Identity is an ongoing process of establishing relationships, of finding a sense of self and belonging, and of maintaining a place in a world that we ourselves constantly change (Diaz-Andreu and Lucy 2005). Identity requires continued renegotiation. It is also enabled and enhanced by the material world. We have a fascinating relationship with this world in that material things make us as much as we make them. But never is this relationship as immediate as with dress and personal adornment which allow us to materialise, display and communicate our ideas of our self directly on our bodies.

In current perspectives on the mechanisms that carried identity-forming processes in the early Anglo-Saxon period, the act of adorning and displaying the dead has been conceptualised as a collective strategy of reinforcing understandings of affiliation and social boundaries (Härke 1992; Lucy 1998; Sayer 2013; Stoodley 1999). Increasingly, ideas about how people identify themselves with and through daily used objects also influence how elements of early Anglo-Saxon bodily adornment are now being addressed in artefact studies (Brunning 2013; Felder 2014; Inker 2000; Martin 2012). These focus on identity as something experienced and created by the individual in everyday contexts, and on the production, use, use-adaptation or disuse of items of dress as fields of individual agency. Cemetery- and artefact-based studies of early Anglo-Saxon mortuary evidence have often addressed the link between dressing, human agency and identity from different angles: highlighting different human actors (collective and individual), domains of bodily display (funerary and daily worn dress) and levels of archaeological context (the cemetery, the single burial and the artefact). Recent studies of early Anglo-Saxon identity that have focused on single artefact classes have begun to develop frameworks that integrate all of these themes and types of evidence (Felder 2014; Martin 2011).

Studies like the ones named above are founded on the idea of identity as created and transmitted through social communication. However, there is a need for explicit approaches that conceptualise and develop methodologically how the dynamics and effects of this communication can be investigated through the material record. Communication is central to sharing knowledge about social meanings and creating notions of identity. The present paper thus argues that in the archaeological study of identity as a human experience, modes of communication and transmission, and particularly their materiality, need to be considered more explicitly. The conceptual outline it presents is built on the understanding that social communication is a fundamentally material practice—it takes place through material culture. Particular focus is on the transmission of encoded meanings within practices of production and consumption of dress. Artefact meaning is viewed here as created at all stages of the object lifecycle, and the communication of this meaning as a process in which many people are involved. The way in which artefact-analytical methods and methods of contextual cemetery analysis are combined in this approach is influenced by the network concept which has attracted increased attention in archaeological studies over recent years (Knappett 2011, 2013b). The analytical concept of the network is useful in two ways: for coming to terms theoretically with the complex structure of the human interactions that are part of the process of identity formation; and as a conceptual step towards the application of network-analytical methods which offer innovative ways of investigating such interactions through network analyses of distributed material culture (a task that will have to be reserved for a future paper).

As a case study, this paper uses the early Anglo-Saxon girdle-hanger1. This type of object occurs only selectively in the burial record and has traditionally been regarded as a symbol of a specific female identity. The conceptual considerations presented in the first part of the paper are applied to the data presented here by beginning at the level of the artefact and working gradually through different social contexts of production, use and disuse. In each of these contexts, the focus is on the ways in which notions of personal and social meaning were visibly assigned to girdle-hangers by different people and specific practices. The aspect of materiality and visibility of these practices had a significant function. It allowed the meanings of these objects to be non-verbally communicated and transmitted across space and over time—enabling people to internalize this meaning, and girdle-hangers to become a means of identification with a commonly shared cultural value.

Girdle-hangers derive their form from functional iron keys with T- or W-shaped terminals known since the Iron Age (Manning 1989: 90). At the lower end of a narrow shank the object typically extends to a horizontal base element which forks upward on either end to form two prongs (Fig. 1). Girdle-hangers most often carry a moulded element below a transverse suspension loop at the upper end, and their body surface commonly carries punched, incised and notched decoration. There are many indications that girdle-hangers were purely non-functional imitations of keys. Lock mechanisms typical for T-keys required the insertion of the prongs into fitted holes in wooden bolts which had to be lifted or slid aside by the key in order to release a bar (Manning 1989: 90). No traces of wear corresponding to use in a technical construction have been observed on girdle-hangers, and the build of the 100–160mm long flat cast copper-alloy objects whose terminals were rarely thicker than 2mm also makes such a function unlikely. Prongs of girdle-hangers often end in extended decorated lobes which in many cases are connected with the central shank through horizontal or tilted extensions (see Fig. 1), making them unusable for lock mechanisms like the ones described.

Girdle-hangers, commonly occurring in pairs, are always found associated with female jewellery in graves of the 5^th and 6^th centuries AD. Where bone material could be aged and/or sexed, 83% of the individuals were identified as adult, 90% as women. Girdle-hangers were suspended by a construction with U-shaped hoop and horizontal bar and mostly worn on the left hip (see examples in Fig. 1; Lucy 2000: fig. 3.7 for a reconstruction). Despite early comparisons to continental openwork discs with chain elements (Smith 1852), there is no evidence so far that girdle-hangers were used to suspend further items. The total number of grave contexts with girdle-hangers recorded in the present study comprises 80 inhumations and 12 cremations, which altogether contained 155 girdle-hanger specimens. An additional 98 unstratified specimens recorded in 19^th-century antiquarian collections and during modern cemetery excavations, despite missing a grave context, can be regarded as conjectural grave goods based on their good state of preservation, often as intact specimens, and on their recording in pairs.

The database assembled for the author’s study comprises just over 600 (complete and fragmented) single specimens of girdle-hangers. Approximately 300 of these were found by metal-detectorists. These are finds with known findspot data, but missing contextual data, recorded in the Historical Environment Records (HER) and the Portable Antiquities Scheme (PAS) database. Such finds are not suitable for contextual analysis, but there are clear benefits in including metal-detected finds in studies of early Anglo-Saxon artefacts (Chester-Kadwell 2007; Martin 2011; McClean and Richardson 2010). They provide, for instance, a much more comprehensive data basis for technological, typological and spatial analysis2.

Girdle-hangers were recorded in 39 cemeteries with 5^th- and 6^th-century material across the east and north-east of England. Finds from antiquarian collections and modern ploughed-up material indicate that the original number of burial sites with girdle-hangers will have been higher. However, overall girdle-hangers are comparatively rare. Girdle-hangers are found in 0.5 to 5% of all graves with female jewellery at a cemetery. Only at the cemetery at Morningthorpe, Norfolk, a higher number of female gendered graves with girdle-hangers (10%, n=9) was recorded. This, however, was shown to be a locally specific exception (Felder 2014).

Ideas about the possible meanings of girdle-hangers began to take shape from the mid 19^th century (Smith 1852: 235; Neville 1852).3 Their enigmatic imitative character and their infrequent occurrence in the archaeological record prompted ideas of a specific, not immediately accessible, symbolism. Smith was reminded of Victorian châtelaines (Smith 1852: 235), a construction worn by women to suspend small household items which often included keys. In another context he suggested that “keys may be particularly considered as insignia of the Saxon women, as they were, to a comparatively late period, of the English housewife“ (1856: xli). Similar connections were also drawn between female key possession and the domestic authority of the housewife in the Roman period (Pitt-Rivers 1883: 15). This was taken as evidence for the possible origins of this concept in pre-Anglo-Saxon times, and provided a closer historical analogy for a corresponding interpretation of early Anglo-Saxon girdle-hangers (Lethbridge 1931: 5; Steuer 1982: 204). Historical analogies were then further pursued in the later 20^th century with a stronger focus on medieval source material. It was shown that the key as both physical symbol and metaphor of the married woman’s role in managing the household and its material resources was prevalent in the wider Western European medieval world by the 11^th century (Gräslund 1999; Hines 1997: 263; Smith 1856, xli; Steuer 1982). Particularly significant for early Anglo-Saxon contexts is Fell’s (1984) study that traces this meaning in the early 7^th-century laws of Aethelbert of Kent.

Susan Hirst’s publication (1985) of the early Anglo-Saxon cemetery at Sewerby, Yorkshire, and her discussion of girdle-hangers from the site introduced a corresponding interpretation of girdle-hangers into modern Anglo-Saxon grave good studies. In addition, her discussion of grave 49, a prominent wealthy burial at Sewerby which contained girdle-hangers (Hirst 1985: 38–43), had a decisive influence on ideas of the girdle-hanger as occurring in higher-status contexts. The girdle-hanger as a symbol of a woman’s economic authority in the household and an indicator of higher status thus still prevails in descriptions of girdle-hangers in modern publications (Evison 1994; Haughton and Powlesland 1999; Leahy 2007; Penn and Brugmann 2007; Sherlock and Welch 1992).

This very specific meaning had been assigned to girdle-hangers primarily through analogical inference, specifically from historical sources which were situated in either culturally or chronologically differing contexts. The need for a detailed archaeological study of the artefact group had repeatedly been expressed through the years (for instance, Hines 1992). However, while the last decades of the 20^th century saw extensive studies of the more regularly occurring early Anglo-Saxon artefact types (Dickinson and Härke 1992; Härke 1992; Hines 1993, 1997; Mortimer 1990; Swanton 1973 among others), archaeological observations on girdle-hangers remained anecdotal. More recently, increased interest in the girdle-hanger was reflected in topical choices of University dissertations (Felder 2009; Kerr 2009; Knox 2006),4 and the artefact group began to be included in quantitative cemetery evaluations (Leahy 2007) and chronological studies (Penn and Brugmann 2007). Current trends in early Anglo-Saxon artefact studies show a growing interest in dressing and wearing as a social field of action (Brunning 2013; Martin 2011, 2012; Weetch 2013).5 It is this discourse in which the present author’s study of girdle-hangers is situated.

Archaeology has long embraced the view that the human experience of identity, belonging to a group and distinguishing oneself from others, is inseparably tied to material culture (Diaz-Andreu and Lucy 2005; Insoll 2007; Sofaer 2007). The foundation of identity in material practices is thus one of the many fields where archaeology can particularly contribute to wider fields of social and anthropological study. Indeed, significant early contributions to a theoretical framework that integrates identity, material culture and the archaeological burial record have come from within Anglo-Saxon studies (Pader 1982; Härke 1992; Lucy 1998). These have characterised the early Anglo-Saxon burial ritual as a complex, socially regulated material performance where ruling social norms were enacted and group identities shaped and reinforced.

Aspects of collective identity and negotiations of cultural and social affiliation are particularly strong themes in the story of the two centuries following the end of Roman Britain in the early 5^th century AD. In this period, Britain’s cultural and material landscape underwent a large-scale reconfiguration. Views of these socio-cultural changes have changed considerably in recent years. Traditional notions which saw the larger-scale material-cultural zones known from the 5^th- and 6^th-century cemetery record as static territories of historically attested Anglian and Saxon tribes have long come under critical scrutiny (Hills 2003). Our knowledge of the varying demographic scales and regionally distinct courses of early Germanic immigration is continuously growing (most recently, Hills and Lucy 2013), enabling us to draw increasingly enhanced pictures of a landscape of complex localised material traditions. However, larger-scale zones of affiliation, apparent in broad distribution patterns of characteristic items of female costume, do remain a focus of interest. One zone, for instance, is emerging in the overlapping distributions of wrist-clasps (Hines 1993; Lucy 2000: fig. 5.5(d)), later 5^th- and earlier 6^th-century cruciform brooches (Martin 2011: table 3.17, maps 5–8) and girdle-hangers (Fig. 2). Based on the current state of research this is now understood as a broader region of collective identity newly emerging in the later 5^th century.

Importantly, however, the broader cultural affiliations within such larger zones intersected with multiple other identities. Through multivariate analyses of parameters such as grave good provision, body position, grave construction and space use (Carver et al. 2009; Hakenbeck 2011; Härke 1992; Lucy 1998; Malim and Hines 1998; Pader 1982; Sayer 2009, 2010; Stoodley 1999), many studies of early Anglo-Saxon burial sites have shown the great diversity in how communities materially dealt with aspects of social, gender, ethnic, regional, local and household/family identity. Such practices could differ considerably between and even within sites, shifted over time and, most importantly, were constantly at work at once. This places considerable challenges on interpretations of the rich early Anglo-Saxon burial record. It invites us to reconsider how we can conceptually draw together these multiple scales and practices of identity and investigate them with adapted methodologies.

The act of negotiating identity takes place on different social and phenomenological scales, “from the personal to the political” (Dobres 1999: 17), performed by individuals as well as entire societies. Analytically it can be looked at on different levels of data resolution, from the individual body to larger-scale spatial distribution patterns. None of these dimensions exist in isolation; they are all connected, making identity an immensely complex social phenomenon. The archaeological study of identity is faced with the challenge of unravelling these multiple intersecting scales and, at the same time, making the links visible.

For some years now, network-based approaches have been gaining in popularity in archaeology as a way of understanding complex interrelations in the archaeological record and accessing human interactions in the past. Networks provide thinking tools and analytical techniques for conceptualising, visualising and explaining the interdependencies between different variables within these interactions. Formal analytical approaches in archaeology have been influenced by Social Network Analysis (SNA), developed in the social sciences (Wasserman and Faust 1994), and network-scientific advances in physics and complexity science (Knappett 2013a).

Archaeology inherently works with incomplete material datasets that have to serve as proxy for past human activity. Archaeological network approaches thus require critical reflection and adaptations of the available models and techniques in order to fill the gaps between material representations of human activity with actual people and their actions (Brughmans 2010; Knappett 2011, 2012, 2013a). The most important insight is that human interaction and the building of social networks never take place between humans alone, but are fundamentally constituted by material culture. Material culture is a variable that creates links and relations between people; it can be a precursor for human actions and interaction as much as a result of them. In Actor-Network-Theory, this notion is represented in the characterisation of both humans and objects as equivalent actors in networks (Latour 2005: 70–82). The way to understanding human interaction and social relations according to Latour (2005: 5) is to follow the trail of associations between these actors. Archaeology has particular expertise in identifying material associations between sites, assemblages or artefactual traits (see especially artefact analysis below). By following these associations across space and time it is able to trace human interactions where humans are mostly physically absent.

Formal applications of network analysis in archaeology are not without innate methodological difficulties and limitations (see Brughmans 2012; Isaksen 2013; Knox et al. 2006 for discussions), but they are proving their potential to enhance knowledge of past human interaction in a growing number of archaeological studies. They have been shown to be particularly useful in studies of regional interaction (Knappett 2013b) and especially larger-scale spatial analyses, for instance in studies of long-distance trade and contacts, centrality and (proto-) urbanism (Fulminante 2012; Knappett et al. 2008; Sindbaek 2007; Riede 2014). In archaeology, these fields have a longer tradition of using the term network to describe larger-scale interaction spheres such as trade networks and political networks. In these the distributions of artefacts and assemblages across wider geographical spaces have readily been understood as connectors between network nodes such as settlements and markets, and encouraged the development of purpose-built applications of network analysis.

Smaller-scale, more intimate dimensions of human interaction on the other hand are still addressed less frequently (Knappett 2011: 9–11, 2013a: 12). Different from studies of larger geographical scope, they can be viewed as taking place in the social space of everyday, face-to-face interaction. Here, the network is particularly useful as a thinking tool for coming to terms with the social and mental spaces of human interactions which are archaeologically more elusive (cf. Evans and Felder 2014). The present study focuses on these smaller-scale interactions.

Identity is created, performed and transformed in this social space of everyday, face-to-face interaction. It fundamentally builds on the exchange of ideas about what things, people and practices meant for identity—it is formed in interaction. Material culture is built into this concept as a medium of human communication, suited for the transmission of information and meaning (Jessen and Jensen 2012; Wobst 1999). Accordingly, humans and objects can be thought of as forming a network in which material-cultural meanings were shaped and transmitted (see, for instance, Knappett’s (2012) exploration of semiotic networks in a study of Minoan pottery). The following addresses how this idea can be incorporated into the archaeological study of identity and its specific materialities by considering the function of dress in social communication.

We identify with a group by using outward devices, certain artefacts for instance, which both we and other members of the group understand as a representation of group membership. Dress is an ideal context for conveying notions of affiliation directly on the human body, affiliations which would be difficult or even impossible to communicate efficiently without the help of an instantly visible material medium (Wobst 1999: 120). In characterising this role of worn material culture, authors have in fact often adopted terms that describe conversational processes—aspects of identity are talked about materially, communicated non-verbally through style, or reinforced through repeated citation (Gilchrist 1999: 82; Wiessner 1990: 106–107; Wobst 1999: 120).

Visibility plays a crucial role in making the communication of identity work (White 1989). This is especially prevalent in the early Anglo-Saxon period where identity is conspicuously performed on the adorned human body, especially by women who wear socially and regionally specific costumes which can include ostentatiously decorated dress fasteners. Equally, identity relies on readability, considering that the meanings conveyed by dress are subtle and indirect rather than instantly apparent from the artefact (see Knappett 2012: 87–88 for an expanded discussion of direct and indirect perceptions of meaning). This refers, for instance, to the indirect meaning conveyed by the design of a brooch or where it is worn, as opposed to the directly perceivable meaning that the brooch has in its function as a dress fastener. Readability of such indirect meanings relies much more on continuous reinforcement through social dialogue and material performance.

The majority of early Anglo-Saxon archaeological finds come from burials, and so the study of identity in this period is inseparable from the investigation of the funerary record. Where this focuses on the social event of the funeral, the material enactment of identity is viewed from the perspective of the group. Items of dress are investigated at the last stage of their material and social lifecycle, used by mourners as grave goods. The burial is viewed as a social arena which communities used to describe an idealised version of society. Techniques such as dressing the dead, placing and constructing the grave, laying out the body and gathering for the event of the funeral functioned as narrative aids and mnemonic devices (Sayer 2013: 155; Williams 2006) in the conveyance of this social message to the community. Statements of affection, affiliation and social position were directed at recipients on different personal, local and regional levels, and single burials could convey statements about multiple nested identities of the individual (Hakenbeck 2007, 2011).

However, death and burial formed only one of many contexts in which body ornamentation was used to create images of identity (White 1989: 214). Taking place directly on the individual human body, creating identity through dress is always an immediate physical experience of the living. Fowler (2004) has argued that funerary practices form no parallel universe, but rather bring idealised and lived images of human identity together, encapsulating “all of the relationships that sustained and composed the person” (Fowler 2004: 84; similarly Sayer 2009: 155). “It would seem unlikely that death and the dead formed an isolated sphere of interaction which was ideologically at odds with other everyday social relations” (Fowler 2004: 99). In his view, personhood is construed in both life and death according to the same principles. Consequently, funerary dress tells as much a story about the wearer’s self as it tells of collectively sanctioned rules and traditions (cf. Dobres 1999; Wobst 1999). This has significant bearings on how we interpret the assembled material record of the buried body.

Fowler (2004) is against a rigid conceptual separation between funerary costume and real-life cos­tume, arguing that burial practices pick up on identities which are the subject of real-life social discourses, in which people perform identity with meaningfully charged dress in daily contexts. Choices made by mourners must inevitably reflect daily discourse because only the meanings formed through this daily discourse can be understood by those who are addressed by the funerary tableau. Thus, even though we are dealing with burial evidence the investigation of performances of identity through dress cannot be limited to studying the collective practice of the funeral. It also requires the investigation of the dress found in a grave as a matter of lived experience and choice. This means the analysis of patterned burial practices in the early Anglo-Saxon cemetery record needs to be complemented by analyses of artefact wear and use and costume composition (see below). We cannot reveal the exact modalities of dressing in daily practice, but burials form a useful snapshot of these practices.

Meanings can also be subtly, often secondarily, encoded in artefact style. Importantly, this communication of meaning through style does not begin with the wearer’s choice, but is fundamentally determined by the maker’s choice (Wiessner 1990; Wobst 1999). Early Anglo-Saxon non-ferrous dress items provide a highly complex, dense material basis for analyses of style (see Hines 1997 and Martin 2011 among others). Studies of artefact meaning and identity should therefore also consider the makers of items of dress, and include technological and stylistic analysis. “Style always talks loudly about individuals” (Wobst 1999, 121), and style variation and change provide a fruitful ground for exploring how these individuals translated and communicated meaning. A maker will aim to give an object the physical appearance that will be readable to its anticipated recipients and adequately convey the social message intended with the object (Wobst 1999: 124). At the same time, artefact design is conditioned by a maker’s individual experience, knowledge, skill and vision (DeMarrais 2004). What has been said above about the assemblage of funerary dress equally applies to the style of single items of dress: it tells as much a story about the maker’s self as it tells of collectively sanctioned rules and traditions. As a further consequence, style inevitably changes, intended or unintended by the craftsperson. Each new artefact potentially reshapes the template for following artefacts, resulting in multiple, stylistically related sequences. Constructing systematic typological sequences is of course not new to archaeologists. However, the angle towards typological creation and development taken by Wobst and others (similarly Inker 2000 and recently Knappett 2012) changes significantly how they are socially interpreted.

In summary, in the study of identity and specific artefact meanings via the burial record it is essential to combine established methods of cemetery analysis with more specific, individualised analyses of artefact style, wear and costume composition. Such an approach has the potential to reveal how people and things could form semiotic networks (Knappett 2012: 104)—here, generating meaning in the social dialogue between makers, wearers and mourners. This is the approach adopted in the author’s study of the early Anglo-Saxon girdle-hanger (Felder 2014). The following will illustrate its practical application with selected data from this research.

In the present study the above approach is applied by looking at how the social and personal significance of girdle-hangers was perceived, communicated and reinforced by individuals and collectives in different social contexts. Shared ideas of the meaning of girdle-hangers are understood as (a) cumulating by people’s continuous engagement with this meaning at all stages of the artefacts’ lifecycle, in manufacture, use, repair and discard (cf. Fowler 2004: 65); and (b) created in people’s exchange of ideas in ongoing daily discourse. Knappett (2012: 104) has described these dimensions of meaning generation as vertical and horizontal networks of meaning. The following describes how the engagement of makers, wearers and burying communities with girdle-hanger meaning has been approached in the material record of girdle-hangers.

Our understanding of identity as a complex, changing human experience of belonging and affiliation confronts the archaeological study of identity with particular challenges. It seems that we have found it easier to look at single dimensions of identity, single social fields where identity is being performed, and single material settings where it takes physical form, than to tell the whole story of making identity. With the aim to convey the dynamics of this story, this paper explores a methodological approach towards early Anglo-Saxon burial evidence that combines different strands of archaeological analysis. At its heart lies the principle that semiotically charged items of dress, individuals and different groups of people were tightly knitted into networks of meaning generation. Within these, shared ideas about identity and its material representations were formed in an ongoing dialogue in which the visible engagement with the object played a significant role. In recent decades, the study of early Anglo-Saxon identity based on the adorned buried human body has strongly focused on meaning transmission in the practice of burial. It has been argued here that identity needs to be viewed as a dialogue between many more social actors—makers, wearers/users, and burying groups alike. The example of the girdle-hanger, a symbolic key and symbol of the identity of a specific group of early Anglo-Saxon women, was presented as a case study to help define different types of archaeological evidence that can reveal different forms of human engagement with material culture worn on the body and its role in communicating notions of identity.

The approach proposed here has been influenced by current archaeological thought on networks as a useful concept for investigating past human interactions. Defining the above types of evidence is also a first step towards developing formal network-analytical approaches to early Anglo-Saxon burial material. These will have the potential to offer new insights into social communication in this period by enabling an innovative graphic visualisation, analysis and interpretation of complex interrelations within the material record. Regretfully, such an analysis has to be the subject of a future paper. However, starting by thinking in network terms about social interaction in the early Anglo-Saxon period is already offering a useful tool for conceptually linking separate types of archaeological evidence and analysis. It can help us understand identity, collective knowledge and shared social values as a product of complex social communication processes—highlighting above all the smaller-scale daily practices which took place within rural communities and were carried by their individual members.