Challenging the bias against local agency

It is essential to address critically a series of postulates argued by the Roman-centric approach, which has denied local agency in relation to these constructions. According to this, the saunas—but also the sculptural and decorative art—would be a consequence of Romanization, going back to either the Augustan period (27 bc–ad 14) or the Julio-Claudian and Flavian periods (27 bc–ad 96) (Calo Lourido Reference Calo Lourido2010).

This approach dates back to the 1960s. At that time, regional archaeology was moving away from the old diffusionist paradigm by adopting the perspectives of New Archaeology and traditional Romanization theory. While diffusionism proposed origins ranging from the Aegean (Mycenaean influence) to central Europe (Celtic influence) (Cardozo Reference Cardozo1949, 497; Hawkes & Cardozo Reference Hawkes and Cardozo1958; López Cuevillas Reference López Cuevillas1953; Sarmento Reference Sarmento1899), the new archaeology reduced the Castro culture and its constructions to a Romanized phenomenon. The reason is simple. Many hillforts had continued to be inhabited after the Roman conquest. While the new excavations were carried out using scientific methodology, they were limited to the Roman occupation levels, thus underestimating the earlier phases of indigenous occupation. Notably, Roman material found on the surface was sometimes considered sufficient evidence to date the entire lifespan of a settlement (see critical analysis in Rodríguez-Corral Reference Rodríguez Corral2012, 84–6).

According to traditional Romanization theory, saunas were adopted as an example of Roman provincial architecture (Almeida Reference Almeida1986). Along with statuary−stone warriors−and material symbols−shared by stone warriors and saunas, among others−they would have emerged within a society subjected to ‘strong acculturation by Rome’ (Calo Lourido Reference Calo Lourido1994, 806–7; Reference Calo Lourido2010, 260–65; Ríos González Reference Ríos González2017, 368–9). The adoption of this approach led to a strong rejection of any explanation in which the indigenous, symbolic or religious played any role. Under this logic, all features that deviated from the Roman style, rather than reflecting local logic (and therefore contradicting this Roman-centric view), were explained as the result of local technical limitations in imitating the Roman style. With regard to the symbolism present in the interior, despite its originality, this view ignored any role other than the purely decorative, intended to embellish a space dedicated to mere bodily hygiene (Calo Lourido Reference Calo Lourido1994).

It seems undeniable that Millett’s (Reference Millett, Blagg and Millett1990) concept of progressive emulation, already deconstructed in the 1990s by various authors inspired by post-colonial intellectual currents (see Gosden Reference Gosden2004b; Mihajlović Reference Mihajlović and Smith2019), is present in these approaches. However, this is not the only problem, as these approaches are also based on a chronology that is difficult to accept. The materials used to date the saunas to a Roman period lack contextual relevance. This is the case, for example, of the saunas at Sanfins (Pacos de Ferreira, Porto), Santa María de Galegos (Barcelos, Braga) and Monte Padrão (Santo Tirso, Porto). The ceramic materials found in these structures are very scarce (only three fragments in Sanfins) and all appeared in disturbed layers, abandoned or without context (Almeida Reference Almeida1974, 161–2; Silva Reference Silva1986). In fact, at Monte Padrão (Moreira Reference Moreira2010–11), they were found mixed with medieval ceramics.

Likewise, the biography of places cannot be ignored. The construction of the Tongóbriga sauna (Freixo, Porto), which was completely excavated out of the rock, does not appear to have been influenced by Roman culture. A Roman bath was built right next to it in the first century ad (Dias Reference Dias1997; Silva Reference Silva2019, 211). It is reasonable to assume that the building lost its original function when the settlement was Romanized. At Monte da Saia (Barcelos, Braga), two niches containing Roman sculptures of clear funerary significance were found inside the building. These were undoubtedly added later, as the space was reused for a different purpose (Silva Reference Silva1986).

Further evidence has been posited which suggests that the utilization of these structures is in fact associated with local communities that were in existence prior to the incorporation of the north-west within the Roman provincial system. At the Sabroso hillfort, the presence of Roman fragments must be attributed to the emergence of trade networks, rather than being indicative of a process of Romanisation (Ríos González Reference Ríos González2017, 145–8). The community that inhabited the hillfort appears to have followed a clearly indigenous lifestyle, and it can be deduced that the sauna must have functioned under the same logic. The excavations of the saunas at Sanfins (Silva Reference Silva1986) and Briteiros (Cruz et al. Reference Cruz, Antunes and Costa2010–11, 94) provide further evidence for this idea, as their foundation levels date back to the second and first centuries bc. In the Braga sauna, the archaeological evidence is also consistent with this chronology. The artefacts found on the sauna floor are local and all dated to the Late Iron Age. This enables us to establish a definitive terminus ante quem. Moreover, beneath the foundations of the sauna, a deposit consisting of two Early Iron Age pottery sherds has been revealed, potentially indicative of a foundation offering or a remnant of an earlier ritual structure (Lemos et al. Reference Lemos, Leite, Bettencourt and Azevedo2003, 324). In either case, this find powerfully reinforces the site’s deep cultural resonance for the local community, as well as the enduring temporality of the practices enacted within its landscape.

It is also noteworthy that, in the northern region, saunas such as Chao Sanmartín (Grandas de Salime, Asturias), Coaña (Coaña, Asturias) and Pendía (Boal, Asturias) have been radiocarbon-dated to the third and first centuries bc. All of them were abandoned during the initial phases of the Roman occupation (Villa Valdés Reference Villa Valdés2012). Despite the absence of pedra formosa, the dating of these saunas shows that steam baths were a prevalent practice among local communities in northwestern Iberia during the Early Iron Age. The information transmitted by Strabo (3.3.5) appears to reinforce this hypothesis, as he points to steam baths as a practice typical of local communities. This statement is most likely based on information obtained from Posidonius (135–51 bc), who may have obtained it during the expedition undertaken by the consul Decimus Junius Brutus in 136 bc (García Quintela & Santos-Estévez Reference García Quintela and Santos-Estévez2015; contra Ríos González Reference Ríos González2017, 289). If this is the case, it would once again prove that steam baths were a common practice prior to the arrival of Rome in the region.

This chronological framework allows us to rethink the saunas in context. Rather than emerging in isolation, these contractions with pedra formosa were integrated into a material entanglement that includes the new fortification systems built during this phase. These systems delimit and isolate the interior of the settlements. Additionally, powerful material symbolism extended within the walls. This entire entanglement emerged and functioned between the second and first centuries bc, a period marked by uncertainty and apprehension about the erosion of indigenous identities. This anxiety arose within a specific socio-political context, as the region grew increasingly unstable and fragmented. The local community experienced mounting pressure from Roman expansion along its frontiers, in a period prior to the region’s integration into the Empire (Rodríguez-Corral Reference Rodríguez Corral2019).

Challenging the bias against material agency

In contrast to Romano-centric perspectives, some scholars emphasize the indigenous significance of these constructions. However, I argue that this indigenist interpretation rests on two problematic assumptions that obscure the intricate interplay of place, perception, and power: a symbolic definition rooted in Saussure’s linguistic model (Reference Saussure1974) and a view of cognition drawn from Searle’s internalist theories (Reference Searle1983). By treating the pedra formosa’s symbolic motifs as signifiers of prior meaning, this approach risks dematerializing them, reducing their materiality to mere signs conveying deeper concepts. This presupposes an internal idea in the creators’ minds, subsequently expressed through external symbols. The process of externalizing symbols raises a critical issue. By treating motifs as linguistic symbols, we assume their connection to meaning is arbitrary and conventional. However, the essence of these signs, carved into enduring stone, does not lie within the motifs themselves, but in the vanished intersubjective horizon in which their significance was woven. The arbitrary relationship between symbols and their meanings creates a gap that is difficult to bridge. As a result, scholars are confronted with this challenge, with some describing a ‘profound symbolic content’ (González Ruibal Reference González Ruibal2006–7, 576), while others acknowledge the difficulty of approaching the symbolic-religious significance of these objects (Silva Reference Silva2019, 208) or concede that their meaning ‘remains elusive’ to archaeological research (Villa Valdés Reference Villa Valdés2016, 103).

The notion of deep significance has ultimately given rise to inconclusive debates. In response, certain scholars have proposed an alternative approach, which is, however, no less problematic. They explore this form of decorative art as a manifestation of iconicity, despite its high level of abstraction (Fig. 2). By establishing a correlation based on similarity (iconicity) between signs and real-world referents, these motifs are endowed with non-conventional referential meaning, which could potentially aid their interpretation. Various interpretations have emerged from this approach. For example, complex compositions have been interpreted as stylized human figures (Cabré y Anguló Reference Cabré y Anguló1922) or schematic female idols (Jordá Cerdá Reference Jordá Cerdá1969). Triskeles and rosettes have been given a solar significance: Cardozo (Reference Cardozo1931), for example, suggested that the arrangement of triskeles on the Briteiros stone depicts the sun’s journey from dawn to dusk. García Quintela and Santos-Estévez (Reference García Quintela and Santos-Estévez2015) have revisited this interpretation, arguing that the absence of triskeles within the third circle signifies either the setting of the sun or a lunar representation. Another author sees the signs engraved on the pedra formosa as a ‘representation of transit in a strictly symbolic sense’, referring to a mystical landscape (Brañas Abad Reference Brañas Abad2000, 103). Similarly, Silva (Reference Silva2019, 208) has suggested that these symbols reflect an Indo-European cosmology centred on elements such as the sun, fire and water. According to this perspective, triskeles, circles and closed interlaced patterns have astral connotations, while triangular interlacing and circle-based designs symbolize fire. Rope motifs represent strength, and curvilinear or chain-like S-shaped patterns evoke dynamic elements such as water, steam and smoke. According to Silva, these signs embody ‘elements of life and the functioning of these baths’.

Figure 2. Sauna and pedras formosas: (top) sauna from the hillfort of Briteiros (Guimarães) (modified after Cardozo Reference Cardozo1931); (below, left) from Eiras, Vila Nova de Famalicão (Braga); (below, right) pedra formosa from As Eiras de Arcos de Valdévez (Viana do Castelo) (drawings 2 and 3 modified after Queiroga Reference Queiroga2003).

In my view, this approach faces a significant limitation: iconicity rarely reveals the full depth of meaning, particularly its functional significance, even when the link to a real-world referent seems to be clear. The warrior statues at hillfort entrances provide a compelling case, as their clothing shares symbolic elements with the pedras formosas, as will be explored later. As icons, these sculptures unequivocally depict armed men, enabling them to be interpreted straightforwardly as warriors. However, their specific identity—whether as deities, heroes or eponymous princes—remains unresolved, as their iconicity offers no definitive answer. Moreover, their purpose goes beyond mere representation, raising complex questions about their symbolic and spatial relationships. These include their positioning at hillfort entrances, their connection to the defensive walls and their association with the saunas (Rodríguez-Corral Reference Rodríguez Corral2012). If ambiguities of this kind persist in the case of sculptures, where referentiality is relatively clear, they are even more pronounced in the geometric and abstract decorative schemes of the pedras formosas, where referentiality is totally unclear.

The persistence of this interpretive approach can be attributed to the deep influence of representationalism within Western art theory, which exerts a considerable influence on modern thought (Alberti Reference Alberti2016; Gell Reference Gell1998, 73; Malafouris Reference Malafouris2013; Thomas Reference Thomas2004; Boivin Reference Boivin2008). As Robb (Reference Robb2017, 596) has pointed out, ‘not that all societies have some version of “art”; it is distorting to assimilate other people’s powerful symbols to our “art”’. In structuralist archaeologies, for instance, representational elements are frequently reduced to discrete units of meaning, which are then combined to form complex messages. The method entails the extraction of motifs—defined as minimal units of meaning—from their material context, and their subsequent presentation in typological frameworks (Meirion & Cochrane Reference Meirion and Cochrane2018). Research into the symbolism of the sauna, for instance in the form of Silva’s semiotic analysis (Reference Silva2019), adheres to this pattern, while drawing on models such as Eco’s (Reference Eco1979) for Western art and overlooking non-Western anthropologies of art (Gell Reference Gell, Coote and Shelton1992; Morphy Reference Morphy2007, 166). Consequently, the pedra formosa is regarded as a passive medium, serving merely as a substrate for inscribed symbols. These symbols and patterns are conceptualized as a ‘language’ that requires decoding to reveal its meaning. While this approach acknowledges the symbolic weight of these rituals, it neglects the agency of the participants, that is to say, their performative capacity actively to shape social and ritual interactions.

In any case, to understand the role and significance of these places, we must consider two key issues. Firstly, we must recognize that symbols in non-Western art serve functions that extend beyond merely encoding meanings. Secondly, we must consider the relationships these symbols establish with the physical structures in which they are located, such as saunas, walls and hillforts.

Saunas as performative places

Traditional approaches to the study of these structures are framed within what Alberti has described as a ‘conventional archaeological ontology’ (Reference Alberti2016, 167). This archaeology, deeply rooted in modern dichotomies, draws an ontological divide between things and individuals, limiting our understanding of these structures to essentialist and static interpretations. My intention here is to transcend this reductionist view by assuming an approach that recognizes the dynamic interdependence between human beings and the material world. People, in their social lives, generate material culture, but this materiality, in its physical and symbolic dimensions, is not merely a passive product: it actively shapes the conditions for the existence of society itself, moulding human cognition and behaviour. Mind and action do not reside exclusively in the individual but are distributed across a network of human and non-human agents, intertwined in a process of co-creation (Malafouris Reference Malafouris2013; Reference Malafouris, Newen, De Bruin and Gallagher2018). As Malafouris and Röhricht (Reference Malafouris and Röhricht2024) argue, performativity constitutes the active mechanism that facilitates this distribution, while agency emerges as its consequence, a relational capacity that arises in encounters between people and things.

The following sections examine the saunas from a performative perspective, focusing on four dimensions that clarify their role and significance within local communities. First, the distinctive character of these sites will be considered, with the topographical setting, material composition and sensory qualities defined and assessed in terms of their impact on human experience. Secondly, the integration of saunas within the wider hillfort will be explored, viewing them as components of an interconnected network of material and social relationships, rather than isolated entities. Thirdly, the symbolic potency of the large, decorated stones that mark these saunas will be examined. These stones serve as thresholds that mediate access to the heated zone and imbue the space with cultural meaning. Finally, the article will explore how sauna architecture has the capacity to influence human movement, perception, and interaction, functioning not as a passive framework but as a dynamic agent in shaping encounters with the material world. Collectively, these four dimensions furnish a framework for comprehending the performative function of saunas, elucidating the manner in which material agency is engendered and disseminated through the intertwinement of individuals, artefacts and the topography within the cultural milieu.

The special character of the saunas in the Duero-Miño region

Two distinct traditions emerged whitin the Castro culture territory: one in the Cantabrian region and the other in the Duero-Miño region (see Figure 1). While the constructions in both regions share the common function as saunas, those in the Duero-Miño region show a number of features that distinguish them from those in Cantabrian region, revealing greater socio-material complexity. One of these characteristics is their location. The saunas of the Duero-Miño region are the only constructions to be found outside settlements in the entire territory of the Castro culture. Some authors have argued that this unique location should not be exaggerated, arguing a purely ‘functional’ (as opposed to ‘symbolic’) cause. In contrast to the Cantabrian area, the Duero-Miño saunas were directly connected to a water source. This would explain, according to them, an extra-mural location, where supplies would have been more readily available (Calo Lourido Reference Calo Lourido1994; Ríos González Reference Ríos González2002; Reference Ríos González2017). However, in my view, a more detailed analysis reveals that there were more complex factors at play. For instance, while the sauna at the Sanfins hillfort was close to a spring, it was also near the the wall and the entrance area of the settlement, where a stone warrior was positioned on a rock (Silva Reference Silva1986). This suggests that the sauna was situated in a space imbued with symbolic significance (Fig. 3). Consequently, it is more likely that the liminal space associated with the edges and access areas of the hillforts was the underlying cause. Indeed, there are instances where the saunas, which are clearly associated with this liminal space, received water from the springs through channels that spanned considerable distances. An example of this is the saunas of Briteiros II, which are located in the outer area next to the wall and 400 m from the granite outcrop from which the water flowed into the settlement. If proximity to the spring was the primary factor, it makes sense that the sauna would be in the settlement. However, while the relationship to water sources shows variability, saunas consistently seem to have been located outside the walls and access areas of the settlements, even if these were supplied by springs within the walls.

Figure 3. Sanfins hillfort (Paços de Ferreira, Porto): (A) modern reconstruction of the stone warrior on the outcrop at the entrance, where the lower part of the original statue was found (Photograph: Sole Felloza); (B) sauna (Photograph: author.)

A second distinctive feature of the Duero-Miño saunas is their semi-subterranean architecture, characterized by low mounds which blend seamlessly into the surrounding landscape. The interior has been meticulously designed to provide a distinctive sensory experience. A comparison with the Cantabrian saunas allows us to highlight a number of critical issues in this sense. While both types are spatially associated with walls and entrances, they are integrated differently into this ‘theatrical’ setting. In the Duero-Miño region, there is a radical break with the settlement and its architecture. Here, the saunas are located outside the walls, using large orthostats and taking the form of tumuli, while the Cantabrian ones are located within the walls, adopting the same construction style as the buildings in the settlement, that is, slate masonry structures. The use of orthostats and semi-subterranean construction in the Duero-Miño region has no parallels or precedents in the region, contrasting radically with the architecture of the hillforts, which used granite masonry (Fig. 4). In some cases, such as at the hillfort of Tongóbrica, the sauna was excavated directly into the rock. These architectural solutions give the saunas a megalithic style. Therefore, their peripheral location outside the walls, as well as their tumular and megalithic character, reveal their exceptional nature. The construction recalls the prehistoric building concepts of the region and seeks to integrate into the outside world, that is, the wild space.

Figure 4. Sauna of Briteiros (Guimarães). (Photographs: author.)

The third defining feature of the saunas of this region is the rich decorative art, which is absent in the buildings of the Cantabrian region. On occasion, this absence has been attributed to the type of stone used. According to these explanations, granite, used in the Duero-Miño region, facilitates engraving, whereas slate, used in Cantabrian saunas, makes it difficult (García Quintela & Santos-Estévez Reference García Quintela and Santos-Estévez2015). Slate is indeed not the optimal material for engraving. Nevertheless, I think that focusing solely on the materials used is misguided. There are other considerations that must be addressed. In this context, it is important to note that in this region, slate was replaced by granite in this region when necessary. This can be observed in both the domestic environment of the settlements (where the mortars are made of granite) and in the saunas themselves (where the pools at the entrance are also made of granite). It thus appears reasonable to argue that, if symbolic motifs were of significance in the Cantabrian, a solution to the problem could have been sought.

It is my contention that the presence or absence of decorative motifs is an issue that transcends the boundaries of the sauna space itself. Rather than concentrating solely on the type of stone used, a more productive approach lies in examining the natural form of the material—rock itself—as a cultural and symbolic medium. The Duero-Miño region, characterized by a granitic landscape and numerous rocky outcrops, held deep significance for prehistoric communities. This is evidenced not only by the ritual sites discovered in the vicinity since the Bronze Age, but also by the rich tradition of rock art associated with them. These rocks were for a long time a space of activity, a material proxy for thought, ritual and symbolism that helped to construct a sense of ancestry (Bradley Reference Bradley and Fábregas1998; González Ruibal Reference González Ruibal2006–7; Rodríguez-Corral Reference Rodríguez Corral2019). In other words, communities in the region accumulated a rich iconographic tradition. During the second and first centuries bc, when saunas were in use, rocks continued to be of great importance. As previously mentioned, the stone warriors, contemporaries of the saunas, were placed precisely on them, in the context of walls and entrances.

There appears to be a recurring relationship between saunas and outcrops at different levels. In numerous instances, the water that feeds these saunas originates from the rocky outcrops. The orthostats of the saunas also come from these outcrops. Notably, some saunas were even carved out of the rock of these outcrops (Dias Reference Dias1997). These subterranean constructions appear to allude in some manner to the natural rocky spaces themselves as sites of exceptional significance. The association between rocks and saunas seems to reflect the material symbolism of the region.

It can be argued that the presence of decorative art in the saunas is simply a reflection of its existence outside of these spaces. In contrast to the Cantabrian region, the Duero-Miño region has evidence of two distinct but related phenomena: (1) the existence of a symbolic and iconographic tradition linked to the rocks, and (2) the emergence of a complex decorative system that is repeated in different areas of the space of the hillfort during the Late Iron Age. The decoration was not confined to the interior of the saunas, but also extended to the bodies of the stone warriors on the rocks adjacent to the walls, the seated sculpture and the constructions within the settlements. This extensive art allows for the sauna room to be connected to other areas of the communal space. If this idea is accepted, it is then reasonable to conclude that if there is no documented decoration inside the Cantabrian saunas, it is because there is no decoration outside, in the hillforts. And I think this is key to understanding this symbolism.

Entangled symbols: architecture, statuary and saunas

In the saunas, the decoration is largely concentrated on the pedras formosas. While in the majority of cases the motifs and patterns are distributed over the entire surface (e.g. As Eiras, Briteiros I or Calvos) or limited to the contour of the opening (e.g. Sanfins or Tongóbriga), in some other cases they seem to be absent (e.g. Monte Castro). However, this apparent absence of decoration must be qualified. Firstly, it cannot be ruled out that some of the motifs were simply painted. Despite the poor preservation conditions in the region, painted motifs have been identified outside the saunas. The most evident example is the trisquele (a recurrent element in the pedras formosas) discovered in San Cibrán de Las (Ourense) (Fig. 5). Nevertheless, painted remains have also been identified on the walls of houses in settlements (Calo Lourido Reference Calo Lourido1994, 798). It is therefore highly likely that the paint would have been for use in the saunas. Secondly, it should be noted that some of these stones have suffered erosion, which may suggest that the decoration has been lost. For example, the bas-relief decoration on Briteiros II is perfectly preserved, while the engraved decoration has almost disappeared and is difficult to see today. If this is the case with the engraving, it is similarly likely to be the case with painting. Finally, in some instances, the decoration is not directly visible on the pedra formosa, but rather on elements that are attached to it. In the saunas of Santa María de Galegos (Barcelos, Braga), a block decorated with geometric motifs was discovered in situ, attached to the pedra formosa. It is postulated that this block was originally positioned in this manner, as it was carved with the distinctive handle observed on other pedras formosas, to assist a person in entering through the narrow opening.

Figure 5. Painted and engraving trisqueles, San Cibrán de Las, Ourense. (Photographs: author.)

This decorative art is not limited to the interior of saunas. Its field of action includes other spaces such as sculpture and the interior architecture of the hillforts (Figs. 6 and 7). Thus, it can be affirmed that, in the Duero-Miño region, these symbols are present in the saunas to the extent that they exist outside them; in the same way that, in the Cantabrian saunas, they are absent to the extent that they do not exist outside them. Consequently, in my view, the ultimate sense of the pedras formosas is clearly related to the appearance of a symbolism distributed throughout the entire area of the hillfort in the Late Iron Age.

Figure 6. House doors at the hillfort of Âncora, Caminha (left) and the hillfort of Briteiros, Guimarães (right). (Photographs: author.)

Figure 7. Copy of a stone warrior, Lesehno Hillfort (Vila Real, Portugal). The stone warrior is positioned where it is believed to have stood, on the rocks next to the walls. It is visible to visitors approaching the hillfort. (Photographs: author.)

The use of decorative art is often observed to increase in contexts of conflict and resistance, as scholars have noted in the ethnographic record (Gell Reference Gell1998; DeMarrais & Robb Reference DeMarrais and Robb2013; Hanks Reference Hanks and Boric2010). Saunas, statues and symbolism emerged during the Late Iron Age. As already pointed out, these centuries correspond to the final phase marked by insecurity and anxiety, as a consequence of the pressure exerted by Rome on the territory, demographic growth and the increase in inter-group violence (González Ruibal Reference Morphy, Perkins, Morphy and Perkins2006–7). In this context, it is evident that there is a fear of the dissolution of identities (Rodríguez-Corral Reference Rodríguez Corral2019). When this occurs, the construction and definition of community space often becomes a battlefield where social reality is negotiated, among other, on a symbolic level. In these contexts, the physical boundaries of the community become critical lines to act upon. This is not only through their demarcation with large walls or the performance of rituals of various kinds, but also through decorative art.

In the context of the potential dilution of local identities, strategies of self-affirmation may have become necessary. The warriors emerge from the rocks in the liminal context of the entrances. They wear torques, bracelets and clothing decorated with the geometric motifs and symbols that appear on the pedras formosas. While the geometric motifs cover the warrior’s body, the trisquele occupies a less visible part of the clothing: the back of the belt. In any case, the association of these symbols with the warrior bodies—as well as with torques, in a context of hillforts entrances, where ideas such as ancestry, heroism and the defence of the community are negotiated—shows the importance of this decorative art. Thus, it appears that there is a strong connection between the pedras formosas in saunas and the stone warriors that emerge from the rocky outcrops located right at the entrance to certain hillforts.

Another group of sculptures that we might consider is the so-called seated statues, of which five examples are known. They wear the same clothes and prestige objects (torques and viriae) as the stone warriors; however, they do not carry weapons. Instead, they hold a kind of bowl on their knees in the position of an offering. Of particular interest is the statue from Pedrafita (Ourense) (Luis Reference Luis1997) (Fig. 8). The sculpture represents an individual seated on a seat decorated with geometric motifs and with the trisquele represented on his back. As in the body of the stone warriors, this symbol is represented outside the visual field, on the back. In this case, however, it is not depicted on the body, but on the back of the seat. While the exact meaning of this symbol is uncertain, it can be associated with concepts such as protection, enthronement and heroization. In any case, the potency of the symbol does not lie in its overt presentation to an audience, but rather in the quiet resonance of its relational meaning, which is woven into the lived experience of place and materiality.

Figure 8. Seated statue of Pedrafita, Ourense.

The context of the seated statues is less well documented than that of the stone warriors. However, the statues from Xinzo de Limia (Ourense) allow us to establish a relationship with the watercourses. They are two statues located in an enclosure of approximately four square metres on a wooden plank structure in a muddy area adjacent to the Limia riverbed. The site is of interest due to the funerary importance of the rivers on the Atlantic coast. In Castro Culture, this is evidenced by the deposition of objects in them and the absence of cemeteries. Strabo’s (3.3.5) allusion to this river as the Lethes allows us to associate this river with an ultramundane topography. This is an access route to the Otherworld through which warriors embark with all their equipment, never to return (García Quintela Reference García Quintela and Bermejo1999). The profound significance of this river is once again evident in the historical accounts of authors such as Livy (Per. 55), who describes how, during the expedition of Decimus Julius Brutus (137 bc), Roman soldiers, having heard the ancient stories of the local communities, refused to cross this river for fear of losing their memory (Alfayé Villa Reference Alfayé Villa, Ferrer, Lozano and Mazuelos2010). The two seated statues in the riverbed can be interpreted as evidence of a connection with water and funerary or heroic rituals. This is exemplified by the diadem, a remarkable piece of Iron Age goldsmithing from the region. Its iconography has been interpreted as both a funerary and heroic representation. Like the warriors depicted in stone at the entrance to the hillforts, it shows a group of warriors wearing torques and carrying a small round shield (caetra-type). According to Marco (Reference Marco, Alvar and Mangas1994), Alfayé Villa (Reference Alfayé Villa, Ferrer, Lozano and Mazuelos2010) and Rodríguez Corral (Reference Rodríguez Corral, Ferrer Albelda and Sánchez Domínguez2024), among others, the warriors are crossing the waters towards the afterlife.

It seems reasonable to maintain that there was a relationship between this decorative art and the liminal spaces of these Atlantic communities. The same symbols and decorative motifs appear on the bodies of the warriors, on the walls of the houses and on the walls of the settlements. It can be proposed that the same symbolism operated at the threshold of the body (individual), at the threshold of the house (family) and at the threshold of the hillfort (community). It is also possible to relate these symbols to watercourses, in line with the aforementioned discussion about the seated statues. Atlantic communities imbued rivers with deep meaning, viewing them as liminal thresholds to the Otherworld. These fluid landscapes were not just backdrops, but active participants in rituals involving offerings and the dead. The deposition of artefacts or human remains marked these waters as conduits to the supernatural. Such carefully chosen sites structured the sacred geography of these communities: in other words, a liminal, complex space in which the world of the living and the world of the dead (ancestors) overlap (and confuse). It can be argued that all the symbols and decorations functioned simultaneously in the construction of the ‘being’ of these communities at the end of the Iron Age, where the fear of the dissolution of identities encouraged ancestral connections.

The pedra formosa as a ritual mid-point

At this point, we must explore the importance of the pedras formosas. I think it is important to consider the role of geometric motifs in art beyond mere ornamentation. It is also essential to avoid reducing this art to a visual codification that seeks to communicate deep meanings. Instead, it would be more productive to explore its active capacity to affect, mediate and create connections between people, things and ideas (DeMarrais & Robb Reference DeMarrais and Robb2013; Malafouris Reference Malafouris2013; Morphy & Perkins Reference Morphy, Perkins, Morphy and Perkins2006). Material symbols have the power to influence the world and transform human cognition. Ultimately, they affect who individuals are and play a critical role in how they become themselves (Malafouris & Koukouti Reference Malafouris, Koukouti, Meyer, Streeck and Jordan2017; Malafouris & Röhricht Reference Malafouris and Röhricht2024). In contexts of insecurity, fear of the dissolution of identities often encourages socio-material actions and strategies that seek to reinforce boundaries. As previously stated, the emergence of decorative art is a common phenomenon. Hodder (Reference Hodder1982) demonstrated in his ethno-archaeological studies that the negotiation of identities is intensified at these margins, becoming palpable in decoration. However, it is important to recognize that motifs do not merely represent a previously constructed identity on a visual level; they actively contribute to its creation and maintenance.

In the context of our case study, it can be observed that the phenomenon of decorative art operates at the borders of communities on at least two levels. Firstly, at the geographical level. The distribution of the pedras formosas, statues and decorative art is concentrated in the Duero-Miño region. This is the area that was most affected by the processes of synecism and was most exposed to external influences and dangers, being the first to suffer the pressure of Rome (González Ruibal Reference González Ruibal2006–7). This phenomenon can be attributed largely to the concentration of decorative art and sculpture in this region, with a notable absence in the remainder of the territory. The second level is constituted by the limits of the settlements themselves, as previously discussed. Outside the walls, there are no built spaces—such as open settlements or cemeteries—with the exception of saunas. The hillfort is the only model of habitat, so its architecture is the anchor on which the whole landscape is built. This implies the existence of a specific ontology of space, whereby the walls differentiate the interior space from the exterior. The intramural and extramural space are two distinct realms. However, the classical notion of the ‘opposition’ of spaces is eschewed. The concept of superposition of spaces or entities, which generates ontological uncertainty (Deleuze Reference Deleuze1968), is a more stimulating approach. The liminal space, defined as an overlapping, intense and heterogeneous space, is a space where complex processes converge and require attention. The ritual practice observed on the slopes and on the wall appears to indicate this (González Ruibal Reference González Ruibal2006–7). Similarly, the material imagery also evinces this phenomenon. In this area, stone warriors emerge from the rocks and direct their performative power towards individuals approaching the hillfort (Rodríguez Corral Reference Rodríguez Corral2013). Furthermore, the saunas with pedras formosas, which are nourished by the water that emerges from this rocky landscape, also play a role in this area.

At this point, it is necessary to return to the function of saunas in this context of superimposition. As we have just seen, these constructions are places segregated from the ordinary, from everyday life. Their marginal position, outside the walls, places them in a liminal sphere, the ideas of inside and outside being mixed and confused. This overlapping is not only physical, but also ontological. In this context, its underground nature and its interior architecture create specific conditions of somatization. Upon entering the sauna, the most apparent effect is a sensory alteration that contrasts with the ordinary phenomenon of synaesthesia. This process eliminates the distracting elements and creates the conditions for what are commonly referred to as rites of passage. These rites entail a period of separation from everyday life, a transitional state in which the subject’s status becomes ambiguous as they pass through a symbolic domain, and a process of reintegration into the social order with a new position (Van Gennep 1909). The ritual commences with the subject undressing and bathing in cold water. Inside, the movement of the body is completely restricted by the pedra formosa. It seems reasonable to argue that this large stone—due to its arrangement, characteristics and decoration—acts as a midpoint or focal point in the ritual process. To pass through it, a series of steps must be followed. Firstly, one must lie on one’s back on the ground. Secondly, the subject puts their hands inside and grasps the handle located inside the large stone. Finally, they push inwards. Consequently, the individual’s attention is focused on his or her body, and the individual becomes aware of the changes occurring. The passage through the small opening results in a transition from light to dark, from cold to hot, and from dry to wet. Additionally, the individual experiences olfactory changes (sweat and fat) and acoustic changes, such as the internal echoes produced by the chamber formed by large orthostats. It is precisely at this juncture that patterns and symbols become operational, thereby precipitating the occurrence of synaesthetic changes.

The manner in which this decorative art works must be understood to be very different from what our Western sensus communis would encourage. As DeMarrais and Robb (Reference DeMarrais and Robb2013) have pointed out, art often invites participation, providing a focus or medium for relational action. The adornment of warriors with decorated costumes transforms their identities. The power that arises from individuals’ relationships with the symbols and decorative patterns covering their bodies must have required complex ritual processes. It is reasonable to assume that these rituals took place in saunas. In this sense, it could be argued that the pedras formosas are ‘quintessentially participatory art’. The interior space of saunas immerses people in altered states of reality. It is difficult to comprehend the mechanisms by which its efficacy is achieved. As Gell (Reference Gell1998) has pointed out, the effectiveness of decorative art is cognitive in nature. The complexity of design—multiplicity, symmetry, repetition, etc.—is a consequence of skill and knowledge that is hidden in the final result. The visual properties of repetition and symmetry inspire the appearance of animation. Consequently, those who enter its field of action are caught up in the spectacle of a virtuosity that they cannot comprehend. However, the agency of this decoration, while attractive and captivating, can also be agonistic and defensive. In this sense, Gell has theorized that decorative symbols and patterns are ‘potential weapons in conflict situations’ (Reference Gell1998, 83). In many societies, motifs and symbols serve as protective devices, defensive screens and obstacles that keep out spirits and protect community members from pollution. He has described these ornaments as demon traps, or demonic flypaper, to which demons irrevocably attach themselves. This metaphor thus clarifies the operational mechanics of this phenomenon: the drawings provoke and attract, but ultimately block potential dangers (Gell Reference Gell1998, 86). Gell’s perspective establishes a clear differentiation between decorative art and that which falls within his field of action, assuming that they are distinct entities. However, in ethnographic contexts, alternative, more fluid understandings of decoration are documented. In numerous instances, geometric symbols and patterns have been observed to possess the capacity to permeate and saturate the body. Furthermore, these symbols and patterns are integrated into the deceased’s soul permanently (Ingold Reference Ingold2015, 91–2). The function of these symbols as mediators with the ancestral world has been repeatedly highlighted by Morphy (Reference Morphy2007). Thus, it could be argued that these symbols and patterns could have the capacity to influence, empower, transform or protect people, in communities where personhood is perceived not as static, but as fluid (Brück Reference Brück2019, 224–36).

Be that as it may, a strong connection between material symbolism, somatisation and purification can be argued. Crossing the pedra formosa could have implied a powerful transformation in the individual’s body. This is purified, ‘dressing’ the symbols of power. It is therefore understandable that, outside, on the rocks, the bodies of the warriors appear also ‘dress’ the same patterns and symbols. The pedra formosa and the stone warriors share a common symbolic entablement. First, decorative patterns such as those depicted on the clothes; second, the corded patterns depicted on the belts and torques of the warriors. It should be noted that some feature a string-work decoration around the opening in the stone, which may take the form of a torque. Finally, the key symbols are represented: trisqueles, present on the belts of the warriors and on the throne of one of the seated figures. The very position of the saunas and warriors could be revealing an important issue.

The purpose of these semi-subterranean constructions and the decorative art is difficult to determine precisely. However, a number of non-exclusive possibilities can be put forward. First of all, they could be spaces for purification and preparation of the body of the individuals, which allowed them access to the interior of the settlement. Within this underground space, individuals performed a ritual of reintroduction into the community. In addition to symbolizing the power of the community, individuals also crossed the threshold of the pedra formosa, which may have functioned as a material metaphor for the community house due to its decoration and formal similarity. This could have been a common practice when crossing the liminal space or simply in specific situations where individuals, for whatever reason, needed purification processes (e.g. war, contact with other communities or crossing contaminated spaces). It is important to consider the association of this decoration with the bodies of stone warriors, as ritual purification processes could be limited to a group of armed individuals or to the aristocracy involved in potentially polluting activities such as warfare or travel. In this context, the purifying function could have a key apotropaic and propitiatory value in the protection and construction of the social.

However, we cannot ignore another possibility. This superimposed space, due to its liminal character, could have been favourable for the performance of other rites of passage. Thus, another possibility would be to see these constructions as a space for rituals related to the change of status in the biography of certain individuals. In other words, the passage through the pedra formosa could be seen to imply a process of purification in which the symbolism of an elite, perhaps sacred, is incorporated—if we consider its association with the warriors emerging from the rocks and the seated figures. Finally, if we take the saunas as a place of transformation, a third option could be considered. The relationship between warriors, saunas, rocks and water would suggest that heroization rituals were performed in these underground enclosures. In any case, it seems that, whatever the case may be, there is a clear potential for decorative art to function as a system of action, intended to transform the world rather than encode symbolic propositions about it.