Theodore Reeves | University of Birmingham
Hillforts are some of the most striking remnants of prehistory in Britain. First emerging in the later Bronze Age (shortly after 900 BCE) but predominantly seen as a feature of the Iron Age (800 BCE until the Roman conquest), these monumental earthwork enclosures are typically located on hills or elevated positions and dominate the surrounding landscape. There are, however, a group of sites which are instead located in low-lying, and sometimes wetland, positions and which do not conform as readily to this criteria. Over the last few decades, academics have criticised the use of the term “hillfort” for its focus on militaristic function, but what about the “hill-” part? If neither half of the term is relevant, should we use it at all? Further discussion is undoubtedly required to unravel the design and function of these sites, with careful consideration as to how we use them to inform and reshape our classifications into something more meaningful.
Several alternatives have been suggested in the past. “Defended enclosures”? Well, this surely has the same problematic focus on warfare that “hillforts” does, stemming from Roman descriptions of violent barbarians aimed at justifying their own imperial expansion. Bringing civilisation to the natives – you can imagine why early archaeologists accepted these ideas without much critical discourse. What about “enclosed places”? Much more neutral... but is not a sheep pen or courtyard garden an enclosed place? They are hardly comparable with the 19-hectare site of Maiden Castle, with its seven-metre deep ditches. And so, “hillfort” remains the preferred term, out of convenience as much as anything.
But what of sites that are not on hills? Hills form a key aspect of hillfort identity. They exploit the natural topography to exaggerate the anthropogenic earthworks. There does exist another type: marshforts. As the name suggests, these sites can be found situated in wetland landscapes. The best documented of these is Sutton Common, South Yorkshire, whose excavations were first published in 2007 and is the first notable application of the term. The site comprises two enclosures, the larger of which is set on a gravel island surrounded by a peat wetland. The wetland may serve a similar purpose to the hill, using geology in place of topography to create a natural obstacle which hinders and controls access. The only way in is via a causeway from the small enclosure.
Another, perhaps more recognisable example can be found in Stonea Camp, Cambridgeshire. There, the peat fen is recorded against the southwest side of the site. With entrances through the earthworks identified on the north sides of the enclosure, this would likely be the original route of access. The earthworks seen today are a reconstruction built in 1991 to the condition they would have been thirty years prior before it was ploughed. They represent the site in its final state, with it having grown progressively larger over multiple phases, originating as a single concentric bank and ditch. What is notable, however, is that these later additions are all focused on the northern sides. The result is that the enclosure appears to utilise the wetland as a natural obstacle, similar to how promontory forts utilise the topography around them, focusing the construction of earthworks towards the direction of access.
But what of the sites not on hills and not by marshes? Absent either a topographical or geological advantage, what is there to distinguish these sites from causewayed enclosures or other enclosure sites? Surely, if we stick to tradition, these sites should be called “forts”? But although “hillfort” has arguably lost its focus on militaristic function, coming to mean more than the sum of its constituent parts, the term “fort” is unavoidably still entrenched in this association, one which does not help establish a common understanding of these sites.
Cherbury Camp, located near the village of Charney Bassett in Oxfordshire, provides one such example of this. This site is situated in a flat landscape: not then a hillfort. While lots of past scholarship has commented on the marshy ground around the site, recent coring of the surrounding landscape has revealed a shallow plough soil straight onto the sandy geology. There is no evidence of wetland beyond the peripheries of the nearby stream to the west of the site. Not then a marshfort either. This raises an important question; the site does not appear to offer any topographical or geological advantage, so why was Cherbury Camp built here?
The multivallate enclosure encompasses an area of 4.6 hectares. The earthworks comprise three sets of concentric banks and ditches, forming an imposing spectacle viewed close-up or from the air. Viewed from the ground at any distance, though, the earthworks blend with the surroundings, or at the very least, their full extent and complexity would have been hidden behind the first bank. Archaeologists discussing more traditional “hilltop” hillforts have suggested that not only would layers of concentric earthworks have provided a practical defensive function, but seen from afar, they would act as a deterrent, demonstrating the power of the people that built them. The natural and, importantly, human resources needed to build such monumental earthworks would have been considerable, and the ability to divert them from everyday tasks such as agriculture demonstrates substantial wealth and power, whether by an individual or community. If Cherbury Camp was intended as a display of power though, it was not a very visible one. John Bradford, who excavated the site in 1939, suggested this may have been a deliberate design intention; to hide the enclosure for defensive purposes. If this was the case though, it would be exceptionally unusual in comparison to other “hillfort”-type sites. In the later Iron Age, there is the emergence of oppida. These are often low-lying earthwork enclosures, substantially larger than hillforts, often including multiple farmsteads within the earthworks. The earthworks there have been suggested to demark the boundaries of the settlements, but also to act as a limited means to restrict access and in particular control access to wheeled transport such as chariots. They do not, however, tend to have multiple sets of banks and ditches as closely together as we find at Cherbury Camp and are typically much larger. The morphology of Cherbury Camp therefore seems far more consistent with hillforts, though its location does not.
A geophysical survey of the interior in 2007 revealed the potential remains of at least a dozen roundhouses and several clusters of pits. This suggests the site may have been occupied by an Iron Age community. “Hillfort”-type enclosures were not just used by the people who built roundhouses in them. Many likely served much wider rural communities that farmed the surrounding landscape. The pits identified in the interior may provide evidence for people living outside the enclosure bringing their grain to Cherbury Camp to be stored safely, a common practice in the Iron Age. It remains to be seen, therefore, quite how invisible a site of this size and acting as a hub for a wider community could have remained. Excavations have uncovered the remains of a wooden structure in the eastern entranceway, controlling access to the site. These excavations were limited, though, and little is still known about the full extent of how the enclosure was used.
What, then, is Cherbury? An enclosure – yes; defensive? - yes, but we cannot say intentionally so; a settlement – probably. Do we then classify it as “an enclosed Iron Age settlement”? That would make sense. But does it capture the imagination? It does not invoke the same imagery that a “hillfort” might of Maiden Castle, Danebury or British Camp – monumental earthworks which dominate their surrounding landscapes. There we highlight a key issue of classifying archaeological sites. While academics may prefer carefully thought-out, specific terminology, archaeology has a much wider interest. “Hillforts” is the go-to term for sites like this used by the likes of English Heritage and the National Trust to describe them to visitors. How, then, should we hope to disseminate our research if the terminology we use does not capture the imagination and create an instant impression the way “hillfort” does?
The problems of classification and terminology have been long-standing, and it is unlikely that we shall come to a clear solution that pleases everyone. What we can establish, though, is a greater understanding of the complexity and diversity of sites that we currently box away as hillforts with little nuance to the exact nature of these sites and their relationships with their respective landscapes. The point that this discussion is being held at all proves itself. Cherbury Camp was for many years called a hillfort; upon the establishment of the term marshfort, it was quickly lumped too into that category. While we can say there are undoubtedly similarities in the architectural style of these sites, to suggest that Cherbury Camp is the same as Sutton Common or Stonea Camp is to neglect some of the most distinctive characteristics of these sites and ultimately to murky the waters of our understanding. Whether we can agree on the preferred terminology or not should not deter us from rectifying the way in which they are grouped, to begin with. While more work is undoubtedly required to address these issues, acknowledging the nuances of these sites is surely the first step.
Further Reading:
Gary Lock and Ian Ralston, Atlas of the Hillforts of Britain and Ireland (Edinburgh: Edinburgh University Press, 2022).
Ian Brown, Beacons in the Landscape: The Hillforts of England, Wales and the Isle of Man, 2nd edn (Oxford: Windgather Press, 2021).
Robert Van de Noort, Henry Chapman, and John Collis, Sutton Common: The Excavation of an Iron Age 'marsh-fort' (York: Council for British Archaeology, 2007).
Theodore Reeves is a PhD student in Archaeology at the University of Birmingham, funded by AHRC Midlands4Cities. His thesis explores British Iron Age marsh-forts, using a combination of GIS-based analysis and fieldwork to examine the relationship between site and landscape and establish a more nuanced classificatory framework.
Twitter: @TheoDigs