One of the primary proto-urban centres of the early medieval world in Northern Europe was without a doubt Hedeby. Hedeby was situated on the border between Scandinavia and Continental Europe, connecting the North Sea with the Baltic Sea by a portage. Its success as a trading hub is inseparably connected to the destruction of the emporium Reric – situated in an area controlled by the West Slavic Obotrites – by the Danish king Godfred in 808 (Reference Tummuscheit, Pestell and UlmschneiderTummuscheit 2003). In order to control and tax the ongoing trade, Reric’s merchants were relocated to Hedeby. However, while in the contemporary historical sources Reric was addressed as an emporium, Hedeby was rather referred to as a portus (Reference KalmringKalmring 2010a: 42–7). Hedeby’s continental denomination Sliaswīk, though, includes the element -wic, derived from the Latin vicus (Reference LaurLaur 1992: 575). It is not until the report of its destruction in 1066 (Adam of Bremen book 3, chap. 50, scholium 81) and in later sources (cf. Helmold of Bosau, c.1167) that Hedeby is referred to as a civitas. Hedeby’s counterpart Birka was situated on Björkö in Lake Mälaren, a small island situated not only in the border area between Uppland and Södermanland but also at the southern point of the borders between the provinces (OSwd. Hundari; cf. OEng. Hundred) Attundaland, Tiundaland and Fjärdhundraland of Uppland. In terms of transport geography, it was favourably situated along the waterway Fyrisleden (Reference AmbrosianiAmbrosiani 1957), leading from the Baltic Sea via Södertälje and Birka to Gamla Uppsala and Vendel. Birka is referred to as both a portus and a vicus in contemporary written sources but also as a civitas, an urbs and even an oppidum (Reference MohrMohr 2005: 98–101; for discussion, see Reference KalmringKalmring 2014/15: 283). The descriptions of both of the sites characterised by these terms are indeed quite similar: Hedeby is described as a Danish port where ‘merchants from all parts [of the world] congregated’ and the attendance of Frisians and Saxons from Dorestad, Hamburg and Bremen guaranteed that ‘an abundance of goods converged there’ (Rimbert chap. 24).Footnote ¹ According to Rimbert (chap. 20), Birka was also frequented by Frisians, complementing Adam of Bremen’s (book 1, chap. 60) description of visits by Danes, Norwegians, Slavs, Sambians and ‘Scythians’ (see Chapter 8, Footnote Note 2). Rimbert (chap. 19) also stated that it ‘contained many rich merchants and a large amount of goods and money’. What constituted these two almost similarly ascribed Viking-age towns, and what were their roles in the vast communication and trade network sometimes described as the ‘Northern Arc’ (Reference McCormickMcCormick 2001: 562–4, 606–12, map 20.4), which stretched from Western Europe to Central Asia? What made it possible for these towns to be able to rise above their regional settings and attract foreign merchants from afar, resulting in abundance and significant wealth? What was the spark that led to the first urban entities emerging so far beyond the borders of the Roman world? Why does the late Iron Age in Scandinavia witness the emergence of Viking-age towns at this specific point in time?

Both sites, Hedeby and Birka, possessed comparable geographically accessible and advantageous locations near borders – that is, where it was easy for people and goods to converge – making them attractive sites for visiting merchants and resulting in considerable economic prosperity. As described in the previous paragraph, there were even similarities in the way they were denoted in contemporary written sources. But can the latter provide further insights into the closer nature of these sites? The labels used in the continental sources for describing such trade sites have been the subject of considerable academic discussion, closely linked to the vast research field that deals with ‘early stages of the European town’ and urbanisation in Central and Northern Europe in general. Apart from perhaps the Scandinavian Christianisation process, there is hardly any other field of early medieval archaeology and Viking-age studies that has been debated as intensely and with so much controversy. Hence, it is no coincidence that there is still no generally accepted and straightforward interpretation. This chapter offers a general orientation of the bewildering discussion about the terminology and associated concepts that have been put forward in order to capture – in distinction to the definitions formulated by historians – the specific nature of this Viking-age urbanisation on the eve of the classical medieval town: on the one hand, through the initial attempt to find a suitable definition along with a corresponding conceptual denomination (the what; the focus on the result) and, on the other hand, through the application of central place and network theories that focus on the interconnectivity of Viking-age towns as ports for maritime trade and urban production (the how?; the focus on the process). By describing and defining these theoretical processes and concepts, as well as their association to a number of specific features of Viking urbanisation, urbanism and urbanity,Footnote ² as a point of departure this compilation at the same time makes it possible for readers to identify the most prominent features of the Viking-age town.

2.1 ‘Seehandelsplätze’, Proto-towns, Emporia and Viking-Age Towns

The concept of Seehandelsplätze (‘maritime trading places’) was developed by Reference Jankuhn and MayerJankuhn (1958) as a way of recognising maritime trading sites along the coasts of the North and Baltic Seas as a separate group and different from the early medieval vici, which occurred inland. He characterised the Seehandelsplätze through their connection to maritime trade and their roles in long-distance trade networks, as well as by their appearance in contemporary written sources and their abandonment during the tenth century, when most of them were replaced by new towns (cf. Reference Steuer, Beck, Geuenich and SteuerSteuer 2005). Following the ideas of Reference PolanyiPolanyi (1963), these types of sites are also subsumed under what is today the somewhat outdated term ‘ports of trade’, which once more stresses their maritime component. His concept describes trading sites that already had some sort of administrative organisation of trade by an authority in societies that had not yet developed markets. Moreover, these sites were usually located on borders between different political and economic entities, where they could exist without belonging directly to a specific territory. Thus they were able to work as gateways in between various political or economic forms of organisation as well as hubs for the exchange of goods between their respective hinterlands and those of foreign traders (cf. Reference Steuer, Beck, Geuenich and SteuerSteuer 1999: 567–74 fig. 78; Reference Steuer, Beck, Geuenich and SteuerSteuer 2003).

In her influential historical study on Die europäische Stadt des Mittelalters (The European Town of the Middle Ages), Reference EnnenEnnen (1972) suggested an application of a flexible Kriterienbündel (‘bundle of criteria’) instead of defining one single, and thus quite inflexible, criterion for a town. Without defining exactly which criteria would be indispensable, or how many verifiable criteria were needed for defining a town, she instead addressed the appearance, inner structure and function of medieval towns (Reference EnnenEnnen 1972: 11–12 and references therein). In archaeology, this pragmatic approach has later been applied by Reference Biddle and WilsonBiddle (1976: 100) for discussing the various stages of urban development in Anglo-Saxon towns, emphasising aspects of defence, street planning, markets, mints, legislative autonomy, their role as central places, the presence of relatively large and dense populations, diversified economies, ‘urban’ plots and houses, social differentiation, complex religious organisation and, finally, their role as judicial centres. These concepts were also discussed at the international conference ‘Vor- und Frühformen der europäischen Stadt im Mittelalter’ (‘Proto- and Early Stages of the Medieval European Town’) in 1972 at Göttingen, where some of the aims were to clarify conceptual problems, finding a common definition for medieval towns that were subject to research and developing a terminology for earlier sites where such definitions were not yet appropriate (Reference Jankuhn, Schlesinger and SteuerJankuhn et al. 1973: 8–9). In this context, G. Reference Dilcher, Jankuhn, Schlesinger and SteuerDilcher (1973) elaborated on the historical aspects of judicial matters in high medieval towns, particularly through four main elements: municipal peace, municipal freedom, town law and civic constitution. Based on these judicial indicators, he concluded that there was no obvious constitutional distinction between the town and its inhabitants and the surrounding countryside in the early towns of the ninth through eleventh centuries (Reference Dilcher, Jankuhn, Schlesinger and SteuerDilcher 1973: 24–7). In the same volume, Reference Schlesinger, Jankuhn, Schlesinger and SteuerSchlesinger (1973: 262) introduced the terms Vorform (‘proto stage’) and Frühform (‘early stage’) of towns. However, he also clarified that such stages should not be understood in any evolutionary sense and were not essential for a town to become ‘fully developed’ later in the late or High Middle Ages. At the same time, even abandoned urban proto- or early-stage settlements should be considered towns. Despite the pluralistic theoretical approach seen in the Swedish medeltidsstaden project (‘The Medieval Town’ project), its inability to break free of the historical definitions is obvious. This is exemplified by Reference AnderssonAndersson’s (1979) contribution, which suggested that the nature and location of a town are defined either by its function within a spatial network, the topography of the settlement’s internal layout or its specific judicial and administrative structure, that is, its town privileges, municipal law, town council, mayor and coat of arms. Depending on the specific emphasis of the approach, his definition would allow for testing whether a site was a ‘central site’, a ‘population centre’ or indeed ‘a town in a formal sense’ (Reference AnderssonAndersson 1979: 6–7; my emphasis).

The term most frequently used for this type of site is emporium, initially defined by Reference HodgesHodges (1982: 50–2; cf. Reference HodgesHodges 2000: 76–92; Reference HodgesHodges 2006: 63–71), who distinguished between three types of ‘gateway communities’ and was the first to classify phases of their development processes. A common feature of these gateway communities is their location on or near territorial borders and coasts. Type-A emporia constitute the earliest category and refer to fairs, used for short periods or only seasonally. The ‘classical’ type-B emporia may have evolved from type-A emporia and these sites emerge from 725 onwards. They were deeply involved in long-distance trade, featured centralised planning in terms of streets and dwellings and had mercantile communities of foreign traders as well as native crafters. In the cases where type-B emporia were not abandoned due to declining trade, such trading places could develop into type-C emporia, where their roles became more associated with the regional economies. Hodges characterises them by their administrative functions, their fortifications and their partially commercialised level of production. Since its introduction, this classification has certainly been challenged (cf. Reference Scull, Hårdh and LarssonScull 2002: 315; Reference Loveluck and TysLoveluck & Tys 2006: 153) – particularly the evolutionary succession from seasonal type-A to permanent type-B emporia – but with minor revisions this classification as an ‘instrument for explaining the rise and fall of the emporia, as well as their differences’ (Reference HodgesHodges 2012: 99) is still valid and a very useful tool (Reference HodgesHodges 2012: 97–9; cf. Reference Hodges, Gelichi and HodgesHodges 2015: 276–7). On a more general level, contemporary publications (Reference Verhaeghe and StoryVerhaeghe 2005: 260–1; cf. Reference HodgesHodges 2012: 94) rather try to avoid the documented yet ambiguous Latin terms found in written contemporary sources, as they

Unfortunately, despite this assessment, occasionally there are still some elaborate scientific attempts to work with this body of historically documented terms (cf. Reference KleingärtnerKleingärtner 2014: 177–91, tab. 10, fig. 30; Reference MalbosMalbos 2017: 11–16).

In 1994, Johan Callmer suggested that too much of the discussion dealing with the early urbanisation in Northern Europe was still left to historians and that – based on the constantly growing body of archaeological source material – archaeologists should try to formulate arguments and models of their own. In a comparison from around the Baltic Sea, he was able to develop an elaborate model for the different phases of ‘trading sites’ and ‘early urban sites’. Callmer started off with the emergence of a ‘non-permanent trading-site phase’, which developed through a North Sea impetus (700–50), after an indistinct break with late Roman Iron Age/Migration period conditions. It was followed by a ‘consolidating phase’, which led to an increase in size and a growing number of permanent or semi-permanent sites (750–800). This resulted in a ‘heyday phase’ for early urbanisation, when a hierarchy between sites was developed, as well as closer control and standardisation of the craft production (800–50). Callmer then envisaged impending collapse when second-level trading sites lost their importance or were abandoned due to ‘political instability’ (850–900), which in turn led to an engagement of local elites at the remaining major sites and a reorganisation of trade in terms of strong regulation and control (900–50). Nonetheless, most of this old system seems to have collapsed (950–1000), leading to a ‘complete breakdown in specialised production’ (Reference Callmer, Ambrosiani and ClarkeCallmer 1994: 72), which is recognisable in considerable parts of the Baltic Sea area. At the very end of the tenth century, the system was being replaced by royally founded towns based on Western European models. Callmer chose to end his reflections with a phase of continued reorganisation and separate development that occurred on the southern Baltic Sea coast (1000–50), which he describes as ‘a different branch of European urbanisation’ (Reference Callmer, Ambrosiani and ClarkeCallmer 1994: 80), and led to denser populations and the introduction of brick buildings (1050–1100).

Reference Näsman, Lyse Hansen and WickhamNäsman’s (2000) contribution to the debate is not so much in a clear terminology – he applies ‘central places of rural character’ in opposition to ‘central places/trading sites’ and ‘proto-towns or early towns’ (Reference Näsman, Lyse Hansen and WickhamNäsman 2000: 42) as well as ‘new economic centres, the emporia’ (Reference Näsman, Lyse Hansen and WickhamNäsman 2000: 62) – but rather in his discussion, which deals with the premises for the initial urbanisation. The point of origin for his reflections is the end of the Saxon wars, which led to the first direct border between the Carolingian and Danish realms and to a ‘rapidly growing political and economic impact’ (Reference Näsman, Lyse Hansen and WickhamNäsman 2000: 37) of the Carolingians and Ottonians on southern Scandinavia. While he alluded to the increasing number of central places and landing sites emerging in the fifth century as ‘non-urban centres’ (Reference Näsman, Lyse Hansen and WickhamNäsman 2000: 53), he also suggested that the first proto-towns, established in the eighth century, were shaped after Western European prototypes (e.g. Quentovic, Dorestad or Hamwic/Southampton). In addition, he differentiated between proto-towns developing into early towns (Ribe, Hedeby) and coastal markets attached to manors or central places, which were rooted in the Migration/Merovingian period (i.e. Åhus with Vä). According to Reference Näsman, Lyse Hansen and WickhamNäsman (2000: 64–8), the early towns were developed to be both royal strongholds and gateways to the advanced monetary economies of Western Europe, while the lingering central place system would have remained deeply rooted in non-monetary barter markets. Näsman also suggested that the establishment of Ribe and Hedeby in the south-western parts of Denmark resulted in the first urbanised economy in the realm, subsequently turning the region into the main province of Viking-age Denmark. He argued that the kings actively supported these towns as alternative royal power bases in order to finance military and political expenses. However, it was not until the tenth century, through the conversion to Christianity and a more direct royal rule, that a new phase of urbanisation took place.

In his book Towns and Trade in the Age of Charlemagne, Reference HodgesHodges (2000) introduced to this debate the term ‘non-places’, coined by the French anthropologist Reference AugéMarc Augé (1995). Augé argued that ‘places’ were ‘relational, historical and concerned with identity’, while ‘non-places’ did not qualify for such anthropological definitions but were rather products of ‘supermodernity’, which had no relationship to earlier places (Reference AugéAugé 1995: 77–8). To Hodges, the ‘Dark Age emporia’ of the post-classical period, which predated the pre-capitalist marketplaces of the later Middle Ages, represented the ‘embryonic state’ of a developing European post-classical economy. They differed from ‘places’ as central places (see Section 2.2) that had well-established cosmological roots, while the novel emporia instead depended on their links to maritime commerce, which introduced new economic practices, and could only develop in marginal locations outside the bounds of the traditional society. Moreover – and despite their large populations and major public installations, such as harbour facilities – they lacked ritual components and monumental architecture and were hence unable to impart any sense of ‘the sacred or history and memory’ (cf. Reference Kalmring, Koçak, Schmidts and VučetićKalmring 2020a). Consequently, as such ‘non-places’ these emporia barely appear in the contemporary historical sources at all (Reference HodgesHodges 2000: 69–71; Reference HodgesHodges 2012: 93–4, 110). Bearing this in mind, and with reference to a discussion article by Reference TheuwsTheuws (2004: 134, 137), Reference HodgesHodges (2012: 93–5, 115) introduced yet another term, the so called mushroom towns, which refers to the distinctly transitory nature of emporia as ‘non-places’.Footnote ³ To him, they were ‘Like mushrooms, these experiments in urbanism grew quickly, some faltered, some were re-configured, and being without monuments, most then vanished or were transformed. In vanishing, most entered “history”’ (Reference HodgesHodges 2012: 115).

With reference to the works of Reference ReynoldsReynolds (1977, Reference Reynolds and Holt1987, Reference Reynolds1992), current research attempts to break away from this bewildering discussion on terminology (Figure 1) and instead tends to refer to these early medieval urban settlements simply as ‘towns of their time’ or as ‘Viking-age towns’ (Reference Clarke and AmbrosianiClarke & Ambrosiani 1991: 3–4; Reference VerhulstVerhulst 2002: 91; Reference Verhaeghe and StoryVerhaeghe 2005: 261; Reference SkreSkre 2007: 45–6). In the particular context of the study, however, it is most remarkable that the latter definition becomes quite restrictive when applied. Reference SkreD. Skre (2007: 453–5 fig. 1.4), for instance, argued that:

Figure 1 What describes a Viking-age town? Conceptual denominations borrowed from contemporary written sources and/or within the academic debate

(S. Kalmring).

Later, he concluded: ‘With what is known now about the specialised sites for trade and craft production in Scandinavia at the beginning of the Viking Period therefore, only Ribe, Birka, Hedeby and Kaupang can be classified as towns’ (Reference SkreSkre 2007: 455).

2.2 Central Place and Network Theories

Another approach to the problem is borrowed from human geography and deals with the geographical concept of a town through a spatial/functional perspective. Its use of the term ‘central places’ is rooted in the old model of central place theory developed by the German geographer Reference ChristallerWalter Christaller (1933; cf. Reference Denecke, Jankuhn, Schlesinger and SteuerDenecke 1973; Reference Schenk, Ludowici, Jöns, Kleingärtner, Scheschkewitz and HardtSchenk 2010), which is identifiable through its characteristic Thiessen polygons. Less sophisticated attempts to reconstruct the catchment areas of major Viking-age towns are based on site-catchment analysis (Reference RandsborgRandsborg 1980: 77 fig. 20). However, ‘central places’ or ‘central place areas’ are settlement agglomerations influenced by many factors, such as power and protection, resources and craft and trade and cult, that are contained within hierarchically structured spaces. The concept of ‘central places’ is primarily applied to sites dating from the Roman Iron Age to the Migration period (e.g. Reference Fabech and FabechFabech 1999; Reference Larsson and HårdhLarsson & Hårdh 2002; Reference Ludowici, Jöns, Kleingärtner, Scheschkewitz and HardtLudowici et al. 2010) that are often characterised by a remarkable chronological continuity. In turn, these ‘central places’ impact surrounding settlements that have fewer functions and are of lesser – supra-regional, regional or local – significance. However, if all functional criteria are fulfilled, Reference Steuer, Beck, Geuenich and SteuerSteuer (2007: 882) argued that it may be valid to recognise ‘central places’ as correlating to ‘early towns’, while Reference SøvsøSøvsø (2020: 16) suggested that it would be more appropriate to compare them to later manorial or estate centres rather than towns.

On the basis of a slowly but constantly increasing number of previously unknown trading sites around the Baltic Sea, Reference SindbækSindbæk (2007a, Reference Sindbæk2007b) has tried to discern between local markets and trading places of an urban character engaged in long-distance trade by applying complex network theory. Relating to Reference Hohenberg and LeesHohenberg and Lees’ (1996: 165) network urbanism as a point of origin, he recognises trading places as traffic junctions within a network. Through quantitative analyses of the amount of imported goods and crafts in relation to the size of miscellaneous excavations, he argued that with Ribe, Kaupang, Birka, Åhus, Truso, Reric and Hedeby only a few sites, or ‘nodal points’, were primarily and predominantly involved in long-distance trade (Reference SindbækSindbæk 2007a). In such a small-world trading network with a scale-free architecture, the few nodal points possessed a multitude of external connections to dependent local markets along the major sea routes. The local markets, in turn, were closely connected to the hubs but had only a few connections themselves (Figure 2). Reference SindbækSindbæk (2007b) consequently argued that this dependency on a few ‘nodal points’, or indispensable hubs, would have made the entire long-distance network very fragile in times of crisis or when under threat.

Figure 2 Viking towns as nodal points with interrelated local markets in complex network theory, based on the distribution of various artefact types across southern Scandinavia

(adapted from Reference SindbækSindbæk 2007a: fig. 5; with kind permission) (S. Kalmring).

Sindbæk’s initial ideas underwent further development and were used to investigate ‘maritime network urbanism’ in the ‘Entrepôt’ project, which analysed global maritime exchange patterns in the early Middle Ages, comparing contemporary developments in Northern Europe, the Mediterranean, Eastern Africa, India and Southeast Asia (Reference Sindbæk and TrakadasSindbæk & Trakadas 2014; cf. Reference HodgesHodges 2012: 108–9). In the context of urbanity, however, ‘entrepôt’ is a rather vague term, generally describing ‘a port, city, or other centre to which goods are brought for import and export and for collection and distribution’ (Reference StevensonOED 2010: 586). Consequently, in the exhibition catalogue, which presented the project’s output, the editors chose to use the term emporium instead (Reference Sindbæk and TrakadasSindbæk & Trakadas 2014). A comparable project in Lund has also investigated the causes behind early urbanisation, focussing on contemporary places in Northern Europe, East Africa and Southeast Asia, in order to gain new insights for this bewildering and paradigmatic subject (Reference MogrenMogren 2013). As a result, Reference MogrenMogren (2013) argued for a deconstruction of the ‘central place’ concept and for the use of a method that highlighted function, agency and communication instead of morphology and structure as essential parameters for ‘urban’ and further emphasised aspects relating to the initial self-organisation of trading and craft communities. Most importantly, however, he used a model with orthogenetic and heterogenetic sites as a conceptual pair (cf. Reference MiksicMiksic 2000). Here, the orthogenetic sites were associated with stability, whereas heterogenetic sites were related to change and entrepreneurship. While the first type of site would characteristically have been ‘placed in areas of surplus agricultural production, are always in some respect central, are populated by a civil, religious and military bureaucracy, might have an aristocratic/royal presence, and are distinguished by some form of monumentality’ (Reference MogrenMogren 2013: 80), the latter is often, as in the case of Hedeby and Birka, ‘liminally placed, between ecological zones, ethnic or political regions, or different transport zones … and often lack[s] monumentality’ but has ‘production and trade of commodities … [as its] most distinguishing feature’ (Reference MogrenMogren 2013: 80). Independent of Mogren’s approach, Reference SøvsøSøvsø (2020) recently came to a similar conclusion and argued for a differentiation between the two quite dissimilar strands that have shaped classical medieval towns: towns that grew out of civitates and those that grew out of emporia (Reference SøvsøSøvsø 2020: 12, 16–18, 253–5). In this case, the civitas category refers to settlements evolving from Iron Age central places, which functioned as religious (pagan) and political centres and were the homesteads of chieftain dynasties. Such places developed into an idealised version of the ‘Heavenly Jerusalem’ at the turn of the first millennium. These civitas sites correspond well to Mogren’s description of heterogenetic sites. Søvsø’s emporium sites were based on the nodal, later royal trading places within the maritime trade networks and thus resemble Mogren’s orthogenetic sites. In the eleventh century, however, even the emporia – if they survived their transformation – were likewise remodelled into Christian towns.

Regardless of how these sites are defined academically, the question of the specific purpose of Viking-age towns remains elementary to the problem. What is at the core of this discontinuous phenomenon at the dawn of the high medieval town – beyond a mere ‘trade and production’ as a commonplace – that these particular places represent? This contribution does not want to introduce yet more new terminology into this already perplexing debate. Instead of posing questions about the what? (definitions, focussing on result) or the how? (network theory, focussing on process), the aim here is to approach the core problem and to ask about the why?, which is at the heart of the golden circle model. The debate on orthogenetic and heterogenetic sites suggests that there is some fundamental difference between traditional centres of power (central places) and the suddenly emerging Viking-age towns. Thus, in order to ultimately define the purpose of these novel Viking-age towns and to draw further conclusions about their societal benefits at this specific point in time, it is first necessary to examine how the surrounding traditional, rural Viking world was structured.