Appuis-tête en terre crue dans la sépulture collective de Mas Rouge (3050-2890 av. n. è. ;Montpellier, France) – version abrégée

1La sépulture collective néolithique finale (3050-2890 BCE, culture de Ferrières) de Mas Rouge à Montpellier (Hérault, France) a pour particularité essentielle une architecture semi-enterrée (que l’on pourrait qualifier de pseudo-hypogée), pour la réalisation de laquelle le matériau terre domine sous toutes ses formes (Tchérémissinoff et al. 2018 ; Leal et al. 2019 ; Leal et al. ce volume). Elle a accueilli plus de 180 individus, aussi bien des femmes que des hommes, ainsi que des immatures de tous âges (voir note 1 de l’article développé). Ils sont répartis dans la tombe en fonction de leurs états, certains espaces sont dévolus aux dépôts primaires, d’autres aux translations et d’autres, enfin, sont des espaces de réduction, voir des espaces ossuaires.

2En plus des éléments structurels propres et des dotations funéraires attendues dans un tel contexte (tels que le mobilier céramique, les parures, les poinçons en os ou les outils lithiques), elle a livré une importante quantité de mobilier en terre crue, de grands contenants fixes pour l’essentiel - type de mobilier par ailleurs rarement décrit (voir note 2) - mais également deux pièces exceptionnelles, des appuis-tête, sujets du présent article (voir note 3).

3Ces dispositifs funéraires en terre crue sont au moins au nombre de deux (reconnus) et sont localisés dans la partie orientale de la sépulture. L’un se situe dans un horizon de dépôt sub-primaire, l’autre est en position secondaire, plus haut, à la limite entre les secteurs A et D, à l’aplomb d’une aire dédiée à la translation (dépôts temporaires) (cf. fig. 1).

4Ainsi, le premier (TC144) a été découvert assez haut. Il n’y a pas d’ossements dans sa proximité directe. Sa fouille et son prélèvement l’ont malheureusement quelque peu endommagé et les prises de vue de terrain restent les meilleures. Il est incomplet et sa surface abîmée. Le second objet (TC420) est bien mieux conservé. Découvert plus bas, il est rattaché à l’horizon du plancher n+0 situé en sous-sol, en secteur A (cf. fig. 7), dédié aux dépôts primaires.

5Le second, l’élément TC420 a chuté avec les cadavres, un environnement fortement organique, et son plancher enduit, mais de moins haut que le précédent, puis s’est retrouvé bloqué par les effondrements supérieurs de la condamnation. De ce fait, il est resté en position logique, que l’on peut même considérer comme fonctionnelle sous le crâne (le cou) d’un sujet adulte. De cette situation particulière, peut être déduite sa fonction comme étant celle d’un appui-tête, interprétation que nous appliquons également au précédent objet en terre TC144 en raison de leur similarité.

6Les deux pièces TC144 et TC420 sont remarquables à plusieurs titres. En premier lieu par leur forme, plus précisément par leur géométrie, qui relève d’une composition complexe (infra).

7En effet, on a affaire ici à un travail sur le volume très abouti avec haut degré de finition impliquant une technique de la ronde-bosse. Il s’agit dans ce cas précis d’un volume prismatique triangulaire oblong, quasi-identique sur les deux pièces. Une décomposition géométrique de la pièce la mieux conservée (TC420) montre que le volume est formé à partir d’une base triangulaire sphérique et qu’il intègre par ailleurs une réduction discrète d’un de ses côtés de manière à générer deux faces trapézoïdales (cf. fig. 12). Les extrémités triangulaires sont très légèrement bombées (conformément au principe du triangle sphérique) et les deux faces de contact conservées sont à profil légèrement concave. Les arêtes sont adoucies ou coupées. La surface est soigneusement lissée à l’eau (il n’y pas de couche d’enduit identifiée en tant que telle).

8En ce qui concerne le matériau, la composante principale de ces deux pièces est un limon carbonaté peu argileux jaune beige. Il inclut une fraction sableuse fine à grossière (avec des grains de quartz et calcaire allant jusqu’à 3 mm de diamètre). Le malaxage est très abouti, produisant un matériau très dense et compact. Des inclusions végétales sont visibles mais très discrètes. Par ailleurs, l’altération thermique est très irrégulière et rien ne distingue ce matériau de celui mis en œuvre à l’échelle de la tombe. On est dans la même famille de sédiment, ce qui est à souligner en termes d’économie globale. Toutefois, seule l’illite et la smectite sont présentes dans la composition du matériau de la pièce TC420.

9Dans le village concomitant (La Cavalade), la découverte de deux objets comparables dans des fosses domestiques pose donc la question de l’ubiquité de tels dispositifs dans le contexte qui nous concerne.

10Sans être partout présents (ou reconnus), on recense des appuis-tête dans les cultures néolithiques de presque tous les continents. Qu’ils soient en terre ou en pierre, ils sont évoqués dans les description des contenus de tombes sur de nombreux sites, mais sans jamais beaucoup de précisions.

11Le site le plus ancien et le mieux renseigné dans la large zone géographique qui nous concerne est le site catalan de Can Cambús (Allièse et al. 2014). Daté entre 3900-3500 cal. BC, ce site est intéressant pour des observations réalisées dans le cadre de sépulture de Mas Rouge à plus d’un titre, puisqu’il a livré de manière récurrentes (mais pour des tombes individuelles) des revêtements de terre isolant tant le défunt que le mobilier, et des appuis-tête qui apparaissent sous la forme de vestiges d’amas végétaux rectangulaires ou alors indiqués comme « amas de sédiments » entre le bloc cranio-facial et le fond de la fosse.

12Au-delà, au Proche-Orient, ce type dispositif est évoqué dès le début de la période PPNB (site syrien de Mureybet ; Cauvin 1974, p. 49 cité dans Cotenson 1992, p. 186). Évidemment, le cadre égyptien livre de nombreux témoignages dès le 4^e millénaire. Une étude approfondie de ces objets a d’ailleurs été menée dans le cadre d’une thèse de doctorat (Perraud 1997). On peut ici mentionner par ordre chronologique : le site d’Adaïma (4^e millénaire ; Hochstrasser-Petit 2005, p. 53, fig. 11a) ; le site d’Abusir (V^e dynastie 2500-2350 cal. BC ; Dulikova et al. 2015, p. 69, fig. 3, fig. 4) ; ou encore le site de la nécropole Memphite de Giza (de 2500 à 2160 cal. BC ; Lebedev et al. 2018, p. 112, tab. 1 et 2 ; Dulikova et al. 2015, p. 71).

13Dans le cadre des données textuelles et ethnologiques, les appuis-tête interviennent dans la thématique de la protection de la tête tant physique que symbolique comme l’indiquent les nombreuses inscriptions qui ont été relevées dans les exemples égyptiens les plus récents et sont sans équivoque associés aux rêves en tant qu’inter-monde entre les vivants et les morts (Hellinckx 2001 ; Perraud 1997 ; 2002 par exemple).

14Pour notre cadre géographique et chronologique, même élargi (Europe occidentale), les référentiels restent à constituer. Dans cette perspective, les objets décrits ici pour la sépulture collective de Mas Rouge constituent des jalons importants et plusieurs points doivent être soulignés tels que leur qualité technique et leur esthétisme. Leur position dans la tombe et l’association à un squelette en particulier documentent également leur fonction particulière, comme appui-tête.

15De plus, dans la mesure où la partie village du site a livré des objets comparables, faut-il aussi envisager son usage dans le monde des vivants ? Leur rareté en contexte ouest-européen est-elle réelle ou induite par les techniques et/ou contexte de fouille ? Serait-il, par ailleurs, lié à une notion de statut individuel ?

16Enfin, à l’instar de l’architecture de cette sépulture, l’usage de la terre crue est ici clairement privilégié au détriment de la pierre, le choix s’étend donc bien ici à tout le processus conceptuel de la tombe.

⁂

1. Context of discovery and location of the devices

17The Late Neolithic collective burial (3050-2890 BCE, Ferrières culture) of Mas Rouge in Montpellier (Hérault, France) stands out in particular by virtue of its semi-subterranean architecture (which may be qualified as a pseudo-hypogeum) in which raw earth as a building material predominates in all its forms (Tchérémissinoff et al. 2018 ; Leal et al. 2019 ; Leal et al. this volume). The collective burial contained 180 individuals, both adult women and men as well as immature individuals of all ages1. These were grouped in the tomb according to their states; distinct spaces were thus assigned to primary deposits, others to translations, i.e. temporary transfers and deposits, and still others are reduction spaces or even ossuary spaces.

18In addition to the specific structural elements and grave goods expected in such a context (for example pottery, personal ornaments, bone awls or lithic tools), the tomb yielded a large amount of raw earth furniture, in most cases large immovable containers – a type of furniture which in addition is rarely described (Leal, Jallet et al. 2018)2 – but also two exceptional pieces which are the subject of this contribution3.

19At least two of these funerary devices made from raw earth were identified in the eastern part of the burial. One was located within a layer of a sub-primary deposit (TC420), the other in a higher, secondary position (TC144), at the transition between sectors A and D and directly above an area dedicated to translations (temporary deposits) (fig. 1).

1. (Oriented) zenithal view of the burial showing the location of the two objects.

20The first artefact, TC144, was discovered in the upper part during stripping 25 (removed in stripping 26) (fig. 2). It was lying flat in a secondary position on the broken remains of a collapsed plastered floor (fig. 3 and 4). It can therefore be attributed to a layer n+2 of the burial, maybe corresponding even to a ground-level layer (at the junction of sectors A and D). There were no bone remains in its immediate proximity. Unfortunately, this object was somehow chipped during its excavation and removal and the photographs taken in situ therefore remain the most accurate. The item is incomplete and its surface damaged.

2. Position of headrest TC144 during stripping 26.

3. Close-up view of headrest TC144 in situ during stripping 26, oblique view from the west.

4: Close-up view of headrest TC144 in situ during stripping 26, frontal view from the south.

21The second artefact, TC420, was discovered in a lower position during stripping 32 (removed during stripping 34/35), and it was attributed to a layer of the lower, sub-terranean floor n+0 in sector A (fig. 5). It was well preserved and only one of its faces was damaged. In the same area and on the same floor as headrest TC420 the presence of an additional earthen device can be assumed under the neck of an adult individual but it was unfortunately not identified as such in situ (fig. 6 and fig. 5).

5. Location of sector A during stripping 32.

6. Close-up view of the possible additional headrest during the removal of the bones of the associated cervical rachis during stripping 33/34, lateral view from the north (the close-up view initially documents the state of preservation of the cervical rachis).

22The element TC420 collapsed together with the bodies that were “skeletonised” at various degrees (in a strongly organic environment) and with the earth-plastered floor from a less high position compared to the first element. It therefore got stuck in the upper collapsed elements stemming from the definitive closing of the burial. It thus remained in an explainable position, which can be considered as a functional one in relation to the associated bone remains. It was indeed discovered under the skull of an adult individual lying on his/her back with semi-flexed or flexed limbs, still supporting one part of the cervical vertebrae, whilst the neck regions and the head had slipped sidewards (fig. 7 and 8). This particular location makes it possible to interpret the function of the object as a headrest4, an interpretation that is also taken into account for the first earthen artefact TC144 because of their similar shape.

7. Location of headrest TC420 during stripping 34.

8. Close-up view of headrest TC420 in situ during stripping 34, view from the east/southeast.

2. Description of the objects

23The two pieces TC144 and TC420 are remarkable for several reasons. First because of their shape, more precisely their geometry, which reveals a complex composition (infra). Indeed, both pieces testify to highly accomplished work as regards the volume with a high degree of finishing using a full relief technique. In this particular case we are dealing with two almost identical three-dimensional pieces that take the shape of an oblong triangular prism.

24The piece TC144 is 16 cm long and has a triangular section 10 cm high and 10 cm wide. The object weighs 1.110 kg (fig. 9a, and 11a). This piece is not, however, in the most informative state of preservation.

25By contrast, the piece TC420 provides us with better understanding these objects. This piece has a length of 20 cm at its base and a height of 17 cm, with a triangular section of 12 cm by 12 cm and a weight of 2.750 kg (fig. 9b, 10 and 11b).

9.

10. Headrest TC420.

11.

26The geometrical decomposition of the best-preserved piece (TC420) shows that the volume is formed by a triangular spherical base with in addition a slight reduction on one side in order to create two trapeze-shaped faces (fig. 12). The triangular ends are minimally bulging (in accordance with the principle of the spherical triangle) and the two contact surfaces have a slightly concave profile. The edges are softened or bevelled. All the surfaces were thoroughly smoothed using water, and headrest TC420 was moreover given an additional very thin plaster coating.

12. Geometrical structure of TC420.

27Despite their similar dimensions the weight of the two pieces differs considerably with a ratio of 2.50. This is because of the incomplete state of piece TC144 after its removal (the ratio of the preserved volumes being 2.56).

28As regards the raw materials, the main component of these two pieces is a carbonated, slightly clayey, yellow beige silt. It includes a fine to coarse sand fraction (with grains of quartz and limestone up to 3 mm in diameter). The mixing was very accomplished and produced a very dense and compact material. Plant inclusions are visible. The analyses carried out on the material5 during the restoration6 of these objects indicate that the material of TC144 is strictly identical, as regards its clay composition, to the material of the mudbricks used to build one of the load-bearing walls of the tomb (wall MR45 of which a section has been stabilised ex situ), which was discovered collapsed inwards. This material more precisely contains kaolinite, illite and smectite (swelling clay). In contrast, illite and smectite are absent in the material composition of piece TC420.

29These pieces are among the artefacts of the tomb exhibiting the most accomplished level of finishing. Although the material of the headrests and that of wall MR45 stem from the same raw material7, a comparison at the scale of the tomb makes it possible to assess the degree of investment or adaptation of the material between the various portable or immobile elements, the first not being correlated to the second.

30Thus, comparative macroscopic analysis of these two categories of remains correlated to the micromorphological data of this wall section (the headrests were not subject to destructive physical analysis) makes it possible to reveal a specific selection of the material used for the headrests with a proportion of the coarsest mineral fraction reduced and calibrated to a diameter < 5 mm, a mixing including a (low) charge of fine plant materials8 on the one hand and a very accomplished mixing with a shaping to a plastic state on the other.

31In short, the raw material used to shape these objects, which were created outside the structure does not differ from the raw material used inside and at the scale of the tomb: it belongs to the same type of sediment, which must be highlighted with regard to the global economic aspect.

32The exposure of these objects to strong heat is very unequal as is shown by distinct parts (and not only the heart of the piece) that remained raw with a beige yellow colour next to the surface or that were strongly affected by heat with a brick red to black coloration and a carbonised surface of a maximum 2 cm thickness. These observations apply to the entire furniture and the earthen constructions. The analyses carried out on the clay fraction made it possible to evidence that the temperature did not exceed 400 °. Although the context linked to the sealing of the tomb by a fire makes it impossible to evaluate a possible stabilisation of these objects due to fire, the indication of a maximum temperature of exposure, however, compared to the ambient impacts (including on the human bones) indirectly document the degree of exposure of the associated layer.

33The identification of such devices in a distinct sector of the collective burial (i.e. sector A dedicated to primary deposits), which makes it possible to assign a function, also raises the question as to whether these were possibly present in the village9, given that this settlement has also yielded a large series of objects made from raw earth. So, at least two objects may indeed correspond to this same type (fig. 13a and 13b).

13.

34These two objects are quite damaged and were both recovered from two undetermined domestic pits (ST1044 and ST188) dated to the Ferrières culture, which means that they belong to the same time span. Unfortunately, their context of discovery was badly documented. Nonetheless they were not linked to a funerary context.

35The shape of the first object (ST188) (or at least the preserved remaining part as the excavators had chipped the raw earth part) is clearly identical to the headrest TC420 of the tomb of Mas Rouge. The preserved dimensions, i.e. a length of 15 cm and a height of 10 cm, are very similar.

36The other piece, which was discovered in pit ST1044, is also very fragmented. In this case, the interpretation is less clear. However, there are similar characteristics as regards the material, the random character of the heat damage, and the dimensions as it measures at least 16 cm in length for a height of 10 cm and has a preserved width of 7 cm. Its surface finishing is also quite similar, even typical.

37At this stage of the analysis the question arises as to whether such devices were widespread in the context with which we are dealing.

3. Comparisons and discussion

38Although they are not present (or identified) everywhere, headrests are known almost worldwide in Neolithic cultures. They may be made from earth or stone materials and they are mentioned in the descriptions of the contents of burials for numerous sites, but never in much detail. And in most cases, they were not sampled. In these cases, the authors describe headrests made from ‘perishable’ materials which is misleading in cases in which these are made from raw earth.

39Despite these mentions such devices are very rarely documented as regards Western European Neolithic contexts. The earliest and best documented site within this large geographic study area is the site of de Can Cambús (Allièse et al. 2014) in Catalonia. It is dated between 3900 and 3500 cal. BCE and it makes it possible to draw an interesting parallel with the case of the burial of Mas Rouge in several aspects because it has yielded a large number of elements made from earth or organic materials linked with individual burials:

• earth plastering that makes it possible to isolate both the deceased and the grave goods and that is interpreted as funerary beds and that is in some cases linked to wooden devices interpreted as floors;
• headrests are frequently identified (and included in the taphonomic analysis in 51% of the cases) although their material traces are subtle. They appear in the form of rectangular vegetal heaps or as “heaps of sediment” occurring between the block that is formed by the skull and the face bones, and the bottom of the pit.

40This type of device is mentioned for Near Eastern sites as early as the beginning of the PPNB period. At the Mureybet (Syria) site such a device was identified as a “cushion” made from red clay observed under the skull of the complete skeleton of an adult interred “beneath the floor of a house” in a burial dated to the 7th millennium (Cauvin 1974, p. 49 cited in Cotenson 1992, p. 186). On the floor of the same house five isolated skulls were lined up against the walls and an identical device was discovered as it is reported that “several [skulls] were placed on lumps of red clay brought in from the exterior, which served as a support” (ibidem).

41In Egypt such devices are mentioned as early as the 4th millennium BCE. A detailed study of these objects was carried out as part of a doctoral thesis (Perraud 1997), which made it possible to detail the semantic evolution and their diffusion routes.

42Our purpose here is not to provide an exhaustive inventory but to propose some points of reference, for example the renowned predynastic site of Adaïma (4th millennium). The western necropolis of the site more precisely yielded a burial (S341) that contained an adult individual that had a cushion made of plant material in the form of a mat (the assemblage and dimensions of which are not mentioned) placed under his/her skull (Hochstrasser-Petit 2005, p. 53, fig. 11a).

43Headrests (except the elaborate headrests designated as ‘chevet’) are increasingly mentioned in the specific literature regarding tombs (indicated as being modest burials) from 3000 cal. BCE on. As regards the Abusir site (2500-2350 cal. BCE) it is noted for example: “The light brown mud brick (13 × 10 × 8-9.50 cm) was apparently used as a headrest. The face of the deceased was originally looking east, arms bent at the elbows. The fingers of the left hand were situated under the mud brick, while the right hand was placed in front of the face” (Dulikova et al. 2015, p. 69, fig. 3, fig. 4).

44There are also numerous examples dated between the 5th dynasty (2500-2350 cal. BCE) and the First Intermediate Period (2200-2160 cal. BCE), for example, of the presence of headrests made from mudbricks identified in nine tombs (Lebedev et al. 2018, p. 112, tab. 1 et 2; Dulikova et al. 2015, p. 71) of the Memphite necropolis in Giza. The necropolis of Qau in Middle Egypt has yielded a series of at least eight tombs, also including headrests made from bricks among which three (tombs n^o4806, n°4842 and n°4866) could be dated to the First Intermediate Period, i.e. 2200-2160 cal. BCE (ibidem).

45Beyond these archaeological examples, headrests are mentioned in relation to the physical and symbolic protection of the head as indicated by numerous inscriptions that were noted for the latest Egyptian examples and were clearly linked to dreams as a connecting world between the living and the dead. They are connected both with thanatopraxis and with rituals. This interpretation precedes more theoretical and complex concepts that will be progressively developed later on. Nonetheless, certain inscriptions, among the most ancient ones, indicate the aim of protecting life and preserving health (for example Hellinckx 2001; Perraud 1997; 2002).

46We will not analyse this issue in further detail in order to avoid any anachronism or anatopism that would distance us from our subject. The aim here is to develop a perspective between the space of the living and the space of the dead as well as the shadowy space of sleep.

4. Conclusion

47As regards the geographic and temporal framework that was defined and even enlarged (to Western Europe), data repositories still need to be built.

48From this perspective, the objects presented here for the context of the collective burial of Mas Rouge are significant benchmarks and several aspects must be highlighted such as their technical quality and their aestheticism. As a matter of fact, they were manufactured with particular care and their troubling similarity indicates formal reproduction. Although our knowledge about objects made from raw earth has globally increased over the last decade (modular objects, above-ground storage pits, stationary containers, etc.), to our knowledge these objects have no counterparts as regards their shape for the periods concerned. Their position in the tomb, more specifically in relation to a particular skeleton, also documents their particular function as a headrest. This is moreover the case in sectors inside the tomb, which are dedicated to primary deposits (sector A) or to elevated floor-level spaces (sector D), in which the bodies were stored temporarily or displayed.

49In addition, given that the corresponding settlement has yielded comparable objects, were these items used in the world of the living? Is their rarity in the West-European context a reality or is this inferred because they were not identified or preserved because of the excavation techniques and/or the excavation context? Were they linked to the status of the individuals? In ancient Egypt headrests made from bricks were linked to the poorest burials. However, in the pre-state contexts we are concerned with, other sociological or cultural criteria may have mattered. Lastly, the similarity of the three pieces presented here, with a complex geometry/shape, makes it possible to assume that we are dealing with a standardised production for this type of object, at least at the scale of the village.

50Lastly, as is the case for the architecture of this tomb, raw earth materials were clearly preferentially used compared to stone materials. However, at least as a head support, the use of an unworked stone or a plant pillow such as it is known from Late Neolithic burials in the region would have been the simplest solution in terms of investment. This demonstrates again that the material chosen is not the cheapest option, but is rather a deliberate and renewed choice. A choice which in the present case applies to the entire conceptual process of the tomb.