unless stated otherwise. Please contact us for permission to reproduce any text or
Caerdroia - the Journal of Mazes & Labyrinths
Labyrinths in Pagan Sweden
Stone labyrinth, Landsort, Sweden
Photo: Jeff Saward/Labyrinthos
Reprinted from Caerdroia 21 - 1987 - pp.12-24
The vast majority of stone labyrinths in the Nordic countries do not date from prehistoric
times. They must be less than 1000 years old, indeed some are as late as the 18th,
19th or even the 20th century.
There is however a small group of stone labyrinths in Sweden that might be of more
impressive age. Unfortunately there is no real evidence showing them to have been
built in Pagan times, but there are several clues pointing in this direction. From
necessity, these conclusions have largely been based on guesswork.
In southern Sweden there are 17-18 labyrinth sites which seem to belong to this oldest
group. Some still exist, others have been destroyed long ago, and some are only indicated
by place names. In contrast to the large group of more recent labyrinths, these apparently
old ones are not usually situated along the coasts, but in the heart of the oldest
farmland. They are as a rule solitary, while the coast labyrinths are usually found
in dense groups.
These older labyrinths are usually built on high ground, for example, on top of small
hills close to important river crossings, or on top of eskers created by the gravel
of rivers flowing beneath the glaciers that covered Sweden 10,000 years ago. The
entrances of these labyrinths are often orientated towards the west. In seven cases,
they are situated in prehistoric grave-fields. At four places they are found very
close to churches dating from the early Middle Ages. Five are situated at places
where towns developed in the early Middle Ages.
The Goddess in the Labyrinth
The pattern of distribution provides us with an interesting clue. These labyrinths
seem to be spread very evenly over the oldest settled parts of Sweden. In the most
densely populated areas they appear at a distance of 20-40 km from each other. It
seems as if they reflect a system of old communities, where each pagan tribe had
a labyrinth and most of the inhabitants lived within one day walking distance from
It is probable that labyrinths played an important role in the pagan cult of such
prehistoric communities. They were used in spring for religious games or ceremonies.
Surviving lore from different countries indicated that on these occasions a girl
played the role of the mother goddess and took her place at the centre of the labyrinth
(=nether world). One or two men played the roles of the sky god who would liberate
or abduct the mother goddess from her prison castle of the netherworld. When she
had been abducted from the labyrinth she probably joined with her liberator in a
spring wedding (see my discussion in The Goddess in the Labyrinth, Åbo Akademi, 1985).
These small prehistoric communities were probably ruled by sacral chieftains (‘kings’
or ‘queens’) who based much of their authority on their roles as priests and priestesses.
Most of these petty kingdoms disappeared long before Sweden was Christianised. They
were united into larger provinces, so-called landskap and langsagor. However, some
of the very old territories survived as fairly independent communities into the early
Middle Ages (Värend, Finnveden, Tjust and Dala) and the names of a couple of others
have been preserved (Rek and Tör). But the rest of Sweden's old petty kingdoms have
so far remained a mystery.
The places with labyrinths give some idea of where the petty kingdoms had their pagan
cult centres. This is not enough to reconstruct the map of pagan Sweden, but it gives
us a starting point.
The Sacred Places
Much valuable information can be gained from place names which indicate a use as
pagan cult places. In these place names, well-known gods from the Viking Age like
Oden, Tor, Frej and the goddess Freja are combined with words like ‘vi’ (sanctuary),
‘harg’ (cairn or altar), ‘lunda’ (grove), ‘vin’ (meadow), ‘åker’ (field), ‘tuna’
(enclosed area or fence), ‘berga’ (hill), ‘ö’(island), ‘sjö’ (lake). These elements
form place names like Odensåker, Torstuna, Frösvi, Fröberga, etc.
There are also traces of a group of even older and more obscure gods among the Swedish
place names. “Ull” was presumably a sky god and “Njärd” was probably identical with
the mother goddess “Nerthus”, described as terra materby the Roman historian Tacitus
(c. 100 CE). This divine couple of a sky god and a mother goddess from the early
Iron Age has attracted a lot of interest from the experts. They had discovered long
ago that the cult places of Ull and Njard often appear together in pairs. Another
group of place names may come from the cult places of an old goddess, “Skädja” (Skadevi,
Skädharg, etc.) but this interpretation has been met with more scepticism amongst
I have compared these place names with the labyrinths, and I have come to the conclusion
that the cult places of Ull and Njard can usually be combined with those of Skädjaand with the oldest labyrinths. Together they form groups of cult centres, each containing
one labyrinth and at least one cult place of each of the gods and goddesses, Ull,
Njard and Skädja.
Let me give an example from my home town of Västerås, which is situated close to
such a group of pagan cult places belonging to a prehistoric community. Close to
the town, in Badelunda parish, is a large stone labyrinth preserved on top of an
esker. Earlier there was also an extensive prehistoric grave field on the same spot,
now destroyed by gravel extraction. Only 800 metres north of the labyrinth lies Anundshög
mound (probably the largest in Sweden), 12 metres high and 60 metres diameter, that
has traditionally been the site of ‘ting’-proceedings. Adjacent to the mound are
several impressive grave fields with stone settings in the shape of large ships;
altogether the complex contains more than 230 visible prehistoric graves. 1 km to
the east of the labyrinth is a farm called Närlunda (the goddess Njärd's grove) and
some 4 km SE of Närlunda lies a farm called Ulivi (Ull's sanctuary). In the same
concentrated area, at Tuna, less than 1.5 km east of the big mound, was a very interesting
grave field (86 graves) with one unusually rich grave of a woman from c.300 CE, and
8 boat graves, also for women, from the Viking Age (c.800-1070 CE). One theory is
that these women were local priestesses. Some distance away, 10 km west of the labyrinth
is a parish called Skärike, the modernized version of Skädharg (Skädja's cairn).
The Recurrence Hypothesis
No one can fail to understand from the description of the neighbourhood of Vasteras
that here was the centre of a prehistoric community, a pagan congregation, dating
back to the early Iron Age (c.500 BCE -400 CE). In this example, the pattern is
unusually clear and easy to interpret. However it is possible to find similar combinations
of labyrinths, tingmounds and cult places of Ull, Njärd and Skädja at several other
places and use them to reconstruct Sweden's earliest political map.
The key roles in this reconstruction are played by the place names including Ull,
Njärd and Skädja. They give us a fairly good idea of Sweden's territorial organisation
in the early Iron Age. It is more difficult to select the labyrinths because we can
usually not be quite sure of their age. The correct tingsites (‘ting-mounds’ or
‘tingstad’ type place names) are also difficult to select because in the sources
at our disposal, they are mixed with a large number of ting sites used in the later
territorial organisation of ‘härader’ and ‘hundaren’ (Viking and Middle Ages). There
is no simple, reliable method of selecting the older tingsites from the later, but
it reasonable to guess that ting-names, which have never, to our knowledge, been
connected with the later territorial organisation, may belong to the older group.
These tingsites usually have names of the type ‘Tingstad’ (stad=place) or ‘Tingshög’
Combining these elements , it is fairly easy to sketch the centres of the old communities.
The cult places of Ull, Njärd and Skädja are combined into groups which form skeletons
of the old communities. The selected labyrinths and tingsites usually fit perfectly
into the pattern. There can hardly be any doubt that this method leads us in the
I call my method form the reconstruction of pagan communities the ‘recurrence hypothesis’.
The simple idea is that prehistoric labyrinths or place names belonging to pagan
cult places do not recur in the same ‘old’ community. When you move from one place
called Ullvi to another Ullvi you cross an old border, just as you cross the border
between two Christian parishes when you drive from one church to another.
A Map of Pagan Sweden
The exact borders are difficult to determine. Borderlines in our sense of the word
were unknown in these days. The different communities were often separated from each
other by deep forests, marshes or other uninhabited area. In such cases it is often
easy to determine the extension of the different communities. Another method is to
borrow borderlines from later territorial units like the ‘härader’ and ‘hundaren’,
in the hope that they have preserved something of the older territorial division.
One way to check what is a reasonable guess is to measure the distances and presume
that a farm usually belonged to the community offering the shortest and most convenient
route to its cult places. With these tools it is possible to draw a map of pagan
Sweden at the time when the gods Ull, Njärd and Skädja were worshipped. There is
no space here for a detailed description of the different communities. Let me only
say that the cult places of the three oldest gods give a good picture of most of
Sweden, except in the far south, where these kinds of place-names are missing, and
Gotland where they are too few to give any reliable guidance (see maps below).
A careful study of the later group of place-names referring to Oden, Tor, Frej and
Freja confirms much of the older territorial division. Only in one case (Fjädrundaland)
does it seem as if an older community was split into two pagan communities during
the later period. In at least three cases (Närke, Östergötland and Gotland) it seems
as if older communities were united to form larger units, with common cult-places
for the worship of Oden, Tor, Frej and Freja. These steps towards larger religious
units reflects the creation of Sweden's more modern territories, its ‘landskap’ or
The distribution of the labyrinths seems to fit in better with the communities defined
by the old gods Ull, Njard and Skädja than with the later units. This indicates that
the labyrinths played an important role in the cult of the early Iron Age (c.500
BCE - 400 CE). Even so, some labyrinths are also situated on the (supposed) borders
of the oldest communities which were later united to form larger ones (Vasby in Närke,
Viby in Östergötland, Visby on Gotland). These examples of borderline labyrinths
might be explained as later labyrinths, built for common use by two older communities
which were now united and had created a new meeting place on the old border for their
It is probably safe to assume that labyrinths have a long history of pagan religious
use, and have remained in use well into the early Middle Ages. Thus it is probably
correct to guess that they still belonged to the pagan cult during the late Iron
Age. But this little study of the old petty kingdoms of Sweden rather points to the
Iron Age as the time when labyrinths played a crucial role in the pagan cult.
Anundshög, in the distance, and some of the numerous stones at Badelunda.
Photo: Jeff Saward/Labyrinthos.
Map of the Västerås area with the Trojienborg labyrinth at Tibble in Badelunda parish.
In the neighbourhood of the labyrinth is a concentration of grave-fields from the
Iron Age and three big mounds, one of them, Anundshög, is the largest in Sweden
The town of Västerås (Västra Aros) belongs to a more recent period than the prehistoric
monuments. It probably originated as a market place at an important ford, close to
the mouth of a river which also served as a good harbour for deep-draft ships.