The identification of rhetoric with argument has a venerable tradition. The "new rhetoric" has in some cases extended this identification (Perelman), and sometimes revised it to include other methods of proof (Booth). But maybe "proof," however broadly we define "proof," is still too narrow a conception of what rhetoric can mean. Rhetoric can also mean exploration of ideas with no attempt to prove anything. (This is most in line with Oakeshott's view of philosophy as an eternal conversation rather than an argument.)
This distinction is important to any attempt to find a place for hypertext. If we focus on rhetoric as argument (what philosophers would call "philosophy" rather than rhetoric), then it is extremely difficult to see how it can be conducted in the more radical forms of hypertext that eliminate the default path or line of argument. Since the scholarly conversation is entirely about staking competing claims, it will not be easy for scholarly publishing to assimilate this form of text. Perhaps this is why there are so few native hypertexts on the WWWeb, and perhaps why there may never be.
However, maybe this identification of rhetoric with argument and the establishing of claims is nothing but a papyrocentric attitude based on centuries of conditioning by print. If Oakeshott is right that argument only constitutes local passages in the conversation of mankind, then we can see a role for hypertext as a means of exploring ideas without necessarilty seeking to establish a claim.
If we deprivilege the staking of claims and allow exploration without final claim at least equal weight--if we stop handing texts back to students with "thesis unclear" scrawled across them in red ink--we will have to make a radical adjustment in the ways in which use hypertext in the classroom.