Introduction

1The site of Saint-Martin-le-Bas has been the subject of extensive research carried out within the framework of the Collective Research Programme “Les ports antiques de Narbonne” since 2011, the aim of which has been to determine the nature of the occupation and the use of this coastal area over the long term. In addition to the installations of the High-Empire, probably linked to the port system of Narbonne (Sanchez, Duperron 2021), the site contains a funerary occupation developing from late Antiquity to the Middle Ages.1 Based on the recent discoveries made in 2021 and the results of radiocarbon dating of several individuals, two main lines of thought have emerged: How is the logic of the initial distribution of burials defined in time? And how can the analysis of the funerary area allow us to better characterise the nature of the occupation of the coastal site of Saint-Martin-le-Bas in the Narbonne area during Late Antiquity?

1. The coastal zone of Saint-Martin-le-Bas across time

2The coastal zone of Saint-Martin-le-Bas is located around a dozen kilometres to the south-east of Narbonne in the current town of Gruissan. It occupies a strategic position atop an island situated within the marshlands of Narbonne close to a canal connecting the lagoon to the sea.

3The ancient occupation extended over more than a hectare (fig. 1) and archaeological studies have revealed about 3,000m² at present which are spread out over two areas: the first corresponds to the central base of the roman port area, while the second, more recently occupied, sits atop the highest part of a small rocky outcropping situated next to the edge of the lagoon. The site also contains a large and consistent fresh water source which bolstered its strategic value.

Fig. 1. Location of the excavated sites at St. Martin-le-Bas.

4The establishment of the settlement took place during the Late Republican period towards the end of the 2nd c. BCE and the site developed for more than a century It underwent a significant restructuring at the beginning of the 1st c. CE, marked by the construction of a monumental complex organised around a courtyard with porticoes in its central area (fig. 2; Duperron, Mauné 2016 ; Sanchez, Duperron 2021). To the south of this, near the coast, a second complex of almost 1,000m², currently being excavated, was built in the same period. It was organised around a large structure and boasted rich decoration; it also had two large underground cisterns on its periphery, no doubt constituting reserves of potable water for ships. The analysis of these remains tends to suggest that they belonged to a complex of a public nature, most likely in connection with the system of ports in Narbonne. The use of these large installations is well documented until the end of the 2nd c. CE, but the nature of the activities carried out in this area during the 3rd and 4th c. remains poorly understood at present.

Fig. 2. The two excavation areas and the Early Empire settlements.

5A new occupation then developed progressively during the 4th c. and, according to survey data, expanded to at least 1ha from the 5th c. onwards. The excavations made it possible to understand two different aspects of the Late Antique occupation: the first was the living areas and the zones of economic activity, and the second was a funerary area (fig. 3). The inhabited parts appear to have been on the northern side of the site near the fresh water source during this period, but this area has not yet been investigated.2

Fig. 3. The configuration of the site in Late Antiquity according to current knowledge.

6On the central parcel, inside the Early Empire building with porticoes, the Late Antique occupation is notably characterised by the installation of a silaging area consisting of about 30 silos and 20 dump pits, mostly of large dimension. The number of these structures and the abundance of material found reflect the existence of a large population and intensive economic activity during the 5th and 6th c. The investigation of these buildings revealed critical data on the lifestyle and the economic activities of the site’s inhabitants (Duperron 2020). The exploitation of the coastal resources appears to have been of paramount importance, as shown by the large number of seashells in each of the deposits and the frequent discovery of hooks and net weights, as well as the many fish bones and vertebrae of marine mammals. The raising of livestock is also well attested and the archaeobotanical data indicates that grape vines and olive trees were cultivated as well. Finally, the amphorae and other ceramics testify to a great deal of long-distance commercial connections around the Mediterranean. The consumption of these imported products shows a population with a significant income. In this way, the material data points to a population that was able to exploit all of its local resources, but also had access to imported goods.

7While there is some indication in the material record of an occupation that continued into the 7th century, this cannot be as readily seen as that of Late Antiquity.

8The object of our study was the occupation established in this period of Late Antiquity on the rocky outcropping in the immediate vicinity of the coast, about 50 m to the south of the central parcel (fig. 4). A multi-year excavation has made it possible to investigate the funerary complex associated with this Late Antique occupation. In the southern area, the occupation continued until the Middle Ages. In fact, in the 10th c., several buildings, including an imposing tower, were erected in the centre of the excavation area, after extensive earthworks. These works destroyed most of the built structures and post Early Empire occupation levels in this area. It is also possible that they destroyed some of Late Antique burials, particularly those less deeply buried (possibly perinatal individuals in particular; fig. 5). Others large scale earthworks, during the later Middle Ages and the Modern period, have had a profound impact on the earlier structures, partially destroying the Late Antique stratigraphy and cutting into the earlier assemblages in the southern part of the excavation area. In this southern portion, the upper parts of the funerary structures have therefore suffered damage: slab lids or formwork walls have been torn away and skeletons have sometimes been partially disturbed. Several burials have also been looted, making it impossible to date them any more precisely than to the Middle Ages or the Modern period.

Fig. 4. General view of the site: Early Empire structures, Late Antiquity and medieval burials.

Fig. 5. Impact of medieval and modern construction.

2. Occupational phases of the funerary area

9The funerary occupation, presents a total of 38 primary burials and one secondary ossuary deposit (FS5179) plus bones in secondary position in three stratigraphic units (FS5034, US5647, US5719). The occurrence of looting as well as the damage to some of the funerary structures at the southern edge of the site, combined with an almost total absence of grave goods, prompted us to carry out radiocarbon dating on the bone fragments in order to establish a precise occupational chronology. At present, 11 samples have been analysed, indicating a long funerary occupation (fig. 6).

Fig. 6. Phases of occupation of the funerary area.

10The first burials were established in the south-western part of the excavation area at the end of the 3rd or 4th c. for SP5014 and towards the second half of the 4th or 5th c. for SP5214. The excavation operations in 2021 made it possible to locate and expose all the funerary structures present at the site. In particular, a series of burials located in the southern part of the area, until then unknown, required us to completely rethink our interpretation of the different occupational phases of the funerary area. Indeed, we found that after the first graves of the 3rd and 4th c., the occupation seems to have moved slightly towards the east after two sarcophagi were put in place during the 5th or 6th c., which soon attracted other burials around them (conf. infra chapter 4).

11In the 7th c. (and perhaps still in the 8th century?), the funerary occupation seems to extend again, this time in the north-western part of the site, west of wall MR 5443, which was not built until the end of the 5th c. at the earliest. This wall seems to mark the eastern limit of this third phase of occupation. Finally, in a fourth phase lasting into the 13th c., the funerary activity in the area seems to have undergone a final topographical displacement, this time in the eastern part of the site, east of the tower that was erected in the 10th c. Only two graves belong to this phase and are not contemporary with each other: SP5545 dates from the end of the 10th c. or the first half of the 11th c., while SP5500 dates from the 13th c. (for both the individual in primary position and the reduction).

3. The burials: burial methods and biological data

12It can therefore be seen through these four phases of occupation that funerary activity at the site, following the conversion of the buildings at the port during the 3rd c., continued to develop and to sustain itself for almost a millennium. It is therefore a human occupation that seems to have remained significant after the Early Empire and a likely large complex settlement which remains for us to explore.

3.1. Burial methods

13Burial methods vary across the site according to developments in funerary practice over time and the various discourses they represent.

14During the first phase of occupation at the end of the 3rd and into the 4th c., the burials consisted of wooden coffins or coffrages, or structures made of stone and wood (fig. 7, no 1). These types of perishable materials were used only during this first phase and were not found in later burials. The coffrages made of stone slabs became increasingly common later on and were used throughout the site until the funerary area fell out of use (fig. 7, nos 2-3). The second phase of the occupation presents a particularity: it was probably in the 5th or 6th c. that the two sarcophagi SP5760 and SP5770 were used (fig. 7, no 2). The use of such special architecture for these two burials reflects their special topographical position within a building and testifies to the elevated status of their occupants.

Fig. 7. Types of burial used in the different phases.

15While burial methods can vary over time or indicate a particular position of the dead can be observed, there is nevertheless a homogeneity in the orientation of all the burials, in the similar positions adopted for the deposit of the bodies and in the scarcity of furniture. The only slight variations in the position of the bodies are the result of pragmatic adjustments, for example in the case of an architecture ill-suited to the size of the dead (notably SP5572, SP5545). In all of the structures, all oriented east to west regardless of the period, the dead were deposited without any artefacts, excepting a bracelet in SP5699 and three rings made of copper alloy in SP5014, SP5563 et SP5699. These items were worn and therefore part of the clothing of the dead. In fact, all the phases of use show individuals with a supple envelope (highlighted by the taphonomic study, which does not allow us to distinguish between clothing and shroud in these cases) and/or footwear. In detail, material elements confirm the presence of shoe nails for SP5014 (3rd-4th c.), a belt buckle for SP5039 and a bronze fibula for SP5041 (7th-8th c.?). In addition, burials SP5572 and SP5500 (7th-8th c. and 13th c. respectively) revealed small fragments of tissue still preserved in contact with the bones. The taphonomic analysis revealed that 5 individuals (SP5572, 5545, 5122, 5222 and 5319) wore soft shoes and 7 other individuals (SP5104, 5195, 5214, 5017, 5500, 5572, 5319) wore a supple envelope (garment and/or shroud).

16Thus, at least 13 individuals are likely to be clothed and/or shod: 7 individuals clothed with a soft envelope (garment and/or shroud), 4 individuals wearing soft shoes and 2 individuals with both a soft envelope and soft shoes (SP5572 and SP5319).

17There was no difference in terms of the burial methods adopted when it came to the sex of the subjects or their age at the time of death: children and adults, women and men were given the benefit of a proper burial, which remained the same throughout the site and across time: elements of clothing and/or shroud, deep graves, developed funerary architectures, massive and systematized covering system and very likely signage elements.

3.2. Population sample

18The population sample currently uncovered in the primary burials consists of a total of 26 adults and 14 immatures, i.e. 65% adults and 35% immatures, which is consistent with what is expected for an ancient population (tabl. 1).

Table 1. Sex and age at death of individuals for the total sample.

19Of the adult individuals, eight were female, ten were male and eight subjects remained without a sex attribution (inconclusive result or coxal bones too deteriorated to allow diagnosis). The sex ratio is therefore balanced (44.5% female and 55.5% male) and shows that there was no selection by sex (tabl. 1). As expected with age estimation methods, which are necessarily imprecise in order to remain reliable (Schmitt 2001, 2005, 2008), there is a strong representation of mature individuals (eight individuals between 30 and 59 years of age, i.e. 44.5% of the individuals whose age could be estimated). Nevertheless, six subjects are young (20-29 years) and four are old (over 60 years). The representation of young subjects is therefore normal for an ancient population (33.3%). It should be noted that there is a significant presence of elderly individuals (22.2%), contrary to the generally accepted idea of the populations of historical periods (tabl. 1).

20The distribution of non-adult individuals in the different age classes is in line with the expected mortality curves for an ancient population as far as individuals over 5 years old are concerned (tabl. 1). On the other hand, it is abnormal from a demographic point of view for the youngest individuals (fetuses, 0-1 year old and 1-4 years old): the sample includes only one fetus, 2 subjects who died before 1 year old and none who died between 1 and 4 years old, even though these age groups are the ones in which children would die the most. This deficit in the youngest subjects may be due to several factors. It is not uncommon in ancient populations for very young children to be buried elsewhere. On the other hand, as mentioned above, the disturbance of the highest stratigraphic levels by medieval and modern earthworks may have led to the destruction of these burials, which were often less deeply dug. It is not possible to know which of these two hypotheses prevails here, one not excluding the other.

21The study of the health status of the individuals revealed them to be in the same overall condition with very few of the usual indicators: trauma and degenerative diseases (associated more with the age of the subjects when present) were rare and infectious diseases were almost completely absent, as were markers of stress. The only notable indicator was the marker of activity on several male subjects: the horseman syndrome. What must be noted is how this overall state of health was maintained during the entire occupational period of the funerary area: there is no deterioration in the health of the subjects over time, between Antiquity and the Middle Ages. This element reinforces the idea of a perennial occupation whose importance is maintained over time, ensuring a stability of resources over more than a millennium.

22If we look at the population sample of Late Antiquity, i.e. the subjects of the first two phases of occupation between the 3rd and 6th c., we see that the distribution between adults and non-adults is less balanced (tabl. 2). Indeed, out of a total of 17 individuals, only four are non-adults (23.5%) while 13 are adults. Nevertheless, this rate remains acceptable and is perhaps due to the smallness of the sample and the destruction of certain burials of young children, as mentioned above, since only one child under the age of 4 is present (tabl. 2). The number of non-adult subjects is too small to allow further analysis.

Table 2. Sex and age at death of individuals for the Late Antiquity sample (phases 1 and 2, 3rd-6th c.).

23The sex of three adult individuals could not be determined; for the other ten, there are five females and five males, a perfectly balanced sex ratio, showing that in these phases as in the later phases, there is no selection by sex (tabl. 2). The distribution according to age at death also shows that there was no selection according to this criterion, as young subjects are largely present and one elderly subject was also identified (tabl. 2).

4. The Late Antiquity occupation: some founding privileged burials?

24As mentioned above, after an initial phase of settlement of burials during the 3rd and 4th c. that seem to have been confined to the south-west of the excavation area, a new phase of the occupation developed to the south of the site. The second phase could begin with the installation of two sarcophagi—burials of exceptional quality in comparison to the others at the site. Apart from this unusual funerary arrangement, it is the topographical and stratigraphic position of these graves that helps us to better understand the origin and the evolution of the funerary occupation during the Late Antique period.

25In particular, a first sarcophagus SP5760 was appended to the south of a wall dating to the Early Empire. This type of burial was not used in the previous phase and therefore carries particular significance due to its nature and its installation in a small rectangular building. Its north wall is made up of blocks probably installed during Late Antiquity on a base of a wall from the Early Empire (fig. 8). The western and eastern walls could be recognised from their foundations or their negatives; only the southern extent of the structure is unknown. The dimensions of the building are 3.80m east-west and probably about 5.50m north-south. In addition to its probable elevation in large blocks, its internal space has a masonry floor made of tiled concrete installed on a raft. SP5760, which seems to be the founding grave of the 5th-6th c. occupation, quickly attracts other burials of an exceptional nature: a second sarcophagus SP5770, of large dimensions, is adjoined to the south of this first, its installation pit intersecting the pit of SP5760. Then a third pit, much wider, intersected with the pit of SP5770 to the south in the same way. This received a double coffrage SP5762-SP5812 (the two containers were installed at the same time and their lids carry the same sealing mortar). The desire to be as close as possible to the two sarcophagi cannot be mistaken because the northern wall of the double coffrage is in fact constituted by the southern side of sarcophagus SP5770. Moreover, the lid of SP5812 covers that of sarcophagus SP5770. These four tombs constitute a complete set, enclosed within the rectangular building and covered by a masonry floor, that lies at the origin of the development of an extensive occupation during the 5th and 6th c. in the southern part of the site.

Fig. 8. Detail of the sarcophagi area, with the extent of the early monument.

26In this occupation, the intersection of burials can only rarely be seen, despite the long period of use of this burial are, suggesting that a memory of the graves may have remained, probably on account of a perennial sign or marker, now gone.

27One of the interpretative hypotheses currently being considered is that this building may have led to the creation of a place of worship in the immediate vicinity of these privileged tombs, unless the graves were installed shortly after the foundation of this place of worship, a case perfectly documented in Sagone in the 5th c. (Istria 2021; Corbara et al., in this volume). Two hypotheses are envisaged for the location of this place of worship. The first places it around the rectangular building, which it would then encompass and with an internal burial zone. The second possibility envisages it immediately adjacent to the north of the early rectangular building, which would explain the central area containing no burials, with an external burial zone (fig. 9). This second hypothesis is also based on the existence of wall bases from the Early Empire belonging to a quadrangular and perfectly oriented building. It seems quite possible, although difficult to demonstrate, that this building could have been opportunistically reused as a church, without requiring significant architectural modifications. All the elevations and associated levels of occupation would have been subsequently destroyed when the 10th c. tower and associated buildings were erected.

Fig. 9. Spatial organisation of the 5th-6th c. occupation: possible locations for a church, which may or may not include the burial space.

5. The funerary area of Saint-Martin-le-Bas in its geographical and chronological context

28The recent attempts at investigation of the site require the comparison of our burial area with other contemporary cemeteries in both rural and urban contexts, in order to propose new interpretative hypotheses.

29The funerary occupations discovered at Vinassan (Le Pech de Tardieu, ASF aire Nord, Aude, 5th to 7th c.; Gaillard et al. 2013), only 15km from Gruissan, as well as those at Malbosc in Montpellier (4th to 5th c.; Blaizot et al. 2008) are not very extensive funerary areas, whose organisation in rows is the result of small family-type communities, attached to a small rural settlement (fig. 10). This differs markedly from the St-Martin-le-Bas site. They do not reveal the use of sarcophagi, consisting only of burials in amphorae, wooden coffins or coffrages and stone slabs coffrages.

Fig. 10. Comparison with some rural sites.

30While much denser rural funerary occupations are known, such as those unearthed on the outskirts of the small rural settlement of Lunel-Viel, at les Horts (6th to 7th c.) and Verdier (late 3rd to 6th c.), the spatial organisation of the burials differs from that of Saint-Martin-le-Bas: the burial areas were organised into clearly identifiable rows (Raynaud 2010: 98-99). At the site of les Horts, they seem to have followed at first the ancient development of the road and ditch systems (fig. 10). This type of spatial organisation cannot be found at our site. When it comes to burial methods, among the numerous graves at Verdier site, arrangements are various in the 4th c., including wooden coffins, triangular section tile coffrages (“coffrages en bâtière de tuiles”) and burials in amphorae. There followed a slow evolution towards the widespread use of wooden coffins and coffrages during the 5th and 6th c. While two cases of wooden coffrages were found at Saint-Martin-le-Bas (SP5014 and SP5214), triangular section tile coffrages and burials in amphorae were completely absent at Saint-Martin-le-Bas during the same period. Moreover, the coffrages made of complete container of slabs and the sarcophagi came into use much later at Lunel-Viel: were completely absent at Verdier, and do not appear until the 6th c. at les Horts and Saint-Vincent sites (Raynaud 2010: 118-119).

31Although geographically more distant, the development mechanism of the funerary space of Vindrac in Tarn over the long term (6th to 13th c.) seems similar to that of our site, although the overall organisation is not identical (Guyon et al. 2015: 15-105). An important villa-type rural settlement of the Early Empire, the site was then reinvested for funerary purposes. In fact, at the very beginning of the 6th century, it was marked by the construction of two, perhaps three, small buildings around which a large funerary space quickly developed (fig. 10). Some of these small buildings are the result of a complete construction, while others have taken over entire sections of the antic villa in their architecture. This occupation took hold and lasted until the end of the 13th century, being enveloped by later structures that grew increasingly large. Among the 150 graves, many sarcophagi were used and are typologically similar to those at our site.

32The layout of the Maguelone site is, for its part, singularly similar to that of our site (fig. 10). Located on an island at the interface of a lagoon and the sea, this occupation corresponds to a rural agglomeration and is characterised by the initial foundation of a large Late Antique church around which a funerary area developed from the 6th to the 9th c., containing several hundred burials. The original function of the building here is cultic; the funerary function came later. It should be noted that the interior of the building was only slightly occupied by burials: only 25% of the tombs are found there and they were installed late (from the middle of the 7th c.). The organisation of the cemetery seems to have been dictated by the church and by the later privileged burial SP2175, which was installed in a small building close to the chevet (Garnotel et al. 2019: 189). This is the only proper mausoleum in the burial area and corresponds to the founding tomb of the cemetery. The attraction of the initial nucleus of burials in its immediate environment has indeed been revealed thanks to radiocarbon dating. Finally, in terms of typology, as in Saint-Martin-Le-Bas, a great variety of burial types can be observed, with seven different architectures: wooden coffins, triangular section tile coffrages, rectangular section tile or stone coffrages, etc., but also the more exceptional presence of sarcophagus elements, a large part of which was probably looted later.

33These comparative examples lead us to question the definition we wish to accord to the qualifying term “rural”. Should this term be applied to everything which is not found in the context of a large city? The sites that we just mentioned are qualified as rural because of their location in the countryside, but large variations in terms of location and spatial organisation can be seen among them.

34An examination of the Late Antique sites of Narbonne helps to broaden the question (fig. 11). The site of Saint-Félix contains a basilica as well as graves which were put in place during the 5th c. inside an area that was already being used for funerary purposes in the 4th century. This cultic and funerary occupation follows domestic activities. As at Gruissan, no tegulae coffrages were observed, nor were children's burials reusing amphorae. The Late Antique funerary area is mainly composed of sarcophagi installed within three southern annexes of the basilica, while the interior of the church is empty of burials (Ginouvez 1999: 25-46). In contrast, the basilica in Clos de la Lombarde, also erected in the 5th c., contains graves within the building (Solier 1991). This type of installation, which follows a pre-existing one dating to between the 3rd and 4th c., evokes that of our site, which also had a core area of graves prior to the development of a burial area, during a second phase, within a building containing sarcophagi around the 5th century.

Fig. 11. Comparison with some urban sites of Narbonne.

35The development of larger buildings at later dates within the Narbonne sites leads us to the conclusion that such a hypothesis could hold true at Saint-Martin-le-Bas.

36It should also be noted how the use of sarcophagi is generally found, given the place and period concerned here, preferentially within urban contexts and is often associated with an early cult site of the funerary basilica type. While the sarcophagus found some success from the 2nd c. onwards amongst the urban population, it was absent from rural cemeteries during the 4th and 5th centuries (Raynaud 2010: 37). They only appeared from the 6th c. onwards, as in the funerary spaces of Lunel-Viel where sarcophagi, absent from the Verdier site (3rd-6th c.), numbered seven at les Horts (6th-7th c.) and represent 30% of the garves in the church of Saint-Vincent (6th-19th c.). Without going so far as to say that we are dealing with a privileged population at Saint-Martin-le-Bas, but rather a group with a local elite, the data must be reinterpreted in light of the excavation of the funerary occupation. It shows that we are closer to what we find in urbanised areas of the region or in secondary towns such as Maguelone rather than in modest rural settlements. The proximity of the lagoon site of Saint-Martin-le-Bas to the large urban centre of Narbonne may have played a role in influencing local elites in terms of the topographical and typological layout of the burials.

Conclusion

37At the Saint-Martin-le-Bas site in Gruissan, in the light of the latest results of radiocarbon dating and the latest discoveries, the analysis of the topography and the development of the burials has made it possible to propose a hypothesis on the dynamics of the distribution of the tombs. An initial nucleus located to the south-west of the excavation area reveals the first burials, installed at the end of the 3rd or 4th c., on the ruins of an Early Roman occupation. Then in the 5th or 6th c., graves of an exceptional nature, insofar as they can be clearly identified from the rest of the funerary occupation, were put in place in the southern part of the site. These graves exerted a strong attraction because a very clear desire for later burials to be gathered around the earlier sarcophagus can be observed. This exceptional nature and power of attraction were reinforced by the location of the four graves within a small rectangular building that was based on an ancient wall. The assemblage was clearly at the origin of the aggregation and the later development of the southern burial area dating to the 5th and 6th c. that clustered around the structure. Further research will allow us to determine whether this southern burial area was open or found within an even larger structure.

38Later, a group of burials was established to the north-west of the site and of the hypothetical place of worship during the 7th c. (and perhaps the 8th c.?). Finally, two much later burials (10th and 13th c.) were installed on the eastern fringe of the occupation, chronologically sealing the burial use of the site.

39Moreover, the study of this funerary area is of major importance, in that it contributes to the reflection on the status of the site of Saint-Martin-le-Bas during Late Antiquity. The existence of privileged burials at the origin of a perennial and relatively extensive burial area suggests that we are not in the presence of a simple rural settlement. Indeed, in the light of the important discoveries made on this site, we can now envisage the existence, during the 5th-6th c., of a grouped settlement of some importance, as well as the presence on site of a local elite. In fact, several examples of sites of this type emerging in the 5th c. are known in the Languedoc (Schneider 2007), in particular on the coast with the emblematic case of Maguelone, which received a bishop's seat from the 6th c. onwards (Garnotel et al. 2019). It now remains to determine the exact status of this site: the possibility that it corresponded during Late Antiquity to a domanial centre, rather than a simple rural agglomeration, must be considered. To confirm or deny this hypothesis, the exploration of the contemporary habitat of this burial space is essential. More broadly, an even more extensive approach to the Late Antiquity settlement of Saint-Martin-le-Bas is likely to significantly enrich our knowledge of the modalities of occupation and control of the Narbonne coastline during the transition period between Antiquity and the Middle Ages.