How to build a medieval castle today

It's coming back to me now. I remember this clearing in a forest in rural France. I recall the peace and quiet, broken only by the tapping of masons' chisels. This time too, I can hear the clip-clop of a cart horse, the clang of a blacksmith's hammer. There's still no roar of diesel engines, no creaking of cranes – not even the buzz of a circular saw.

Some things have changed. I pass car parks full of retirees arriving on their day trips, school children streaming out of coaches. Back in 2001, it was more or less just me, in a quarry, chatting to a bunch of passionate but obsessed makers starting out on what seemed like a crazy and perhaps impossible dream: to build a full-scale 13^th-​century castle from scratch, using nothing more than the tools, techniques and resources of the time. Now, when I look up at the towering castle wall and turrets, it still seems outrageous, but they've proved it's possible.

It could be the 13^th century. But this is not a film set, a piece of theatre or a performance of any sort. It's not even a demonstration or a museum. This is all for real.

Guédelon is an actual construction site. "We obviously respect all modern health and safety law," says spokesperson Sarah Preston, but "we're able to work without high-vis jackets. There are hard hats, but they're covered with straw or linen, so nothing kind of sticks out."

Keeping workers safe is not easy when they're working with High Middle Ages heavy machinery like the 'cage à écureuil', or 'squirrel cage', a human-powered treadwheel used to winch materials up the castle walls. "When you came, the introduction of the treadwheel winches was a really live issue," Preston explains. "We were very aware that as the walls grew, we were going to need to find some solution. The local inspector of works didn't want to give us permission to use these, even with additional safety features. For example, we've got a braking system, we use modern ropes with known breaking strains. But their argument at the time was this is below human dignity to use this machine," she says, "and the team petitioned them saying we need them, we can't actually continue working. The alternative is carrying heavy wooden buckets up and down ladders!"

The site is open the public too: "There are some slight compromises that we've had to make in terms of copying the architectural model and having a site that's safe for visitors," says Preston. But these are very minor things like adding a few centimetres to the crenelated walls. "It doesn't somehow take you out of the experience, even though we're not pretending," she adds. "The costumes are not identical 13^th-century costumes and we're always very clear to point that out when visitors come. People are dressed up but safety and comfort is the overriding concern."

Guédelon's journey is more important than its destination: a project in experimental archaeology. By putting themselves in 13^th-century workers' shoes, the makers see similar problems and perhaps find similar solutions, discovering how it may have been done back then, and perhaps solving a few archaeological mysteries in the process.

When I visited last time, Guédelon was already solving its first mystery, regarding the meaning of masons' marks on stone blocks. It was known some marks were for positioning "so the mason who is going to fetch the stone, he understands directly where it's going to go, which place in the building", says Toendra Schrauwen, head banker mason (the banker is the wooden table they work on). Other marks were a mystery, but the masons realised they were simply to identify the maker for accounting, as the medieval masons were paid per stone, and for traceability.

Yet at Guédelon, solving one mystery often just opens another. "It's still quite a mystery why on certain buildings we can still see them," says Schrauwen. "You have lots of beautiful monuments where you see the mason's mark. It's too obvious. Why?"

Guédelon has since made many more discoveries about the 'how': the engineering of the building. The keystone is the last stone to go into an arch, but not so a vault, Guédelon discovered. "We build up the walls of the cylindrical tower and then once your walls are high enough you can then fix your corbels," explains Florian Renucci, master mason. "The corbels are the stones which are the beginning of the ribs.

"That's when we must install the centring [the temporary formwork], place the key, build the ribs up towards the keystone, so your skeleton is then in position. But at this point the skeleton of the ribs is still not under load. It's still being supported by this wooden structure," he adds.

When they are ready to strike the centre, the carpenters go back underneath and each branch of the centring is placed on wooden wedges. "We then hammer them out very gently and all at the same time so that each branch comes down at the same time," says Renucci. "As the wood comes down you start to see daylight between the stone and the wood. That's when you cross your fingers, and five minutes later there's a cry of joy and relief, because your skeleton is under load! It's only now that we can build a cover: the web stone." If it's not done in these stages, he says, it goes wrong. He is amazed this had to be rediscovered. There are so many books about medieval art, Renucci points out, and none of them have anything to say about how a vault is built. "It's incredible," he adds.

"These are techniques that haven't necessarily been lost in the world of construction, but have been lost in the world of, say, history of art or archaeology," says Preston. "The archaeologists that we work with or the architects that we work with still today are learning from us that the key goes on first."

Working out how to build something doesn't necessarily mean you can be sure that's how it was really done, but archaeological evidence can help.

Raising roof trusses is a puzzle: lay the trusses on a temporary floor at the eaves level, then swing them upwards into place, but when you near the last ones you begin to run out of room. Guédelon's solution was to build a horizontal rail between the gables, hang trusses on it and then slide them into position like a clothes rail or filing cabinet. "The day that we'd finished doing it, we came back to the office, opened a book and found a picture of a gable wall with this bar sticking out from it," says Preston.

Other solutions change our historic understanding more widely. As the window spaces went in, visitors started asking if they'd have stained glass. But glass would have been out of the reach of a modest nobleman like the Lord of Guédelon. Instead, the scientific committee worked out that they'd have painted waxed cloth, as they did at Windsor Castle up to the 15^th century. Only the frames survive from the time because the cloth itself was fragile, but there are clues in financial records. Experiments, some of them unsuccessful, led to the panels they'll be installing in the chapel this year.

We take glass for granted, but Preston suggests cloth may have been used everywhere, changing our perception of how a medieval town looked. "In terms of experimental archaeology, we are right at the heart of it here. Sometimes people imagine," she explains, "that there are experts who can come and tell you how to do it, and the whole point of experimental archaeology is that people don't know, nobody knows how a lot of these things were done. Guédelon is a meeting place between the practical skills of painters, carpenters, stonemasons, and theories of people who were carrying out research into these areas."

We walk towards the castle, and as we round the corner, I can see how dramatic the progress has been.

When I last visited, just a corner of the base of the walls was built. Now they've completed a whole central chapel, several corner turrets and most of the largest tower, as well as nearly finished an impressive gatehouse that will soon get its portcullis.

I can see the completed roof, which is providing inspiration for the restoration of the destroyed Notre Dame cathedral roof in Paris. It's a fifth of the size but the technique is similar, using tall thin trees rather than the big trees that feed modern sawmills. "Notre Dame required 1,000 oak trees at the time, which sounds like a lot, but the oak trees that were used are what would in modern forests be taken away as firewood, because forest management has changed enormously," says Preston. "Seven former employees at Guédelon are currently working on the site at Notre Dame. Five are hewing oak beams and the other two are making the axes and side axes needed to hew those beams," she adds.

We also see the impressive double-drum treadmill winch, which is lowering the cradle ready to lift a load up the castle wall. "As they walk, the wheel turns; there's a central axle with a rope wrapped around it," explains Preston. "You can see the rope goes up the central mast, around the first pulley, that then goes up to the top of the jib around a second pulley and then back down to the cradle, which is loaded with stones. So just by walking, just a fairly gentle pace, we can then hoist up to 500kg."

A delegation from French engineering giant Bouygues, looking for more environmentally friendly building methods, took a close interest in this piece of medieval heavy plant. "It's something that we couldn't have imagined when we started out," says Preston. "It goes beyond the experimental archaeology, understanding the past, into seeing how the past then can offer potential solutions for the future."

Bouygues was also interested in the lime-​based mortars of the 13^th century, which Guédelon has learnt how to make using its own lime kiln. "It's slow-drying, it'll take 600, 700, 800 years for this lime mortar to completely dry at the heart of the thickest walls where there's no air," Preston explains. "The advantage of using this slow-drying mortar is that there will be a suppleness, an elasticity to the structure. If we were to use quick-drying cement between the stones, at some point there would be so much pressure that either the stones or the cement would crack. This slow-drying, non-hydraulic lime mortar allows each stone to just slowly settle into place." However, to be used in modern building, the world's building regulations would have to be updated to allow drying times of longer than 24 hours.

On my first visit, Guédelon said the build would take 25 years. But that was to underline how they'd take their time, rather than to set a hard target.

"A castle of this scale would have had a team of about 50 people, which is what we've got, 40 to 50 people, but they would have taken half as long, 12 to 15 years, because this is the first castle we've built," she says, whereas in the 13^th century there was building everywhere. "Every major town in Europe has a cathedral under construction, churches, castles, fortified towns being built. So, all these skills can be taken from place to place. This is an experimental site, so there will be times when we must stop and consult the archaeologists we work with, where we'll try different techniques. Sometimes they work, sometimes they don't.

"And the main reason is that we are open to the public," she continues. "At least half the time, certainly today with the number of schoolchildren on site, my colleagues are putting their tools down in order to explain the work they're doing."

Like most construction projects, it's running behind schedule, but nobody seems to mind. There's no client waiting to move in. It's pulling in visitors (enough to fund it) and everyone keeps learning from it. Does completion even matter?

"We really want to see it finished," says Schrauwen. "It's more satisfying. Anyway, we know that there are going to be lots of projects after Guédelon, like the little chapel for the village, the village itself, you can still build extensions of the castle so it's an adventure that's never going to finish."