Early medieval houses in Ireland: some perspectives from archaeology, early Irish history, and experimental archaeology


Sunday 28th May 2017

Aidan O’Sullivan, Brendan O’Neill and Eileen Reilly

Early medieval houses in Ireland, as elsewhere, were the places where people slept, worked on crafts, prepared and consumed food, gathered together at night, and where a household extended hospitality to kin and neighbours. Early medieval houses and dwellings were key venues for the enactment or performance of social identities of ethnicity, social status, gender, kinship and community. As in most house societies worldwide, these buildings were effectively the places where social identities were created—as children were socialized through routine life, and through watching their elders, learned how one interacted with people of different genders, ages, kin groups and rank.

Ireland has amongst the richest archaeological and historical evidence in Europe for understanding early medieval houses. We have excellent archaeological evidence for house architecture, in terms of their building form, sizes, constructional materials and in a general sense, their spatial organization, although as we shall see there are great gaps in knowledge. Archaeological excavations have uncovered evidence in the form of postholes, wall slots, fireplaces, stone walls, and very occasionally wooden materials. This physical evidence for house shape and size, for construction materials, floors, hearths, storage and domestic occupation can be used to begin to relate house form to status, function and use.

 What we also have in Ireland is very good contemporary documentary and literary evidence for early medieval houses and their uses, which can be used by us to imagine what these houses might have been like. Early Irish laws, narrative literature, and hagiographies, provide a range of general and anecdotal detail about house forms and activities within houses. Early Irish laws from the seventh to eighth centuries provide a significant account of early medieval houses, arguably more realistic than the literary sources as they deal with status and property and what people would be expected to own and what fines are due for damaging their property. The early eighth-century law text Críth Gablach provides a remarkably detailed discussion of the size of houses, their construction details, and the types of tools and domestic equipment used within them. A significant theme in the early Irish laws is the careful ordering of objects within dwellings, with a predictable emphasis on social status and to a lesser extent gender as organising principles. It is clear in any case that people were expected as a given to know where to sit, move, and work, using such fixtures and features as doorways, hearths, and seating arrangements, to orientate their movements around houses.

The early medieval narrative literature, particularly the echtrae (adventure tales) and imm rama (voyage tales), provides occasional vivid descriptions of houses. There is little or no archaeological evidence for the more fantastic of these structures, we may assume that they are imaginary or metaphorical houses, symbolising various things, clearly owing more to the demands of moralising and story-telling than to any real-life dwellings. Nonetheless, these stories give a strong sense that people were concerned with appropriate social behaviour and actions within a house and so also provide intriguing insights into contemporary social norms and early medieval mentalités relating to houses. So, for example, some of these tales describe remarkable houses with multiple doorways. Doorways, being literally the location of thresholds, are often symbolically significant in the tales. In the ninth-century Imm ram curaig Máele Dúin (‘The voyage of Máel Dúin’s boat’) the hero finds a house by the seashore, with one door facing the sea and one facing the land. In Togail Bruidne Da Derga, a house is described as having seven doors, with a shutter to block the wind from whichever direction it blows. These mysterious houses with their cauldrons of food, fine textiles, and silver and gold brooches hanging on walls are also often venues for various social encounters, where heroes eat sumptuous feasts or meet with sexually predatory women. In the ninth-century tale Tochmarc Becfola a man and woman first eat a magical meal and then lie chastely together in a fine house with both cubicles and beds (amra an teg hí-sin itir irscartal dergudha) which implies that they were perceived social and practical differences between cubicles and beds, which of course begs the question, how were they distinguished? It is not something we can see in the archaeological evidence, but does it mean that some roundhouses effectively had small ‘rooms’ within them?

However, neither the archaeological or the documentary evidence (as tantilising as the latter can be) give us a sense of what these houses looked like, their physical appearance, their use in a day-to-day sense, or how they would have performed in terms of light, smoke, heat and other practical matters. There are literally no standing early medieval houses today that give us a sense of these things.

One option then is to build one, using the best archaeological and historical evidence, and using architectural principles, ethnological sources, and some interpretative leaps to fill in missing gaps in the evidence. This is where experimental archaeology comes in, as the reconstruction or more accurately the construction of replica buildings is increasingly used around the world as a way of both thinking through the evidence, but also as a means of testing hypotheses or asking questions of the original evidence. Experimental archaeology can be defined as the reconstruction of past buildings, technologies, environmental contexts, and objects, so as to create a better understanding of the character and role of material culture in people’s lives in the past.

Screen Shot 2017-05-28 at 12.54.39

In 2012, UCD School of Archaeology established a Centre for Experimental Archaeology and Material Culture on its campus, one of the only dedicated university facilities of its type in the world. It, and its associated research offices, laboratories, seminar rooms and equipment, are used for research, teaching and public outreach, and are currently home for a range of Postdoctoral Fellows, PhD scholars and students enrolled in UCD’s MSc in Experimental Archaeology and Material Culture. A range of projects are investigating the transformation of various materials in the past, including stone, flint, bronze, iron, textiles and wood, and features created during, for example, iron-smelting, can be compared with features found on archaeologically excavated sites. Experimental archaeology can also be used in a transformative teaching and learning approach, that encourages students to learn through all their senses (including for example, touch, sight, smell, hearing, heat/cold, etc). Students can also design, implement and analyse the results of their own experimental archaeology projects, providing them with experience and skills in project design, problem-solving, creativity, resourcefulness, intrinsic motivation, and time management. Our students identify their own research project, read the archaeological literature, define a problem/question, think practically about their methodology, get and prepare their raw materials, make or do something, and analyse and interpret their results in the context of wider archaeological knowledge. They often find out new things about the past.

With regards to houses and buildings, experimental archaeological projects have typically investigated a number of aspects of their construction, form and use. The can investigate the architectural and technological aspects of building, asking how did people build houses, what raw materials, technologies and skills did they need and use. Once built, they can explore how the might have performed in terms of light, warmth, heat, smoke, in different seasons and weather conditions. They can also explore the sensory or experiential aspects of buildings, as in what it would be like to inhabit and move around structures, and how their spaces were spatially organized using physical settings, such as hearths, beds and other features. Long-term experiments can investigate the temporal and chronological aspects of houses, posing such questions as how long it takes to build them, how do they deteriorate across time, how long they last, and why and how do they collapse, and what is the likely impact of their collapse and abandonment on the archaeological record?

The Early Medieval and Viking Houses Project is an ongoing research programme situated at the UCD Centre for Experimental Archaeology and Material Culture, focused on the reconstruction and interpretation of an early medieval roundhouse, dated to c.AD 700, and of a Viking Dublin house, dated to c.AD 1000. The first construction is complete, the second is currently in progress. The early medieval roundhouse at UCD is based on a house uncovered at the early medieval rath at Deer Park Farms, Co. Antrim, a site excavated by Chris Lynn and Jacqueline McDowell in the 1980s, and recently sumptuously published as a well-researched and illustrated monograph. The site was located at the top of an Antrim glen.

It was occupied between c.AD 650 and AD 1000, and built up into a mound over time as people brought in clay and gravel to cope with waterlogging at the base of a slope. Consequently at the lowest levels of the rath, in Phase 6A, conditions of preservation were excellent due to the anaerobic, waterlogged nature of the soils. Of the houses at these lowest levels, not only the postholes, but the vertical posts and their horizontal woven wattles survived, along with a wide range of other organic remains. The houses were circular structures, 6-8m in diameter, and were built of double post-and-wattle walls, with an inner and outer wall, with the spaces between them filled with insulating, organic material such as grasses, mosses and bracken. There were no internal roof supports of vertical timbers, so the walls were the only structural feature upon which a roof could be constructed. The house doors were framed with upright oak jambs and lintels, and may have been remarkably small (not more than 1.2m in height). Beds were constructed of wattle, and were layered with poles, branches and organic materials. The early medieval roundhouse chosen for reconstruction at UCD was Structure Epsilon, one of the smaller buildings at 5.3m external diameter.

Reconstructing the early medieval house firstly required careful review by the project team of the environmental evidence to ascertain precisely what building materials would be required. The houses at Deer Park Farms were built almost entirely of hazel, with oak used for door frames. Heather, bracken and moss were abundant in the wall fills, and grass sods may have been used as scraw in the roofs, and also as filling materials in the beds. The floors and bedding areas also had heather, bracken, brushwood, leaves, bark, dry flax stems, a ‘hay-like’ vegetation (i.e. dry grass) that had been cut from wet and dry meadows (the latter source being particularly used in the beds), with rushes and weeds caught up in this material. The insect evidence was also important for reconstructing habitats and from that, a better understanding of living conditions and materials used in the house. There were beetles and insects associated with decaying ‘hay-like’ vegetation, particularly in the bedding areas. There were also some beetles typically found in heather/moorland, especially in the cavity walls, the middens around the site, and on the ground. These might indicate the importation of heather into the site from the neighbouring uplands.

What was then required was some decisions about what architectural form the reconstructed house should take. In general, roundhouse reconstructions in archaeological open air museums, in heritage parks, and in experimental archaeology projects all tend to be quite similar in design; round buildings, with pitched, conical roofs. This is despite the fact that it is a form generally unknown in the European vernacular house tradition. In designing the early medieval reconstruction at UCD, the team decided to move away from this type, and to construct a domical (rather than a conical) structure, not unlike the early medieval clochain found in the west of Ireland. Ethnographic analogies were also sought across the world, particularly for domical houses built of light materials, with no internal roof supports. The Sarakatsani nomadic goat-herders in northern Greece build circular houses with steep domical roofs, constructed of narrow branches and thatched with grass and straw. The Wichita First Nations in Kansas/Oklahoma also built domical houses of a ribbed frame of branches, with no distinction between roof and wall. In architectural terminology, these are termed “undifferentiated structures” (i.e. there is no differentiation between wall and roof, the wall curves up and over into a roof continuously, before curving the far side down into the wall). These buildings have the advantage of being immensely strong in all meridians, despite their light materials.

Screen Shot 2017-05-28 at 12.54.18

Building the early medieval roundhouse at UCD commenced with the digging of 50cm deep postholes to set the vertical oak door frames into, and the completion of the door ope. Once the door was in place, the construction of the wall could commence with the setting out of a circle of closely spaced hazel stakes, using a rope of known length tied to a central temporary post, and simply walking it around. This circle of stakes measured c.4.8m in diameter. Horizontal hazel wattles were then woven in and around these stakes to create a low, ankle-high fence, effectively the foundation of the house. Further verticals could then be placed and hammered into interstices in the wattles, to create a second row of uprights. The weaving of the house wall then started, using the distinctive weaving technique known from Deer Park Farms, which effectively is a type of braid with three wattles weaving around each other, and the verticals, to create an immensely strong, rope-like weave. As the inner and outer walls were raised up, heather, bracken (collected from the Wicklow mountains) and grasses were stuffed into the insulation space. As the wattle walls reached head height, the verticals were brought curving inwards with ropes, so the whole structure started to curve inwards into a dome. Further verticals inserted into the weave continued the structure up and in, until it was finally brought together at the top. The house, now a completed wattled dome, was then thatched with c.1.5tonnes of ling heather (Calluna vulgaris) cropped from a raised bog in Co. Roscommon, in keeping with the insect evidence for upland vegetation and heather at Deer Park Farms. This was stitched with hempen rope to wattle using a 19th century box stitch, by a professional thatcher.

Screen Shot 2017-05-28 at 12.54.03

In total, the raw materials used to build the house included c.250 vertical hazel posts, c.1,600 horizontal hazel rods, 4 oak timbers for the door frame, 4 oak planks for the door itself, and the insulation in the cavity being made of c.20 bags of bracken, along with heather fronds, sticks, moss, grass, and weeds from around the site. Brendan O’Neill, who built the house, took about 30 days to construct it, and the thatching took 9 days work. These estimates of raw materials and time are not provided with any sense that we can meaningfully replicate the tasks of experienced workers, but Brendan O’Neill suspects that experienced early medieval house builders, if they had the raw materials ready and to hand, could have probably built a roundhouse in 10-15 days.


Screen Shot 2017-05-28 at 12.53.05

The early medieval roundhouse at UCD is currently being investigated in terms of living conditions and environment. Specifically, a range of instruments including thermocouples, carbon sensors, luxmeters, humidity and wind velocity sensors, and infrared cameras are used to test temperature, smoke pollution, light levels, dampness and draughts and heat-loss from the walls and roof. It is often thought that these houses would be dark, damp and smoky, with high levels of carbon monoxide, particulate matter, benzene and nitrogen dioxide coming from wood fires. Preliminary results of testing of the UCD house while fires are lit suggest that carbon levels are actually quite low even when fires are lit, and are not much higher than 550ppm (and so well within EU health and safety levels which are set at 1,500ppm). It is apparent that the smoke drifts out through the roof and thus dispels it quickly.

Screen Shot 2017-05-28 at 12.53.16Thermal assessments during winter and summer suggest that a 10-15 degree Celsius difference between inside and outside the house can be quickly achieved with a fire, though it is striking how much physical labour and time is required to keep wood chopped and a fire alight. On the other hand, once the fire goes out, temperatures plummet inside the house quickly, and within 30-40 minutes there is little difference between internal and external temperatures. It is possible of course that the UCD house is losing heat too fast, and that better wall insulation and a thicker thatch would retain more hear (though it would also retain more smoke).

Screen Shot 2017-05-28 at 12.53.30

Light levels inside the house are gloomy. It is obviously brighter at the door, especially in the mornings with the rising sun flooding its light into the house. However, light is also provided by the centrally placed hearth. One of the striking things about the rectangular bed is the way it introduces corners into a circular structure, and also creates a brighter and darker part of the house on either side. The back of the house, behind the bed is perpetually in darkness, and could only have been used for storage. The human experience of the house’s spaces is also being assessed.

Screen Shot 2017-05-28 at 12.53.50

All visitors on the surprising feeling of the distinctive space within the building, and emotionally all people enjoy the fire when lit in the central hearth. This fireplace provides the key hub for any social interactions, as people sit around it on low stools, chatting or falling silent looking into the flickering flames. The house is also an important teaching tool for archaeology students, and is frequently used for cooking stews and baking breads in our replicated early medieval souterrain ware pottery – but that’s another story.

In conclusion, reconstructing an early medieval roundhouse provides us with a range of tools for understanding architecture and living conditions in the past. Through engaging with the historical texts and the archaeological data, in a practical, embodied sense we can gain insights into them, but also become aware of new questions that can be asked of them. We can think about practical building technologies and strategies, the choice and use of materials (and the importance of coppiced hazel), as well as the time and labour invested in their construction. Indeed, on that point, we can think about the ease and rapidity with which these houses can be built, and we don’t know as yet how long these buildings might have lasted, but we suspect that they may have been replaced inside ten years. We can begin to investigate the realities and experience of living conditions inside an early medieval building, in terms of light, heat, smoke and dampness, and to think about social life around the fire, about its vital heat for personal warmth and cooking, about how light acts dynamically inside the house and enables and creates a constraint on human activity. The UCD experimental archaeological reconstruction also offers us at least a glimpse of what daily life might have been like inside an early medieval roundhouse, over a thousand years ago.


This paper is the text of the Robert Farrell Memorial Lecture, delivered by Aidan O’Sullivan, at the International Congress of Medievalists at Kalamazoo, Michigan, on Thursday May 12th, 2016, for the American Society for Irish Medieval Studies.

The lecture was based on the co-authors’ collaborative research project, entitled “The Early Medieval and Viking Houses Project” at UCD Centre for Experimental Archaeology, which was funded by UCD Research Seed Funding Scheme.

O’Sullivan, A., McCormick, F., Kerr, T.R. and Harney, L. (2014) Early medieval Ireland, AD 400—1100: The evidence from archaeological excavations. Royal Irish Academy, Dublin.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s