29 Domestic and Settlement Organisation in Iron Age Southern France
Abstract
Archaeological research in the southern French regions of Languedoc-Roussillon and Provence over the last 30 years has extensively excavated a significant number of Iron Age sites. Most are settlements, while cemeteries remain much less well documented, particularly during the Late Iron Age and in the eastern Languedoc. In the absence of ancient documentary sources and given the scarcity of funerary evidence, settlements offer the best information about Iron Age society. Houses in particular provide crucial information about protohistoric households and society, as they constituted the focus of daily life and stood at the centre of economic, cultural and social activities. The analysis of domestic architecture from the seventh century BC to the time of Romanisation has shown that houses became more varied over time in terms of both typology and function. I discuss the connections between changes in indigenous architecture and external influences, and consider the significance of households and social distinction for the transformations of houses in southern French Iron Age society.
Introduction
This chapter concerns domestic architecture and its occupants in Mediterranean France during the Iron Age. In geographical terms, the study area is defined by the Mediterranean Sea and the mountain ranges of the Pyrenees, Alps and Cévennes, which roughly coincide with the modern regions of Languedoc-Roussillon and Provence (Figure 29.1). As for the chronological period, the French Iron Age begins in the seventh century BC and runs until the Roman conquest in the late second century BC. This period is usually divided in two phases – the Early and Late Iron Ages – but there are no agreed beginning or end dates for these two phases (Dietler 1997: 275–76; 2005: 29). In this chapter, I will follow the chronology proposed by Michel Py (1993: 21), who dates the Iron Age between 675 BC and the change of era, and the transition from the Early to the Late Iron Age between 525–425 BC.
Figure 29.1. Map of southern France with the location of the main sites mentioned in the text (map background: Michel Py).
In the last 30 years, extensive excavations have been carried out at many sites, most of which are settlements. We are consequently much less informed about cemeteries, in particular during the Late Iron Age and in the eastern Languedoc. Because the evidence for settlement types, urbanism and domestic architecture is now rather good, in this chapter, I will focus on houses to investigate the protohistoric societies of Mediterranean France.
In the study area, Iron Age houses consist of one or more spaces with combined functions; their domestic nature is evident from the remains of food preparation and consumption such as hearths, food, cooking pottery, querns and so on, but other activities such as storage and productive tasks were also usually carried out in these spaces. The spaces of a house may be connected or not, and adjoining yards may be present as well.
The specific character of each house emerges from a detailed analysis of domestic features and artefacts that make it possible to identify activities carried out in each room and to interpret their functions. Even if some scholars have called into question whether it is possible at all to recognise protohistoric houses (Boissinot 1995: 75), it is my view that Py (1996) has convincingly demonstrated with his research in Lattes that functional and spatial analyses do allow the identification of protohistoric houses in southern Gaul. There remain nevertheless important questions to be resolved, in particular concerning numbers of occupants and household composition, and these matters make up the topics of discussion in this chapter.
Late Bronze and Early Iron Age Traditions: Wattle and Daub
Houses of the early Iron Age (675–575 BC) show numerous and strong continuities with Late Bronze Age traditions. Even some caves continued to be occupied, but open-air settlements represent the most frequent type. They consisted of groups of huts (cabanes) or houses made of perishable materials (Michelozzi 1982: 19–34; Dedet 1987: 177). The sites do not appear to be planned, and there are no indications of a distinction between private and public spaces. The houses are not oriented in any particular way, although they are separated by open areas, where hearths, pits or granaries have occasionally been recorded. It is possible that new huts were built and organised in accordance with kinship relationships and the domestic cycle of households, as has been observed in contemporary African contexts, but clear evidence is lacking (Herbich and Dietler 1993; 2009). The settlements tend to occupy less than one ha, although there are significant exceptions, such as the site of Carsac (Carcassonne), which measures around 25 ha (Guilaine 1986: 179).
The houses that make up these settlements had very simple plans, varying in shape from round to oval, rectangular and irregular, and were mostly smaller than 20 sq m, as for instance Le Laouret (Aude; Gasco 1995: 57) and La Liquière (Py et al. 1984). In his analysis of 248 huts of this period, Bernard Dedet (1999: 325) proposed an average size of 11 sq m for the Languedoc. Recent research at Le Traversant (Mailhac) and Ruscino (Perpignan) has not only confirmed this variability in size and plan but also in the functions of the structures.
During the Late Bronze and early Iron Ages, Le Traversant is made up of absidal houses, some of which were defined by postholes, hearths and pits, while others had been built with earthen walls (Figure 29.2). The biggest house had a surface of 50 sq m (Gailledrat et al. 2000: 176; 2006–2007: 41). A similar house plan, dated to the end of the sixth century BC, was excavated at Ruscino (Marichal and Rébé 2003: 144–59). This one measured 48.3 sq m (10.5×4.6 m) and had a double absis. Its perimeter was defined by postholes, and four larger posts had stood in the central area to support the superstructure. Traces of charcoal and clay suggest that the walls were made from perishable materials and covered with mud plaster.
Figure 29.2. Plan of the structures excavated at the site of Le Traversant (Mailhac, Aude), and restitution of the early Iron Age house plans (Eric Gailledrat).
The only traces of these houses are the postholes and ditches that outline their perimeter, while the wattle and daub used to make walls and roofs has left few remains. There were no internal partitions, but the distribution of domestic features and other remains suggests that certain areas were dedicated to distinct activities: there is usually a hearth in the central part of the building, an area for domestic tasks and a zone reserved for food storage, mostly at the rear. This appears to have been the organisation recognised at Laouret (Floure, Aude; Gasco 1995: 57), while the house in Ruscino saw most activities concentrated in the northeastern part, where three clay benches, an oven and two hearths were found (Figure 29.3). Although nothing had been preserved at the opposite end, the central position of the two hearths and the benches could point to a similar use of space as in the house at Laouret. The archaeological remains mainly consisted of food refuse and pottery related to food preparation and consumption. There was not much storage pottery (Marichal and Rébé 2003: 182–88; Gasco 1995: 50–61).
Figure 29.3. Internal view of a reconstructed hut at Ruscino (Perpignan, Pyrénées Orientales) (Maria Carme Belarte).
These houses have been interpreted in general terms as occupied by small groups of people. How these groups were constituted is unknown, as little or no attention has been given to kinship and domestic structures in protohistoric southern France. While building techniques and domestic features have been described and studied in great detail, domestic activities and the people involved have hardly been topics of investigation. It is nevertheless widely assumed that the protohistoric house was the residence of a nuclear family (Py 1993: 70). At least in the early Iron Age, the size of the main area of the house points to four to five occupants, who could make up a nuclear family.
The domestic features and other finds from the houses do not indicate social differences, despite the variation in size. Overall, the evidence suggests that most economic activities were carried out inside the huts or immediately outside, that is, in a domestic setting, while the absence of large storage structures attests an economy without surplus (Garcia 2004: 35). Even so, there would always be several activities that went beyond the house and that would have been carried out in shared spaces, in some instances collectively as well. The ovens, hearths and pits recorded in the open areas between the houses probably represent such shared usage and joint work. The large enclosures and ditches around some settlements such as Carsac offer evidence of collective labour (Guilaine 1986: 179).
Building Techniques and House Plans: Colonial Contacts and Indigenous Traditions
One of the major changes in protohistoric domestic architecture of southern Gaul is the appearance of houses with rectangular plans, load-bearing mud walls on stone foundations, shared party walls, and often with the rear built against the rampart of the settlement. This kind of architecture is closely associated with the first urban plans that consisted of regular housing blocks of houses separated by streets. Construction of these houses began by the end of the sixth century BC in coastal Languedoc and Provence, while they appeared in the interior from the fifth century onwards (Dedet 1999: 315–21).1 The earliest houses coexisted with those built in wattle and daub.
The appearance of new settlement patterns, architectural forms and building techniques have often been interpreted as related to colonial encounters (Michelozzi 1982: 85; de Chazelles 1995: 51–54; Dedet 1999: 317), because the first contacts between the indigenous communities and other Mediterranean peoples (Greeks, Etruscans and Phoenicians) took place during the transition period from the Bronze to the Iron Age (750–675 BC; Dietler 1997: 277–91; 2005).
The site of Lattes is of particular interest in this regard, as its foundation layers, documented in zone 27, are closely linked to the Etruscan trade in south France. In its earliest phase around 500 BC, the settlement was already surrounded by a stone rampart with houses arranged perpendicular to it, that were built of stone and mud-brick to rectangular plans with shared party walls (Figure 29.4). Two of these houses have been explored, and each has yielded such a large amount of Etruscan pottery, mainly amphorae, that an Etruscan presence in this area of the site seems likely. After the first phase had been destroyed and abandoned, the whole block was rebuilt in the early fifth century BC, but the houses were constructed with load-bearing posts, and Etruscan pottery was not very common (Lebeaupin and Séjalon 2008: 45–64; 2010: 138).
Figure 29.4. View of the destruction layers about 600 BC in zone 27 at Lattes (Hérault) (Michel Py).
This situation is similar to that documented in other Mediterranean areas where colonial contacts had been intensive. In Catalonia, the indigenous site of Sant Martí d’Empúries, where later Emporion was to be founded, is a good case in point. In the second half of the seventh and the beginning of the sixth centuries BC, when the first Etruscan and Phoenician imports appeared, houses were built with mud walls on stone foundations and shared party walls. In the next phase (625/600–580 BC), the houses were replaced with new ones supported by posts of the Bronze Age tradition (Aquilué 1999: 126).
These examples show that the early Iron Age (late sixth to fifth centuries BC) was a period of experimentation, that saw the first rectangular houses with roof-bearing walls and urban plans, while building techniques still varied considerably and shared party walls were not consistently used. Even if external influences can be detected in the building techniques and Greek elements have been recognised in their metrological system (Tréziny 1989), the resulting houses and urban layouts were largely based on similar conceptions, and use of space and indigenous traditions were to persist in later centuries. It has even been suggested that Greek settlers in Marseille might have adopted indigenous building techniques rather than providing new ones for the local communities (Dietler 2005: 134).
The houses with rectangular plans and roof-bearing walls were built using mud-bricks, cob or rammed earth, i.e. using mixtures of mud and clay, sand, clay and straw or heavily compacted gravels, sand and clay. Mud-brick in particular is usually thought to have been introduced from elsewhere in the Mediterranean and distributed in southern France from Marseille, while rammed earth has long-standing indigenous roots. Mud-brick walls on stone bases gradually replaced traditional building techniques, however, and became the dominant way of wall construction by the Roman period. Roofs were made of a wooden frame that was covered with branches and earth; roof tiles were practically unknown. Those tiles found in Lattes have in fact been argued to have been used for gutters and culverts to channel rain run-off (de Chazelles 1996: 280). The house surfaces (walls, floors, roofs) were finished with earth plasters, often complemented with pebbles, flagstones and (crushed) pot sherds.
Houses were normally grouped in regular blocks, and their plans and any changes were closely related to the overall layout of the protohistoric towns. The earliest compact houses coexisted with others that included open spaces such as courtyards. Good examples of the latter type, consistently separated by open areas, have been documented at Le Plan de la Tour, Gailhan (house 1: Dedet 1987: 15–38; Figure 29.5), Montlaurès, where they are dated to around 500 BC (de Chazelles 2005: 248), Pech Maho (ca. 540 BC: Gailledrat and Solier 2004: 375) and Lattes (from 450 BC onwards; Roux 1999: 31–48 and fig. 28; Belarte 2008: 103; 2010: 204; Belarte et al. 2010). Analysis of the open spaces between houses of the fifth and fourth centuries BC in Lattes has shown that they were mostly used for food processing and cooking, as is evident from the ovens and hearths but also from generous quantities of refuse such as ashes, charcoal, faunal remains, seeds and so on (Figure 29.6). Sheds, annexes and other minor buildings continued to be built in perishable materials (Belarte 2010: 205).
Figure 29.5. Plan of house 1 at the site of Le Plan de la Tour, Gailhan (Gard) (Bernard Dedet).
Figure 29.6. Plan of zone 1 at Lattes (Hérault) during the phase 1P (450–425 BC), with houses separated by open areas (Maria Carme Belarte and Eric Gailledrat).
Settlements organised in regular housing blocks appeared in Provence from the early fifth century BC onwards. At Saint-Pierre-les-Martigues, L’Arquet and L’Île (Martigues), for instance, housing blocks were separated by streets running parallel to the ramparts (Chausserie-Laprée 2005: 98). The houses themselves were simple single-roomed buildings without internal partitions, measuring between 10 and 20 sq m. During the fifth and fourth centuries, house plans remained quite simple with mostly just one or two rooms (e.g. Dedet 1987; 1999; Py 1996; de Chazelles 1999; Belarte 2008). At Lattes, one of the most extensively excavated sites, houses measured between 17 and 87 sq m, with an average of 40 sq m in the fifth century BC (Belarte 2010: 213–15)
Use of Space and Social Implications
The evidence discussed indicates a substantial continuity in the use of space from the Late Bronze Age onwards. The size of the houses suggests that they were occupied by nuclear families, while the open areas between the houses point to strong links between their occupants, who shared these spaces for domestic activities. Private and collective spaces were initially not strictly separated and the subsequent transformation of open areas into enclosed domestic spaces may be explained through kinship links.
In terms of social organisation, the first sites with urban plans do not show clear evidence of significant social differences, as no large or otherwise prominent houses have been identified. Different house sizes at major sites such as Lattes nevertheless suggest that a process of social differentiation was probably underway by the fourth century BC. It is plausible that this development related to decisions about the distribution of space, the size of house plots and the definition of areas for common use, as such decisions could be collective in a mainly egalitarian society, but it is on balance more probable that an elite was emerging behind these decisions.
Around the same time, the use of space was becoming increasingly specialised, as meal preparation, consumption and storage took place in different rooms. Cooking was mostly concentrated in the front rooms, close to an open space, but there is also frequent evidence of cooking in unroofed areas such as yards and patios, whilst storage, limited as it was, was usually to be found in the back rooms. If houses were not subdivided, these activities were carried out in different areas of the same room, as at l’Île (Martigues), where the destruction layers of the one-roomed houses have yielded concentrations of hearths, ovens, silos and pottery (Chausserie-Laprée 2005: 132) (Figure 29.7).
Figure 29.7. Photo of the destruction layers of house BI, in the fourth century BC, at Martigues (Bouches-du Rhône) (Jean Chausserie-Laprée).
The First Complex Houses of the Fifth Century BC
Even if most house plans and sizes were quite regular and similar, more complex layouts have been documented in Arles and Béziers. In Arles, housing blocks of the early fifth century BC have been brought to light in the ‘Jardin d’Hiver’ area, where houses were made up of a large courtyard of 60 and 70 sq m, around which several rooms were situated. In the first half of the fourth century BC, the courtyards became much smaller and measured just 7–9 sq m. Strong Hellenic influences have been noted at this site, as the architecture was based on the same metrological system found in Greek colonies such as Agde and Olbia, while Mediterranean pottery imports made up as much as 70–85% of all finds during the fifth and the fourth centuries BC (Arcelin 1990: 196; 1995: 329–30; 2004: 253).
In Béziers, house 1 has been dated to the second half of the fifth century BC. The preserved area measures around 115 sq m, but the original size has been estimated at at least 150 sq m (Olive and Ugolini 1997: 96). It consists of eight separate spaces, including an interior courtyard surrounded on three sides by covered rooms. According to the excavators, both the courtyard-based house plan and the use of roof tiles are without parallel in protohistoric southern France, with the exception of Arles, and may thus denote Greek influence (Olive and Ugolini 1997: 97–98).
In both cases, the organisation of space around a central courtyard, the number of rooms and the overall house size are clearly different from contemporary indigenous houses in Languedoc and Provence. In Arles, the use of space was moreover notably specialised by the fourth century BC, as most rooms appear to have had specific functions. In Béziers, the excavators interpreted one room of house 1 that was situated next to the courtyard as a cooking area, but a lack of finds did not allow further consideration of the functions of the excavated rooms.
Late Iron Age Traditions: Courtyard Houses
After the first period of urban experimentation during the late sixth and fifth centuries BC, houses became more uniform and compact, tightly organised in housing blocks separated by streets, which in turn were laid out more or less concentrically within the sites (Figure 29.8). Their inhabitants did not entirely forget traditional building types and techniques, as is shown by a fourth century house in Lattes that was built of perishable materials, even if it was also part of a rectangular housing block (Roux and Chabal 1996).
Figure 29.8. Aerial view of Lattes (Michel Py).
By and large, houses generally remained unchanged in every site throughout the fifth and the third centuries BC. Plans, building techniques and finishing of the houses were relatively homogeneous within any given settlement. At the same time, houses gradually became larger, as the number of rooms increased and spaces fulfilled ever more specialised functions; as a corollary, it became increasingly less common for several activities to take place in one and the same room. In Lattes, the variability of house types that was already evident in the fifth and the fourth centuries BC became more marked in the third century BC, as certain neighbourhoods were given a distinct appearance, such as zones 30–35 with their shell decorations, and house 301 in zone 3 with coloured stone incrustations (de Chazelles 1990: 118; Belarte and Py 2004: 387–88).
The increased diversity implies social differentiation, especially because it may be related to a growing disparity in settlement types and functions. From the fifth century BC onwards, new and different kinds of settlement appeared, the significance of which has perhaps not been sufficiently appreciated. Among the Late Iron Age sites, a distinction may be made between major urban centres that measure around 10 ha, such as Arles, Lattes or Nîmes, and small towns such as Verduron near Marseille of 3000 sq m (Garcia 2004: 140–42) and small farms such as Coudounèuof just 1000 sq m (Verdin et al. 1996–97: 167). This suggests a hierarchical organisation of settlement and economically specialised sites that were probably controlled by a specific group of people, who presumably occupied the larger houses.
It has also been proposed that pre-established measurement modules were used in some settlements such as Lattes and Les Castels de Nages in order to distribute the available space evenly among households (Py 1978: 149; Tréziny 1989: 39 and 41; Garcia 1996; 1999; 2004) (Figure 29.9). That does not mean, however, that it was impossible to extend a house at the expense of public space (Belarte 2009: 240). In Lattes, for instance, there were several houses of one- or two-roofed rooms, where an adjacent part of the street was transformed into a yard, which would eventually be turned into a covered room of the house, thus transforming public into private space. House 406 is a case in point, as it was transformed into house 409/410 in the fourth century BC (Lebeaupin 1994: 35–62; Py 1996: 177–83) (Figure 29.10). In Nages, all houses were expanded in the early second century BC by roughly the same proportion (Py 1978: 153–55). While this did not change local social differences, the transformations do denote a desire for more space, perhaps because of a changing conception of domestic life that saw previously collective activities transferred to the private domain.
Figure 29.9. Schematic plans of blocks 30–35 at Lattes (a), second village of Martigues (b) and sector A of Nages about 175 BC (c) (Maria Carme Belarte 2004).
Figure 29.10. Transformation of house 406 at Lattes (Hérault) from a two-roomed dwelling with a front yard into house 410, a house with three-roofed rooms, during the fourth century BC (Michel Py).
From the third century BC, rooms and spaces in general were used for increasingly specialised functions, in particular in long-occupied sites such as Lattes and Pech Maho. In the aforementioned house 301 at Lattes, for instance, dated to 225–200 BC and with an overall size of 147 sq m, the four rooms fulfilled complementary functions: one large room was dedicated to food preparation, a second one with three benches to food consumption and the other two were used as living rooms (de Chazelles 1990: 115–25; Py 1996: 170). House 58A-58B-58E at Pech Maho, with a total floor surface of 133.4 sq m and of the same date as house 301 in Lattes, is made up of three rooms, where 58A is a multifunctional space, 58B is a storage room also used for milling or roasting grain and 58E is a space for cooking and eating, i.e. with domestic and social functions (Gailledrat and Belarte 2002: 601; Belarte et al. 2011: 65–66) (Figure 29.11).
Figure 29.11. Plan of house 58A-58B-58E at Pech Maho (Sigean, Aude) (Eric Gailledrat).
Alongside these changes of house size and specialised use of space, building techniques and domestic features such as hearths, ovens, benches, storage pits and so on remained unchanged, which underscores the basic continuity of traditions throughout the Iron Age, even if some Mediterranean influence may be noted in some houses (Belarte 2004: 383). Basically, the protohistoric house was and remained a centre of daily life throughout the Iron Age, where domestic activities such as meal preparation and consumption took place, together with economic activities such as storage, and small-scale craftwork, social life and occasionally ritual celebrations. That does not mean that nothing changed, of course, and one variation that has been noted is that gradually more space was allocated to storage. In Lattes, this is first evident in the third century BC and even more so in the second and first centuries BC, when it became one of the main domestic functions alongside cooking and consumption (Belarte 2004: 382). Several houses even included a room exclusively dedicated to grain storage. Since such quantities clearly exceed those of household consumption, it is likely that much of the grain was exchanged (Py 2009: 214–15). This transformation therefore not only points to an increase and intensification of agricultural production but presumably also made the household wealthier and gave them more means to acquire other goods. More generally, the third century BC is a period of increasing specialised household production (Belarte 2004: 363).
Another development, also well attested in Lattes, is the gradual disappearance of cooking ovens that had been common throughout the fifth and fourth centuries BC (Py et al. 1992: 280). They were usually situated in open spaces and front yards, or even in a front room close to the door, which both served ventilation and made it easy to share the oven and its products. The reduced number of ovens and the relocation of existing ones to enclosed spaces such as a central courtyard underline once more the increasing importance of private space. It might also be another instance of specialised domestic production.
Mediterranean Traditions: Courtyard Houses in Southern France
Courtyards or more generally uncovered spaces constituted an important feature of domestic life in southern Gaul from the early Iron Age onwards. They took on a new dimension in the third century BC, when they were integrated into large dwellings, appropriately known as courtyard houses, which differed from other houses in size and plan. They are basically compact buildings with an internal courtyard surrounded by several covered rooms and with the courtyard as the central domestic space, it is evident that they were based on a rather different concept of domestic life than that of the earlier houses with a frontal yard.
Courtyard houses are usually regarded as derived from Hellenistic models elsewhere in the Mediterranean, Italy in particular (Py 1996: 249). Early instances are the houses in Arles and Béziers already discussed, and they became widespread in southern Gaul between the third and first centuries BC, with examples recorded in Lattes, Glanum, Ensérune, Entremont, Saint-Blaise, Marseille and probably Nîmes. Based on plan, origins and formation processes, three main groups of courtyard houses may be distinguished (Belarte 2009: 247) (Figure 29.12).
Figure 29.12. Schematic plans of courtyard houses (type 1: 1–2; type 2: 3–4; type 3: 5–7): 1. Béziers, 2. Marseille, 3. House 5201 at Lattes, 4. House A in Block 10 at Ensérune, 5. House 901 at Lattes, 6. House 3501 at Lattes, 7. Block X at Entremont (Maria Carme Belarte).
Type 1 courtyard houses were built in a context of strong Greek influence, as the previously described house 1 at Béziers (Figure 29.12: 1) and the houses of the ‘Jardin d’Hiver’ in Arles that date to the fifth century BC. A much later example datable to 250–50 BC has been documented in Marseille, and its excavators insist on the Greek origins of the house plan (Conche 2001: 134) (Figure 29.12: 2). The building has three wings surrounding a courtyard, covering an area of almost 400 sq m. The southern wing, made up of four rooms, had an artisanal function, probably metalworking, while the western one included a room with a paved opus signinum mosaic floor that has been interpreted as an andron or ‘men’s room’; the northern wing fulfilled domestic functions. Although Marseille has yielded little evidence for the fifth and fourth centuries BC, it has been assumed that older predecessors of this building are likely to have existed.
Type 2 courtyard houses are found in indigenous settings. They were vast buildings with regular plans and large courtyards. The oldest examples have been recorded in Lattes, where they are dated to the third century BC. The most extensively excavated ones are buildings 52101 and 54101 that feature indigenous building techniques such as stone walls, mud-bricks, earthen or pebble floors (Figures 29.12: 3, and 29.13). All rooms explored were used for a range of activities, as they included both hearths and pits as well as domestic waste, although there was also one storage room with circular pits for supporting storage vessels (Dietler et al. 2008: 122).
Figure 29.13. Virtual reconstruction of courtyard houses 5201 and 5401 at Lattes (Michel Py).
Several type 2 courtyard houses are known in Glanum, including the so-called ‘House of the Antae’ that was first built in the second century BC (Van de Voort 1991). It was built with large, well-cut stone ashlars, which are without precedent in indigenous architecture. It occupied nearly 800 sq m, including a 240 sq m internal courtyard. At Ensérune, house A of housing block X is a later example of this type that was built in the second half of the first century BC. It covered 525 sq m and was organised around a central courtyard that included a large cistern and that was surrounded by a covered porch through which the rooms of the house could be accessed. The roof was covered with Roman tiles, but the walls, floors and plasters were all executed in indigenous protohistoric traditions (Gallet de Santerre 1968: 41–56) (Figure 29.12: 4).
Type 3 courtyard houses are the result of existing houses that were reorganised and combined. They are accordingly characterised by irregular plans, and an interior courtyard is more or less the only feature that houses of this group have in common. The courtyard is even not always located centrally in the house. In Lattes, three examples of this type have been recorded, namely 901, 1605 and 3501 (Figure 29.12: 5 and 6). They are dated to the second century BC and were created by combining two buildings that were part of two different housing blocks. They were originally separated by a street, which was partly transformed into the courtyard. These houses are likely to have been inspired by older type 2 houses in Lattes and may therefore be regarded as denoting similar changes in lifestyle. Another example of this type of courtyard house has been documented in Entremont, where housing block X (site 2), dated to 150–130 BC, includes a house of several rooms around a small central courtyard (Figure 29.12: 7). There is evidence of regular domestic life throughout this house but also of several specialised activities (Arcelin 1987: 71). Housing block 8 at the same site possibly had a similar organisation. At Saint-Blaise, finally, several houses in housing block 2 dated to 175–125 BC also had their rooms laid out around a courtyard (Bouloumié 1992; Arcelin 2004: 254–55). A defining feature of the type 3 courtyard houses is that the rooms appear to be evenly distributed around the courtyard, while there was actually a hierarchy of spaces as a result of the history of the house (Pinon 1999: 255). The type 1 and 2 courtyard houses were rather more regularly organised.
The appearance of courtyard houses has usually been interpreted as the result of acculturation under Greek or Italic influence (Py 1996: 250). Because many Greek houses between the fifth and third century BC included an open courtyard surrounded by rooms (Zacaria Ruggiu 1995: 291; Nevett 1999: 23–24; 2005: 3), it has generally been assumed that this concept had been adopted in the Greek colonies or indigenous sites with foreign inhabitants in southern France. As already noted, Greek influence seems quite plausible for the type 1 courtyard houses.
Greek or Italic influence could also explain the appearance of the type 2 and 3 courtyard houses from the second century BC onwards, as they could have been adapted from the atrium of contemporary Italic houses. There were in any case many local adaptations, as the symmetry and axiality that are typical of the Italic houses (Zacaria Ruggiu 1995: 358; Fernández Vega 1999) is unknown in the southern French ones. From a functional perspective, it is equally difficult to see clear connections between the indigenous courtyard houses and their Italic counterparts. The construction techniques, finishing and minor domestic features of the southern French houses are moreover predominantly indigenous (Py 1996: 250); the only evident exception is the so-called ‘House of the Antae’ in Glanum with its ashlar walls of a more obviously Greek type. It is therefore just the exterior form (plans) that was adopted from elsewhere and adapted to the conception and use of space of the indigenous world of southern France.
To sum up, the complex house types of southern Gaul developed in two phases: in the first one, beginning in the fifth century BC, several courtyard houses were built under Greek influence; in the second phase, indigenous versions were created that followed either Hellenic or Roman models. In Lattes, the large type 2 houses were first built at the beginning of the third century BC, and they cannot therefore be attributed to Roman influence. It is, however, possible that they were adapted from Hellenistic houses in Marseille, as Lattes maintained close contacts with the city.
The Use of Space in Courtyard Houses
The central courtyard houses differed from their protohistoric counterparts in both ground plan and conception of space. In the latter tradition, courtyards opened onto streets or were shared by two houses, and some domestic activities took place in these (semi-)public or collective spaces. The courtyard houses were by contrast inward-looking, as the courtyards sheltered the occupants from the outside and thus provided privacy.
From a functional point of view, one might expect a more specialised use of space in the larger courtyard houses with more rooms, but in most cases it has been difficult to assign clear functions to rooms because the available evidence is mostly rather limited as a consequence of either incomplete records of old excavations or simply poor preservation. It is nevertheless relevant to note that most courtyards had access to good light and ventilation and often had water management structures such as gutters and cisterns, whereas the front yards of the older houses were mostly used for cooking and had generally been more closed off; the presence of hearths confirms that point. By and large, the use of space varied considerably between houses, and most rooms appear to have been multi-purpose ones. In Lattes, for example, domestic features such as hearths and pits and similar artefacts occur in more than one room and suggest that similar activities took place in most rooms. Storage was the only exception to this apparent rule. The conclusion, therefore, has to be that the adoption of a new house type with a different spatial organisation did not necessarily entail a different use and functional organisation of domestic spaces, at least as far as the available archaeology evidence suggests.
Complex Houses and their Occupants: Households and Social Organisation
The variability of ground plans, sizes and increasing segregation and specialisation of space in the Late Iron Age are arguably related to social changes. One interpretation is that Iron Age society in southern Gaul was egalitarian, given the homogeneous forms and dimensions of domestic architecture (Arcelin 2004: 241). In this view, it is only by the Late Iron Age, in particular from the third century BC onwards, that a process of social differentiation began, initially prompted by Italic trade and later, during the second and first centuries BC, influenced by Roman presence.
The evidence of domestic architecture suggests, however, that social differentiation and elite formation predated Roman contacts. At Lattes, for example, the first complex houses indicate social differences already in the third century BC, before trade with Italy became important. Economic or political factors would therefore seem to have played a more important role, depending on the nature of the sites concerned. Social changes during the Late Iron Age are also evident in the different relationship between public and private spaces and indeed the ability of elites to occupy parts of public space (Belarte 2004: 383; Dietler et al. 2008: 120). The latter is made most evident by the type 3 courtyard houses, as most of these blocked off a street; but it is not confined to these houses, as it has also been noted in fourth-century Lattes and Nages, where streets were narrowed if not entirely closed by expanding houses.
Comparable developments have been observed further south along the Iberian coast, where complex dwellings were built in Valencia as early as the fifth century BC and in Catalonia during the fourth and the third centuries BC. The largest of these has been brought to light in Ullastret: it is a house of about 800 sq m, which was created by combining two pre-existing houses. Its construction also required a street to be closed off and even appropriated access to the rampart. This house has been interpreted as the residence of a large aristocratic household (Martín et al. 2004: 266).
Complex houses may well reflect external influences, but that in itself does not adequately explain why they were built in the first place, and changing use of space is more readily related to social complexity (Kent 1990). From this perspective, I would therefore propose three distinct, if not exclusive, interpretations of the variability of house plans and the emergence of complex houses, which intersect with my view of Iberian houses (Belarte 2013: 87–89).
The first one is that larger houses such as the courtyard ones were occupied by more people. Combining two pre-existing buildings, as in the type 3 courtyard houses, substantially increased the available domestic space, and the implication of a wish or perceived need to do so implies that the household itself had expanded and required more space for its members. The increase of space could, however, only be proportional to the increase of household members, if each and every household member always had access to the same space – and ethnographic evidence suggests that that is usually not the case – because large households include dependants, or some occupants have more priviliges, for instance through polygamy (Van der Berghe 1979: 163; Netting 1982: 657). We have, however, no textual or archaeological indications that this was the case in southern Gaul.
Another possibility is that the larger houses belonged to wealthier people, and ethnographic evidence indicates that correlations may exist between house size, number of rooms and wealth of occupants (Kramer 1979; Horne 1991; Netting 1982). In southern Gaul, however, the archaeological evidence of the large courtyard houses is unfortunately insufficient to decide whether their inhabitants were richer or were involved in different economic activities than the rest of the population. It is in the end only the house form and sizes that make them stand out.
The two interpretations may also complement each other, as the households with more members, even if some of them were dependants, could well also have been the wealthier ones, who also lived in larger and complex houses. Ethnographic evidence suggests once more that elites often have larger households because they control more resources and take care of specialised functions of an administrative, social or economic nature (Van den Berghe 1979: 163; Flannery 2002: 425). It would therefore not be unreasonable to interpret the complex houses in the major sites of southern France as belonging to large elite households.
A third view is that differences in the size and structure of houses reflect different stages in the domestic cycle of the household occupying the house (Tourtellot 1988; Gerritsen 1999: 81–82; Goodman 1999; Normak 2009: 239). The simpler dwellings may thus correspond to households in the initial stage of their formation, while the more complex ones could be the result of rooms added to accommodate new household members. Combining two pre-existing houses to create type 3 courtyard houses may for instance also imply the joining of two previously separate households through marriage and the birth of a child.
Neither Classical texts nor archaeological evidence can, however, inform us adequately about family structure and socio-economic organisation of protohistoric communities in southern Gaul. We are therefore left to conclude that the increased domestic space in courtyard houses would have accommodated a larger number of occupants, and that the many rooms suggest that the household was divided into smaller units. It is thus not too far-fetched to propose that the courtyard houses were occupied by extended households that in turn were made up by two or more nuclear families. This has been argued in some detail for the courtyard houses in Lattes (Dietler et al. 2008: 122) and for certain Greek houses in Delos (Luce 2002: 86–87). The increased significance of privacy in these houses might also be related to cohabitation of distinct families.
The appearance of complex houses and courtyard houses in particular should also be examined in the light of social and economic organisation of indigenous communities. The third century BC is a key moment in their development, when social and economic differences began to emerge and local elites first asserted themselves politically and culturally (Arcelin and Tréziny 1990: 29; Py 1993: 153–57). These elites were based in the major settlements that played a central role in the wider region, and building large residences is likely to have offered a strategy to distinguish themselves (Arcelin 2004). The roots of this presumably predate the third century BC, and the changing domestic architecture of Lattes should probably be seen in this light (Garcia 2004: 89).
Final Remarks
This chapter is the result of ongoing research of domestic contexts in southern Gaul. Despite significant external influences, indigenous architectural traditions of domestic housing, urban organisation and the use of space were preserved in one way or another throughout the period examined. The changes observed in architectural forms and use of space were related to wider social transformations, and the general increase in the complexity of domestic architecture of the Late Iron Age seems to reflect increasing social complexity. At the same time, the appearance of social differences and probably of certain changes in household organisation may have influenced the use of domestic space, as is suggested by the creation of new plans and changing attitudes towards private space. This process is directly related to the emergence of local elites and the creation of indigenous sociopolitical polities.
Acknowledgements
I would like to thank Michel Py, Eric Gailledrat, Bernard Dedet and Jean Chausserie-Laprée for the illustrations provided for this chapter.