Unlocking Prehistoric Houses

In my experience, there is nothing in archaeology quite like tracing someone’s home. My first experience of this was at Roman Verulamium in Britain. This walled city was strategically placed on Watling Street, and one to which Queen Boudicca and the Iceni made a beeline when trying to kick out the Roman invaders. When you dig into a Roman city, the chances are high that you will come across houses. Think of the streets of Pompeii, the fast-food outlet recently identified, or the room with two victims of the eruption still lying contorted where they died. At Verulamium in 1959, we uncovered a fabulous mosaic floor. The centrepiece was a lion with the severed head of a stag in its mouth. My brother John was lucky enough to be on the team exposing it.

On examining the day’s images of excavations at Non Ban Jak, we made out faint white lines. What was going on here?

I was elsewhere on the site, working on a shop that fronted the Watling Street. Crowds must have passed by, on their way to the theatre just 50m away, or by turning right, to the forum. I worked in the back room behind the shop front. No mosaics here, but successive thin clay floors, which peeled off one by one. In the corner, there was a pit, where the shopkeeper had swept the rubbish. In one afternoon, we found about 30 coins, all low denominations, small change lost to the owner. With each one, we rubbed and cleaned the surface to reveal the face of the emperor and inscription of his name, and so we were able to date the building.

Since archaeology is necessary destruction, we worked our way down through the house to its foundations and beyond, and came across a thick layer of bright-red burnt daub. The first Roman settlement hove into view. No coin under the burning layer post-dated AD 60; we had uncovered the remains of Boudicca’s sacking of the city. The emperor Claudius, anxious to gain a military triumph, had ordered the seizure of Britannia under General Aulus Plautius in AD 43. As in most instances of unwarranted invasion of someone else’s land, there was fierce resistance, and the heroine Boudicca rose to the challenge. So here, we had just 17 years of Roman occupation to open up with our trowels. The first houses bore no resemblance to the fine residences that came later, with their concrete walls, mosaics, and hypocausts to warm the inhabitants during the cold winter months. They were quite ephemeral, with wooden foundations and small rooms, and they certainly went up in smoke when the Iceni warriors poured in.


Two years later, James Mellaart came to Cambridge to talk to the Field Club about his remarkable season of excavations at a Turkish Neolithic site on the Konya Plain, called Çatalhöyük. None of us in the audience knew quite what to expect, so you can imagine our astonishment when he described the crowded houses entered via a trapdoor into the roof, with their walls painted with vigorous hunting scenes, and the graves concealed under the floors.

That same year saw several who listened to that lecture in Macedonia, digging another Neolithic site called Nea Nikomedia. When scraping clean freshly exposed surfaces, we encountered walls and floors of the first farmer’s homes.

Almost 60 years later, I was back in Greece for my son Tom’s wedding. During our stay, Tom and his wife Katerina took us to visit her grandfather, who lived in the village of Zarkou. It was there, at Platia Magoula Zarkou, that the clay model of a Neolithic house was unearthed. Like a modern doll’s house, it was equipped with furniture and a family of eight.

Over several decades, and a lot of months at many sites, I never anticipated the possibility of entering prehistoric homes in Thailand. At the back of my mind was always the models of houses in bronze, found in the amazing elite tombs of Shizhaishan and Lijiashan in southern China. Dating to the last centuries BC, they present a newsreel of life at the time. There is a party in progress in one house. Peer into an upstairs room and you can see a young couple having a romantic exchange. A groom sleeps on the ground while his master attends a big meeting. Almost needing a magnifying glass, you can see a rat climbing up one of the posts that raised the house above ground level to avoid wet-season flooding. And there lies the rub. It has always been assumed that houses in Southeast Asia were elevated on posts, as so many are to this day, to counter the downpours and flooding that come every monsoon season. All I ever found over the years were countless post holes that once supported houses. The problem was that wooden posts decay and were constantly replaced, so that they do not reflect a clear plan.

The early houses had been destroyed by fire. It was quite a job revealing the fragile remains of the walls.

Everything changed during our first season at the Iron Age settlement of Non Ban Jak in 2011. Like Çatalhöyük, this site is a mound with two distinct peaks. In our first season, we worked on the eastern rise, and began to delve down into the late-prehistoric layers. My daily routine involved downloading all my images in the evening and making sure I backed them up after reviewing each one. I was not sure if my eyes were deceiving me or whether I should blame the stiff gin and tonic just downed one evening, when I thought I could discern very faint parallel white lines in one of my images. I asked my co-director Rachanie to have a look, and we agreed that there was definitely something there that needed careful exploration. The following day, I climbed down into the square and we began to peel away the dark fill to reveal white clay that, indeed, followed geometric principles. Manifesting as bands about 30cm wide, they ran parallel to each other and turned right angles. It was a town lane. And to add to this surprise, on either side of the lane we began to expose thick clay wall foundations that turned into the square rooms of houses. The foundations had embedded within them the circular post holes to support the walls, and within each room, there was a solid clay floor. My mind reeled back to that distant lecture on Çatalhöyük and the similar walls we uncovered at Nea Nikomedia. For the first time in Southeast Asia, I was walking not only on a prehistoric floor, I was also walking on air.

As we uncovered our first room, measuring about 5 x 5m, we came across a rectangle of dark and soft fill partially covered by a collapsed wall. This was soon followed by a second, smaller rectangle and then a third, this time tiny, near the corner of the room. Down we probed into each and, before long, another surprise. Each contained a human skeleton, ranging from an adult woman to a child and, finally, a new-born infant. They were burying the dead under the floorboards! Then, in the very corner of the room, nestled against the walls, we uncovered the lid of a pot, which still covered the complete pot below. Was this a form of votive offering? We turned to the other corners, and each had in place identical lidded pots equidistant from each other. Our square was only 8 x 8m and I itched to expand it to try and map an entire house, but our priority was to go down rather than sideways, so down we went. And more walls, floors, and burials appeared over the ensuing weeks.

Just as at Verulamium, as we reached the earliest occupation layers, we came across evidence for a conflagration. This centred on a circular pit filled with burnt clay, ash, and charcoal. What was this, we wondered. The lumps of burnt clay had rounded contours, as if they had once been in the form of a dome. Probing into the pit, we came across half a dozen complete pottery vessels. And here was another first for the region: a pottery kiln. Accompanying the pots, we uncovered a large socketed iron implement. My Thai village workers crowded round, and pronounced it to be a ploughshare. This was a third first. Ploughing is hugely significant as a new and efficient form of agriculture. You can cultivate a far bigger area with a plough than with a hoe.

It might have been the proximity of the kiln to housing that set this part of the town alight: abutting the kiln, we began to trace very thin burnt walls and floors. This was very demanding digging, because again like Verulamium, we were down to the first occupation phase and the pioneers who founded the town did not make the same splendid houses as their successors. We soon revealed a complete house with very small rooms. One of these contained, in the centre, the grave of an infant. I could walk through the house into the kitchen, where the fire that destroyed this quarter had left in place a pot over the stove, and rice grains littering the floor. What better sample could you find for dating than a grain of rice? The result, about AD 250. But even this was not all. The burnt house had replaced another, on a slightly different orientation.

I am a great believer in persisting when you have the good luck to find something really exciting, so we returned to Non Ban Jak for another five seasons. The site lies some distance from the nearest village. If you studied Google Earth images taken over the duration of our work there, you could just trace the pathway that developed between our excavation squares and the fully flushing loo that Rachanie had installed. We had our tea breaks under the spreading shade of a large tree and, as always, had the luxury of our canvas roof that covered an area of 10 x 10m. Progressively, we revealed more lanes, houses, and kilns. We found complete human burials within the rooms, sometimes two adults interred alongside each other, and houses that contained graves all on the same orientation; men, women, and infants bedecked with gold, glass, bronze, and agate jewellery. We have begun to return this extinct society from oblivion, and there is much more to come, because we have extracted DNA from the dead.