But isn’t the original document contents treated as basically a single image, and therefore a single layer? I am not an expert on it, which is why I am asking.
It depends. :)
If you scan to a raster file - jpeg, tiff - and stay raster/image editor, then, yes, one layer of pixels. However..
I just took a photograph jpeg, one flat layer of pixels, opened it in Preview and saved it as a PDF, then opened the PDF in Illustrator.
In Illustrator I have: Layer 1, under that Group Layer, Image Layer, Clipping Layer. That's for a photo jpeg. For an image scanned or printed to PDF, depending on the settings, program and software… you can get a lot of layers you never made.
Here, to me, is a good test:
If we've discovered the original forgers layers, then:
1) We likely would see layers that aren't named as the software automatically does (Layer1, Group, Image, Clipping Path). It would be a nightmare to keep track of the layers without more than this set of names. We don't see this. We see what the software usually does.
2) We would be able to cleanly ungroup, release paths, etc. and get to each piece in some usable manner for manipulating. We can't - the software isn't perfect, it separates what it thinks it should and puts in on a layer, then repeats. This is a mess to try to alter.