The Future of Periodization. Dissecting the Legacy of Culture History

Abstract

This paper discusses the future role of periodization in the wake of recent critiques of culture-historical chronologies concurrent with the rise of high-definition radiocarbon dating. It is argued that periodization has two distinct facets, a narrative function and a dating function, which should be separated. Archaeology may eventually be able to abandon the latter, but not the former. However, the two aspects are closely intertwined and the goal of this paper is to disentangle them and, through a case study of archaeological periodization in Iceland, demonstrate the need to re-engage with culture-historical taxonomies by reverse engineering their construction. Only in this way will the utility or poverty of such culture-historical periods be exposed to proper scrutiny and the ground cleared for building new, narrative periodizations.

Introduction: the legacy of culture history

Periodization remains one of the most fundamental operations in archaeology. It was essential to its very establishment as a scientific discipline with the adoption of prehistory and deep time through the acceptance of the Three Age System (Rowley-Conwy 2007), and as a general method, it remains a key means of ordering time. Is there a region on earth that does not sort its archaeological remains into a sequence of chronological divisions? Most certainly these divisions are subject to revision, even complete overhaul in some cases, but their utility remains largely unquestioned. However, this is not to suggest that archaeologists are ignorant of the problems they entail. One of the central critiques has revolved around their association with culture-historical archaeology and its legacy of taxonomic thinking. As Seren Griffiths has recently argued in the wake of the Bayesian revolution in radiocarbon dating:

Employing period-specific taxonomic models of culture within a linear, sequential narrative of social change has a tendency to collapse time in a quest for simple narrative structure. Terms such as ‘Mesolithic’ are not neutral when employed in a chronologically successive model; they are parts of an iconic analogue for how we understand societies and the nature of change over time. By glossing variability in the archaeological record, we risk abstracting time, reifying change into binary flips between binary lifeways, and imposing predetermined archaeological concepts onto myriad material evidence. (S. Griffiths 2017, 9)

In other words, if we have the tools now at our disposal to create fine-grained chronologies using calendrical dates, why adopt this kind of taxonomic model of change that periodization embodies, at all? Surely it can be abandoned as an outmoded element of archaeological thinking—essential at a time when there were almost no methods of independent dating, but half a century after the adoption of radiocarbon, it has become more of a millstone than a crutch. And yet the matter is not that simple.

When radiocarbon dating first entered archaeology more than half a century ago, it was used mainly to calibrate the different existing periodizations (Renfrew 1978), and for the most part its continued use has not made such relative chronologies redundant so much as fine-tuned them and provided an absolute framework within which they can be slotted. The rise of what has been called the third radiocarbon revolution (Bayliss 2009) has thrown this issue into sharper relief, but in itself it cannot really solve the problem which, as Griffiths identifies, relates to the taxonomic model of change which dominates archaeological thinking in the form of periodization. A related problem with periodization, which Griffiths also emphasizes, is the way it condenses all change into singular moments, ‘binary flips’ from one period to the next. This point has also been taken up by Rachel Crellin in her critique of what she calls the ‘block time’ model of change used by archaeologists (Crellin 2020): in this model the past seems to consist of long periods of relative stability and continuity, but ruptured by brief periods of change marked by period transitions or boundaries. For example, the British Neolithic or the Viking Age are constituted by a set of coherent traits, which are then replaced by another set in the succeeding period. For both Griffiths and Crellin, the legacy of this kind of taxonomic thinking restricts our ability to tell more nuanced stories. For Griffiths, this is largely a dating issue; it is not about using radiocarbon dates to refine our culture-historical periodizations, but to bypass such periodizations altogether (also see Bayliss & Whittle 2019; Griffiths et al. 2023). For Crellin, it is a more theoretical issue about narrating change as continuous, rather than condensed into short bursts marked by period boundaries (Crellin 2017).

The subtle differences between Griffiths's and Crellin's discussions of periodization point towards an important issue that needs to be fully dissected: what is it that periodization does? We suggest it does two very different things, which are often conflated. One is where periodization is grounded in a narrative framework or skeleton, enabling us both to delimit where our stories start and end, but also how they break up into pivotal segments. Such narrative periodization is simultaneously both a way of creating some coherence within a span of time (e.g. the Viking Age), but also a means of marking what are regarded as key historical changes (e.g. the Three Age system). The narrative function of periodization extends back thousands of years; it is a way of giving meaning to history, from Hesiod's five ages of Man (sic) to the conjectural histories of the Enlightenment. Whether it is tied to royal or imperial dynasties, evolutionary stages/ages of civilization, or historical eras, it acts to provide an order to history and change, whether this is directional or not, and regardless of the particular moral or political stance behind it. The other function of periodization, however, is more recent and in fact more specifically archaeological (but also geological and palaeontological, from which archaeology derived its use); this is periodization as a dating tool. Before the advent of scientific dating, periodization was effectively the only tool archaeologists had, at least for prehistory. The basic idea is that through seriation and stratigraphy, certain artefacts could act as ‘index fossils’ for a period (O'Brien & Lyman 2002), and this is how it is still used today. When we excavate a site, the typological characteristics of the recovered finds provide the first and often last way to date the site. Even when radiocarbon or other methods are used, although that gives the assemblage or site more chronological specificity, it is also usually just reinforcing the broader periodization.

As a dating tool, periodization was inevitably bound up with its narrative function in the early part of our disciplinary history. In a sense, culture-historical archaeology of the late nineteenth and first part of the twentieth century was defined by this dual, intertwined function, epitomized in the Three Age System, which was both a narrative structure and a dating methodology. The first radiocarbon revolution arguably did not alter this relationship; it is only now, in the midst of the third revolution, that real anxieties are being voiced about archaeological periodization, as we have seen. But these anxieties have tended not to distinguish the two functions of periodization we have identified here, but rather conflate them. This is the legacy of culture-historical archaeology. In the rest of this paper, we want to dissect this legacy and ask: what future does periodization have in archaeology?—both as a narrative device and as a dating tool. To help explore this question, we will draw on our experiences with archaeological data from the thousand-year history of Iceland. But first, we need to understand in more detail the complex relationship between these two functions of periodization in archaeology.

Disentangling the narrative and dating functions of periodization

A feature common to both narrative and dating periodization is that they have scalar properties insofar as they work at different spatio-temporal levels, from the universal to the local. A universal periodization is one that applies to the whole of humanity/human history, or at least a large part of it; typical examples in archaeology derive from evolutionary stages of political development such as Morgan's (1877) Savagery-Barbarism-Civilization model (Childe 1951) or the Band-Tribe-Chieftain-State models of neo-evolutionism (Service 1962). More localized periodizations are those that pertain to a specific region, such as the Indus Valley sequence of Pre-, Early, Mature and Late Harappan, while schemes such as the Three Age System occupy a middle ground, i.e. applicable across most of Europe but quickly breaking down the further it is extended. In general, periodizations in archaeology have occupied the whole spectrum from the local to the universal, and while one might certainly suggest the more universal periodizations have come in and out of fashion, they show no real signs of dying out (e.g. see Graeber & Wengrow 2021, 446–9).

But another aspect of periodization is the extent to which a periodization constitutes an exclusive system of temporal units so that for any given time period, only one periodization is possible. Although some of the more universal periodizations have claimed (implicitly) a monopoly, in practice it tends to be the more local periodizations that battle for primacy. We can turn to the Indus Valley chronology again as an example, where various competing periodizations are on offer (Manuel 2010), but what is important to note here is that these are not offered as equally viable alternatives, but exclusive choices. If you adopt one, you cannot really use another without causing confusion and inconsistency. The reason why it is the local periodizations that tend to be more exclusive is because these are also the ones most entangled with the dating function of periodization. The connection between the exclusivity of a periodization and its function in dating is understandable, given that the whole purpose of chronology is to provide a singular framework within which to order archaeological remains. Having multiple periodizations would be as confusing as using multiple calendrical systems. But more substantively, there is also an implicit assumption that such chronologies, built on material culture sequences, embody a set of roughly synchronized changes to an ensemble of material traits resulting precisely in the block time and binary flips identified by Crellin and Griffiths. Hence it is also at the level of local periodizations that the culture-group concept has most purchase—a set of traits delimited in time and space, defining a community with shared customs and practices.

It is perhaps here that the anxieties around periodization discussed at the beginning of this paper are most felt. What is really at stake is not periodization per se, but the synchronicity and exclusivity implicit in periodization in its function as a dating tool. It is useful here to step outside archaeology and consider the way periodization functions in history, because it may help to give some perspective. Unlike archaeology, history never seems to have had a problem with adopting multiple, ad hoc and contingent periodizations; even though they may work within some very broad blocks like the Middle Ages or Early Modern period, in practice most historians are happy to adopt whatever periodization serves their narrative—which might just as easily cut across these larger divisions as not. The reason they can be more flexible is also the reason why archaeologists cannot: historians do not use, or indeed need, the other function of periodization that archaeologists still depend upon: periodization as a dating tool. Given the way their sources either come with their own date stamps or that they can be dated in ways independent of their use, periodization in history only has one function: to provide narrative structure.

For some archaeologists, the third radiocarbon revolution offers an opportunity to make archaeology more like history (e.g. see Bayliss & Whittle 2019). Although the distinction is not made between these two functions, one could assume that by freeing archaeology from its dependence on periodization as a dating tool through radiocarbon, the presumption is that archaeology will then also be freed from its ties to the exclusive, block-time periodizations implicit in the taxonomic model of culture history. Yet this is not necessarily the case, not unless we make an explicit attempt to disentangle the dating and narrative functions of periodization as suggested above. It is not enough to single out the dating function of periodization and argue for its superfluity; this still leaves the narrative function intact, which may still carry the burden of many of the traits derived from chronological periodization. Archaeological narratives will almost certainly continue to draw on some kind of periodization, implicitly or explicitly; the question is, how to disentangle this need for narrative periodization from the specific connotations of exclusivity and synchronicity which are an implicit part of culture-historical taxonomy.

To explore this, we want to use a case study of the archaeological record of Iceland. Besides the fact that both authors know this material very well, the case study has the advantage that it covers a relatively short chronology (1000 years), which means the amount of data, though still massive, is manageable. It also has the benefit that there exists no deeply entrenched periodization within the discipline; indeed cultural unity and homogeneity has been considered as a given. It is an island society, settled in the late ninth century ad by a presumed, relatively homogeneous cultural group with its own language and customs linked to western Scandinavia. After the first period of settlement, there was minimal movement of people into or out of Iceland until the late nineteenth century.

Periodization and dating: a case study from Iceland

Icelandic history has traditionally followed a, largely implicit, tripartite periodization: the Settlement Period (landnámsöld, 870–930 ad); the Free State or Commonwealth period (þjóðveldisöld/goðaveldisöld, 930–1262 ad) and later periods (seinni aldir, 1262–present), which may take subdivisions depending on the subject, e.g. the English century (enska öldin—the fifteenth) in the context of trade history, or the Age of Learning (lærdómsöld—1550–1770) in the context of intellectual history (see e.g. the divisions employed in the authoritative Saga Íslands series: Líndal 1974–2016). The focus on divisions in the earlier centuries reflects an entrenched, nationalist ideology which glorifies the first centuries when Iceland was an independent country before it came under the kingdom of Norway in 1262. The lumping together of most of what we would call the medieval and post-medieval periods reflected a sense of a long dark age where nothing really changed. This is a periodization that is implicitly or explicitly guided by a clear sense of what counts as relevant change: political autonomy on the model of the modern nation state. Although archaeologists have partially drawn on this historical periodization, there is actually no commonly used scheme, and in practice, the main division really occurs around 1000/1100 ad, which marks the end of the Viking period. In effect, there is Viking and post-Viking archaeology, although again this is more implicit than explicitly stated (Vésteinsson 2015, 216–17).

One of us has recently critiqued this tacit periodization by both arguing that there is no real or meaningful chronological break around this time based on the archaeological record and contesting the assumption that nothing really changes after this break (Vésteinsson 2015). Specifically, can any arguments be made, either for continuity or major discontinuity, from the archaeological record? Building on this work, and also another study on periodization from the perspective of a single site in Iceland by the other author (Lucas 2019), we want to interrogate these issues in more detail and question both the assumption of discontinuity at the end of the Viking Age and continuity within the subsequent centuries.

The Viking package

As a spin-off from his book A History of the World in 100 Objects, Neil MacGregor collated 10 objects emblematic of the Vikings in connection to an exhibition at the British Museum (MacGregor 2014). Included in the list were weighing scales, a brooch, a silver hoard, a boat burial, an Ulfberht sword and the great long ship from Skuldelev. These objects are clearly special and, although certainly evocative of some of the key traits associated with the Vikings in popular and scholarly imagination—trading and raiding—they are hardly representative of what you might find in excavating a Viking settlement. Archaeologically, the ‘Viking package’ might more typically include items such as steatite vessels, glass beads, oval brooches, ring-headed pins, ‘Thor's hammer’ pendants, composite antler or bone combs, schist whetstones, iron knives, plain swords, axes and spearheads. However, this ‘package’ is very unlikely to be found together in one place; indeed the range of items found in burials is often more iconic in this regard (jewellery and weaponry), whereas on settlements, the finds are generally less spectacular, but also more diverse. Moreover, there are also subtle differences in assemblages from different part of the Viking area; Iceland, for example, lacks cremation burial, and some early scholars like Shetelig even characterized Iceland as a very impoverished version of the mainland corpus (Shetelig 1937, 210; also see Pétursdóttir 2009). In general, the impoverishment of the Viking package among the early settlers of Iceland seems to have been a recurring theme among scholars ever since, reinforced by Eldjárn in his survey of Viking burials in Iceland from the 1950s (Eldjárn 1956). More ambiguously, it also informs more recent work in framing the material culture associated with landnám sites of the ninth and tenth centuries through the concept of a ‘settler kit’ which is essentially everything a self-respecting Viking needs to transplant their society in a new land (Batey 2011; Forster 2004). Regardless of whether we view this settler package as ‘Viking culture stripped to its bare essentials’ or not, it does nonetheless offer an interesting opportunity to explore the idea of the Viking assemblage in almost ready-made laboratory conditions. To what extent does the archaeological evidence associated with this period actually constitute a coherent and confined assemblage, temporally speaking?

There are essentially two types of sites that have produced evidence for Viking Age material culture in Iceland: furnished burials and settlements. The burials have a research history stretching back to the nineteenth century, providing the backbone for Iceland's Viking Age culture history. Already in the nineteenth century the burials were understood as a subset of Scandinavian burial custom (Kålund 1882) and subsequent debate for a long time concentrated on defining the degree of variance and refining the chronological limits of the grave-good assemblage. The absence of spectacular finds was an early concern, as we have already noted, but in his seminal study of this material from 1956, Kristján Eldjárn came to two principal conclusions: on the one hand the assemblage was reflective of Norse ethnicity and on the other it was overwhelmingly from the tenth century, conforming admirably, in his opinion, to the historical events that framed this material: the settlement of Iceland beginning in the 870s and the conversion to Christianity in 1000. Subsequent finds have not changed these conclusions; the typologically sensitive objects are indeed almost entirely from the tenth century, with only a handful of outliers from the late ninth and early eleventh centuries. In Eldjárn's view, echoed by all authorities until very recently, the grave goods reflected Icelandic culture as a whole from c. 870 to c. 1000. What has changed is that it now seems clear that these artefacts formed a part of ritual practices that were only followed by certain sections of society and only for a limited time within the tenth century. In fact, it looks like a short-lived fashion, peaking in the mid to late tenth century at the same time as other, unfurnished, burial practices were also being introduced (Vésteinsson 2020; Vésteinsson et al. 2019; see also Leifsson 2018). The fact that the majority of furnished burials only have non-diagnostic artefacts, previously understood as a marker of poverty, underlines that the ‘cultural package’ represented by the diagnostic artefacts is only an element of Icelandic tenth-century culture.

The grave-good assemblage has some overlap with the material culture of settlements, but its most conspicuous elements (i.e. those that have received most of the attention even if they are not necessarily the most numerous) are also those least likely to be found in domestic contexts (e.g. brooches and swords). It is those same kinds of artefacts that are mostly absent from later periods, contributing to the sense that there is a marked difference between Viking Age materiality and that of subsequent centuries. A closer look reveals that this is by no means a straightforward matter. There are changes in use, in style, in frequency and in visibility which affect our perceptions of cultural change. When examined in detail it becomes apparent that many different things were happening at different times for different reasons, so the sense of change is primarily created by the grave-goods and their abrupt disappearance is a result of changes in burial custom. It is instructive to look at a few examples.

The most significant change that undoubtedly took place is stylistic. The zoomorphic styles which make the material culture of the Viking Age so distinctive were replaced from around 1100 by the Romanesque, bringing Norse stylistic expression in line with the rest of Catholic Europe. This change is conventionally linked to the conversion to Christianity and seen to mark the post-Viking Age integration of Scandinavian politics and culture with those of its European neighbours. This makes sense as the general context, but the events unfolded over a long period, not at all in an obvious or logical sequence. Furnished burial ceased a century or more before the last of the Viking Age art styles (the Urnes style) lost its popularity in the early twelfth century, and it seems that the female costume, characterized by oval brooches but also several other distinctive jewellery items, became old-fashioned sometime in the interim. These changes can all be seen to relate to a general trend away from separate Scandinavian ways of doing things towards European sensibilities and solutions, but this is also a narrative and point of view which is largely suggested by these very changes. In fact, they all had their separate reasons and historical contexts. It is not at all clear, for instance, that the female costume became unacceptable as ‘Sunday dress’ because of its pagan connotations—this type of dress may have had a good run long after Christianity had begun to permeate and shape people's everyday lives (Sanmark 2004) and its eventual obsolescence was likely due to something entirely different. The thing is that once furnished burial ceased, our knowledge about female dress evaporates, contributing to a sense that there must have been an abrupt change—but this is not clear at all.

Several quite distinctive types of jewellery did eventually go out of fashion. These include oval brooches, trefoil brooches and bead necklaces, but jewellery as such continued to be used, even if its archaeological visibility is sharply reduced once it was no longer deposited in graves. Similarly, beads, which are found in significant numbers both in burials and Viking Age settlement contexts (Hreiðarsdóttir 2010), are an example of an object category where there are changes in use which affect frequency but the artefacts as such continue. Bead necklaces had gone out of fashion by the twelfth century, but beads continued to be used as part of dress and in rosaries. In later centuries they are frequently found on church sites, and their numbers increase in the early modern period (Hreiðarsdóttir 2007). Weapons, similarly, are much reduced in visibility by the discontinuation of furnished burial, but there is good evidence that their use continued. Endemic violence, warfare and individual weapons are described in twelfth- and thirteenth-century chronicles, leaving no doubt about the ubiquity of weapons in post-Viking Age Iceland, and occasional chance finds of weapons back this up.

Steatite vessels, found in burials but much more frequently in settlement contexts, are often seen as the quintessential element of the ‘settler's kit’, utilitarian implements which the settlers are seen to have brought with them at the time of colonization, but which were not replaced by further imports and are as a result mostly found in late Viking Age contexts reused as spindle whorls or gaming pieces. By the twelfth century it seems that the steatite stock had been completely worn out and the material is hardly ever found in domestic contexts thereafter. Steatite vessels occur, however, in later medieval deposits at high-status sites, suggesting that availability as such was not what caused the phasing-out of this material (Forster 2004). Combs occur in both burial and domestic contexts in the Viking Age and continue in domestic contexts (Gísladóttir & Snæsdóttir 2019). They change in style, but their frequency seems not to change. Coins are rarely found in burials and occur, along with hack silver, more frequently in settlement contexts or as chance finds. The number of finds peaks in the late eleventh century but cease completely thereafter, and coins are not even found at later medieval trading sites. This change is consistent with a general trend of reduced coin finds in Scandinavia after 1100 (e.g. Jonsson 2009).

Apart from burials, practically all evidence for Viking Age material culture comes from farmsteads. There is good evidence for widespread settlement in both coastal and inland areas in the decades around 900 (Vésteinsson & McGovern 2012) and it used to be thought that the distinctive three-aisled halls with central fireplaces and bow-shaped walls were built by the first wave of colonists, as this was the house type prevalent in the regions they came from in Scandinavia and the Scandinavian settlements in the British Isles. It now appears that this kind of building is almost nowhere securely dated to the earliest period and that it only becomes common in the mid to late tenth century, in the same period as furnished burial, in fact (Vésteinsson 2014). The same pattern is emerging in the Northern Isles of Scotland (D. Griffiths 2023), suggesting that Norse architectural styles were only adopted by the second or third generation of immigrants from Scandinavia. There are very few securely dated dwellings from the earliest period of settlement in Iceland, but at the only site with fully excavated architecture from before 940 (Sveigakot, in northeast Iceland) people lived in pit-houses and these are widely seen as likely first-phase housing at many sites. Pit-houses are, however, most frequently found as ancillary buildings contemporary to the halls. They are clearly associated with weaving although they may also have served as dwellings for a part of the household (Milek 2012). The pit-houses rarely have more than one phase and generally only have tenth-century dates. Their disuse is widely connected to changes in the design of the farmhouse, the three-aisled halls, which start to sprout annexes already before the end of the tenth century. Some of these new rooms were for food storage and cooking, but the biggest change involves the so-called ‘stofa’, often glossed as ‘living room’, where weaving was at least a part of the functions, and this space has therefore been seen as successor to the pit-houses. There is considerable variation, between individual sites as well as regions, in how and when these changes happened, but overall, a development can be seen from farmsteads made up of a single hall plus several detached ancillary buildings towards a more nucleated farmhouse divided into several interconnected rooms (Bolender et al. 2011; Ingólfsson 2023; Vésteinsson 2010). This change seems to be taking place between c. 950 and 1150, although the pit-houses appear to go out of use in the earlier part of that period. In the later part the halls become narrower, 3 m instead of 5 m or wider, and their walls become straight instead of bowed (overview of developments in Stefánsson 2019).

These changes in architecture and farmhouse layout have traditionally been seen in terms of adaptation: that the settlers brought with them Scandinavian housing conventions but soon learned that they were inadequate for conditions in the new country. Colder climate and lack of timber for building and fuel are then seen as the drivers (Roussell 1953; Stefánsson 2019). Changes in economic strategies have been similarly explained: an initial emphasis on cattle giving way, by the twelfth century, to more sheep has been seen as an environmental adaptation: with plentiful rangelands but short summers for fodder production, Iceland was more ideally suited to sheep rearing. Shrinking proportions of pigs and goats from the eleventh century have been explained with reference to decreasing woodland cover, a repercussion of human-induced changes to the environment (McGovern et al. 2007). Barley cultivation was widespread in the tenth and eleventh centuries but thereafter limited to the south and west, probably at a much reduced scale compared with the earlier period (Mooney & Guðmundsdóttir 2020). Similarly, there is good evidence for iron production in the tenth and eleventh centuries, but more limited thereafter (Smith 2005). Some of these changes may be more apparent than real: from the Viking Age there are multiple fully excavated farm sites with large animal bone and artefact assemblages, but for the subsequent centuries the datasets are much fewer and smaller with long periods very patchily represented, if at all (Vésteinsson 2004). This contributes to a sense that the things that characterize the Viking Age and are not observed in later, often much later, centuries, must have ceased at the same time as the Viking Age, but this is an unsafe inference. In some cases, other explanations than adaptation have been offered which then require adjustments to the timeframes, and, interestingly, the patchy state of our datasets seems to allow this. Thus the decline of barley cultivation is for some scholars most obviously a result of climate change (as opposed to the difference in climate between the middle of the North Atlantic and Scandinavia) requiring it to have happened after 1200 when the cooling trend set in. Also, growing sheep numbers have been related to increased textile production for export, a response to growing reliance on imports (Hayeur Smith 2015), which can be seen to have been driven either by internal or external forces. Which formulation is chosen affects in what date-range the change is seen to have occurred.

Although there is widespread evidence for human presence and significant human impacts on the environment, the character and material culture of the first one or two generations of settlers is diffuse. This stands in clear contrast to the mid to late tenth century, when there is plentiful material both from burials and settlements. It is interesting that, at the same time as furnished burial had its heyday in the mid to late tenth century—a custom seen to represent the old and soon to be outdated—other innovations were also being introduced. Churches were being built and Christian-style burial began to be practised (Vésteinsson forthcoming)—widely perceived as the beginning of the post-Viking Age—and a system of earthwork boundaries was being constructed, regulating, it seems, property divisions and separating home-ranges from commons (Á. Einarsson 2019; Einarsson et al. 2002). These disparate, and some might say conflicting, developments mark the late tenth century as the single most visible and heavily studied period in Icelandic archaeology. Both the furnished burials and the earthwork boundaries were short-lived phenomena, but their outstanding, although very different, materiality makes them weigh heavily in perceptions of the Icelandic Viking Age at large. To us they look more like fads, not signals of one period ending and another beginning.

Compared to later centuries, the Viking Age looks like a period of rapid change. This perception is created in large part by the much better resolution of the data, compared to the subsequent centuries, and it is given a boost also by the poorly dated changes which tend to get lumped with the end of the Viking Age, even if they may in fact have occurred later. It is furthermore given apparent substance by the introduction of written documents around 1100, a major transformative change that gives a completely different perspective on the before and after. It seems to us that the contrast is to a large extent artificial, an effect of a combination of research bias and new types of sources. The things that most stand out in the Viking Age are primarily those associated with furnished burial, the kind of context that is simply not represented in later centuries. It should be clear from the disparate cases recounted above that changes were occurring throughout the Viking Age; changes to different things, at different times and for different reasons. It was clearly not a case of everything getting imported at the end of the ninth century and staying more or less the same until end of the period. Rather, change was continuous, which begs the question: what is the period made of?

Post-Viking continuity

In contrast to the Viking Age, the later Middle Ages and the early modern period have received comparatively little archaeological attention. Within this much longer time-span the bias in research has been towards high-status sites (monasteries, episcopal seats, manors; e.g. Kristjánsdóttir 2023) and, recently, towards the eighteenth and nineteenth centuries (Lucas 2012). In historiography, the emphasis has been on continuity or decline and the measures of such, like foreign exchange and production. Interestingly there have been very limited efforts to use archaeological approaches to throw light on the thirteenth and fourteenth centuries, the golden age of literary production in Iceland—which in many ways remains the murkiest period within the underexplored middle part of Iceland's chronology.¹

The one narrative thread that connects the archaeology of all periods is that of the development of the Icelandic farmhouse. We saw in the previous section how the Scandinavian-inspired three-aisled halls acquired additional rooms from the late tenth century onwards, but a major change occurred before 1362, when a volcanic eruption buried two farmhouses—Gröf (Gestsson 1959) and Bær (B. Einarsson 2020)—which are characterized by a central passage that is standard in all later farmhouse architecture down to the early twentieth century. The central passage completely realigned traffic within the farmhouse, but the functions of the individual rooms remained similar. Two important changes have also appeared in these fourteenth-century buildings: a separate kitchen replacing the central hearth in the hall as the main cooking area, and the appearance of a ‘baðstofa’, a kind of sauna which later got enlarged and became the principal sleeping and working space in the farmhouse. How this change from sauna to bedroom unfolded is unclear and has occasioned much learned debate from Sigurðardóttir (1966, 69–79) to Stefánsson (2019). This is interesting because the innovation of the separate kitchen is much less commented on, yet it clearly had major implications for the status of the housewife and for health as people no longer had to sleep around an open fire. This is an example of a change which has not been accorded narrative significance, whereas the baðstofa issue has long been seen to reflect that things were going badly for the Icelanders after the union with Norway in 1262. The move into the sauna is then seen as a symptom of scarcity and misery: people no longer being able to keep their houses warm and huddling into one small room, the only one they could afford to heat (Stefánsson 2019). This theme of scarcity makes sense in the nationalist narrative, which was keen to find evidence for decline and hardship in the periods of foreign domination.

Narratives of continuity stress the permanence to the settlement structure: the accumulation of farm mounds where farmhouses were slowly rebuilt and regenerated on the same spot for a millennium (and frequently nowadays topped by a modern farm building) and the basic elements of the economic system: animal husbandry subsidized by fishing and, to a smaller extent, hunting. These matters seem to represent the basis of the societal structure and in comparison, the changes that can be enumerated can seem trivial. Similarly, in the twelfth century a network of parish churches was erected, which held regular services and had attached cemeteries that served the whole parish. This is in contrast to the Viking period, where small churches with household cemeteries were built at every other farm in the late tenth and early eleventh centuries, although they had already started to decline before 1100. These small churches also lost their burial rights at the same time or soon after, but the buildings continued to be used for prayer and household services and only reduced in number very gradually over many centuries—a fair number were still in use in the early eighteenth century (Vésteinsson forthcoming). Around the same time as the establishment of the parish system, several religious houses were founded, lasting until their dissolution in the sixteenth century (Kristjánsdóttir 2023).

And yet there are changes discernible in settlement patterns. From at least the 1700s, a number of settlements show signs of increasing population density, especially trade ports and fishing sites. Much of this is linked to the policy and increased presence of merchants and fishermen/whalers. Architecturally, the changes are very notable at the trading sites, which were scattered around the coast of Iceland, some of which had been in operation since the Middle Ages. Since at least the thirteenth century, if not earlier, trading sites were only seasonally occupied as foreign merchants camped in temporary shelters marked by low turf and/or stone walls during the summer months (e.g. Gásir, Gautavík). These temporary booths are a common feature on seasonally occupied archaeological sites, not just at trading posts but also fishing settlements and assembly sites, and they display a long continuity of form since the tenth century. Although many of the foreign merchants and fishermen used such structures, there are occasional instances of foreign-style timber buildings on these trading sites. The Hanse imported prefabricated German-style buildings and a copper-roofed church at Hafnarfjörður in the sixteenth century (Gardiner & Mehler 2013: 7). Over the late seventeenth and early eighteenth centuries, Danish merchants also built timber structures at some of their sites, but it was not until the law prohibiting year-round trading was lifted in 1759 that one sees major transformations in the physical appearance of these trading sites. They become more planned, with various permanent buildings, and often enclosed by a perimeter wall/fence (Martin 2022, 236–46).

These changes coincide with a mid-eighteenth-century attempt to stimulate the Icelandic economy by various ideas connected to expanding economic production, from salt-making and quern production to horticulture (Beck 2020; Guðmundsdóttir 2012; Lucas 2010b; Róbertsdóttir 2014; Sigurðsson 2013). All of these initiatives clearly fall within changing ideologies of improvement which are well known archaeologically in other places (e.g. Tarlow 2009). Among these initiatives in Iceland was the establishment of a joint stock company which came to be known as Innréttingar (‘New Enterprises’). It focused its efforts towards promoting industrial textile production and set up a factory on the farm of Reykjavik, the remains of which have been excavated (Nordahl 1988; Roberts 2002). Although it only lasted about half a century, it was the catalyst for the development of the first urban settlement in Iceland, transforming Reykjavik from a farm to a town when it was granted its charter in 1786. The changes evident in the merchant centres also impacted the range of goods coming into Iceland, as evidenced in changes to the artefactual assemblages across ordinary farms from c. 1700.

Ceramics occur exceptionally on Viking Age sites and are not generally seen as a part of Viking Age materiality. Their numbers start to increase in the thirteenth century, and they can be expected in very small quantities on ordinary farmsteads from the fourteenth century, although they do not become common until the seventeenth, at least on higher-status sites. It is, however, only in the nineteenth century that pottery can be characterized as at all abundant on most sites (Lucas 2010a). Glass is very rare in medieval contexts, primarily associated with ecclesiastical and high-status sites (Lucas 2020), but becomes more common in household refuse from the seventeenth century, although like ceramics, it is not until the later nineteenth century that it occurs in any substantial numbers on most sites. The increase in both glass and pottery vessels from the seventeenth century is clearly linked to changes in styles of dining, especially as whole new vessel forms appear. A large part of the pottery found from the thirteenth to the sixteenth century are vessels associated with serving and drinking wine (chiefly Rhenish stoneware jugs/tankards), with very little in the way of cooking vessels or other dining wares. It is only in the seventeenth century that a wider repertoire of forms is consistently present, including vessels linked to the consumption of colonial hot beverages, i.e. tea, coffee and chocolate. It is also at this time one sees the arrival of clay smoking pipes for another colonial product, tobacco; the earliest instances in Iceland date from the early to mid seventeenth century and thereafter they are fairly common up to the end of the eighteenth, becoming rarer again in the nieteenth century as snuff and snus become more popular (Lucas & Jónsson forthcoming).

Equally significant was the introduction of knitting in the early sixteenth century (Guðjónsson 1985). This made possible the mass production of small woollens—socks and gloves—by a wider range of the workforce, especially children, and may have spelled the end of the vertical loom which was a fixture of every household down to the end of the Middle Ages but becomes more difficult to trace thereafter (Guðmundsdóttir 2023). Horizontal looms were introduced in the eighteenth century, and so was the spinning wheel, marking major technological advances as well as organizational changes in the household-based textile industry. Although not as prominent as ceramics, glass and clay pipes, these changes in textile production are still visible archaeologically (Hayeur Smith 2012).

Yet while all these new kinds of objects (and others such as bricks and stove tiles, or dress items like buttons) mark a major change in the archaeological record of Iceland c. 1700, many others show longer continuity from earlier periods. A classic example are stone hammers to beat stockfish which appear around 1300 and become a fixture of all artefact assemblages thereafter, down to the nineteenth century (Árnason 2018). A simple enough innovation, they reflect the growing importance of fishing for household consumption at the same time as its importance as an export commodity was also increasing. Similarly with whetstones used to sharpen blunt blades of scythes, turf-cutters and knives: made from imported Norwegian schist since the tenth century, they are a regular occurrence on sites into the nineteenth century and change little in form (Hansen 2011). Another practice which exhibits long-term persistence is the bi-perforation of metapodials, which appears only in very late or post-Viking Age animal-bone assemblages and continues throughout Iceland's history, even into the early twentieth century (Bigelow 1985; McGovern 2009). Also found in the Faroes and Shetland, they seem to represent north Atlantic cooking methods, where marrow was extracted from leg bones of sheep by boring a hole rather than cracking the bone, perhaps reflecting a preference for boiling rather than roasting meat. Conversely, other items largely disappear by c. 1700, if not earlier. Norwegian baking plates of schist appear in the twelfth century and continue to the fourteenth, representing an element of Norwegian cuisine—crisp bread (Gísladóttir & Snæsdóttir 2011)—and may even have lasted into the seventeenth century. Stone spindle whorls, which are a common find from the ninth century, gradually disappear by the fifteenth, to be replaced by wooden ones which do not survive as well, although examples are known (Snæsdóttir 1981). Similarly, lamps for burning whale/fish oil were made in stone from the ninth/tenth century, but lamps from iron and copper alloy take over from the fifteenth/sixteenth century when candleholders also start to appear.

If the Viking period seems to be a caricature based on an inflation of a much shorter burst of intense social change in the tenth century centred around burials, the post-Viking period has the opposite character: a reduction and eclipse of various changes into one long, bland continuum. Thus the post-Viking period, c. 1100–1900, seems like a very broad canvas, ignoring very classic breaks defined elsewhere in Europe, especially the one dividing the medieval from the modern. At first blush, it seems like a dumping ground for everything in Icelandic history after the golden age of settlement, epitomized in the very term post-Viking, however unofficial that label is. Although one might be tempted to attribute this to a nationalist ideology pervading historical periodization as alluded to earlier, it does not quite fit, as the key cut-off on that basis would be the mid thirteenth century, not the early twelfth. Although certainly an element of nationalism is at play here in terms of the ‘making of a nation’ being linked to the settlement period, of equal import is probably the fact that the twelfth century is also when written sources become available, as mentioned above. In other words, the difference between Viking and post-Viking therefore could equally be labelled as that between prehistory and history—or, more precisely, prehistoric and historical archaeology, as illustrated by debates in Iceland during the 1990s (B. Einarsson 1994; Friðriksson 1994). And although the days have long passed when it was considered that the proper domain of archaeology was its investigation into periods before written sources, anything else being ‘an expensive way of telling us what we already know’, there is no doubt that the legacy of this kind of thinking may still play a role in this division of the Viking and post-Viking periods, especially insofar as far more archaeological time and effort has gone into the excavation and analysis of Viking sites and finds in Iceland.

At the same time, if we wanted to break up this long post-Viking period, it is hard to argue, from the archaeological evidence, for a break at the conventional division of the medieval and modern, i.e. at c. 1500. But then, despite its entrenched position in academia, both historians (e.g. Barraclough 1955; Gerhard 1956; Green 1992; Le Goff 2015) and a few archaeologists (e.g. Courtney 1997; Fagerland & Paasche 2011) have voiced concerns about this divide, pointing to the eighteenth century as a more important century of change. Given that our discussion of the post-Viking period in Iceland has highlighted the eighteenth century over the fifteenth as a major point of change, we may be in good company. And yet one must be cautious here; not everything changes at this time—we have already noted some bigger changes in the nineteenth century and some earlier, in the fifteenth/sixteenth century. Moreover, given the fact that there has been a lot more archaeological work recently on the eighteenth and nineteenth centuries, our understanding of the archaeological record of the thirteenth to sixteenth centuries is comparatively poor. Maybe the eighteenth-century break is an artifact of research bias as much as anything, enhanced by the explosion of written sources that occurs at this time.

Based on this survey of Viking and post-Viking periods in Iceland, it appears that assuming both radical changes at c. 1100 ad and continuity in the succeeding centuries encounters serious problems when examining the data in detail. Summarizing the narrative descriptions in this section, we can roughly map out the chronologies of individual elements in a time chart (Fig. 1), and if we look for ‘natural’ breaks, they are simply not there. This is admittedly a rough sketch. The data are very variable in quality: the dating is often very broad for some items, and added to that, we often have to extrapolate from a few sites or assemblages to the whole country. Furthermore, this chart adopts only a crude presence/absence of elements, whereas frequency might be more subtle. Nonetheless, given these caveats, the chart reinforces our scepticism about Icelandic periodization as it is currently in use.

Figure 1. Time chart of key changes for different elements in the archaeological record of Iceland c. 870–1900 ad.

Conclusions

So where does this survey of Icelandic periodization leave us, and what larger lessons can we learn from it? Tying this back to the general argument in the first part of the paper, the problems in this case can all be summed up by the implicit drive to maintain a periodization which is exclusive and presupposes a sequence of synchronized changes. Such exclusivity or monopolization and synchronicity has never been explicitly stated or even argued for; they are rather assumptions, carried over from the association between the dating function of a periodization and its narrative function. Periodizations like Viking and post-Viking conflate the dating and narrative functions of periodization at a period in our disciplinary history where no such conflation is necessary. Does that mean, then, that terms like the Viking period or post-Viking period should be abandoned?

Yes and no. There are two key take-home messages we would like to propose. The first is that periods like the Viking or post-Viking only work at the narrative level; as dating tools, these terms are meaningless. If we want to retain the use of the term Viking period, at least for Icelandic archaeology, it should probably be confined to a much shorter moment in the tenth century and one that revolves specifically around burial customs; it has much less validity when expanded to refer to a whole culture-historical period. ‘Viking’ would then come to represent something more like a burial fashion or cult which had a brief floruit in the tenth century ad, and we can certainly tell its story. But to use it as a means to define a whole period of time exclusively for Iceland is grossly misleading. The post-Viking period is equally problematic, but for different reasons. Instead of inflating a more specific phenomenon like a short-lived burial rite, this brings to the fore specific elements which evidently do traverse the whole period—but at the expense of backgrounding or ignoring others which do not. Continuity of vernacular architecture and farm mound formation, continuity of selective artefacts like whetstones and nails, can all be conscripted in to produce a narrative of continuity from c. 1100 to 1800. But such a narrative needs then to argue why these things are more important than pottery or other trade goods, which clearly do undergo change over this period.

The second take-home message is that if we want to retain any role for the dating function of periodization, then it has to be concretized more than it has been. It needs to be tied to typological changes in specific artefacts, or changes which are largely site specific. Of course, whether such typological or site-specific changes should even be called periodization is perhaps debatable, insofar as this term is generally tied to culture-historical sequences. But the point we want to make is that, even if independent dating makes culture-historical periodization superfluous as a dating tool, this does not extend to the basic methods of relative dating which underpinned cultural historical periodization—namely, typology and stratigraphy. Indeed, by their very nature they embody very concrete material changes in the archaeological record.

Thus, if the graph in Figure 1 shows anything, it is that the changes in different aspects of material culture rarely synchronize to form major boundaries or discontinuities; and so, while typological changes may still be a useful dating tool, they cannot necessarily be used to define broader narrative periods. The same would apply to changes visible on individual sites, where major episodes of re-building or abandonment are unlikely to be synchronous with similar changes at other sites. This does not mean, for example, that the term ‘Viking’ cannot be applied to a type of artefact found widely distributed in time or space; it just means one cannot automatically extrapolate from such instances to a larger, abstract entity such as a period and claim that all such instances belong to the ‘Viking period’.

This last point raises the question of the ontological dimension of periodization, which is not an issue we have engaged with in this paper so far. If we wish to retain the term ‘Viking’ as a narrative periodization—in the restricted sense outlined above—to what does it refer? Is it purely an analytical construct, a rhetorical ‘fiction’ we use to frame our narrative, or does it claim to have any historical reality? Such questions lead us into old debates in history on the relation between res gestae and historia rerum gestarum (history as what happened versus history as the retelling of what happened) and much of the discussion on narrative in the 1970s and ’80s can be seen as a conflict between those who upheld a sharp separation of narrative from reality or everyday experience and those who did not (e.g. see Meretoja 2014; Strawson 2004). We do not have the space to engage with such issues in this paper, but we do think it useful to open up the question, especially in light of the two points made above. For example, what is the difference in the term ‘Viking’ when used to refer to a period of time as opposed to an object or practice? In the latter case, it clearly has a very concrete and material referent, and some objects are distinctly recognizable as ‘Viking’, such as a trefoil brooch, while others are not, like a schist whetstone; yet both can occur together in a ‘Viking’ assemblage or site and so belong to the ‘Viking’ period.

From this, it might seem as if the Viking period is an abstraction from more concrete, Viking objects and assemblages. But in fact it is the other way round; we only view a trefoil brooch as Viking because it has been found on sites designated as ‘Viking Age’. The term ‘Viking’ therefore—and indeed any period descriptor such as Neolithic or Iron Age—cannot be derived from or attributed to specific objects, but rather refers to a constellation of objects occurring over a certain period of time and region of space. This is, ironically, almost a definition of culture history, but the key difference is that, where culture history interpreted these constellations as manifestations of a culture group, we would make no such assumption. Indeed, what such manifestations refer to are precisely the subject of empirical study; maybe they are a burial cult (as suggested for the term ‘Viking’ in Iceland), maybe they are a trade network, a mode of production or a religious cult. This is ultimately how the notion of narrative periodization still has any utility: whether one should still call such constellations ‘periods’, though, is certainly questionable, although they will always have a temporal quality. For insofar as a narrative is defined as a story unfolding in time, then the history of these constellations can be understood as a product of narrative periodization.

Our goal in this paper has been to try and disentangle the complex legacy of culture history as it pertains to periodization, especially at a time when advances in independent dating, notably the third radiocarbon revolution, seems to have engendered a debate about the utility of such periodizations. We have argued that it is too easy simply to urge their abandonment as it conflates two different functions such periodization fulfils: as a dating tool and as a narrative framework. We agree that the role of periodization as a dating tool is long overdue an obituary; however, its use-life as a narrative device will probably never expire. But in disconnecting the dating and narrative functions of periodization, we also need to be more attentive to how this narrative function might work in the future. Part of this involves abandoning the need for periodizations to carry any exclusivity or monopoly. We simply should not feel the need to perpetuate the model where specific regions or areas have their own culture-historical sequence which are reproduced in textbooks, syntheses and even disciplinary specialisms, especially as most periodizations also have implicit political implications (e.g. Maynes & Watner 2012; Morgan 2016; Orser 2013). Doing this does not invalidate the continuing and productive role typological sequences or stratigraphy play in building our narratives; it just acknowledges the fantasy that we should expect these sequences to synchronize in a manner that can be scaled up to create culture-historical periods like the Viking Age. But the other part of this work involves a more concerted deconstruction of the narrative content of our conventional periods, as for example we have suggested here for Iceland. We need to unpack or ‘reverse engineer’ what exactly is involved in our narrative periodizations, which may often be a legacy from a focus on a very narrow part of the archaeological record, like burials, or selective traits, like settlement continuity. Ultimately, in retaining the notion of periodization as a narrative tool, we should be more attentive to a question of ontology: to what is a period referring, if not a culture group? Any continued use of periods necessitates an explicit answer to this question. It is not enough to argue that scientific dating will make our periodization superfluous, not when our periods embody a complex entanglement of chronological and narrative functions.