As I read over a bit more of Landow's Hypertext 2.0 again, I remembered what makes hypertext fit particularly well with postmodern thought:
Hypertext values source information at least as highly as the research that uses it.
So if I write a piece of research, it's useless until people start linking to me. But my writing isn't valuable because it's true or right (as someone might have asserted in the past), but because people like, hate, and reference me. Nothing has to be true; it's just talk, just more to read. And one can never read it all.
Hypertext assumes that an issue is never settled. No single person can come to a crushingly-correct idea. The tempting opportunity to revise, to disagree, to undermine is always there, and is encouraged by the medium.
Hypertext assumes that we are all Oedipa Maas, and not much more. This would be a surprise for Vannevar Bush, who conceived his Memex to be a device that would encourage people to build on each other's ideas. While this can happen, it is by no means the default setting of people who use the Web.
Humanity has in the past assumed that forcing ideas into a logical structure is the best way to test their veracity. If hypertext truly decenters heirarchy (I'm not sure it does), then it decenters logic.
Not that this matters. People seem to have abandoned logic and causality without hypertext. The Link has become the most common rhetorical tool I have recently heard, whether it's the link between Al-Queda and Saddam Hussein, the link between Cheney and Haliburton, etc etc etc, blah blah blah.
Trust, if there is any, seems to come from the structure of information these days (which is why I'm not sure hierarchy has been truly decentered).
As nice as Landow's (Marxist? Populist?) hypothesizing sounds in the book, I will say this: the signal to noise ratio in human communications has really plummeted since the Web came round. True, we hear more thoughtful voices, but we have to slog through a more flamewars and greater residue of human depravity to get to those voices.