Archaeological knowledge production: reading mortuary reconstructions

Abstract

Research on archaeological knowledge production emphasises the contingent nature of understandings of the past. In practice, however, levels of uncertainty and conjecture can easily become less than obvious in interpretations, perhaps especially visual ones. The authors interrogate multiple textual and visual accounts of a Viking-Age burial to demonstrate how selection processes—what to portray or omit—highlight the contextual nature of the knowledge claims in these images. Arguing that the circulation of reconstructions shorn of textual nuance leads to misperceptions, the authors call for transparency in the creation of these images. Rather than definitive depictions of archaeological fact, these reconstructions offer tools for archaeologists and the public to think with.

Introduction

Few archaeologists today would question that we understand the past subjectively through the lens of the present, constructing interpretations influenced by both overt and more subtle processes. Decades of debate on the production of archaeological knowledge have generated a rich literature (for a selection, see: Edgeworth 2006; Lucas 2012; Olsen et al. 2012; Roberts et al. 2020). Whether focused on the construction of scientific facts, externalising networks of observation, ethnographies of practice, reflexivity, entanglement, or a range of other orientations, ultimately these studies concentrate on the reciprocal relationships between knowledge production processes and the types of knowledge claims that result (Editorial Collective of FKA 2016). Acknowledging the importance of epistemological discussion, we enquire whether the current media landscape of archaeological publication risks an inflation of the degree of certainty attached to archaeological knowledge claims.

We explore this question through the discussion of mortuary reconstructions and their place in archaeological interpretations. Specifically, we reinvestigate a Viking-Age inhumation burial complex excavated under demanding circumstances at Kaupang, Norway. The original text-based publication, which highlights a number of uncertainties around the interpretation of the site, has subsequently been supplemented by a series of reconstruction drawings. The wide circulation of the latter has made them influential on the collective understanding of the site and helped make it representative of not only the Kaupang cemeteries but also Viking-Age mortuary behaviour in general. Taking these captivating visual interpretations out of context in this way, however, allows the challenges and nuances of understanding this particular burial complex, as well as Viking-Age mortuary behaviour more widely, to recede and unwarranted certainty to come to the fore. With this case study, our aim is to highlight how evocative archaeological reconstructions, which circulate within both scholarly and popular media, can become established knowledge claims in their own right, with the implicit power to shape understanding and to direct academic focus. Our purpose is not to denigrate the use of such images, which are invaluable tools for the production of academic knowledge and for public dissemination. Rather, we seek to highlight their inherently interpretative nature and, through this, to contribute to an awareness of how knowledge is (re)produced and promoted.

An evolving interpretation: burial complex Ka. 294–297

In 1951, a mortuary complex was excavated in the Bikjholberget cemetery, one of the burial grounds connected to the Viking-Age town of Kaupang, Norway (Heyerdahl-Larsen 1995: 22–26). The context was challenging to excavate, as revealed by the nickname given by the excavators to the complex—Forargelsens hus (The House of Frustration)—and, as the name also suggests, was initially interpreted as a building (Heyerdahl-Larsen 1995: 22). Containing three adult burials within a single boat (Ka. 294–296), along with lavish material culture and several animals plus a fourth adult body underneath the keel (Ka. 297), interpretation of the context was difficult, to say the least. In addition, the excavation also recovered the remains of what may be an infant, although these were so poorly preserved that the subsequent report could not dismiss the possibility that they belonged, instead, to a small animal (this ‘burial’ was unnumbered in the report; Heyerdahl-Larsen 1995: 23).

The results of the excavation were not published until many years later, and the interments discussed here were initially presented as three distinct burials within a boat, which was, in turn, placed over an older burial (Heyerdahl-Larsen 1995: 22–26). In 2007, a new Kaupang project published a comprehensive overview of the earlier excavations, including Frans-Arne Stylegar's (2007) invaluable evaluation and catalogue of all the finds from Kaupang's cemeteries. Stylegar considered the boat burial to represent an integral event (Ka. 294–296), suggesting that the grave contained a “couple and their sorceress” (Stylegar 2007: 95), thus going beyond the original interpretation in the 1995 publication towards an interpretation emphasising a relationship between the bodies. Stylegar's interpretation and suggested reconstruction drawing (Figure 1), commissioned from the Icelandic artist Þórhallur Þráinsson, quite literally fleshes out the human and animal remains, showing a woman placed at the prow, with her head facing towards the centre of the boat. Placed head-to-head with this woman lies a man, while at the stern is a seated woman accompanied by several animal sacrifices (Stylegar 2007: 100 & fig. 5.20). By choosing to illustrate and discuss this particular grave in detail, Stylegar and Þórhallur have generated a lasting interpretative legacy, establishing it as the best-known grave of this rich and complex burial site.

Figure 1. Ka. 294–296, as presented by Stylegar (2007: fig. 5.20) (pencil illustration by Þ. Þráinsson; © The Kaupang Excavation Project, Museum of Cultural History, University of Oslo).

Subsequent research, focusing on the possible ritual associations of the seated woman, resulted in a more detailed reconstruction drawing by the same artist (Figure 2a). This image reconfigures the sacrificed animals and many of the finds based on a closer reading of the 1995 publication, while retaining the same placement of the three adult bodies (Price 2008, 2010). The two iterations of this reconstruction offer rich interpretative narratives—textual and visual. The accompanying texts, by Price and Stylegar, emphasise that these visual interpretations are of an uncertain nature due to the complexities of the grave. Detached from these texts, however, the images have taken on independent lives and have become accepted knowledge (see the online supplementary material (OSM)). Their frequent inclusion in academic and popular works has arguably created a general perception that the grave is well understood. In truth, interpretations of this burial complex must always involve considerable guesswork due to issues of preservation, excavation and publication, as well as the complexity of the rituals themselves (see the OSM).

Figure 2. A) Ka. 294–296, as presented by Price (2010: fig. 1); B) Ka. 294–296, as commissioned with specific instructions for this article in 2020; C) an alternative reconstruction, commissioned with specific instructions for this article in 2020, where the male body (Ka. 295), along with the horse and accompanying material goods, has not yet been deposited (pencil illustrations by Þ. Þráinsson).

Many aspects of the grave are open to alternative readings, including the placement of the bodies (Pedersen 2014b: 171–72). Although the original publication includes a sketch drawing showing the woman in the southern end of the boat with her head facing north towards the centre of the vessel (Blindheim & Heyerdahl-Larsen 1995: 116), the written account states that her head faced south, towards the prow (Heyerdahl-Larsen 1995: 23–24). The reconstruction drawings created for both Stylegar and Price rely on the decision that the sketch in the original publication was more likely to be correct than the text. Returning to Heyerdahl-Larsen's (1995) written account, alongside the archive of original field drawings (see the OSM) held at the Museum of Cultural History, University of Oslo, instead suggests (though not with certainty) that the woman was orientated the other way, with the male body placed at her feet. In order to illustrate this, we commissioned a third reconstruction drawing, again by Þórhallur Þráinsson (Figure 2b).

Reorientation of the woman placed at the prow makes no difference to the interpretation of the seated woman at the stern and her suggested ritual role (Price 2019: 113–15), nor does it solve the complexity and uncertainty of the grave complex as a whole. What this version can do, however, is to highlight the ambiguity of the archaeological record, not least the evolving nature of archaeological knowledge production. It also subtly shifts our view of the deceased humans and animals and their relationships with each other and the surrounding objects. Specifically, it may diminish the likelihood of the man and one of the women being perceived as a ‘couple’—a heteronormative interpretative mechanism often implicitly reproduced by archaeologists (see Moen 2019). When no longer placed head-to-head, their association may be read differently. Archived catalogue notes also suggest that the male body may have been a later addition (see the OSM). Therefore, the exact sequence in which the bodies were deposited, and the period over which this was done, is open to question. This raises the matter of temporality in reconstructions: which point(s) in the complex sequence of mortuary acts should these drawings depict? In order to illustrate these uncertainties, we commissioned a further drawing, from which the male body and the horse (which has been interpreted as being associated with the male body) are omitted (Figure 2c). Furthermore, some slight adjustments in the placement of the shields have been made in both drawings, in keeping with the original publication's observation that none of the shields were found “connected with” the “male grave” (formulations that can themselves be deconstructed; Heyerdahl-Larsen 1995: 24). The result illustrates how interpretative choices can produce different visual narratives. Moreover, we note that none of these reconstructions consider the body beneath the keel (Ka. 297), and all of them choose to depict the indeterminate infant/animal remains as a human child.

These new reconstructions clearly highlight the nested intersection of epistemological issues that underpin archaeological knowledge claims. Archaeological data represent the fragmented remains of a past reality, further affected by variable standards of field recording. In turn, these data are filtered through delays in publication that may have resulted in the loss of information and misunderstandings, and which can also reflect shifts in thinking that develop with time and changing perspective. Thus, even the near-infinite choices facing those commissioning and creating these reconstructions (Figure 3) are dependent on the contingent nature of the evidence, which is itself arguable.

Figure 3. Two sketches from the illustration process leading to Price's (2008) reconstruction, illustrating the many choices involved (illustrations by Þ. Þráinsson).

Our sense of certainty about our understanding of any particular context, find, location, or even era, ought always to be measured against the possibility of the appearance of new knowledge. Consequently, none of the latest drawings of this boat burial should be seen as definitive, but rather further steps on along a path; they can never be its destination.

Reconstruction, bias and ‘accuracy’

The use of reconstruction drawings in archaeology has a long history (e.g. Piggott 1978; Moser 1998), including artists’ self-evaluations (e.g. Sorrell 1981; Davison 1997; Clarys 2014). Yet, when considered as representations of knowledge, they have rarely been examined in the detail they deserve (e.g. Molyneaux 1997; Kyriakidis 2007). Notable exceptions are found in feminist critiques of stereotyped gender roles, as perpetuated through such representations (Gifford-Gonzales 1993; Moser 1993) and in debates surrounding how they present archaeological interpretations of death as static tableaux, rather than dynamic, ritually and emotionally charged constructions (Williams 2009; Giles 2016; Gardeła 2019). One particular quality of reconstruction drawings is that they can easily be shared, especially online. Being so readily available, they are often used not only in academic literature and lectures, but also in a variety of public arenas, leading to an uneasy (if interesting) disconnect from their original contexts. Whether rejected or applauded, they unquestionably influence the production of new knowledge claims.

In this regard, it is worth considering the effect of commissioning multiple reconstructions for different purposes from the same artist. When one form of artistic signature becomes associated with a particular archaeological site, context or even period, it may promote an unspoken understanding of how these things should be seen; it could even be argued that it lends a sheen of authenticity to visual representations by linking with already established representations. Arguably, we cement the public perception of the visual expression of Kaupang's graves, and by extension that of Viking-Age graves more generally, by continuing to commission Þórhallur for our new illustrations. In our wish to contribute to a wider focus on the role of reconstruction drawings in knowledge production, however, we consider that the benefits of continuity in visual expression (and aesthetic quality) outweigh any potential fixing of perceptions. Here, continuity in style demonstrates the mutability of interpretations. More generally, though, we encourage greater diversity in visualisation.

The contextualised production of academic knowledge is relevant to all forms of communication, not just visual. Transparency and accountability are intimately intertwined with the production of knowledge—a necessity that acknowledges the possibility of some knowledge claims being more viable than others (e.g. Haraway 2013 [1988]; Wylie 2015). In archaeology, the solidity and defensibility of our interpretations rest on a combination of information from comparable finds, a broad knowledge of the specific cultural contexts in which they are situated, careful consideration of specific discoveries, and, also, a certain degree of qualified guesswork. Whilst visual reconstructions undoubtedly represent an emotive and accessible means of encapsulating particular interpretations, in terms of content, they may differ little from their written counterparts; thus, they can be considered to be visualised knowledge claims. The reification of the ‘primary record’, as mediated through the archaeological excavation report, can serve to mask what are, in reality, progressive removals of interpretation, commencing with the first act of field intervention. Both text and image are slippery within different rhetorical power structures of communication (Savani & Thompson 2019). The drawings we have shown here illustrate an archaeological journey from field sketches of skeletal remains and partially preserved objects toward more narratively engaging, fleshed-out reconstructions and, eventually, multiple diverse reiterations that correspond with their accompanying textual interpretations.

The chosen style of reconstruction drawings also communicates implicit messages and has potential effects, not least through the choice of the perspective from which the context is pictured (Gardeła 2019: 99). Whilst some academics favour reconstructions as events, complete with attendant mourners and detailed backgrounds, others (us included) employ the ‘view from above’ perspective, including nothing but the burial as envisaged in its final form and drawn based on the empirical remains accessed at excavation. This view seeks to reconstruct the burial as we imagine it may have looked prior to processes of decomposition, though it may bear little relation to how it was experienced by those who orchestrated it. This chosen format cannot account for sights, sounds, smells and other physical impressions (Gardeła 2019); instead, it presents a product of the archaeologists’ imagination, adding flesh to bones and organic materials, while not extending beyond the grave itself. There is reason to suppose that Viking-Age burial tableaux may have been created with the specific vantage point that we employ here in mind, with the viewer standing at the foot-end of the scene, viewing the body and the burial from the feet of the deceased (A. Pedersen, pers. comm; see also Pedersen 2014a).

A question persistently directed towards reconstructions concerns perceptions of their ‘historical accuracy’ (critically explored by: Giles 2016; Gardeła 2019). It is necessary to pay close attention to the reconstruction of specific details, such as the stylistic properties of dress accessories or the particular arrangement of certain objects, in order to create a valid interpretation. It is also essential, however, to confront the distinctions between the faithful representation of what is known, the (in)completeness of the record, and the more speculative construction of what is unknown (see Price 2020: 31–34 on the representation of Vikings in TV drama). A further issue is that archaeological reconstructions by necessity rely, in part, on imagination—both that of the artist and commissioning archaeologist—and of those viewing the resulting images.

Some of the challenges encountered in the production and ongoing distribution and use of reconstruction drawings can be overcome through the use of other media that expand the tactile engagement with the archaeological context, including sounds and smells (in museum displays, for instance). In terms of illustrations, however, a reconsideration of terminology may be appropriate, allowing for explicit acknowledgement of our use of imagination. Does an illustration that involves as much interpretative guesswork as ours really warrant the term ‘reconstruction’? There are distinct differences between, for example, adding the few missing pieces of a brooch of a well-known type, furnishing a grave with organic material assumed to have been there, and creating a scene that we can only vaguely imagine—yet all are customarily labelled as ‘reconstructions’. The use of more precise terminology, differentiating, for example, between a replication, an interpretative drawing and even a re-imagining, would allow for greater transparency in the processes involved.

All types of visual illustrations, however, operate at multiple levels, as media for creating knowledge, but also for disseminating or critically evaluating that knowledge. Furthermore, while these images are contingent on the interpretations of our academic peers, they may be circulated and received very differently by archaeology's many publics. In all cases, ‘accuracy’ takes on a range of meanings that are dependent on situation and reception. These meanings, and their relevance or irrelevance to the subject matter, may be beyond the archaeologist's control: ultimately, a balance of accuracy and creativity is involved in any visual and written presentation of archaeological remains (e.g. Savani & Thompson 2019: 221).

Conclusions

Visual reconstructions and texts both form part of the source-critical process of archaeological analysis. These words and images extrapolate a conjectured whole from (sometimes very) fragmentary material, often via analogies to better-preserved finds of a similar character. When interpretative choices are distilled into a reconstruction drawing, they become eye-catching; perhaps therein lies the impression that reconstruction drawings are somehow more biased than written accounts. And yet, they are an extension of the written accounts from which they are drawn, presenting an iconographic version of textual knowledge claims. In any condensed version, be that a reconstruction or a summary, a number of alternatives are omitted. One specific shortcoming of any individual reconstruction drawing, at least of the type discussed here, is their visual fixing (Savani & Thompson 2019: 216), which limits the opportunity to express the varying degrees of uncertainty that might exist and to demonstrate how several different interpretations may be equally likely (as we have attempted here by comparing drawings; for an early attempt at something similar, see Millett & James 1983: 245–46). In a reconstruction drawing, the ability to disguise our choices behind ambiguous sentences is lost, and we can no longer hide behind academic jargon that saves us from being explicit (Pedersen 2017: 41–42). Herein lies one of the strengths of illustrations: choices must be made, which necessitates confronting details and uncertainties that might be omitted or more subtly concealed in a text (see also Savani & Thompson 2019: 216).

One possible path forward lies in taking control of our work by proactive contextualisation and by admitting what it represents. Thus, as a scholarly community, we might ensure that we provide a variety of visualisations to the public. Multiple reconstructions as visualised interpretations, acknowledged and embraced as such, express the essence of archaeology and lie at the heart of the production of archaeological knowledge. They represent tools to think with and to discuss, based on the material remains that survive. In a subtly different fashion to texts, they make interpretative choices manifest and force scholars to confront assumptions—both their own, as well as those of their audiences. In this, they are inevitably and positively speculative products of the same imagination that, ultimately, empowers any attempt to access the human past.