ROBERT JOHNSTON

PATTERN WITHOUT A PLAN: RETHINKING THE BRONZE

AGE COAXIAL FIELD SYSTEMS ON DARTMOOR,
SOUTH-WEST ENGLAND

Summary. The coaxial field systems on Dartmoor are widely interpreted as

the result of a relatively rapid period of planned land division during the middle
centuries of the second millennium BC. This article seeks to challenge this
notion of a ‘planned landscape’. Using examples from southern (Shaugh Moor)
and north-eastern (Kestor and Shovel Down) Dartmoor, it is demonstrated that
the boundaries materialized existing structures in the landscape which had
emerged through patterns of dwelling and long histories of tenure. In seeking
to present a new narrative for the enclosure of the Dartmoor landscape, it is
argued that tenure was articulated at a local level through the relationship
between occupancy and ancestral ties to the land, and that land division was
only possible because the forms of tenure and perceptions of landscape were
already in place. The coaxial pattern emerged in a reflexive tradition of
boundary construction rather than as part of a transformative plan or a
conscious strategy to reorganize and enclose the moor.

The Bronze Age fields on Dartmoor are amongst the best surviving examples of land
enclosure that occurred across many areas of north-west Europe during later prehistory. While
in most other regions the land divisions survive only in a fragmentary state due to later
agriculture and development, on Dartmoor they remain preserved as substantial stone and earth
banks that can be traced across thousands of hectares of moorland.1 Accompanying the
boundaries there are a great many monuments, houses and enclosures, established in the late
third and early second millennia BC.
Our understanding of the development of this landscape has been dominated by the
work of Andrew Fleming, who focused his studies on the coaxial field systems, otherwise known
as ‘reaves’.2 He interpreted the reaves as boundaries between land use zones, organized within

1 Dartmoor is an area of upland granite moorland situated in south-west England. The archaeological remains
discussed in this paper are mainly concentrated on the lower slopes of the moor above the improved, enclosed
grassland (250–400 m AOD).
2 While many archaeologists have drawn extensively on Fleming’s research, very few have attempted to offer
alternatives to his explanation of the reaves as a planned and rapid reorganization of the landscape. Price’s
assertion that Bronze Age occupation of the moor was principally for the extraction of tin represents one exception
(Price 1985; Price 1988; cf. Fleming 1987b), as do the issues raised by the RCHME following their survey of

OXFORD JOURNAL OF ARCHAEOLOGY 24(1) 1–21 2005

© Blackwell Publishing Ltd. 2005, 9600 Garsington Road, Oxford OX4 2DQ, UK
and 350 Main Street Malden, MA 02148, USA. 1
 PATTERN WITHOUT A PLAN

a number of valley-based territories, laid out in a radial fashion around the moor (Fleming 1978;
Fleming 1983) (Fig. 1). Fleming suggested that the reave systems were a development, during
one brief time span, of pre-existing territories established during the practice of ‘inter-
commoning’ undertaken in the Neolithic (Fleming 1994). This dramatic formalization of

Figure 1
Schematic map of the main areas of coaxial field systems on Dartmoor (based on Fleming 1983, 221). The locations
of the case studies discussed in this article are indicated: (1) Shaugh Moor and (2) Shovel Down. Shaded areas depict
land above 200 and 400 metres.

Holne Moor (RCHME 1997, 8), and Spratt’s brief discussion of the evidence for prehistoric territories (Spratt
1991, 444–5). Most recently, Richard Bradley (Bradley 2002, 72–81) emphasized the historical context in which
the reaves were situated. Rather than being organized in a rigid, arbitrary system, he noted that existing
monuments and houses had an influence on the locations and alignments chosen for the boundaries. The
communities who built the reaves demonstrated an awareness of and a varying respect towards the visible
landscapes of earlier generations.

OXFORD JOURNAL OF ARCHAEOLOGY

2 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

territorial markers was the result of a ‘Commons Dilemma’ (Fleming 1985; cf. Lane 1998;
McCay and Acheson 1987; Shoup 1990). Whether or not it was a decision made by a top-down
autocratic authority, or it was mutually agreed among a collective of communities, the reason
for the reaves was, according to Fleming, an attempt to deal with the problem of collectively
owned grazing land coming under pressure.
Fleming’s argument is persuasive. The regularity of the coaxial divisions, the scale of
each system of boundaries, their interdependency, and the strict adherence to the dominant axis
seem best explained as the result of an agreed strategy adopted in common by all those
occupying the moor. At an empirical level, however, there are aspects of the evidence that do
not fit or are as yet insufficiently understood to support this interpretation. The regularity of the
field systems, and by implication the degree to which their conception and use were conceived
within a single plan, is dependent upon the scale at which they are studied. When represented
at a small scale, the boundaries appear extremely straight, and unwavering in their adherence
to the axis of the overall system. Yet up close, it is often the variability in construction and
irregularity of their course across the landscape that are an equally if not a more striking feature
(Fig. 2). Variability is also apparent in the form of enclosure. The ‘classic’ pattern of coaxial
boundaries is only present in a few areas. Elsewhere, enclosure strategies appear to have been
less consistent. In striking contrast to the axial boundaries, there are a great number of ‘aggregate
fields’, characterized by small irregular enclosures, frequently situated outside, though on
occasions incorporated within, the main concentrations of reaves. Neither the function and
chronology of this type of enclosure nor its relationship with the coaxial boundaries are
understood. The chronology of the reaves is a further issue that has yet to be systematically
addressed. The only fully published radiocarbon dates come from the excavation of one
boundary, on Shaugh Moor (Balaam et al. 1982, 237–40).3 These do not bracket the construction
and abandonment of the boundary, but instead relate to phases of activity during the use of the
feature – calibrated, they span the period 2140–1260 BC.
Fleming has recognized all of these issues in his work, and has sought to accommodate
them within his arguments. Ultimately, concerns over chronology and functional variability can
only be tested through fieldwork. Nevertheless, it is possible, using our existing knowledge, to
offer a different narrative for the enclosure of the Dartmoor landscape. In this article I wish to
challenge the notion of a planned landscape, and to argue that the final form of the coaxial field
systems emerged as the result of independent communities following varied occupation
strategies but within common social and material conditions. This argument parallels that
already made with reference to ceremonial monuments dating to the later Neolithic (e.g. Barrett
1994, 24), in that large-scale building ‘projects’ do not necessarily require either the presence
of an elite controlling authority, or the structuring influence of a commonly conceived plan. The
mismatch between the coaxial pattern and varied character of individual boundaries does not
represent ‘the plan and its variable local outcomes’ (Fleming 1994, 73); rather it demonstrates
the working of individual agencies within the material conditions of an existing socialized
landscape and the ideational structure of land enclosure and division. There was never a single
‘plan’ to enclose the moor. Instead, a long history of tenure can be traced that made land division
a necessary means of inhabiting the landscape.

3 Radiocarbon dates are published for Fleming’s excavations on Holne Moor (Burleigh et al. 1981, 18–19). But
this is only a summary of the assays with limited contextual information.

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 3
 PATTERN WITHOUT A PLAN

Figure 2
The field boundaries on Horridge Common, south-east Dartmoor, follow a dominant axis yet this emerges from a more
fragmented process of piecemeal construction (based on Butler 1991, 45).

an alternative history of a boundary: saddlesborough reave,

south-west dartmoor
Localized variability and an inadequate chronology are both apparent in an analysis of
the excavated boundaries on Shaugh Moor, south-west Dartmoor (Smith et al. 1981, 209–16).
The investigations were focused on a long stone bank stretching across the upper slopes of the
hillside and forming the terminal for at least six axial boundaries (Fig. 3). Of the two main
phases of construction that were identified, the first varied along the length of the boundary: a
wide shallow ditch; a ditch with an accompanying bank, possibly with a fence on top of the
bank; and a free-standing timber boundary with no ditch or bank. In contrast to this, the phase
two boundary was a continuous stone wall accompanied by the silted-up phase one ditch to the
south-west.
It is possible to offer an alternative sequence (Table 1), and to suggest that the final
alignment of the boundary on Saddlesborough was contingent upon many different actions

OXFORD JOURNAL OF ARCHAEOLOGY

4 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

Figure 3
Plan showing part of the system of coaxial boundaries on Shaugh Moor, south-west Dartmoor. Rescue investigations
in the late 1970s focused on a 600 m length of the ‘terminal’ of a system of axial boundaries between the high ground
on Saddlesborough and a boggy area on the lower slopes to the south-east (based on RCHME 1998, 6; and Smith
et al. 1981, 210).

taking place over an indeterminable length of time. It was variously a ditch, recut and remodelled
on numerous occasions, and a routeway used by animals moving between the high ground in
the north-west perhaps to a water source below the lower slopes. In the first instance, the primary
features that were excavated beneath the line of the boundary varied considerably in their

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 5
 PATTERN WITHOUT A PLAN

structure. Not all the trenches showed evidence for a single, unequivocal ditch (Fig. 4).
Elsewhere, notably towards the south-eastern end of the boundary, the ‘ditch’ was wide and
shallow, and consider-ing the cattle and sheep hoof-prints that were discovered in its base further
upslope, these wide linear depressions may have formed as eroded trackways (Fig. 5) – an
interpretation acknowledged by the excavators. It can be inferred from the presence of stones
set into the sides of these features, including the example running perpendicular to the boundary

Figure 4
Profiles through the ditch of the ‘Saddlesborough reave’ (based on Smith et al. 1981 and unpublished section drawings
in the archive of the Shaugh Moor Project, by permission of Plymouth City Museums and Art Gallery).

OXFORD JOURNAL OF ARCHAEOLOGY

6 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

Figure 5
Simplified plans of trackways along ‘Saddlesborough reave’ (based on Smith et al. 1981 and an unpublished plan in
the archive of the Shaugh Moor Project, by permission of Plymouth City Museums and Art Gallery).

in trench AP, that they were deliberately maintained as paths for stock. Elsewhere, the line of
the boundary was recut, further ditches were added, and timber fences were erected. The bank
and later stone wall are similarly variable. The construction of the wall therefore formalized an
alignment that had developed into a boundary.
The excavations along the ‘Wotter reave’, which ran perpendicular to the ‘terminal’
boundary on Saddlesborough, demonstrated that it too was referenced to an existing landscape.
In its initial phase it was constructed as a short length of bank and ditch, perhaps forming one
side of a field. Access to this field was through a gap, flanked by post-holes, at the junction with
the Saddlesborough boundary. The earthen bank was later replaced by a wall constructed from
stone removed from the boundaries of a small adjacent field system and aligned on a different

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 7
 PATTERN WITHOUT A PLAN

axis to the reaves that continued downslope to the south-west. As on Saddlesborough, the stone
wall significantly changed the character of the boundary, in this case it blocked the accessway
at the junction and slighted a previous field system, but its axis was determined by existing
features: the earth bank and ditch of an earlier field.
While there have only been a few other excavations of reave-type boundaries on
Dartmoor, in every case, without exception, there has been at least some evidence for a
historical sequence. Of two parallel reaves excavated on Holne Moor (Fleming 1988, 71–93;
Fleming 1994, 72–3), one consisted of a later wall situated upon an earlier bank accompanied
by a shallow ditch. Further along the same boundary there were no traces of earlier
features. The other parallel reave was preceded by a line of stake-holes in the area close to its
junction with the terminal reave, yet beyond this there was only slight evidence for a possible
fence. Approximately 150 m further on, the reave consisted of a line of granite boulders.
The terminal reave on Holne Moor had no predecessor, although a line of stake-holes was
discovered on the same alignment as the boundary but situated slightly to the north. At
Gold Park, a short length of a silted-up ditch and a bank was excavated beneath a reave-type
coaxial boundary on a different alignment (Gibson 1992). Recent fieldwork on Shovel
Down, north-east Dartmoor, demonstrated that a long axial reave consisted of at least two
large enclosures that were subsequently linked together forming a single boundary (Brück et al.
2003).
The archaeological excavation of reaves shows them to have long and complex
biographies. Some began as the edges of fields, others as trackways. They shared a common
alignment, but even this was the product of a long and complex process. The axes that the
field systems adopt would appear to have developed piecemeal, and to be contingent upon a
variety of occupation practices. When the variability in construction over the 600 m of the
Saddlesborough reave is projected throughout the many hundreds of kilometres of boundaries
on the moor then the complexity of such a process begins to become apparent.

an alternative history of coaxial enclosure: shovel down and kestor,

north-east dartmoor
The accretive character of the boundaries, their potentially lengthy history, and the co-
presence and then replacement of different forms of field systems can all be emphasized in
reviewing the evidence from Shaugh Moor. Doing so, it must be acknowledged, is at the expense
of discussing the regularity of the axial divisions of the reave system, and the fact that the line
of the Saddlesborough reave can be argued to cross the valley of the river Plym and, turning
north-west, continue for a further 7 km before terminating at a hilltop cairn on Eylesbarrow
(RCHME 1998, 5; cf. Fleming 1978, 117). The challenge remains in explaining how these
two scales of observation might be linked: in what way do the apparently localized and
fragmented processes of land enclosure result in such an archaeologically distinctive and
pervasive pattern?
The following analysis of the coaxial field systems on Shovel Down and Kestor offers
an attempt to bridge these scales. Situated on the north-east side of Dartmoor, the fields are
distributed across an area of high ground (350–450 m AOD) framed by the valleys of the North
and South Teign rivers and with up to seven topographically distinctive hilltops, most of which
are marked by a tor of outcropping granite (Fig. 6) (see also Butler 1997, 255; Fleming 1983,
235–6).

OXFORD JOURNAL OF ARCHAEOLOGY

8 © Blackwell Publishing Ltd. 2005
© Blackwell Publishing Ltd. 2005
OXFORD JOURNAL OF ARCHAEOLOGY
ROBERT JOHNSTON

9
Figure 6
The field systems, settlements and monuments on Shovel Down and Kestor, north-east Dartmoor (based on surveys by English Heritage, copyright reserved).
 PATTERN WITHOUT A PLAN

Monuments
Our understanding of the landscape prior to the construction of the buildings and
boundaries is rather limited, as it is elsewhere on Dartmoor. Nonetheless, it can be argued
that by the beginning of the second millennium BC there would already have been an intri-
cate patchwork of places and paths linking various locales associated with monuments and
agricultural activity (Evans 1999, 26–34). There is good evidence for the management of
woodland fringes on the high moor during the Mesolithic, with the coincidence of microscopic
charcoal and a gradual reduction in arboreal pollen between 7700 and 6300 BP (Caseldine and
Hatton 1994, 40). Undoubtedly clearance continued, albeit on a small scale, until the middle of
the second millennium BC by which time some have argued that the lower slopes of the moor
were substantially cleared of woodland.4
This picture of an earlier landscape structured by a focus on particular places is
supported by the archaeological evidence from Shovel Down. Scatters of flint debitage and tools
were recovered from three locations within 1.5 km of the area, and in each case these seem to
represent several millennia of activity. The reasons for visiting these places were no doubt
varied, but it is unlikely to be a coincidence that they are close to groups of ceremonial
monuments. The lithic scatter at Batworthy, for instance, lay only a short way from the five
stone rows built on the eastern slope of the high ground on Shovel Down, while the most recently
discovered lithics, and a possible Late Mesolithic/Early Neolithic pit, are immediately adjacent
to the stone rows (Brück et al. 2003). Although they all follow a roughly north–south alignment,
these monuments show subtle but significant variations in their form and location. This can be
taken to suggest that just as the lithic scatters appear to represent phases of occupation and
abandonment the rows result from a history of building projects rather than being part of a
single, planned monument.
The stone rows on Shovel Down not only served as a focal place, but also as an axis
around which future occupation was structured. They occupied an unenclosed corridor, which
was both the alignment for what seems to be one of the earliest reaves, and also a boundary
between the lower-lying pasture enclosed by parallel field systems and the higher moor where
both parallel and irregular systems of fields were built. The concentrations of monuments at
places such as Shovel Down have previously been interpreted as central places within large
community territories. But it is perhaps their influence upon their local landscape that is most
relevant to understanding how the systems of field boundaries came about.

Buildings
The pattern in the distribution of settlements either side of the stone rows is one that is
likely to have emerged when the monuments were still ‘in use’. Although none of the houses
within the fields on Shovel Down and Kestor have been excavated under modern conditions,5
early second millennium BC radiocarbon dates were associated with the earliest phases of

4 This impression must be tempered somewhat: the main evidence for major woodland clearance comes directly
from archaeological sites on the lower slopes (e.g. Balaam et al. 1982; Maguire et al. 1983), yet the pollen cores
from the high moor are dominated (>50 per cent) by arboreal pollen (Caseldine and Hatton 1994, 43).
5 A roundhouse is currently under excavation at Teigncombe, on the eastern side of the Kestor field system (Gerrard
2001).

OXFORD JOURNAL OF ARCHAEOLOGY

10 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

settlement at Shaugh Moor on the south-west side of Dartmoor (Wainwright and Smith 1980).
Elsewhere, it is possible to date a small number of the more than 300 buildings excavated by
local antiquarians and archaeologists during the late nineteenth and early twentieth centuries by
their association with a style of pottery known as Trevisker Ware, which is considered to have
been in use from c.1700 BC (Needham 1996; Parker Pearson 1990; cf. Radford 1952). A more
unequivocal example of the relative date at which the buildings were constructed is that they
are frequently in a primary relationship to and therefore earlier than the boundaries, either at
the centre of a radial arrangement of small plots or incorporated into the course of one of the
coaxial boundaries. This is not meant as an argument for all buildings being of an earlier period
than the field systems. Rather that in some cases buildings were constructed before the
boundaries, though it is not possible to tell whether over a matter of centuries or as part of the
same period of occupation.
The earliest stone buildings added a further important structural element to an already
complex landscape. It is clear from the excavated examples that a high proportion were occupied
and probably served a variety of domestic functions. They have substantial stone walls, an
entrance that is occasionally elaborated with a porch, hearths, pits and a range of material
culture. These structures served to contain and categorize domestic activities, and were therefore
important in making distinctions between activities and identities within the social group, and
as a consequence access to resources. Buildings might therefore have served as places through
which rights of access and use over resources could be legitimated. Crucially, the buildings also
defined the place and fixed the memories of such occupations within the landscape. They
frequently acted as the nodes of networks of boundaries, or they were incorporated into the line
of a boundary. Assuming that all the buildings and fields were not constructed and occupied
contemporaneously, then abandoned dwellings remained important in structuring people’s
perceptions of their present landscape. The fact that many of the buildings do not appear to have
been robbed of their stone when the boundaries were constructed also lends support to this
argument (Bradley 2002, 79).
A more explicit expression of the ways in which buildings materialized the attachment
between people and place can be found in the related depositional and architectural histories
of houses and cairns. For instance, buildings and cists share a common dominant alignment –
to the south-east (Bradley 2002, 77). There are also examples where it has been argued
that houses were restructured to look like burial monuments (Butler 1997, 137–8). More
frequently we find close similarities between the character of the deposits found within houses
and cairns, particularly ring cairns (Johnston 2001, 150–60). For instance, excavations of
buildings, ring cairns and to a lesser extent burial monuments frequently uncover pits filled
with burnt soil, stone and charcoal (Fig. 7). In the context of the ceremonial monuments
these pits have been interpreted as ritual deposits, ‘eccentric cremations’ and cenotaphic
offerings. In contrast, those discovered in buildings are commonly referred to as ‘cooking holes’.
Based on the published excavation evidence, there is frequently little to distinguish between the
pits found in buildings with those found beneath cairns. They are of similar size; they are
located centrally or close to the inner perimeter of the structure; they are occasionally lined with
stones and/or sealed by a single stone; they contain charcoal; and they are accompanied by
sherds of pottery or the remains of a complete vessel. Examples of such pits are known from
the excavated ring cairns at Shaugh Moor (Wainwright et al. 1979), Metheral (Worth 1937),
Farway Hill (Pollard 1971) and Deadman’s Bottom (Worth 1900); 14 of the barrows listed
by Grinsell, including Broadun and The Croft (Grinsell 1978); and 15 settlements, the best

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 11
 PATTERN WITHOUT A PLAN

Figure 7
Plans showing the location of pits filled with charcoal within houses and ring cairns (based on plans in Baring-Gould
1896; DEC 1894; Fox 1957; Wainwright et al. 1979).

OXFORD JOURNAL OF ARCHAEOLOGY

12 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

known of these being Grimspound (DEC 1894), Watern Oke (Anderson 1906) and Dean Moor
(Fox 1957).
An increasing number of interpretations have drawn parallels between the architectural
and depositional histories of monuments and domestic buildings (e.g. Hodder 1994; Bradley
1998). What seems most pertinent here is the recurring link that is made between burial mounds,
houses and tenure. Neolithic long barrows acted as houses for a community’s ancestors, and as
such embodied a direct link between people and place through the generations interred in the
tomb. Early Bronze Age round barrows were built upon hilltops and ridgelines where they
delimited the boundaries of territories. The linear character of some of the barrow cemeteries
formed a visible expression of the lines of descent and inheritance through which access to
resources and social status was maintained. The construction of round houses redefined social
organization such that smaller kin groups appropriated and held tenure over specific places on
a long-term basis. These ideas are all closely linked in the way that they associate the
construction of lasting structures, which were directly associated with individual and group
biographies, with claims over land and resources. Undertaking rituals at cairns and living in
houses were both key contexts where agents created, sustained and challenged senses of identity
and place.
If, as both Fleming and Bradley have argued, the monuments on Dartmoor formed the
earliest physical markers of a community’s attachment to an area of the landscape, then the
houses represented a continuation of this process. Crucially, with the construction of buildings,
the scale at which tenure operated was now visibly focused around the household unit,
howsoever it was constituted. It brought about a change in how people identified with particular
places, and where the memories of occupation were made permanent in the landscape. It was
during the building, abandonment and revisiting of such structures that the link between
occupation, memory and tenure developed.

Boundaries
The construction of boundaries appears at first to be an altogether different approach
to identity, place, and tenure. The demarcation of distinct parcels of land related as much to how
land was used and access to it controlled, as it did to the formation and maintenance of social
groupings. Yet on Shovel Down and Kestor, the boundaries were built within the networks of
access and rights to resources that had developed over centuries, and which were being mediated
in part at least through the occupancy of buildings.
This is most evident amongst networks of aggregate enclosures which, unlike the
reaves, form irregular patterns of boundaries. In every case the location of the buildings within
these settlement groups structured the pattern of the boundaries built around them (e.g. Fig. 8a).
This situation is less common amongst the coaxial boundaries, where in many cases the
buildings were located within the fields rather than along the line of the boundary. Nevertheless,
there are at least five examples of buildings that appear to have preceded the construction of a
coaxial boundary. For instance, the north-west–south-east aligned boundary that seems to have
formed the initial limits of the coaxial fields on Kestor, before further boundaries were added
to the west of the tor, incorporated a small building into its construction (Fig. 8d, no. 1). The
boundary on the south-east side was kinked slightly so as to join with the wall of the structure.
A further two structures, situated just to the south of Kestor Rock, were both incorporated into
a coaxial boundary extending to the south-west from the north-west–south-east boundary

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 13
 PATTERN WITHOUT A PLAN

Figure 8
Examples of the relationships between boundaries and buildings on Shovel Down and Kestor: (a and b) within
settlement areas; (c and d) buildings incorporated into the line of an axial boundary; (e) a building joined to an axial
boundary with a short length of walling (based on surveys by English Heritage, copyright reserved).

associated with the preceding example (Fig. 8c, nos. 2 and 3). The boundary was built around
one of the buildings, and abutting either side of the other. The small enclosure into which one
of the buildings faces seems to have been built after the construction of the boundary. The
buildings pre-dated the stone boundary but were built on the same alignment that the boundary
would later take.
A physical link was also established when a building was joined to a coaxial boundary
with a short length of walling, such as on Kestor, north-east of the Round Pound; on the eastern

OXFORD JOURNAL OF ARCHAEOLOGY

14 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

table 1
An alternative sequence for the boundary on Saddlesborough

Excavator’s phasing Alternative sequence

Phase 1a Single ditch, ditch and bank, and timber Pre-boundary Trackway, field ditches, possible lynchet
boundary
Phase 1b Some recutting and enlarged bank Phase 1 Elaboration and maintenance of some
lengths of ditch and bank, possibly
also construction of timber boundary
Phase 2 Wall Phase 2 Build-up of banks and some lengths
of walling

side of the north–south boundary forming the ‘spine’ of the coaxial fields on Shovel Down (Fig.
8e); and three instances among the plots and buildings on the northern side of the stream to the
north of Stonetor Hill.6 These short walls may have been part of an enclosure, the rest of the
boundary being constructed from other materials, such as timber. Nevertheless, it is still
interesting that the boundary between the roundhouse and the fields was constructed in stone.
In most cases the gap between the structures is quite small, yet it was clearly important that the
link was substantial and prominent.
The links between buildings and boundaries demonstrate that existing structures had
an important influence upon where people chose to locate field walls, and that there was an
intentional effort to physically emphasize the links between houses and boundaries by joining
them together with short lengths of stone wall or bank.

Discussion: the enclosure of Shovel Down

The synchrony that seems evident from the neat patterns of boundaries, settlements and
monuments on Shovel Down and Kestor is wholly deceptive. Not only can the horizontal
stratigraphy of the boundaries be unravelled, but it is also possible to identify a much broader
sense of process in this landscape as successive actions were structured by, and themselves
continued to structure, the material resources and networks of significance for future generations
of inhabitants.
Inhabitation of the area before the construction of any of the standing archaeology
would have defined places and routeways. Areas had already been cleared of trees for grazing
stock and perhaps for small patches of cultivation. Before and during such deliberate activity,
the movements of animals within their own territories would have created ‘small-scale diversity’
(Evans 1999, 27). The influence of previous dwellings is evident in the way that alignments
between tors and along watercourses and valleys are maintained by the later boundaries. Such
natural features were foci in earlier landscapes (e.g. Tilley 1998). The names and narratives
associated with these places were no doubt reworked over the generations, but as structures they
remained influential.

6 Examples of this relationship were excavated on Wotter Common during the Shaugh Moor Project (Smith et al.
1981, 226–7) and on Holne Moor, site B (Fleming 1988).

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 15
 PATTERN WITHOUT A PLAN

The coaxial boundaries broadly respected the topographic form of the landscape and
the distribution of ceremonial monuments. The north–south axial boundaries were constructed
perpendicular to the North Teign river that flows east to west with a series of 90 degree bends,
framing the moorland of Shovel Down and Kestor. The hilltop tors also served as markers during
the laying out of the fields. Though fragmentary, one of the coaxial boundaries running
perpendicular to the alignment between Kestor Rock and Frenchbeer Rock links together Middle
Tor and Thornworthy Tor, while the principal north-east–south-west boundary on Shovel Down
curves from the main boundary to the west of the stone rows, over the high ground and across
the valley towards Stone Tor. The rivers, tors and relief are of course all interrelated, as they
result from the geomorphological processes that shaped the topography of the moor. But, the
humanly constructed features relate closely to the form of the landscape, even if they may on
first appearances seem to impose a different, more rigid, order.
The stone rows respected existing alignments. Though they seem divisive on a modern
survey plan, it is unlikely that they created boundaries that did not already exist in people’s
experiences of such places. It seems that stone-built roundhouses were first constructed only to
the west of the stone rows. Such settlement areas generated and reproduced structures within
this landscape. They were built in locales that might have already been occupied, but through
their continued occupation they contributed to the creation of new places. The ways in which
these areas were inhabited is as yet unknown even though the character and intensity of
occupation practices and land use must be recognized as fundamental.
Practices were not restricted to spaces formalized by stone walls and boundaries. There
were resources distributed both on and off the moor that structured the movement and tempo
of people’s daily lives. Crucial to the local concern of this study, there were ‘in-between places’
(Rocheleau and Edmunds 1997) on the moor itself where resources such as fuel, wild foods and
medicinal plants might be gathered: small gardens adjacent to houses, the plants and shrubs
growing along boundaries, or the scrubland found along stream courses and on steep rocky
slopes. Divisions along the lines of kin, gender and social responsibility contributed to the ways
in which space was used and organized. Such divisions were reproduced and maintained in the
manner that people lived their daily lives, and such practices in turn structured and were
structured within time-space settings with limits and material boundaries that were defined both
arbitrarily and according to the availability of resources. Added together with the more
formalized expressions of place embodied in the houses and non-axial field boundaries, the
alignments of the stone rows, the tors and the river, the moorland on Kestor and Shovel Down
became a temporally and spatially complex network of inhabitations.
The construction of coaxial boundaries within this inhabited environment was not a
wholesale reorganization, regardless of whether or not the reaves were constructed over a short
or a long time span. They maintained the dominant alignments that had been present since people
first came to the area: notably the tors and the river. They were also built to incorporate pre-
existing features, particularly roundhouses, other buildings and small enclosures. The buildings
were material linkages between people, through their kin associations, and the land with its
ancestral associations. They were, therefore, critical places for negotiating tenure. The coaxial
boundaries were constructed within these existing human-land relations; tenure was not
rewritten. Nonetheless it would have changed; the coaxial boundaries did in themselves create
new material conditions. They had the potential to formalize boundaries that had once been open
to negotiation, or perhaps had never before been recognized discursively.

OXFORD JOURNAL OF ARCHAEOLOGY

16 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

concluding discussion

The argument presented thus far is that the layout of the reaves was not consciously
structured according to an agreed plan adopted by all the communities inhabiting the moor, but
rather that it emerged within particular traditions of tenure and land use. These ‘traditions’, as
we recognize them now, were not necessarily consciously articulated or maintained by people
at the time. Rather the principles that structured how people inhabited and organized their
landscapes existed through those very same actions of inhabitation and organization. This
recursive relationship was dynamic because the individuals and communities using the moor
were agents situated in a history, both of their own life and that of the social worlds that they
inhabited.
Such an interpretation does not preclude the kinds of communal decision-making that
Fleming advocates. But nor is a discursive plan required in order to explain a distinctive pattern
of homogeneous material and ideational conditions. The structures of everyday life, the form
and history of the environment, the geographies of people’s lives, their occupation strategies,
and the networks of social relations, enabled particular forms of land organization and precluded
others. In other words, the conditions and practices that constituted social life during the third
and second millennia BC on Dartmoor enabled land to be divided and enclosed within coaxial
field systems. The regular pattern of the boundaries emerged in a particular though diverse set
of social and material structures. Given the fact that we do not yet know over what timescale
this process took place, it is possible that the initial layout of a few of the fields was of critical
importance in structuring subsequent land use, allotment and enclosure (Fleming 1987a).
It diverges from Fleming’s approach in that it does not consider the coaxial boundaries
as a unitary phenomenon, and therefore the layout of the fields is not necessarily part of a
synchronous system. It is argued that land enclosure was only possible because the structures
were already in place to enable it. Rather than there being a plan or a conscious strategy to
reorganize and enclose the moor, land division emerged as the way that communities conceived
of the relationships between the land and with each other. Furthermore, once coaxial boundaries
were built they became critical in transforming how the future landscape was perceived and
inhabited.
This also relates to scales of community and landscape. The ‘commons dilemma’ that
Fleming believed emerged during the second millennium BC requires tenure to operate within
large groupings. Power over the land was invested in the community – a human collective
occupying valley-based ‘large terrains’. The studies of boundaries and landscape that were
presented in this article, taken together, suggest that tenure was articulated at a local level,
through the relationship between occupancy and ancestral ties to the land.
This relationship was realized as boundaries were built between places and along paths.
On Shovel Down, the ancient and contemporary histories associated with the buildings were
combined with the significance of a house, as a symbol of identity, occupancy and tenure, to
structure where and how physical boundaries could be constructed across the land. Such
boundaries were materializing virtual divisions which emerged through patterns of dwelling and
long histories of tenure that had already left structures in the landscape.
Pathways and the spatial and temporal limits of everyday activities were equally
powerful structures. The routeways that linked places were a strong embodiment of occupancy.
They became more prominent through use, and their identity changed as the localities that they

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 17
 PATTERN WITHOUT A PLAN

joined and the areas through which they passed acquired or lost significance. These trackways
could themselves be virtual boundaries: either delimiting one space from another along their
length, or laterally, as a byway between locales. Ditches, stone banks and fencelines formalized
these trackways, as at Shaugh Moor.
Occupations not only defined places and paths, they also structured areas of landscape.
As woodland and scrub were cleared or as pasture was turned to cultivation, practices generated
their own time-space settings. The limits of these areas were not necessarily arbitrary, but nor
were they consciously monitored and sustained. Where such rights of access or statements of
control did become part of a physical discourse, they are recognizable in the selective recutting
of a ditch, the blocking of accessways or in the construction of boundaries along the edges of
zones of land use.
In all these ways, inhabitation of the land conditioned the location of physical
boundaries. Fundamentally, tenure was rooted in the occupancy of the living. That occupancy
was embodied in the places where people lived, the paths they made and followed, and the time-
space limits to daily practice. The legitimation for occupation lay with those that had inhabited
the landscape previously, embodied in the barrows and cairns and articulated through deposits
left in houses and monuments. The concern with occupancy enabled the knowledgeability that
would have been necessary for a firmer and more localized control over land.
The networks of Bronze Age coaxial boundaries that we can see today dividing up the
land on Dartmoor were built within the social and material conditions outlined above.
The resulting pattern was never conceived as a plan, even at a broad scale. The construction of
the stone banks was contingent upon many different conditions and practices. As they were built,
and the, in places, unwavering regularity of the pattern was formalized, so further boundaries
repeated, mimicked and fitted in. The construction of boundaries was reflexive, as each further
action looked back upon those that had gone before. There is certainly a place in such a process
for what Bourdieu has termed ‘unconscious coordination’ (Bourdieu 1990, 58 ff.), though on a
grand scale. Therefore, rather than identifying a dominant discourse through which the reaves
were planned, it is better to think of the interaction between individuals, families and
communities as a multitude of elements that came about through a variety of different strategies.
To leave it at that would be to suggest that there was nothing consciously intended by
the reaves. The tone of the argument could be taken to mean they were an accident, mere chance
brought about by the concurrence of certain social and material conditions. Or, on the other
hand, by interpreting the changing human-land relations on Dartmoor as a process, I am
suggesting it was somehow inevitable or predetermined. Neither of these is the case. For one
thing, the fundamental reason that land could be divided in this way lay in the fact that tenure
was sustained and negotiated through occupancy. This made the reaves possible. Yet such agency
was worked out in peculiar, contingent and localized conditions. For all the varied reasons that
seemed necessary at the time, there emerged local traditions of boundary building founded on
the close ties between occupancy and land.

Acknowledgements
The ideas developed in this article formed part of my doctoral thesis, and I must first
acknowledge the support and advice of my supervisor, Jan Harding, the examiners, Professor Richard
Bradley and Professor Geoff Bailey, and the funding body, the AHRB. In addition, I am grateful to Fiona
Pitt at Plymouth City Museum for helping with access to the archive of the Shaugh Moor Project and the

OXFORD JOURNAL OF ARCHAEOLOGY

18 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

lithics from Batworthy, Louise Barker for her invaluable assistance during several field visits to Dartmoor,
and Mark Edmonds, Rachel Pope, John Roberts and Helen Wickstead for offering comments on drafts of
the text. An immeasurable debt is owed to Joanna Brück and Helen Wickstead for the many discussions
that we had together while undertaking fieldwork on the moor. This article was written whilst on a period
of sabbatical leave supported by a Leverhulme Trust Research Fellowship.

Department of Archaeology
University of Sheffield
Northgate House
West Street
Sheffield
S1 4ET

references
anderson, i.k. 1906: Eleventh report of the Dartmoor Exploration Committee. Report and Transactions
of the Devonshire Association 38, 102–13.
balaam, n.d., smith, k. and wainwright, g.j. 1982: The Shaugh Moor Project: fourth report –
environment, context and conclusion. Proceedings of the Prehistoric Society 48, 203–78.
baring-gould, s. 1896: Third report of the Dartmoor Exploration Committee. Report and Transactions
of the Devonshire Association 28, 174–99.
barrett, j.c. 1994: Fragments from Antiquity: an archaeology of social life in Britain, 2900–1200 BC
(Oxford).
bourdieu, p. 1990: The Logic of Practice (California).
bradley, r. 1998: The Significance of Monuments: on the shaping of human experience in Neolithic and
Bronze Age Europe (London).
bradley, r. 2002: The Past in Prehistoric Societies (London).
brück, j., johnston, r. and wickstead, h. 2003: Excavations of Bronze Age field systems on Shovel
Down, Dartmoor 2003. Past 45, 10–12.
burleigh, r., matthews, k., ambers, j. and kinnes, i. 1981: British Museum natural radiocarbon
measurements XII. Radiocarbon 23(1), 14–23.
butler, j. 1991: Dartmoor Atlas of Antiquities: Volume 1, the east (Exeter).
butler, j. 1997: Dartmoor Atlas of Antiquities: Volume 5, the second millennium BC (Exeter).
caseldine, c.j. and hatton, j.m. 1994: Into the mists? Thoughts on the prehistoric and historic
environment. In Griffiths, D.M. (ed.), The Archaeology of Dartmoor: perspectives from the 1990s (Exeter,
Devon Archaeological Society), 35–47.
DEC 1894: The exploration of Grimspound: first report of the Dartmoor Exploration Committee. Report
and Transactions of the Devonshire Association 26, 101–121.
evans, j.g. 1999: Land and Archaeology: histories of human environment in the British Isles (Stroud).
fleming, a. 1978: The prehistoric landscape of Dartmoor, part 1, south Dartmoor. Proceedings of the
Prehistoric Society 44, 97–124.
fleming, a. 1983: The prehistoric landscape of Dartmoor, part 2: north and east Dartmoor. Proceedings
of the Prehistoric Society 49, 195–241.
fleming, a. 1985: Land tenure, productivity, and field systems. In Barker, G. and Gamble, C. (eds.),
Beyond Domestication in Prehistoric Europe (London), 129–46.
fleming, a. 1987a: Coaxial field systems: some questions of time and place. Antiquity 61, 188–202.

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 19
 PATTERN WITHOUT A PLAN

fleming, a. 1987b: Prehistoric tin extraction on Dartmoor: a cautionary note. Report and Transactions
of the Devonshire Association 119, 117–22.
fleming, a. 1988: The Dartmoor Reaves: investigating prehistoric land divisions (London).
fleming, a. 1994: The reaves reviewed. In Griffiths, F.M. (ed.), The Archaeology of Dartmoor:
perspectives from the 1990s (Exeter, Devon Archaeological Society), 63–74.
fox, a. 1957: Excavations on Dean Moor, in the Avon Valley, 1954–1956: the Late Bronze Age settlement.
Report and Transactions of the Devonshire Association 89, 18–77.
gerrard, s. 2001: The Dartmoor archaeology and bracken project. Past 37, 8–10.
gibson, a. 1992: The excavation of an Iron Age settlement at Gold Park, Dartmoor. Proceedings of the
Devon Archaeological Society 50, 19–46.
grinsell, l.v. 1978: Dartmoor barrows. Proceedings of the Devon Archaeological Society 36, 85–180.
hodder, i. 1994: Architecture and meaning: the example of Neolithic houses and tombs. In Parker
Pearson, M. and Richards, C. (eds.), Architecture and Order: approaches to social space (London), 73–86.
johnston, r. 2001: Land and Society: the Bronze Age cairnfields and field systems of Britain (University
of Newcastle, Unpublished Ph.D. Thesis).
lane, c.r. 1998: Introduction: overview of the pastoral problematic. In Lane, C.R. (ed.), Custodians of
the Commons: pastoral land tenure in East and West Africa (London), 1–25.
mccay, b.j. and acheson, j.m. (eds.) 1987: The Question of the Commons: the culture and ecology of
communal resouces (Tucson).
maguire, d.n., ralph, n. and fleming, a. 1983: Early land use on Dartmoor – palaeobotanical and
pedological investigations on Holne Moor. In Jones, M. (ed.), Integrating the Subsistence Economy
(Oxford, BAR Int. Ser. S181), 57–105.
needham, s. 1996: Chronology and periodisation in the British Bronze Age. Acta Archaeologica 67,
121–40.
parker pearson, m. 1990: The production and distribution of Bronze Age pottery in south west Britain.
Cornish Archaeology 29, 5–32.
pollard, s.h.m. 1971: Seven prehistoric sites near Honiton, Devon part II. Proceedings of the Devon
Archaeological Society 29, 162–80.
price, d.g. 1985: Changing perceptions of prehistoric tinning on Dartmoor. Report and Transactions of
the Devonshire Association 117, 129–38.
price, d.g. 1988: Prehistoric tin extraction on Dartmoor: a reply to Andrew Fleming. Report and
Transactions of the Devonshire Association 120, 91–5.
radford, c.a.r. 1952: Prehistoric settlements on Dartmoor and the Cornish Moors. Proceedings of the
Prehistoric Society 18, 55–84.
rchme 1997: Holne Moor, Dartmoor, Devonshire: a landscape survey by the Royal Commission on the
Historical Monuments of England (Exeter, unpublished report).
rchme 1998: Shaugh Moor, Devon: an archaeological survey by the Royal Commission on the Historical
Monuments of England (Exeter, unpublished report).
rocheleau, d. and edmunds, d. 1997: Women, men and trees: gender, power and property in forest and
agrarian landscapes. World Development 25(8), 1351–71.
shoup, j.a. 1990: Hima: a traditional Bedouin land-use system in contemporary Syria and Jordon.
(Washington University, unpublished Ph.D. Thesis).
smith, k., coppen, j., wainwright, g.j. and beckett, s. 1981: The Shaugh Moor Project: third report –
settlement and environmental investigations. Proceedings of the Prehistoric Society 47, 205–73.
spratt, d.a. 1991: Recent British research on prehistoric territorial boundaries. Journal of World
Prehistory 5(4), 439–80.

OXFORD JOURNAL OF ARCHAEOLOGY

20 © Blackwell Publishing Ltd. 2005
 ROBERT JOHNSTON

tilley, c. 1998: Rocks as resources: landscapes and power. Cornish Archaeology 34, 5–57.
wainwright, g.j., fleming, a. and smith, k. 1979: The Shaugh Moor Project: first report. Proceedings
of the Prehistoric Society 45, 1–33.
wainwright, g.j. and smith, k. 1980: The Shaugh Moor Project: second report. Proceedings of the
Prehistoric Society 46, 65–122.
worth, r.h. 1900: 19th report of the Barrow Committee. Report and Transactions of the Devonshire
Association 32, 46–54.
worth, r.h. 1937: 13th report of the Dartmoor Exploration Committee. Report and Transactions of the
Devonshire Association 69, 147–50.

OXFORD JOURNAL OF ARCHAEOLOGY

© Blackwell Publishing Ltd. 2005 21