ERIMI LAONIN TOU PORAKOU . A TEXTILE COMMUNITY OF PRACTICE IN MIDDLE BRONZE AGE CYPRUS

: Ongoing excavations at the Middle Bronze Age site of Erimi Laonin tou Porakou (Cyprus) have been revealing substantial evidence for textile production and clues of its economic impact and social implications to this prehistoric community. The analysis of the architecture and artefacts recovered in the workshop complex, in combination with archaeobotanical evidence, has allowed us to identify the presence of different textile activities, including dyeing. An additional textile data-set comes from coeval extra-mural funerary contexts pertaining to the Erimi settlement. As a key-site for understanding textile production during the Middle Cypriot period (MC), Erimi offers a privileged viewpoint on a large array of social dynamics connected to textile work. This paper will discuss the impact of textile production in the transmission of knowledge and expertise, and in the construction of the community mindset by applying the model of Communities of Practice (CoP) as a theoretical background.


Introduction
Recent scholarship has helped us to appreciate the relationship between theories of learning, transmission of knowledge and archaeological analysis. 1 Scholars like Lave and Wenger 2 have claimed that, far from being an abstract and culturally invariant process, learning is indeed a situated activity that develops through a strong social interaction and entails an "increasing participation in communities of practice". 3 The concept of 'Community of Practice' (CoP) has been particularly fortunate in socio-anthropological debate, 4 as its various contemporary applications to different disciplinary fields demonstrate. Its theoretical strength lies in considering the dynamics between social skills and technological skills in terms of mutual enhancement: its main characteristic consists of a "domain of knowledge, which defines a set of issues, community of people who care about this domain, and the shared practice that they are developing to be effective about this domain". 5 At the same time, this transmission of ideas and skills contributes to tacitly 6 create a shared social capital, 7 which makes such transfers socially relevant.
In this respect, archaeologists can benefit from the theoretical underpinnings of these approaches, insofar as they allow them to address two main archaeological problems. The first one concerns the symbolic role of the transmission of skills and practices among ancient communities and the extent to which social strategies of negotiation are visible in the archaeological record. To what extent are transfers of knowledge also meaningful in terms of the symbolic values they represent? Objects are of course more visible than ideas; however, objects might also be vectors of abstract knowledge. From this perspective, it can be argued that transmission processes are often not only bearers of a new technological know-how, but they also embody wider cultural meanings and social value that is attached to the transmission, assimilation and eventual dissemination of such practices.
This interpretative approach might help us to investigate into the concept of prehistoric community. 8 Consequently, the second set of theoretical issues that such a wider interpretative approach allows us to tackle is the connection between symbolic meaning and social change. From this perspective, 8 Sassman and Rudolphi 2001, 407-409;Webb and Frankel 2007, 191-192;Knappett 2010;Knappett 2011. the prehistoric community might be interpreted as a social construct where the division of social spaces into functional areas coexists with a progressive strengthening of common identity between community members within and without vis à vis external entities. 9 Transmission of knowledge, intended 9 Knapp 2003.  as a technological as well as cultural phenomenon, is therefore an important vector of social change, as it defines the role of social actors via their participation and involvement into communal activities, fosters their sense of belonging and facilitates their adhesion to the beliefs, standards and behaviours of the community. Indeed, transmission is interaction and on-going transformation at the same time. For this reason, it is both a synchronic model of analysis as well as a diachronic one. As a matter of fact, the introduction of new values and symbols through the transmission of communal practices is, for contemporary interpreters, a fundamental marker of the development of communities of practice in early societies, as it allows us to appreciate cultural and social shifts occurring in such societies. Against this theoretical background, after having briefly introduced the archaeological contexts of Middle Bronze Age Cyprus and the specific case of the textile industry at the proto-urban site of Erimi-Laonin tou Porakou, in this paper we will make a claim that this site can be actually interpreted as a textile Community of Practice.  (Fig. 1). Different functional areas have been identified within the site, such as domestic and working spaces and three funerary clusters. Furthermore, a planned organisation of the settlement spaces has been recognised. The hilltop area is occupied by the workshop complex (Area A), which has been currently investigated to encompass an area of 30 × 30 metres. 11 The architectural evidence and residual associated material assemblages found in the lower terrace immediately downwards the hilltop seem to indicate the presence of household complexes (Area B, Trench 2). 12 Two distinct funerary clusters (Area E, upper and lower terraces) have been identified in two terraces south to the domestic area, 13 whilst one other (the Ypsonas Vounaros cluster), situated east to the settlement, was rescue-excavated by the Department of Antiquities in 2012. 14 The foundation of a large wall (1.6/1.7 m wide and 0.6 m deep) filled with rubble stones and large stone blocks bound with plaster mortar was discovered during the fieldwork season of 2016. 15 The wall lies in a minor terrace south to the domestic quarter and immediately north of the upper funerary cluster in Area E. This peculiar location and the monumental dimensions of 10 Sciré Calabrisotto and Fedi 2017, 297-299;Webb 2017a, 202-205. 11 Bombardieri 2017a, 21-57. 12 Bombardieri 2017a. 13 Bombardieri 2017a, 73-74. 14 Christofi et al. 2015. 15 Bombardieri 2016, 4-5;Bombardieri 2017a, 363. the structure have led us to hypothesise its function as a circuit wall separating the settlement from the cemetery clusters and the exterior 16 (Fig. 2).

Erimi-Laonin
Textile production at Erimi-Laonin tou Porakou: primary and secondary evidence With regard to the settlement, substantial evidence for textile production has been detected in the workshop complex. Interestingly, at Erimi the study of abandonment deposits has revealed two alternative processes. 17 The first one refers to gradual structural degradation, as mostly observed in the domestic units (Area B) and is connected with the common behaviour of household members to remove objects on the abandonment of houses. 18 The second one involves a rapid destruction of the buildings, possibly caused by fire, which completely sealed the deposits and facilitated good preservation of artefacts (including textile artefacts), as observed in the workshop complex (Area A). Nonetheless, four spindle whorls have been recently recovered from Trench 2, 19 definitively proving the presence of spinning within domestic activities as expected for pre-industrial societies. However, the limited data available to date will not allow us to include the domestic quarter in our analysis. This paper will thus consider the evidence for textile production in the workshop complex (Area A) and the significance of textile tools both in the workshop and in the southern funerary clusters (Area E). In the following sections, selected meaningful contexts and assemblages will be analysed to investigate textile activities and possibly understand their organisation and scale of production. These will be discussed by applying the model of Communities of Practice (CoP) to evaluate the impact of textile activities as a constitutive element of ideology and social aspects of the Erimi community.

Unit SA I and textile dyeing
The roofed unit SA I in the workshop complex is a good starting point to begin an investigation of textile production at Erimi. As will be outlined below, a combination of different sources of evidence has allowed us to detect the presence of textile dyeing as the main productive activity carried out within this room and characterizing its function in the complex 20 ( Fig. 3; 4).
Two main occupational phases are attested in SA I. During the earliest phase (Phase B), SA I is characterised by two specific working installations, a circular hearth (Ft.42) and a large multiple basin (Ft.43) composed of a rectangular basin connected with a deep circular pit. 21 Archaeobotanical analyses revealed a high concentration of peculiar plant remains of recurrent wild species from the filling of Ft.43. 22 These have been identified as nutlets (>100) belonging to the family of Boraginaceae, to be referred to Mediterranean wild plants of genera Echium and Lithospermum (Buglossoides). 23 Because of their high concentration within the intact filling of a relatively small structure, these archaeobotanical remains can be considered as the residual materials from the processing of plants gathered for this purpose. 24 Currently under study, the seeds seem compatible to those produced by E. plantagineum, E. italicum, E. angustifolium, E. glomeratum (viper's bugloss genus) and Lithospermum/Buglossoides arvensis and L./B. tenuiflora. 25 Although most of these plants are known for their medicinal properties, some could have been used also to produce a red/purple dye extracted from their roots. 26 No evidence for the exploitation of these species as dye plants in the prehistoric and protohistoric Mediterranean is available. 27 Nevertheless, in Cyprus, analogous species have been identified at sites where textile production is particularly evident. For example, Echium vulgare (viper's bugloss) was found in City Quarter 3 at Hala Sultan Tekke, where Chrozophora tinctoria (dyer's croton) was identified within a probable "pigment/dyeing processing space". 28 22 Sciré Calabrisotto et al. 2017, 282-283. 23 Sciré Calabrisotto et al. 2017, 282-283. 24 Muti 2017b, 24. 25 Sciré Calabrisotto et al. 2017, 282. 26 Murray 1998, 222;Vassio and Bombardieri 2014, 294. 27 Barber 1991, 227-235;Cardon 2007. 28 Kofel 2018, 571, 574-575,  The basin Ft.43 and the nearby hearth Ft.42 are likely to be connected, as boiling is a crucial part of textile dyeing. It remains uncertain, however, whether the basin was used as a dye-basin or as an installation for preparing the dye (e.g. for maceration). 29 Only a few, albeit meaningful, residual artefacts were recovered from the same context. These include pouring vessels (e.g. jugs), and ground stone tools (e.g. a pounder), along with other smaller installations, such as mortar-like structures. 30 During the most recent phase (Phase A), SA I appears to have been spatially re-organised. A monolithic threshold was placed at the entrance, and the interior space was divided into three sub-units. 31 Eight large pithoi were recovered from this phase context, along with a double-chambered hearth of a different type than that from Phase B. It is worth noticing that these pithoi have a large capacity 32 and allowed for storage or work with a considerable amount of liquids and may contain voluminous materials (e.g. textile fibres or fabrics). 33 A continuation in performing the same activities as those identified for the earlier phase can be hypothesised. Because of the higher quantity of pithoi recovered, their larger capacity if compared with Ft.43, and the presence of a high-performing hearth, textile dyeing in SA I seems to have been carried out on a larger scale in the later phase of the Middle Cypriot. 29 Muti 2017b, 25-26. 30 Bombardieri 2017a, 32-33,   The pithoi, stabilised within their emplacements and in close association to the hearth, could have operated in the same manner as the earlier multiple basin Ft.43 rather than functioning as storage vessels. 34 The residual artefact assemblage is also extremely indicative of this and is characterised by vessels for pouring and drawing liquids, and cooking pots intended for boiling. 35 Remains of plants of the same and other families, some of which could have been used for textile dyeing, have been recognised for this phase. 36 However, the very limited quantity recovered does not allow us to definitively assert their use as dyeing plants.

SA I in context: the organisation of the workshop complex
SA I is not the only spatial unit in the workshop to be characterised by textile activities, but the overall organisation and function of the complex is likely to be related to textile production. The workshop complex comprises semiroofed and roofed units (SA I-X) and a series of open spaces (WA I-VI) with a great concentration of plastered basins of different shapes carved into the limestone bedrock and drainage channels. This suggests that activities involving the use of water or liquids (e.g. washing) were extensive. 37 Such installations and related activities are characteristic of diverse steps of textile production and are particularly related to textile washing and dyeing. 38 The presence of aligned postholes in the open spaces is also relevant. In fact, they are interpreted as supports for informal structures -such as party fences or sheds, and temporary shelters for workstations -which could have been used for drying yarn or textiles. 39 In this perspective, the concentration of open spaces in the northernmost wing of the complex with direct exposure to wind channeled within the river valley is highly significant. 40 Also noteworthy are the two roofed units adjacent to SA I (SA II and SA III), from which homogeneous material assemblages have been recovered. These are mainly characterised by vessels for liquid processing and storage, cooking pots and selected grinding and pounding stone tools, 41 which seem indicative of a series of activities normally performed in textile dyeing. 42 Moreover, during the most recent phase of room a in SA II (Phase A), grinding tools were found being clearly associated with mortar-like installations and benches suggesting that grinding activities were practiced inside the room. 43 34 Muti 2017b, 26. 35 Bombardieri 2017a, 34. 36 Sciré Calabrisotto 2017 Bombardieri et al. 2015. 38 Alberti 2007, 59-60;Alberti 2008, 27, 31. 39 Bombardieri 2017a, 348. 40 Bombardieri 2017a, 348. 41 Bombardieri 2017a, 28-45;Webb 2017a, 132;Webb 2017b, 207-209, 211-217. 42 Alberti 2008, 27, 31. 43 Bombardieri 2017a Textile tools A significant quantity of spindle whorls was recovered both from the workshop complex and from burial contexts. At Erimi, spindle whorls represent the only category of textile tools identifiable a priori and without any relation to their contexts. Their contexts and association occurrences, however, may change their significance, and spindle whorls from the workshop complex and Southern Cemetery clusters will be discussed below in different sections.

Spindle whorls from the workshop complex
Sixteen terracotta spindle whorls were recovered during the 2009-2014 campaigns in the workshop complex. 44 Another five examples were found in the roofed unit SA VI during the last fieldwork campaign of 2017. These last findings have only been subject to preliminary analysis, but their form, decoration and functional parameters seem uniform to those observed for the artefacts from the previous campaigns. 45 Spindle whorls have been recovered primarily from the roofed units, including SA I. A few have been found also in the open spaces. Whorls from the workshop complex can be ascribed to either of the two main occupational phases, although the largest number belongs to the most recent one. 46 The majority are biconical and belong to Crewe's type II. 47 Concerning their decoration, both decorated and undecorated spindle whorls have been recovered from the workshop. Decorated whorls show typical incised geometric patterns filled with a white paste. Notably, most show decorative patterns characteristic of the South Coast region. 48 About 1/3 of the spindle whorls recovered from the workshop are undecorated, which are the most common style in settlement contexts in Cyprus (e.g. Marki Alonia, Alambra Mouttes). 49 Interestingly, a high ratio of undecorated spindle whorls has been found at Pyrgos Mavroraki, in a loosely contemporary complex interpreted as a workshop for metallurgy, textile and other production. 50 In relation to whorl operation, four weight classes have been identified at Erimi, ranging from light (10-30 g) to very heavy (70-90 g). Even if all the weight classes are represented in the workshop complex, the majority of whorls belong to medium or medium/heavy classes. 51 The calculation of the ratio of the speed (r = h/d) 52 has revealed that the whorls from the workshop were fast. 53 Recent studies have shown that there 44 Muti 2017a, 219, 220, is no univocal correspondence between the weight of whorls and fibre types, and the human component plays a fundamental role in the characteristics of the spun yarn. 54 Nonetheless, the absence of very light whorls, normally related to the production of very fine yarn, 55 may be correlated to dyeing activities for which yarn is required to be strong enough to be dyed without damage. 56 Use wear compatible with low whorl spinning has been detected around the narrow extremities of the whorls from the workshop complex. 57 Similar use wear patterns are recurrent on a large quantity of EC-MC spindle whorls from different sites, 58 and signify an extensive and continuous use of these artefacts. Whorl findspots are normally less diagnostic when compared with findspots of other textile tools. 59 It has been widely recognised that spindles and spindle whorls are portable objects, and no designated spaces for spinning were needed. Moreover, spindle whorls are small finds which can be easily lost or moved through taphonomic processes. Nonetheless, two contexts in the workshop seem very suggestive of primary textile working activity. The first one corresponds to a group of six spindle whorls found in an area between the threshold (Ft.10) and a wall (W 3) of unit SA IIb, Phase A. 60 These whorls were found clearly grouped and in association with the particular assemblage of the room, mainly characterised by 54 Kania 2013. 55 Grömer 2005, 109-111. 56 Muti 2017a, 231. 57 Muti 2017a, 228-229. 58 Crewe 1998, 61-62. 59 For loom weights, see, e.g. Barber 1991, 102. 60 Bombardieri 2014, 49;Bombardieri 2017a, 43. grinding tools and storage containers and vessels for liquid processing, as well as a picrolite comb-shaped pendant. 61 Another significant findspot is that of whorl P55. This spindle whorl was found in association with a large shell and some ceramic fragments belonging to large closed vessels within a circular pit carved in the floor of the SA VI. 62 Although the meaning remains unclear, it is interesting to notice that this feature, dated back to the earlier occupational phase (Phase B), was sealed by the thick floor plaster of the later Phase A. 63

Spindle whorls from the Southern Cemetery clusters
Twenty-two terracotta spindle whorls were recovered from five out of twelve excavated pit and chamber tombs in the Southern Cemetery clusters during the fieldwork campaigns of 2009-2014. 64 Another 22 artefacts, currently under study, were recovered from two chamber tombs and a multiple chamber tomb in the most recent 2016 and 2017 fieldwork seasons (Muti, personal observation, 2017). 65 In comparison to the spindle whorls recovered in the workshop, no significant variance in operational attributes are evident, but remarkable differences in type distribution and decoration are apparent. 66   deposits have more complex and more carefully executed decoration and were likely to have been specially selected as grave goods. 68 Spindle whorls from the Erimi cemetery clusters, as well as from coeval cemeteries, show use wear traces. 69 It is thus evident that they were not specifically made as grave goods.
The frequency of whorls per tomb at Erimi is higher than in contemporary funerary contexts in Cyprus, 70 and the most recently excavated tombs not only confirm this trend, but are exceptional in this sense. 71 For example, twelve whorls were recovered from the chamber tomb 429, nine from tomb 464C, and seven whorls from the pit tomb 231. Interestingly, in the whorl assemblage of tomb 231 an object of a very unusual shape (T231.14) was found. It is a 'double' spindle whorl composed of two spherical whorls of different sizes joined at one terminal 72 (Fig. 5). Only a few comparable examples exist for this whorl type: one from Alambra Mouttes (E80), one from Mesoyi (PM3023/4) and four unprovenanced. 73 Thus, the 'double' whorl can be considered as a functional tool, but its rarity and context might also charge it with a special value. 74 One other interesting aspect to consider is 68 Muti 2017a, 231. 69 Crewe 1998, 61-62;Muti 2017a, 228-229. 70 Muti 2017a, 231. 71 Muti, personal observation. 72 Muti 2017a, 232, 235, Fig. 6.10. 73 Bombardieri 2014, 50. As to the unprovenanced examples, three 'double' spindle whorls are stored in the Tsirides Collection, and one was in the ex-Morris Collection. 74 Bombardieri 2014, 49, 52-53. whorls' findspot in the burial space. In several cases whorls were recovered in the space reserved to the deceased (T.231) or in proximity to bodies (T.248; T.428) or concentrated in an area (T.230; T.429; T.464C) (Tab. 1).

Possible representations of textile tools
and special objects from the settlement A few peculiar artefacts recovered from the site seem to have a special connection with textiles, although they are not textile tools, nor they appear to have had a practical function. Some can be considered as valuable objects (stone ornaments, a ritual askos), others are rare and ambiguous, but do not seem to be particularly precious (clay weights).
Two stone comb-shaped pendants have been recovered at Erimi, respectively from unit SA IIb in the workshop context (A.428.1), and from the depositional floor of tomb 240 (T.240.10) 75 (Fig. 6). It is worth reiterating that pendant A.428.1 was found in proximity to a group of six spindle whorls (see above). Different interpretations have been proposed to explain the peculiar form of these pendants. 76 Among these, an influential hypothesis was made by Peltenburg, 77 who interpreted their shape as a reproduction of wool combs. More recently, Knox 78 proposed that they were models depicting beaters for tapestry weaving. If we accept either one of these interpretations, the pendants can be considered as 75 Bombardieri 2017a, 240-241. 76 See, e.g. Bombardieri, forthcoming. 77 Peltenburg 1981, 23. 78 Knox 2012, 142. textile-related symbols and are likely to assume even stronger connotations as "symbolising community identity and status" at Erimi. 79 In the same abandonment context of unit SA III (workshop complex), an unusual group of nine unfired or low-fired conical objects and a special terracotta vessel were found. 80 For manufacture, form and size, the nine conical objects are identical to the most common EC/MC loom weight type. 81 Nonetheless, they are not perforated, and seem not to be functional. Instead, two of them show respectively two carved, non-perforated holes at each side below the apex, where the perforation can normally be seen. 82 This set of pseudo-loom weights has no parallels, and two explanations may be proposed. They might have had a different function from that of weaving tools (e.g. another kind of weights?) or might be representations of textile tools. 83 The last hypothesis would explain the presence of holes instead of perforations. Notably, terracotta models of spindles are present in the EC-MC artistic repertoire. These, however, are exclusively from burial contexts (Vounous), and no examples made of unfired clay are known. 84 A zoomorphic askos shaped as an ovicaprid was found not far from the findspot of the 'pseudo loom weights'. 85 These special containers are normally found in burial contexts, 86 and 79 Bombardieri 2017a, 358. 80 Bombardieri 2017a, 49. 81 Muti 2017a, 236-239, Fig. 6.12. For EC/MC loom weights see Frankel and Webb 2006, 175-177. 82 Muti 2017a, 236-237. 83 Muti 2017a, 237-238. 84 Morris 1985, 253-254. 85 Bombardieri 2017aWebb 2017, 164-165, Fig. 4.31. 86 Webb 2017, 165. possibly had a ritual function. The shape of this vessel and its association in occurrence with the 'pseudo loom weights' seems significant in the light of the textile activities performed in the workshop. 87 The textile symbolic capital of Erimi The evidence presented above suggests that textile production at Erimi was extremely relevant, both economically and socially. At the current stage of research, the Erimi economy seems to have been based on textile manufacture, mainly aimed at the production of dyed yarn or textiles, and probably also for trade. The textile activities performed in the workshop complex were extremely beneficial to this community and seemed to increase its wealth throughout time. 88 In parallel, they contributed to the construction and strengthening the community bonds and the elaboration of a community mindset, shared values and symbolism. 89 These different aspects related to textile production are, in fact, two faces of the same coin. By applying the CoP model to the interpretation of the Erimi data-set, it becomes evident that this community found its focal centre (and raison d'être) not only in performing specialised working activities, but also in their reproduction and transmission. 87 Bombardieri 2017a, 358. No functional loom weights were retrieved from the units excavated before 2017 (SA I, SA II, SA III, SA IV and SA VI). Complete and fragmentary loom weights were recovered from the more recently investigated unit SA V and are currently under study. 88 Bombardieri 2017a, 355-358. 89 Bombardieri 2017a, 353-355, 358. The workshop complex is the first element which may be read through the CoP lens. Because of its prominent position on the hilltop and pivotal role to the community life the workshop can be identified as a 'thematic field'. Within CoP theory, the term 'thematic fields' indicates all aspects which bind the people who participate in them. 90 The well-planned organisation of the workshop indicates that the Erimi community put effort, resources and expectations in seeking out and organising the ideal location for the foundation of a working area. As no evidence for institutionalised or centralised control is visible at Erimi, we deliberately refer to the whole community as agent both to the construction and use of these spaces. This does not exclude the contribution of individuals (e.g. task differentiation), but highlights an extraordinary coordinated effort to pursue a common aim. In this sense, the workshop complex is not only a central space, but a social arena of intertwined collective and individual relationships.
Also -but not exclusively -related to the workshop's spatial features are labour strategies, which can be likely identified as another 'thematic field'. 91 The existence of a designated workplace implies the presence of specific work tasks and schedules. Although the presence of certain steps is problematic to detect, a textile chaîne opératoire can be reconstructed in the workshop complex. The combination of the different pieces of evidence discussed above has allowed us to identify that at least spinning, preparation of dyes, dyeing, washing and drying (not exclusively in this order as washing and dyeing may also occur during the preparation of textile fibres) took place in the workshop. As spindle whorls were also recovered from the settlement, it is logical to think that textile activities were also practiced at a domestic level. Thereby, two textile chaînes opératoires may have occurred at Erimi. Although no evidence in this sense is available, it cannot be excluded that these two chaînes opératoires partially overlapped. For example, it is not unlikely that domestic spinning integrated the production of spun thread in the workshop as the quantity of thread needed is normally high -even higher at an extra-domestic level -and spinning is extremely time-consuming. 92 Cooperation and mutual engagement seem distinctive traits not only of spinning but of all the textile activities performed at Erimi. The growth or procurement of raw materials, and preparation activities occurring before spinning are considered as part of the textile chaîne(s) opératoire(s). The collaboration between community members occurred within and outside the workshop but, in a way, the workshop is always central as a collection/arrival point of the activities performed outside. This massive community effort denotes the elaboration of a deeply rooted sense of community through textile activities as practice and the working 90 Wenger 1998. 91 Bombardieri 2013 Andersson Strand 2015, 46;Olofsson et al. 2015, 96-97. performances as "mutual engagement over time". 93 This is transformed into another aspect of practice, i.e., shared knowledge, negotiation of significances, elaboration of symbols and transmission of a legacy derived from reflection and awareness of activities practiced. 94 Before analysing these cognitive and ideological aspects, a paradox within the thematic field of cooperative labour can be detected. No evidence of institutionalised or centralised control of wealth is attested at Erimi but it is possible to identify clues for the emergence of social differentiation and conflict. 95 Although all the activities in the chaîne opératoire -from the procurement of raw materials to dyeing -contribute to the final aim of producing dyed yarn (or fabrics), not all the 'textile' tasks are (thus perceived as) equal. Some require a large workforce (e.g. procurement, processing, spinning, washing), others (e.g. preparing dyes and dyeing) a considerably reduced labour force. More importantly, certain crucial activities (e.g. spinning and dyeing) presuppose specific knowledge and know-how. 96 The concentration of such activities in certain open or closed spaces of the workshop might not be exclusively related to their functionality. On the one hand, it is functional to set washing and drying in large open spaces. Nevertheless, the fact that textile dyeing activities are concentrated in a single closed space, possibly provided with controlled access, might be also related to permitting access only to selected people. 97 As the other rooms have the same access features, 98 this may be valid for all the closed spaces in the workshop complex.
As anticipated, defining Erimi as a CoP means that the construction of the community identity and the ideology through which this is expressed is embedded with the activities performed by the community. 99 It is not infrequent that specific working activities are represented though the related tools. To investigate this specific aspect, we have looked in particular at textile tools and special objects possibly revealing a connection to textiles.
Spindle whorls appear significant for various reasons, such as quantity (in comparison with other EC-MC sites), 100 formal characteristics, association occurrences and findspots. It is relevant that not only in tombs, but also in the workshop these artefacts might have been charged with an additional value. It is worth highlighting that no other complete artefact but a spindle whorl was selected for burial with the shell in the circular pit in SA VI. Although the shell is under study and the species not yet identified, shells recovered from Area A cannot be related to textile production as they were not 93 Wenger 2013, 230. 94 Wenger 2013. 95 Bombardieri 2017a, 361-362. 96 Cardon 2007. 97 Bombardieri 2017a, 353. 98 Bombardieri 2017a, 28-57. 99 Bombardieri 2017a, 358. 100 Muti 2017b, 24. suitable for producing dyes. 101 However, they were likely to express a symbolic or ritual significance. 102 The group of six whorls from SA IIb seems also significant. These whorls do not show marked differences in functional parameters as to be designated as a set 103 but all show marked use wear. As the abandonment of the workshop may imply closure rituals and exposure to controlled fire, 104 these whorls seemed intentionally grouped as used and probably 'belonging' to the workshop. Their association with the comb-shaped pendant is also evocative of the symbolism related to textile work. 105 Although we cannot definitively prove that spindle whorls recovered from the cemetery were used in the workshop production, the homogeneity of the functional parameter of whorls found in both the workshop and in the tombs may reflect the specific productive goals of this community. Certainly, spindle whorls were used before transforming their significance and their significance changed when transformed into grave goods. At Erimi, the presence of whorls in tombs is higher than in coeval cemeteries, and this seems to be related to the specific importance of this category of objects for the Erimi community. 106 The symbolic significance assumed by whorls as grave goods is not exclusively related to their importance for expressing community identity, but also to personal identity. 107 As observed above, the concept of CoP does not call off the individual component from the model. In particular, it has been noted that in the Erimi burials whorls appear as selected, and often tombs contain artefacts of the same types and showing analogous decorative motifs. 108 This can be considered as additional proof of Frankel's and Webb's 109 observation that the recurrence of certain design patterns is to be referred to family groups, and the diffusion of similar design patterns among neighbouring areas was produced by movements of people. Whilst there is no direct correlation between burials and spindle whorls at Erimi nor in other EC-MC cemeteries, 110 the relationship between individuals and whorls seems to be a means for conveying other elements of personal identity (e.g. belonging to a family group, being socially recognised as a spinner). This is further proven by the whorls' location within the burial space. Their proximity to human remains might indicate a metaphorical proximity to the individual.
The recurrence of other artefacts possibly connected to the domain of textile production but not functional to production (e.g. ornaments, ritual vessels, non-functional tools), is additional evidence in support of the thesis that textile activities permeated and helped forge the community mindset. This may represent a further step of idealisation within the elaboration of textile symbolism. The elaboration of such a complex ideology can be considered as an expression of a community 'code' and, at the same time, suggests the will and energy to transmit both practical and ideological aspects of textile practice.
In conclusion, the application of the CoP model is extremely beneficial in allowing us to gain a clearer understanding of the social dynamics operating within the Erimi community in relation to working strategies. This model has permitted us to explore the level of cooperation and interaction between the community members around common activities and aims, the elaboration and perpetuation of a 'code' of shared technological and ideological behaviours through a tacit transfer of knowledge. The concept of tacit knowledge 111 is itself beneficial to the analysis of pre-literate communities, such as Erimi. In this perspective, the sense of joint enterprise towards textile production, which animated the local community, had a strong impact on the elaboration of its ideological system through time. The transformation of local crafts and tools into important symbols in the expression and maintenance of social identity resulted in the creation a powerful ideological legacy, namely the symbolic capital of this community. 112