Gardner's description is about the general working of a skeleton mold, and doesn't go into the bosses, but it's the basic idea that's important.
I just got Gardner's book from the library. I sat in the parking lot, read the passage, and thought about it. And I think I figured it out. Whether you all agree or not is another question! I have the distinct advantage of having seen a photo of the type of mold this was made with. The bottom diagram on this page I posted earlier gives a general idea, sort of:http://www.glassfairs.co.uk/Articles/burmese.htm
(Another somewhat related visual clue I came across was a photo of a vase blown into a metal frame that remained part of the piece. It was point out the large openings in the frame had large bubbles of glass coming out, and the glass exposed by smaller openings was virtually flat. Not surprising, but a nice little demonstration of how glass will tend to expand.)
The reason it's called a skeleton mold is because it's not enclosed. There are simple versions, with a bunch of rods stuck in wood, and these form ribs on the bubble. The one that made the pitcher is more complex. Gardner describes it like the skeleton of an umbrella (handle up). There is a basal plate holding a bunch of pieces of metal upright in a circle, and they are hinged at the bottom, so they fold toward the middle. Some of them form ribs, but there are also some "arms" that have rings on the end (the one I saw a photo of had heart-shaped ones instead of ovals or circles). I believe the latter move independently of the ribs. A bulb-shaped gather just slightly inflated is put in the mold and the ring arms are contracted, "grabbing" and shaping the bosses while the glass is still thick. As the parison is expanded, these move outward, so there is little pressure on the glass to expand inside those areas. The bubble is inflated until full size, the ribs are shaped by the other pieces of metal, and then the mold is opened, and all the hinged pieces unfold, releasing the item.
Does my description make sense? The idea does, but it's worthless to you if I can't communicate it. I might be able to sketch an example.
I'm no glassblower, but I have watched a fair bit of it, and have a reasonable grasp of the behavior of glass. I think this would work, and it ties in with all the bits and pieces of information I've gathered. And most importantly, it fits the glass. All four examples I've posted in this thread show the same characteristics, ones that can't be explained by any other method mentioned.
Any comments? Feel free to tell me I'm dreaming...I'll just have to find more evidence!