Local research and geophiz suggested the latter, but careful measurements (of the existing walled garden - it was extended) and reinterpretation of what looked like a classic 17th c. garden suggest the former, as this garden may actually be the house, as suggested by the old map (which we were skeptical of).
It's 50/50, so a 5m square test pit is needed to find out. I was not a happy bunny yesterday morning, but now it could be quite exciting to have floor surfaces.
[The drone shot north arrow is wrong]
