RE: 1.3.1 info and relationships

I understand what is being said about the inverse view of the problem not
actually being covered, because there are a number of WCAG success criteria
that don't actually cater for the inverse situation although they really
should. But in this case, with 1.3.1 and the situation raised here, I don't
think there is a problem.

Consider it this way. If I see three paragraphs of text on a page, that's
all there is - three simple paragraphs. I expect that content to be
conveyed to assistive technology as three paragraphs, with no other
semantic structures or other nonsense thrown in.

So if I then turn my screen reader on and start hearing about headings or
tables or other suchlike stuff, that's a fail!