Roskilde – En bygrundlæggelse i vanskeligt terræn

  • Jens Ulriksen
  • Cille Krause
  • Niels Jensen
Nøgleord: Roskilde, bygrundlæggelse


Roskilde – the founding of a town in difficult terrain

Roskilde’s origins do not lie in urban development or in trade networks centred on the fjord and harbour (figs. 9‑10). Intentions back in the late 10th and early 11th century were more along the lines of a royal seat as a tradition-based locus for the power axis between monarch and Church. The area where the town nevertheless sprang up over the course of the next century was characterised by damp scrub, tracts of wet, boggy land, muddy wallows and a generally high water table (figs. 1‑2). It was not uninhabitable, but before houses and roads could be built it was necessary, first of all, to consolidate these wet areas (figs. 3‑8). It was not the most obvious place to found a town and the question is: Why did Roskilde in spite of all that come to be located precisely there?

On the basis of the archaeological data gathered during many years of excavation in the town centre, it is clear that finds predating AD 1000 are exclusively without context and also extremely rare (fig. 11). No remains of buildings with earthfast posts or pithouses from the 8th-10th centuries have been demonstrated within the urban area. Neither are there any records of the flat-bottomed settlement pottery that is so common at the Viking Age settlements of the hinterland. On this basis it can be concluded that Roskilde’s first traces date, at the earliest, from the decades around AD 1000 (figs. 12‑15).

One hypothesis is that the first church, together with the royal residence or kongsgård, was built in a place where there had been a cult site in preceding centuries, in which case this sacred place must have been associated with Roskilde’s many springs. The probable connection can be explained in terms of the significance of springs in Nordic mythology, where three springs are said to flow beneath the ash tree Yggdrasil, the axis of the world. Finding support in the archaeological record for this hypothesis of a holy place is not straightforward. A few traditions handed down suggest that there could have been a very large catchment area for open-air ceremonies in which trees, groves and springs were the focus of attention. Given Roskilde’s striking springs and flushes, it is possible that such events and activities took place there in the Late Iron Age.

If we follow this line of argument, there could have been a continuation of the site’s sacred function into Christian times. The church building was obviously the shrine and the numerous springs that characterised Roskilde would readily have found a place in Christian consciousness. Water plays an important role in Christian belief, for example in the sacrament of baptism and the cleansing holy water (fig. 7). Springs with a fortifying or curative effect similarly form a vital part of Christian conviction, and several of Roskilde’s springs are dedicated to saints.

There is much to suggest that these springs were important to the Christian king. The location of the church and the royal residence appears at least to suggest this – they could hardly have been established closer to the area’s most powerful and productive spring (fig. 16). Roskilde’s first church was probably built on the site later occupied by the cathedral, and the royal residence was located to the west of this. The cathedral was constructed on the most marked feature in the terrain, on the edge of a hill with a steep 10‑14 % downward slope. It is from this hillside that copious springs emerge, sending powerful streams racing down towards the fjord (fig. 8). The royal residence was built on terrain that is more or less level in an east-west direction but on the absolute edge of the slope leading down towards the spring Maglekilde with its impressive deposits of travertine. This location could have been prompted by a tradition associated with the construction of royal halls, as seen at Lejre, Gamla Uppsala in Sweden and Huseby near the trading place of Kaupang in Norway. If the situation of the Lejre halls in the terrain is compared with that of the church and royal residence at Roskilde, it is striking how, in both instances, these buildings were positioned so as to be as conspicuous as possible (fig. 18). This was not, however, on the highest point in the landscape, but on a break of slope, creating an extraordinary visual effect.

Roskilde was founded around AD 1000 with the aim of marking and reinforcing the monarchy and the Church as the incontrovertible ideological, religious and ruling factor in waning Viking Age society. No role as a trading centre was envisaged from the outset and it is therefore unlikely that the initial considerations included the fact that, 50‑100 years later, inhabitants would have to invest substantial efforts in levelling and drying out the terrain because the place had by then acquired an urban character and was expanding rapidly.

Developments similar to those leading to the foundation of Roskilde also took place in Lund and Viborg and – we presume – in Odense. The names of Viborg and Odense indicate the continuation of heathen sacred places, while in the cases of Roskilde and Lund this can only be postulated. As stated above, these four towns possess some common topographic and archaeologically-historically demonstrated characteristics, leading to the conclusion that around AD 1000 they formed part of a greater plan. This exercise involved the establishment of bridgeheads that – mentally and physically – marked the connection between the traditions, practices and customs of forefathers and the young Christian kingdom that undoubtedly contained the seeds of massive social change. The fact that each of these four especially selected places is located in its own ‘country’ is hardly coincidental.

Jens Ulriksen
Roskilde Museum

Cille Krause
Roskilde Museum

Niels H. Jensen
Institut for Miljø, Samfund og Rumlig Forandring
Roskilde Universitet

Ulriksen, J., Krause, C., & Jensen, N. (1). Roskilde – En bygrundlæggelse i vanskeligt terræn. Kuml, 63(63), 145-185. Hentet fra