Reading the prehistoric landscape

Mats Widgren

During the first millennium bc, farmers in Scandinavia intensified their

farming. Stones were cleared from fields to a much larger extent than before. In many areas, as cultivation became more permanent, boundaries
in the form of earthen banks or lines of stones were developed. From
then on, human influence on the landscape was seen not only in the
composition of the vegetation, but also and increasingly in the morphology of the land. The woodlands of the interior part of southern Sweden contain a large amount of clearance cairn fields, which originated
during this period. On the heaths of Jutland (Denmark) and in the pastures and sparse woodlands on the island of Gotland (south-eastern Sweden), it is possible to walk in field systems that originated in the eighth
century bc. Similar field systems emerged in the whole of northern Europe at about the same time. In the British Isles and central Europe, such
traces are even older (Bradley 1978, Fries 1995). For the historical geographer and the landscape archaeologist, these remains open up the
possibility of reading the landscapes of the past.
For Swedish historical geographers, the ancient fields have become
an important source material for understanding past land use and past
societies. This chapter presents some of the results from this research,
focusing on the development of farming and societies during the Iron
Age, which in Swedish prehistoric chronology refers to period between
roughly 500 bc to ad 1050.
69

Ymer 2010

Sweden through Time and Space

The landscape concept

Can prehistoric societies and the spaces that they were operating within
be meaningfully understood by researchers living more than 1000 years
later? In answering that question, I take as my starting point the concept
of landscape. It is a polysemic term, one with many different, but closely
related, meanings. Whereas the main definition of landscape in the English language is as a picture representing a view of natural inland scenery or the art of depicting such scenery, it can also mean the landforms
of a region in the aggregate or a portion of territory that can be viewed at
one time from one place (Merriam Webster Online Dictionary Sept.
2009). In the Swedish language, the concept of landscape (Sw. landskap)
is not so immediately connected only to pictures and the scenic aspect. It
can refer to a portion of territory that can be viewed from one spot but it
also refers to old Scandinavian pre-Christian lands or provinces (Hgerstrand & Sporrong 1993).
In his article Recovering the substantive nature of landscape, Kenneth Olwig showed how the English connotation of landscape as scenery was developed via Dutch landscape paintings from an older
German concept that relates to the territory, the conditions of that
territory and the customs and rules under which the land was governed (Olwig 1996). This understanding of landscapes as lived-in territories is, of course, in line with Swedish usage of the word landskap
when referring to the old, pre-medieval lands that preceded the State
in Nordic countries. Similar words exist in most languages, often alluding to land, soil, earth, and people and nation. On the other hand,
the specific Anglophone landscape concept is missing from, or was
only recently added to, many other languages of the world. In current
usage, the term landscape thus refers to at least three different aspects, as shown in Table 1.
Through the medium of the English language, the visual aspect of
the landscape concept has become a part of Nordic languages. However, we also know that what we see is dependent on what we know
and what we value. The landscape concept therefore also opens up
70

Mats Widgren

Reading the prehistoric landscape

Table 1. Landscape concepts.

Landscape as scenery

Representation
Idea (mental construction)
A way of seeing

Landscape as institution

Customary law
Social order, land rights
A way of communicating, a way of acting

Land as a resource

Land use
Production
Capital

the various ways different people may perceive the same surround
ings or the same image. The Germanic and Nordic landscape concept,
on the other hand, focuses on the people on the land, their territories,
their traditions and the social institutions that govern these territories. Finally, landscape is often used as an equivalent to land, and the
ways in which it has been transformed by labour and served as a basis for both biological production and accumulation of wealth (Widgren 2004).
When we talk about landscapes, we are thus dealing with three different, but closely interrelated, concepts. Together these concepts offer
important insights and analytical tools for considering how we relate
to our physical environment and to social structures. At the same time,
the concepts provide a key to the material analysis of landscapes as
something substantive and materially existing out there.
I have thus argued that the reading of landscapes can be seen as a
powerful approach in social science (Widgren 2006). Sometimes tak
ing a starting point in the landscape can open the way for an understanding of social structures and power relations that are not evident
to social scientists and historians using other sources. However, land
scapes and social structures do not have a one-to-one relationship.
Using landscapes as an entry point into social relations demands careful
consideration of the socio-spatial dialectic. Landscapes are the result of
71

Ymer 2010

Sweden through Time and Space

human decisions and of social relations. However, not all events in society and all aspects of power relations are expressed transparently on the
ground. During certain periods and in certain contexts, individuals and
societies make clear and readable imprints in the landscape, while other
societal changes leave few or faint traces.
(Widgren 2006, p. 57)

A checklist for reading landscapes

The analysis of landscapes can never achieve the same exactness as, for
example, a chemical analysis. Nevertheless, a structured reading of
landscapes is possible. Such a reading can be compared to wine tasting
as it contains several types of observations. Furthermore, as in wine
tasting, it is possible to agree on a structure for these observations,
some of which can be objectively measured whereas other judgements
are more qualitative. A group of experts might agree on this latter type,
in what is called an intersubjective judgement. However, there are other judgements, both of wine and of landscapes, which are truly subjective, individual and not shared among a larger group of persons. A
structured reading of landscapes can follow this path.
Form. Forms are easy to agree upon. The landscape comprises convex
and concave forms, points, linear forms and areas. Projected into the
plan of the map fields can be circular, rectangular or irregular, have a
reversed S-shape, or even be shaped like a banana. Classification of the
morphology of the agrarian landscape is well developed, and in the
1960s a group of European landscape researchers agreed upon an international terminology (Uhlig & Lienau 1967).
Function. The mere classification of forms, however, tells us nothing if
it does not help us understand the function of that form. Similar forms
can have different origins and different functions. A clear classification
of the form helps in talking about features in the landscape, but does
72

Mats Widgren

Reading the prehistoric landscape

not solve the really interesting questions: what purpose did this form
have, when was it constructed and by whom?
Process. Landscapes may seem static, especially when examined in photographs. However, all landscapes are undergoing processes of change.
Sometimes these changes are fast, sometimes slow. In analysing and understanding landscapes, we can make use of the concept of process in
two ways. First, a person with a trained eye, perhaps familiar with the region or with the features in the landscape, can see changes even in a still
photograph or by looking at a landscape where no apparent change is
happening. One example might be a newly cleared patch in the forest or a
formerly arable field in the process of regrowth with bushes and trees.
Another way to see processes is motivated by the signs of past and longsince finished processes. Sometimes labour processes, daily routines,
etc., have been fossilised in the landscape so that age-old practices can
be read from that landscape. Such changes tell us about societal or natural forces that drive the changes. Even the most stagnant landscape is the
result of historical processes. Few landscapes lack revealing signs of the
history they have gone through. Most landscapes are littered with historical evidence, be it a dismantled railway, the curving streets in a medieval town, or the property structure established during enclosures in the
18th and 19th centuries. A close reading of landscapes past and present
must look for these signs of ongoing and past processes.
We can thus conclude that form, function and process are important
constituents in a formalised reading of landscapes. They can even be
seen as creating a straightforward, neutral and scientific way of reading
landscapes. However, all experiences from such readings (showing
landscape photos to students, being exposed to landscapes abroad)
tell us that the imagination is too restricted to understand new land
scapes just by considering these three questions in the checklist. Swedish students may be clever enough in reading and recognising their
native, and also British and American landscapes, but can have prob
lems in understanding, for instance, African rural landscape images.
73

Ymer 2010

Sweden through Time and Space

From one cultural sphere to another, the way people have used their
land differs. Differences in culture, land tenure, class structure and pow
er relations make their clear imprint on the landscape. This is where
the cultural understanding of landscapes as (different) ways of seeing
and understanding the world becomes an important analytical tool.
The checklist, formfunctionprocess, must be completed in association with the concept of context.
Context. Landscapes have developed in different cultural, social and
economic frameworks. To understand landscapes, one must therefore
always ask about the context in which it was developed. This analytical
step involves transcending the obvious (i.e., what may seem obvious to
the observer) to imagining other possible contexts. When analysing
landscapes of the past, it helps to remember the phrase the past is a
foreign country (Lowenthal 1985) Sources other than the landscape
forms have to be mobilised to reconstruct the social and cultural context in which people shaped different landscapes.

Prehistoric field systems in Sweden

The checklist suggested above should ideally work for all types of land
scapes urban, rural, present, and past. In the following part of this
chapter, I give some illustrations of how a close reading of forms in ancient, abandoned, agrarian landscapes can give insights into their func
tions and processes, and their contexts.
One of the most important sources for the investigation of early
agriculture is the land itself. Clearing and cultivation have in many
cases left a permanent and clear imprint on the landscape. Especially in a country such as Sweden, where much of the land consists of
till (unsorted material deposited directly the ice: Sw. morn), these
remnants can now be seen in pasture and forests abandoned by later,
more mechanised farming practices. In many parts of Sweden, the
clearing of stones is a precondition for cultivation and efficient
74

Mats Widgren

Reading the prehistoric landscape

Figure 1. Features in fossil agrarian landscapes. Drawing by Charlotte Rinaldo.

t illage. Therefore, the remains of this clearing in the form of clearance cairns are the most common traces of previous agriculture.
Furthermore, when land was lying bare after tillage, precipitation
contributed to the sculpturing of the fields in the form of accumulation downslope and erosion upslope. In that process, low terraces
(English special terminology: lynchets) have formed downslope. The
contrasting form upslope is usually called a negative lynchet. Similarly, on light sandy soils, wind erosion can have contributed to the accumulation of sand and soil on the banks between fields and also to a
sunken surface in the middle of a field (Figure 1).
Different ways of working the soil can also have contributed to the
morphology of the cultivated landscape. During the medieval period
and later, a fully fledged plough with a mouldboard was used in western Sweden and the fields were then often ridged so that a characteristic pattern of ridge-and-furrow was developed. The plough equipment that was used, before, during the Iron Age was an ard (or scratch75

Ymer 2010

Sweden through Time and Space

plough), which lacked a mouldboard. To achieve a thorough preparation of the soil, the lands were cultivated in a criss-cross pattern, which
was often connected to square fields and to the accumulation of soil on
the banks between fields.
In many cases, stones were used to construct walls that acted as barriers for livestock, preventing them from entering the fields. In the first
millennium ad, on the islands of Gotland and land, houses were often constructed with rather substantial stone foundations. All these
traces form the basis for reconstructions of the prehistoric agricultural
landscape.
During the past 50 years, historical geographers and archaeologists
have documented such areas of ancient fields and settlements in different parts of Sweden, especially in the central and southern parts. This
has been based on a series of thorough and detailed mapping exercises
in the field, in rough pastures and in woodlands. Based on this, four
types of ancient fields can be distinguished, which together give insight
into farming and society during the Iron Age.

Clearance cairn fields in the

interior of southern Sweden
The interior parts of southern Sweden sydsvenska hglandet (the
southern Swedish highlands) exhibit a type of natural landscape that
differs from the valleys and plains that surround it. In contrast to many
more intensively cultivated areas of Sweden, after the last deglaciation
this area was never covered by the sea or by the different lake stages of
the Baltic Sea. These highlands are above the highest shoreline, that
the sea and the different lakes stages formed after the deglaciation. As
a result, the till left by the ice was never wave washed and sorted. Consequently, there are few broad valley bottoms but instead large areas
covered with till, boulders and stones, which has created a landscape
unsuitable for modern farming. On the other hand, the silt fractions in
these soils are still there, rich in minerals and with a capacity to retain
76

Mats Widgren

Reading the prehistoric landscape

Figure 2. Clearance cairn field in Jrparyd, southern interior Sweden. 800 bc to

400 ad. No individual fields are visible. The pattern of clearance cairns reflects
the clearance of several partly superimposed irregular fields over a long period
(from Pedersen & Jnsson 2003).

77

Ymer 2010

Sweden through Time and Space

nutrients. In a typical situation under the highest shoreline, these fractions were washed down to the valley bottoms and plains where they
now form a rich cultivable clayey soil.
These lands were not the most attractive for early farmers, but with
the technology of that time, these stony soils were less of a hindrance
than they are for modern steel ploughs. Farmers cleared the land of
stones and, as a result, large areas of cairn fields remain in the present
woodlands. The remnants of these field systems cover substantial parts
of the present forests, well beyond what later became infields and meadows. These cairn fields often have a long and complex history. Many of
them originated in the Late Bronze Age, from the ninth to sixth centuries bc and some of these areas were still in use as agricultural fields
into the early second millennium ad.
There are few visible boundaries between separate plots in this type of
ancient field, but where it has been possible to reconstruct the form of
the separate fields, they are small irregular, rounded fields of the size of in
the order of a hundred square metres. Some short- or long-term rotation
of fields must have been used on these lands. This, together with the evidence from vegetation history, bears witness to a mosaic-type of landscape with small cultivated fields alternating with secondary woodlands
(Widgren 2003)

Celtic fields on Gotland and in Skne

On Gotland and in the southernmost province of Sweden, Skne, the field
systems were of the same character as in continental northern Europe.
They consisted of extensive areas of Celtic fields small quadrangular
and rectangular fields bounded by sandy or earthen banks. Investigations
of some of the many extensive field systems of this type on Gotland show
that such field systems emerged in the eighth century bc and were still
functioning until the second century ad (Figure 3). Whereas Celtic fields
from this period are common in Denmark, they are less well documented
in nearby Skne, but the scant evidence suggests they were common the78

Mats Widgren

Reading the prehistoric landscape

Figure 3. Celtic fields at Vinarve on Gotland. 800 bc to 200 ad. The quadrangular form
of the fields is related to criss-cross ploughing with an ard (from Lindquist 1974).

re during the first millennium bc. The fields were tilled with wooden ards,
and since about 800 bc, these were equipped with separate wooden shares. From excavations of fields in different parts of Sweden, the characteristic criss-cross pattern associated with ard tillage has been documented (Lindquist 1974, Carlsson 1979).

Stone wall enclosures in eastern Sweden

and on the Baltic islands
In the provinces of Uppland and stergtland and on the islands of land
and Gotland in the Baltic, large systems of stone-walled enclosures were
built in the first part of the first millennium. They epitomise a specialised
type of farming where the raising of cattle and sheep was closely integrated with the cultivation of small, well-manured fields. The stone walls
connected several single farmsteads in a common grazing organisation
(Figure 4).
79

Ymer 2010

Sweden through Time and Space

Figure 4. Collapsed stone walls in Srstad, stergtland. ad 200500. Note the long
double rows of stone walls, which had been used as cattle paths or drove-ways. They
give witness to past processes of daily transport of livestock from common pastures to
the farmsteads (source of illustration) (from Widgren 1986).

On Gotland, this type of stone walls directly overlies the previous

landscape of celtic fields. Enclosed areas of cultivation of crops and
hay meadows intended for the collection of winter fodder became
clearly separated from outlying areas for grazing. Especially on the
islands of land and Gotland, there was at the same time a trend
towards increased sheep farming. It is clear that the whole landscape
changed during the first few hundred years ad. From the form of the
stone wall complexes, we can infer that there was increased competition for land. Land had to be separated more clearly for different purposes. We can also see that the integration of the livestock component
and the arable land became more important. Such long stone-walled
cattle paths are usually found in farming systems where there are daily
movements of cattle, as for milking. The wider political and economic
80

Mats Widgren

Reading the prehistoric landscape

Figure 5. Strip fields in Mnstad,Vstergtland, first millennium ad. These strip fields rep
resent a first division of the land into almost equal parcels. In contrast to what is seen
in the celtic fields of Gotland (Figure 3), the form of the fields can hardly be related to
tillage. The broad strips are most probably remnants of a division of the land based on
a system of rights. Inside some of the broad strips in the western part of the field system can be found small irregular plots that relate to cultivation (from Widgren 1990).

context in which we should understand this intensification must be

sought in the fact that the first 500 years ad in Europe was a period of
increased trade and exchange, which also affected the northern regions, well beyond the direct political influence of the Roman Empire.
It seems clear also that the Scandinavian countries were affected by increased economic specialisation (Widgren 1983, Pedersen & Widgren
1998).
81

Ymer 2010

Sweden through Time and Space

Strip fields in western Sweden

In south-western Sweden, another type of field pattern, which originated
in the first millennium ad, can be found. In the forests and pastures in
Vstergtland and Halland, a large number of localities with long and
straight lines of stones and earthen banks were found in the mid-1980s.
They did not resemble anything that had been found before. The lines
form a regular pattern of several strips but, in contrast to strip fields that
are found on 17th and 18th century cadastral maps, these strips are often
much broader. A closer analysis of the fields documented in Figure 5
shows smaller divisions and signs of cultivation inside the broad strips.
An important distinction in the terminology of forms in the agrarian
landscape is that between ownership parcels and working parcels. In this
case, the broad strips demarcate ownership whereas the actual cultivated plots, the working parcels, were much smaller. We can thus draw the
conclusion that the geometry of the fields in this case was not caused by
the technology of cultivation but rather in the social sphere. There are
strong indications that such fields started to develop in the middle of the
first millennium ad in Vstergtland, but the most recent dates obtained
from Halland point to the late first millennium ad and into the first few
hundred years of the second millennium. It is not easy to reconstruct the
social context of these fields with certainty, but when similar field patterns have been documented later in history and in other parts of the
world, they have often been related to a social context where land rights
are based on kinship and a perceived or real genealogy (Widgren 1990,
Mascher 1993).

The regional pattern

The map in Figure 6 gives a rough picture of where the different forms
of ancient fields from prehistory and the Early Middle Ages in Sweden
can be found. Clearance cairn fields can be found throughout Sweden
but they are best preserved in the interior of southern Sweden above
the highest shoreline. Extensive areas of contemporaneous Celtic
82

Mats Widgren

Reading the prehistoric landscape

Figure 6. Distribution of different types of ancient fields and settlements

from the first millennium A.D. The map was compiled in the mid 1990s
and since then new finds have been made, especially of strip-parcelled
fields in Southwestern Sweden (from Widgren 1997).

83

Ymer 2010

Sweden through Time and Space

fields are found in Denmark, were most probably common also in large
parts of Skne in southernmost Sweden and have been well documented on the island of Gotland. A clear division between east and west is
manifested in the different distribution patterns of the stone-wall complexes and the broad strip fields. The specialised large stone-walled enclosure complexes characterise some provinces in the eastern part of
Sweden, from the Baltic islands in the south to the province of Uppland, north of Stockholm. On the other hand, the broad strip fields
from the latter part of the first millennium have a clear south-westerly
distribution, with some scattered occurrences also in the eastern part
of the southern interior. This characteristic difference between west
and east is a phenomenon that can also be found in other later cultural
and social phenomena in the historical geography of Sweden.

Bibliography
Bradley, R. (1978): www.jstor.org/stable/124286 Prehistoric field systems in Britain and
North-West Europe - A Review of some recent work, p. 265280 in www.jstor.org/action/showPublication? journalCode=worldarchaeology World Archaeology, Vol. 9.
Carlsson, D. (1979): Kulturlandskapets utveckling p Gotland: en studie av jordbruks- och
bebyggelsefrndringar under jrnldern [The Development of the cultural landscape
on Gotland: a study of changes in agriculture and settlement during the Iron Age],
Stockholms universitet, Stockholm.
Fries, J. C. (1995): Vor- und frgeschichtliche Agrartechnik auf den Britischen Inseln und
dem Kontinent. Eine vergleichende Studie. Verlag Marie Leidorf, GmbH, Espelkamp.
Hgerstrand, T. & Sporrong, U. (1993): Landskap p. 265280 in Nationalencyklopedin: ett
uppslagsverk p vetenskaplig grund utarbetat p initiativ av Statens kulturrd, Vol. 12,
Bra bcker, Hgans.
Lindquist, S.-O. (1974): The Development of the agrarian landscape on Gotland during
the Early Iron Age, 632 in Norwegian Archaeological. Review, Vol 7, No. 1. 1974.
Lowenthal, David (1985): The past is a foreign country, Cambridge University Press, Cambridge.
Mascher, C. (1993): Frhistoriska markindelningar och rjningsrseomrden i Vstsveri-

84

Mats Widgren

Reading the prehistoric landscape

ges skogsbygder, Kulturgeografiskt seminarium 93:2, Stockholms universitet, Stockholm.

Merriam Webster Online Dictionary Sept. 2009.
Olwig, K. (1996) Recovering the substantive nature of landscape, p. 630653 in Ann.
Assoc. Am. Geogr 86.
Pedersen, E. A. & Widgren, M. (1998): Jrnlder 500 f.Kr. 1000 e.Kr. p. 237459 in Welinder, S. Pedersen, E. A. & Widgren, M. Jordbrukets frsta femtusen r, Det svenska
jordbrukets historia, Natur och kultur/LT i samarbete med Nordiska museet och
stiftelsen Lagersberg, Stockholm.
Pedersen, E.A. & Jnsson, B. (2003): Rjningsrseomrdet Jrparyd i Rydaholms socken, Finnveden, Smland p. 1170 in Widgren, M. (ed.) Rjningsrseomrden p sydsvenska hglandet: arkeologiska, kulturgeografiska och vegetationshistoriska underskningar, Meddelanden frn Kulturgeografiska institutionen vid Stockholms universitet, Stockholms universitet, Stockholm.
Uhlig, H. & Lienau, C. (eds) (1967): Flur und Flurformen: Types of field patterns = Le finage agricole et sa structure parcellaire. Giessen.
Widgren, M. (2004): Can landscapes be read? p. 455465 in H. Palang, H. Soovli, M.
Antrop, G. Setten (eds) European rural landscapes: persistence and change in a globalising environment, Kluwer Academic Publishers, Boston.
Widgren, M. (1983): Settlement and farming systems in the early Iron Age: a study of fossil
agrarian landscapes in stergtland, Sweden. Stockholm.
Widgren, M. (1986): Bebyggelseform och markrttigheter under jrnldern, p. 1826
in Ymer 1986.
Widgren, M. (1990): Strip fields in an Iron Age context: A case study from Vstergtland, Sweden, p. 524 in Landscape History, vol. 12, 1990.
Widgren, M. (1997): Fossila landskap: en forskningsversikt ver odlingslandskapets utveckling frn yngre bronslder till tidig medeltid, Kulturgeografiska institutionen,
Stockholms universitet, Stockholm.
Widgren, M. (ed.) (2003): Rjningsrseomrden p sydsvenska hglandet: arkeologiska, kulturgeografiska och vegetationshistoriska underskningar. Meddelanden frn Kulturgeografiska institutionen vid Stockholms universitet, Stockholms universitet, Stockholm.
Widgren, M. (2006): Reading property in the landscape, p. 5764 in Norsk Geografisk
Tidskrift Norwegian Journal of Geography, vol. 60.

85