I really love the way screenshotters always use that off-kilter angle to show something that we're supposed to try to puzzle out for them...:P
I'm pretty sure I know what the problem is, especially in context with your edit. Firstly, keep in mind that everything in Minecraft that involves detection of something else (ie, villagers and doors) stems from the foot level of the entity. Secondly, villagers will only detect doors up to 6 blocks above them or up to 3 blocks below them, and I think up to 16 blocks horizontally. This means that all your doors in your screenshot are completely unseen by any of the villagers because there's a full 5 blocks of space between where the villagers are standing and where the doors are placed. Since they can't see the doors, they cannot form a village and therefore there's no bounding box for the golems to spawn in.