it can potentially exist.It is in this way that we have to consider the scale in which the productive and cultural processes developed (Escalona Monge, 2010;Wickham, 2008).Both the frame and the scale were deeply transformed in the post-Imperial period, linked to different dynamics which vary in time and space.It is because of this that statements which equate materiality with singlecause and homogeneous processes should be avoided.Other interpretations based on production processes and their related social relationships should be sought for.
Even if there are some publications which propose these types of studies, considering domestic units and socio-economic characterisations as a whole, in early medieval archaeology of the centre of the Peninsula, Catalonia and the Basque Country (Quirós Castillo, 2011c;Roig Buxó, 2009;Vigil-Escalera, 2007b), there is as yet none overall analysis of the north-western corner of the Peninsula focused on the local scale of social production organisation.The aim of this paper is to be a first step in this kind of analyses.

THE ARCHAEOLOGICAL EVIDENCE: THE SAMPLE
In this paper a gazetteer of 64 sites from the north-west corner of the Peninsula with remains of domestic architecture has been compiled (fig.2).
The close relationship existing in Spain between Commercial Archaeology and the study of early medieval rural communities has already been proposed (Quirós Castillo, 2011a;Quirós Castillo and Vigil-Escalera, 2006), as the so-called «silent revolution» (Quirós Castillo and Bengoetxea Rementeria, 2010).However, the use of these data in general historical discourses has only been possible due to a paradigm shift that has entailed critical analyses of the material culture and previous historical studies.This change of paradigm created a new frame that allowed new historical discourses (Foucault, 1978).In other words, until this record was acknowledged it was not possible to identify it archaeologically and to include its results in scholarly discussions.
The first examples of early medieval domestic architecture in northern Spain were identified in the 1950s and 1960s, although these discoveries were limited to a few examples of well-identifiable stone constructions (Quirós Castillo, 2011a).However, negative stratigraphic units, which sum a significant amount of the material record, have largely been invisible until very recently, largely due to lack of knowledge of specific methodologies and of the scientific developments in other European regions (Hamerow, 2011).While this type of construction has been known in the Anglo-Saxon world from the 1930s (Leeds, 1936;Ralegh Radford, 1957), and in Germany and France from the 1960s (Chapelot and Fossier, 1980;Demolon, 1972;Donat, 1980), their existence in the Iberian Peninsula was only identified in the late 1990s (Vigil-Escalera, 1999), followed soon by the first typological studies on these constructions (Azkarate and Quirós Castillo, 2001;Vigil-Escalera, 2000).
From the late nineties, and especially in the following decade, over a hundred early medieval sites have been excavated in the Iberian Peninsula.However, the results have been greatly constrained by the legislation of each Autonomous Community to Commercial Archaeology (Parga-Dans, 2010).In Castile and León, as in other Communities, the development of commercial archaeology has been focused on peri-urban areas around the main cities (Segovia, Zamora and Valladolid) or on the main road and railroad construction sites.In other regions and provinces (Soria, Palencia, Galicia, Asturias, Cantabria or Northern Portugal) the impact of Commercial Archaeology has been much less evident, thus very few sites with domestic architecture have been identified.This asymmetric distribution should warn us about the relative gaps in our knowledge and to be moderated in our proposals, because the volume and the impact of construction works during the last decade determine the nature of the available material.
While in Madrid (Fernández Ugalde, 2005) or Catalonia (Roig Buxó, 2009) there have been various archaeological excavations in big open areas, very few excavations have been carried out in this way in the rest of the Peninsula.The norm continues to be small area excavations, by means of sondages and trenches dug directly where constructions are going to take place, providing with a very partial view of the sites.This makes historical interpretations and the analysis of domestic units extremely difficult (Ayán Vila, 2003;Vigil-Escalera, 2007b).
Lastly, it should be pointed out that even if Commercial Archaeology has led the research on early medieval rural settlements and their domestic architecture during the last fifteen years, it is also true that there have been some excavation projects led by academic institutions, as in the Basque Country (Azkárate, et al., 2011;Quirós Castillo, 2011b), the Duero Basin (Ariño Gil, 2006;Ariño Gil and Rodríguez Hernández, 1997), or some areas of Portugal (Tente, 2009).The simple quantification of the domestic units found in over a dozen sites from various areas of the Peninsula contrasted with the overall excavated area (fig.3) shows that on average one domestic unit is located every 1.340 m 2 of excavation.Considering these data, it is evident that only chance or a careful diagnosis enables the location of these constructions when digging sites under 200-100 m 2 in extension, which is the case of most excavations.It should be noted that these categories are nothing but a simplification, as it is normally the case that a same construction used different construction techniques and building materials (Brogiolo, 2008: 14), so a detailed typology would require further classification analytical criteria.
The distribution of these constructive typologies in geographic and diachronic terms shows a series of remarkable issues.In the 64 sites analysed in this study, sunken structures have been located in 33 of these, and fully standing buildings in other 33 sites, although it is more common to find that both type of constructions are used in combination (fig.5).
However, the typology distribution analysis is more telling.Thus, up to now long, post-structures with semicircular byres, known as longhouses, have only been found in the Basque Country.Chronologically, these belong to eighth-and ninth-century contexts in sites such as Aistra, Gasteiz and Gorliz.The distribution of sunken-feature buildings similar to Grubenhäuser is also very telling.The presence of similar structures in Protohistoric sites from the Meseta and in areas under Byzantine control in the Early Middle Ages allows us to reject any ethnic interpretation (Quirós Castillo, 2011a).
The use of stone footings and walls is first documented in the Meseta from the sixth century, and in the Basque Country only in the tenth century (Azkárate and Solaun Bustinza, 2003).All of these characteristics indicate a very complex reality.But in order to advance beyond this general characterisation it is necessary to go deeper into detail on villages and peasant settlements.For the sake of brevity, it will not be possible in this paper to discuss the domestic architecture linked to the early medieval centres of power which have been progressively excavated in the recent years (e.g., La Cabeza de Navasangil, El Castillón or Bernardos), although it is necessary at least to mention that these present different characteristic and technical skills to those identified in many of peasant settlements (including traces of urbanism, stone construction, lack of storage pits, etc.).
One of the most representative examples in the south area of the Duero Basin is La Mata del Palomar (Nieva, Segovia), excavated in the year 2002 as a result of the construction of a new rail road (Strato, 2002).The site is very close to the slate quarries of Domingo García, which made the extraction of building material for domestic structures easier.Despite the site's taphonomy and later agricultural impact, which deeply affected the nature of the site (Strato, 2002: 18), the excavators identified 22 sunken-feature buildings in two of the excavation sectors, together with some standing structures very difficult to interpret.Most of these sunken structures present rectangular or sub-rectangular layouts (14 examples), the few remaining had round layouts (7) and a one indeterminate example.Some of them had post-holes, although the majority lacked them.This lack of internal upright supports may be due to the size of the structures (6 m 2 on average), which would have made internal supporting posts unnecessary.At least of the structures had a slate floor, so it may have been the case that supporting posts were put directly on top of the pavement, without the need for a post-hole.
With regards to the non-sunken structures, these are very difficult to interpret because the site's upper levels have been almost completely destroyed, although some slate and mud walls were identified on top of the natural soil, but without a foundation trench.Only one of the structures maintained two walls of this type, and the pavement of this building was also made with slate.
One of the most significant aspects of the site is the internal organisation of the structures, which had two production areas with furnaces and vats, which have been identified as cooling vats for metal smithing.Likewise, it is interesting to see that there seems to have been a wall surrounding the whole settlement as if it were a fence, which may be related to animal herding.
The chronology of La Mata del Palomar goes from the mid-fifth to the early-eighth century.Two absolute dates have been obtained from the site: an individual buried inside a silo has been dated by radiocarbon to the late seventh or the early eighth century7 , and a thermoluminiscence date has been obtained from a brick from structure LVI (a furnace), dated between the eighth and the tenth centuries.However, the current revision of the ceramic material indicates that most of the excavated structures were abandoned at some point between the latefifth and the mid-sixth century, showing perhaps a first phase of the site, and the structures of later phases may be those that have not been excavated.
The site of Lancha de Trigo (Diego Álvaro, Ávila), is a clear example of the type of domestic architecture characteristic of the southwest area of the Duero Basin.This site, excavated in the 1950s, had seven free-standing quadrangular structures, of vary different sizes, ranging between 19,5 m 2 (house number 2) and over 740 m 2 (house number 3) of internal space.The latter has a long rectangular area enclosed by walls made out of a core of rubble, with stone facings.This technique is used in most of the buildings of the site, but it is also known for other sites, such as Monte el Alcaide (Paricio Alonso, 2009), Navalvillar (Abad Castro, 2006) or Congosto.
On the north side of this building there were three or four rooms of unknown function.In the central one (section C) various imbrices were found underneath the slate paving which had been used to level the ground, which suggests a previous tiled phase for the site (Gutiérrez Palacios, et al., 1958: 64).The presence of inscribed slates in the paving of various houses led the original excavators to date the site beyond the eighth century (Gutiérrez Palacios, et al., 1958).Nevertheless, considering the wider regional context, this dating should be pushed further back in time (Quirós Castillo, 2011a: 73) We can observe important differences comparing sizes forms and techniques in different buildings of Lancha del Trigo which may refer to different domestic functions for each one.The lack of silos located, due to the geology of the region (formed primarily of granite, in which is very difficult to construct sunken-featured structures) would mean that either surplus was stored inside each dwelling or else there was some sort of communal granary (or perhaps both).However, the site was excavated in the 1940s and 1950s, so many details and pieces of information were not recorded, as for example a site plan.
In the site of Cuarto de las Hoyas/Dehesa del Cañal (Pelayos, Salamanca), a dozen rectangular and trapezoidal buildings were documented in the 1980s, and the main construction technique of the site was non-mortared squared rubble blocks of quartz (Fabián, et al., 1985).The layout of the site shows how some structures aggregated forming groups, which may indicate some degree of organisation of the dwelling areas, but should not be considered as an indication of urbanism (Quirós Castillo, 2011a: 72).In this site almost all of the structures are internally subdivided into areas with various purposes including storage and cooking.Similarly, animal enclosures have been detected in almost all the buildings.
The settlement of Pelayos is, as that of La Mata del Palomar, enclosed by a wall 130m long, built in the same construction technique as the other buildings, and which the excavators linked to a settlement boundary.Perhaps other explanations, linked to economic reasons, such as enclosing cattle, together with a sense of communal identity could be proposed.However, we only have a superficial record of buildings that prevent us to analyse which of them were contemporaneous and the number of domestic units coexisting in a definite moment.
In the north-western corner of the Peninsula, mostly in Galicia and Asturias, the patterns are slightly different, perhaps due to the minor amount of excavated sites.
The site of A Pousada (Santiago de Compostela) was excavated in 2001 (Ballesteros-Arias and Blanco-Rotea, 2009), and the finds include a building whose foundations were laid on big blocks of local stone, and filled in with smaller rubble.The inside was paved with beaten clay, on which a hearth and a post-hole were found.The room, which could not be totally excavated, measures 15.48 by 3.71m, with an overall area of over 60 m 2 .This construction was dated to the seventh or eighth century, although it was subsequently modified, reorganised and subdivided up to is abandonment between the eleventh and thirteenth centuries.Five silos and several terraces can be ascribed to the first phase.
Free-standing structures on stone footings are commonplace in the north-western part of the Peninsula, and can be found in sites such as Villa Bidualdi, As Pereiras, Montenegro, Padrón or Casas Vellas, in Galicia and La Cavada/San Nicolás de Corao, Campa Torres and the villa of Veranes, in Asturias.
In this last site (Fernández Ochoa, et al., 2005-2006), the remains of a hearth and a small post-hole were excavated in the same room, and the spoliation of the building materials of the villa for the construction of nearby structures was also identified.In the northern courtyard of the villa many post-holes have been found, together with a round sunken hut.This reoccupation of villa spaces in perishable materials has been related to marginal occupation or squatters8 (Lewit, 2003;Sánchez Montes and Rascón Marqués, 2006).However, this relationship is based on the analysis of isolated cases, and not on regional studies, which actually show that these apparent «marginality» is, in fact, the norm.The reuse of old Roman rural elite centres can be related to a self-sufficient economy linked to small communities rather than to the production of surplus and cash crops.
The interesting site of Peña del Mazo (Pajares, Burgos) (Palomino and Negredo, 2011) is located on a hilltop in the upper Ebro valley.This site is located on a terrace within a calcareous rocky formation on the southwest face off the Sierra de Árcena.The first occupation phase of the site consists of four rectangular sunken huts excavated in the rock.They are all rectangular, the most prominent being structure 4, which is 32 m 2 , and it is placed on the easternmost end of the terrace, in which the most remarkable feature is «the perfect levelling of the rocky floor in the inside, into which at least four post holes were carved, lined and in parallel close to the bottom surface of the cut»9 (Palomino and Negredo, 2011: 199-200), which probably indicates a structure built on tim-bers.These constructions were abandoned in the course of the fifth century due to the construction of a cult centre and a necropolis, in use between the sixth and eighth centuries, later heavily rebuilt in the medieval period.
The importance of this site is that the abandonment of the rock-cut dwellings by a cult building, which significantly modifies the social organisation of this space of chief geographical preeminence, perhaps indicating the presence of a new power within the community.

Regional diversity and diachrony in the construction processes
The collapse of the Roman imperial economic system and of the related networks of power which happened in the north and the central Iberian Peninsula during the fifth century (Vigil-Escalera, 2009) caused the fragmentation and adaptation of the local political and socioeconomic processes, on production and social terms.Historical and archaeological studies in the recent years have shown that two of the main consequences of the change of the scale of the system was the emergence of systems of villages and peasant farms, linked to the decentralisation of the processes of production (Vigil-Escalera, 2007b), and the increasing importance of local elites (Castellanos and Martín Viso, 2005).Considering that ultimately in this period the demand generated by the elites is the key which determines the complexity of exchange and production (Wick-Fig.7. Sunken-featured structure excavated in the rock in Peña del Mazo (Aratikos, 2005) ham, 2008), the strong regionalisation of the acting powers in the Peninsula in the Early Middle Ages explains the material culture of this period.Pottery, for instance, serves as a perfect example of the processes of fragmentation of the distribution systems and of the transformations of the production systems, which had to adapt to regional scales (Centeno Cea, et al., 2010;Dahí Elena, 2010;Larrén, et al., 2003;Vigil-Escalera, 2007a).Architecture, just as settlement hierarchies, is another archaeological indicator of social transformations and the articulation of the local elites, as defended by Ch.Wickham (2005), which greatly characterises peasant domestic units.
The archaeological study of early medieval domestic architecture has usually been limited by a series of preconceived ideas with social negative connotations marked by a comparison with the earlier (Roman) period.Some authors, however, have underlined that the use of perishable materials in domestic architecture should not be considered as an indication of social or productive decline, but rather that they require a distinct set of specialised skills and new ways of social interaction.H. Hamerow has recently pointed out that the construction of the great longhouses requires a large social capital (Hamerow, 2011: 129), which would explain the social meaning given to them in some European areas.With no doubt, the noteworthy constructions of Gasteiz (Azkárate Garai-Olaun and Solaun Bustinza, 2009) and Aistra (Quirós Castillo, 2011a) should be understood from this perspective.
Taking these proposals into account, it would seem as if the north-western corner of the Iberian Peninsula held multiple «constructive forms» during the Early Middle Ages, characterised by chronological and regional homogeneity.More precisely, four different areas can be proposed: the centre-south area of the Duero Basin, its south-west area, the High Ebro valley and the north-western corner (fig.6).
1.In the centre-south of the Duero Basin, domestic architecture is based on sunken and semi-sunken structures, as seen in many examples as the already seen site of La Mata del Palomar, but also in La Cárcava de la Peladera (Hontoria, Segovia), Senovilla (Olmedo, Valladolid), El Pleito/La Casilla (Rubí de Bracamonte, Valladolid) or Navamboal (Íscar, Valladolid).The use of over-ground self-standing structures is limited to a few concrete examples, and always with stone footings.Firstly, in the sites of Carratejera (Navalmanzano, Segovia;(2barqueologia, 2007-2008;Strato, 2003), Soto de Tovilla (Tudela de Duero, Valladolid;(Martín Rodríguez and San Gregorio Hernández, 2008) or El Soto/La Verdina (Sieteiglesias de Tormes, Salamanca) (Strato, 2008) these structures are linked to old late-Roman large estates.Secondly, the continuity of stone structure construction continues into the early medieval period, as the site of La Huesa (Cañizal, Zamora), which has a sandstone building built on top of an earlier sunken-feature building (Nuño González,

2003
).Similar structures are also found in those sites with more complex internal domestic structures in the Early Middle Ages, such as La Mata del Palomar or La Cárcava de la Peladera, where it is possible to infer from the functional distribution of the space a communal or suprafamiliar organisation of specialised production.Finally, these techniques are identified in hill-top settlements, such as Bernardos (Segovia) or el Castillón (Tábara, Zamora), linked to relatively stable local centres of power (Castellanos and Martín Viso, 2005), which required long-lasting constructions.
Anyhow, and within this constructive technique, the highest point in the development of these constructions lies within the fifth and the seventh centuries, with a few notable exceptions of Canto Blanco (Calzada de Coto, León; (Strato, 2010) or La Huesa (Nuño González, 2003), which seem to continue into the later centuries.
2. The south-western part of the Duero Basin held a completely opposite constructive form.It consists of a relatively homogenous area, as far as the constructive techniques are concerned, including sites as Soto/La Verdina (Sieteiglesias de Tormes, Salamanca), San Cristóbal/Las Henrenes (Cillán, Ávila), Lancha de Trigo, La Legoriza (San Martín del Castañar, Salamanca), El Cuquero (Villanueva del Conde, Salamanca), or la Serna (Garcihernández, Salamanca).To this list, the fortified hill-top of La Cabeza de Navasangil (Solosancho, Ávila) could be added, due to its close geographical location, which may have functioned as main organising nucleus in this network whose extension cannot yet be archaeologically determined.As we have seen, this area is characterised by the use of stone, at least in the foundations and the footings.The use of various local raw materials makes the final buildings seem more heterogeneous, but the construction technique itself is very similar in all sites.This type of construction is dated between the sixth/seventh and ninth centuries.The main factor that determines this constructive form is the geology of the area mainly formed by granites, which makes the construction of sunken-featured buildings very difficult.This situation has important consequences for the organization of domestic units, as we will later analyse.
3. With regards to the north-western end of the Peninsula, the lack of data prevents us from making a general interpretation of the domestic units.However, the data show a variety of cases, which ranges from the case of A Pousada, with long-lasting constructions, to short-lived farms, as those in Padrón.Whereas in the other models there seems not to be a direct relationship between sunken structures and built structures, in this area there is a combination of both types within the same domestic unit, probably indicating a disperse and rather unarticulated settlement pattern.The monumental architecture of the site of Gauzón, however, shows that from the sixth century onwards there were powerful local elites, continuing into the following centuries.It should not be forgotten that despite the lack of data, the most recent analyses indicate the presence of a complex economic structure, relevant enough to generate local trading networks, as that identified in the town of Vigo (Fernández Fernández, 2011).
4. In the Basque Country and the Upper Ebro valley we find construction techniques which follow wider continental patterns.There are very few known structures datable to the fifth or sixth centuries, although these include post-supported structures as those of Zornoztegi, as well as sunken-feature buildings, parallel to Grubenhäuser, in sites such as Aistra, and cave occupations, both natural (Los Husos) and artificial (Las Gobas).From the seventh century onwards there is more evidence, and there is a wider collection of constructive typologies.In Gasteiz, one of the sites which has provided material for the eighth, ninth and tenth centuries (Azkárate Garai-Olaun and Solaun Bustinza, 2009), the existence both of longhouses of considerable size and of sunken-featured buildings has been identified.And from the tenth century onwards structures on stone footings appear.In Aistra two large longhouses (over 200 m 2 ) have been found, together with other rectangular constructions supported by vertical posts.The site of Zaballa (Quirós Castillo, 2012) confirms the introduction of stone footings in the region in the tenth century, while before that buildings with semisubterranean areas were the norm.

From architecture to domestic units: general considerations
Once the domestic architectures have been analysed and as a result different constructive forms in the north-western Peninsula in the Early Middle Ages identified, what can we learn from their domestic units?In the current state of affairs, three ideal models of domestic units in this region can be put forward10 .
In the centre-south of the northern Meseta we find sites with a high proportion of sunken features over other types of structure; a building technique that we can link to mobile domestic units.Examples such as La Mata del Palomar or La Cárcava de la Peladera would therefore show domestic units with various structures (one or more huts, with one or two silos and, in some cases, a stone footing structure) in which autonomy seems to be the goal for production, distribution, transmission and reproduction, functions which would take place in the various buildings belonging to the same domestic unit.This type of productive autonomy model perfectly adapts to a set of constructions technologically simple enough to enable the mobility of communities or family groups in short cycles (every one or two generations, as inferred from the data from La Mata del Palomar) as the productive resources (arable land, grazing fields, forests) diminish or decline 11 .
In La Mata del Palomar, the of buildings together with sunken huts of a relatively small size suggests a domestic unit similar to those identified at the village of Gózquez (Quirós Castillo, 2011a), which included 10-12 dwelling plots, in which buildings are cyclically built, abandoned and rebuilt.The presence of open plots in the site is a point of contrast; in Gózquez these are thought to have been plots dedicated to cultivation, but in the case of La Mata del Palomar, the bad preservation of the different phases does not allow to identify whether if these plots existed or not.
In this way, we can talk about certain social and settlement structures in the centre-south area of the Duero Basin based on villages and farms, with their respective domestic units, which may have been linked to local centres of power, of which they could or could not have been subsidiaries.Even if the nature of this link between a centre of power and the surrounding settlements cannot be properly assessed, the fact is that the centre of power was not capable of fixing and organizing the settlement in a given space, and also that all this refers back to a productive structure focused preferably but not exclusively to selfconsumption and the storage of the production by the domestic units.
On the opposite corner we would find the domestic units of the south-western Meseta, characterised by stone constructions dated to the sixth/eighth centuries, and by the lack of sunken-featured buildings.These constructions seems to have a more permanent and long-lasting character, as evidenced by the repairs evident in some of these buildings such as in Lancha de Trigo, Monte el Alcaide (Paricio Alonso, 2009) or in San Cristóbal/Las Henrenes (Díaz De La Torre, et al., 2009).The construction of these buildings inside a community, attached to a concrete space, indicates a different concept of the property of the space and a different concept of the territoriality (Escalona Monge, 2010).Even if it is impossible to talk about «urbanism» in sites such as Cuarto de Hoyas (amongst other reasons, because we do not know which of them were contemporary), there is a kind of rational organisation of the space occupied by the domestic units, if compared with the rather unorganised settlement pattern evident elsewhere in the Meseta.
The architecture at Lancha de Trigo or Cuarto de las Hoyas shows domestic units based upon a single building, which would hold most of the productive and reproductive functions of the domestic unit.This different pattern is reflected in the internal subdivision of the buildings into specialised domestic areas.If compared with the relative homogeneity of the domestic units in the central-southern Meseta, this area has different architectures (according to the size or the internal divisions) which may indicate the presence of public areas where to productive (communal storage) and reproductive (socialising) functions rather than the presence of village elites12 .
It is unavoidable to link this region with the distribution pattern of the inscribed early medieval slates, which have been linked to the presence of strong local powers, perhaps either elites linked to large-scale herding or to a religious centre (Martín Viso, 2007a;Velázquez, 1989).Both records (inscribed slates and domestic architecture) indicate a strong social and regional articulation, which would generate a given set of productive patterns which would be the frame in which the aforementioned domestic units would develop.We should consider here the model of the existence of one or various centres of power (Salvatierra de Tormes?La Cabeza de Navasangil?) which acted as receivers and/or redistributors of the agricultural surplus of the villages which in turn would not need to store it mainly in their domestic units.The centres of power would therefore act as the main nodes of a network it is possible to understand the ways in which actual action is influenced by irrational factors of all sorts, such as affects and errors, in that they account for the deviation from the line of conduct which would be expected on the hypothesis that the action were purely rational» (Weber, 1978: 6).
based on an economic structure which worked on the extraction of surplus.
Finally, in the Upper Ebro valley a new type of domestic units is to be found from the eighth century onwards.These units were more socially differentiated and communal production and reproduction functions are more clearly articulated.This is the type of domestic architecture which develops in sites such as Gasteiz or Aistra.The distribution of the domestic space in aisles may indicate that multiple functions were carried out under the same roof, linked to the presence of strong regional powers (Quirós Castillo, 2011a;Quirós Castillo, 2011b).
Evidently, the conceptual characterisation of these domestic units does not imply a strict territorial demarcation.In order to explain the various similarities that can be observed in these territories we must be aware of the presence of small groups of itinerant workers or parttime peasants which would collaborate in the construction of these structures with the settled peasant communities, in parallel to the hand-made ceramic productions identified in the centre of the Peninsula (Vigil-Escalera, 2003b).Indubitably, the organisation required by the construction of the longhouses identified in the Basque Country, which were up to 8/10m long and that required long beams and skilled labour to put them in place, cannot be compared with the Grubenhäuser identified in the Meseta in earlier centuries which in turn reflect a higher articulation of the different communal activities and the domestic units.

CONCLUSIONS
The aim of this paper was to offer an overview of the domestic architecture and domestic units detected in the northwest of Iberia.With this overview we have described four constructive forms, from which we have proposed three ideal types of domestic units that should be used as a general frame for further analyses.
We started the analysis of domestic units based on the hypothesis that the post-Roman world was characterised by the fragmentation of productive and social relations.We think the analyses presented confirm this hypothesis.Domestic architecture reflects the adaptation of domestic units to a new economic and political frame; we can detect this in the simplification of constructive processes, the use of local raw material for the construction and the importance of local geology in final constructive decisions.In the same way we have proposed that not only these issues determine the domestic architecture and the development of domestic units, but also the social organisation of production according to the degree of integration within a socio-political scale.
Having reached this point, three issues arise, which in our opinion, should lead researchers in the following years.
Firstly, it is obvious that we need to analyse in depth the impact of the taphonomic processes and the methodology of excavation in the formation of the archaeological data.As we have seen, these issues determine not only the quality of the data, but also its presence, and in many cases, most of the structures are simply missing from the record, leaving us an incomplete view of the sites that can completely distort further analysis.
Secondly, the minimal social differentiation inferred from the study of domestic architecture should be considered.The archaeological record cannot indicate, for the time being, social stratification in rural communities, neither is it possible to define an elite domestic architecture between the sixth and ninth centuries.The fact that the same construction techniques appear both in hill-top settlements and in rural settlements, for example, may indicate that the elites did not use domestic architecture as an indication of social rank or a way to identify their power.Elite markers are therefore to be found in other contexts, perhaps in apparently isolated religious buildings or maybe in necropoleis, by means of anthropological or ancient diet analyses.Similarly, it should be considered, especially for the early period between the fifth and the sixth centuries, that domestic units existed in a society relatively equalitarian in terms of power distribution and with a very weak social stratification (Tejerizo García, 2011).Only the emergence of supra-regional political entities, which generated at the end of the ninth century and in the tenth century a change in the scale of local interactions (Escalona Monge, 2010), marks the beginning of clearly identifiable architectural elements of power which can be identified archaeologically, like castles, churches and palaces.
Another issue which arises from this paper is that the relationship between rural domestic architecture in the central Peninsula is still largely understudied between the eighth and the tenth centuries.The few available pieces of information for this period come largely from funerary contexts (Martín Viso, 2007b), and the exceptional excavations of San Gens in Celorico da Beira (Tente, 2009), which seem to indicate a restructuring of the landscape and the emergence of supra-regional political entities, as seen in the Basque Country (Quirós Castillo, 2011b).This lack of information is related to the continuity of historiographic paradigms such as the depopulation/colonisation of the Duero basin, which has determined our understand-ing of the material record of the eighth century.Hypothetically, it may be possible that this «archaeological invisibility» is nothing but the continuity of previous regional and local peasant dynamics until the point in which the aforementioned supra-regional power dynamics appeared in the region.
Summing up, this paper has presented a series of hypotheses which reconsider some of the ongoing debates regarding early medieval archaeology, and that should be solved in the near future as new data are obtained, more thorough analyses are carried out, and especially, these issues should be solved with the development of new theoretical frameworks into the archaeological discourse.

Aknowledgements
This paper is part of a thesis aimed at the study of the early medieval rural settlement of the Duero Basin.For this task I rely on the superb collaboration of various companies which have excavated in the recent years in various areas of the Peninsula, and in particular Strato S.L, Área S.C and Aratikos S.L. Similarly, I would like to thank Juan Antonio Quirós for helping on the concept and the development of this work, Alfonso Vigil-Escalera for putting my feet back on the ground and Javier Martínez for translating this paper.

Fig. 3 .
Fig. 3. Correlation between excavated area and number of sunken huts identified in a simple of early medieval sites

Fig. 4 .
Fig. 4. Main construction categories identified in the territory under study

Fig. 5 .
Fig. 5. Domestic structure distribution according to the number of sites in which they are identified

Fig. 8 .
Fig. 8. Constructive forms in the north-west of the Iberian Peninsula