Structural archaeologist Geoff Carter, from Hexham, has a radical theory about how Hadrian's Wall was constructed.
As he explains below, he believes it may have been preceded by a massive timber wall, which acted as a temporary frontier while the stone structure was completed.
The remains of Hadrian's Wall lie almost unnoticed under the streets and buildings of North Tyneside.
So, when these areas are redeveloped, even if holes are being dug in the road, archaeologists carefully check them for any traces of the wall.
Piecing together evidence gathered from places like Wallsend, Byker, and Throckley, archaeologists realised there were three lines of "postholes" running between the wall and the ditch on its north side.
In the past, large timber posts were set in the ground as foundations for buildings and other structures. Even though the posts rot away, their distinctive postholes often survive.
As an archaeologist who has spent 20 years researching postholes, I was intrigued and excited to find they had been discovered in front of Hadrian's Wall - and was not convinced they had once held sharpened wooden obstacles, as other archaeologists thought.
Geoff builds matchstick models to help him understand the evidence
With six holes every 1.2m, they form a distinctive and regular pattern, and, even in age of computing, I find the best way to make sense of clues like these is to build small-scale model.
Placing matchsticks in the same places as the posts, and using others to represent the horizontal timbers, it was possible to work out that the postholes were the foundations of a massive timber wall.
It would have been an ingenious structure, made from standard 10-inch (25.4cm) horizontal timbers slotted between pairs of posts. Known as "box ramparts", these structures were hollow timber walls usually filled with material from a ditch.
This was an exciting discovery, but it had even wider implications for how Hadrian's Wall might have been built.
Archaeologists already knew that the milecastles were built first, and then joined together by lengths of wall - a process that took about six years.
They had also discovered, under Hexham Road in Throckley, that the wall ditch and the postholes curved in towards a milecastle.
Geoff's sketch suggesting the phases of construction at Wallsend
So, it seemed reasonable to suggest that the wall ditch and the timber wall had joined up the milecastles during the years it had taken to build the main wall.
After all, a frontier with gaps in it is not really a frontier, so it would make perfect sense to build a temporary one first.
Timber was quicker and simpler to use and was the standard method of construction at that time. Roman forts like Vindolanda were rebuilt five times in timber before stone was used.
The timber wall and ditch, together with another ditch and rampart to the south of Hadrian's Wall, known as the Vallum, would have formed a sealed military corridor to protect the wall, its builders, and their stocks of materials, provisions, and livestock.
The temporary timber wall would then have been removed as each section of the stone wall was completed - and the wood recycled into other parts of this huge building project.
Further excavation needed
About 200 years earlier, Julius Caesar described how his army built 17km of similar ditches and rampart in just three weeks.
Geoff's vision of a section of Hadrian's timber wall
Hadrian's Wall is nearly seven times longer, but it is not unreasonable to suggest that a temporary wall could have been built in less than a year.
Using a model, it is possible to estimate that it would have required about five million pieces of wood.
This would make the timber wall one of the largest timber structures ever built, only to vanish after half a dozen years.
Only future excavation will be able to test the theory further - and see if the postholes extend coast to coast.