Because of the graph structure behind the scenes, pretty much any data can be expressed in the RDF model and hence in an RDF syntax, although it might get a bit nonsensical when it comes to interpreting the triples. Here is a case in point. There was definitely some sensible discussion of Atom syntax being RDF/XML (but the handful of extra attributes needed were considered too much overhead). But also I vaguely remember (or maybe imagine) HTML/RDF mappings being done pre-Turtle. It just crossed my mind, couldn't resist having a go.
So here's an example:
@prefix : <http://example.org/html9/> .
<http://example.org/hello> a :html ;
:head [ :title "A Page" ] ;
:body [
:h1 "My Page" ;
:p "Hello World!"
] .
The placing of the bnodes is a bit arbitrary, but I rather like the idea of a resource being a HTML. I believe this corresponds to the RDF/XML:
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns="http://example.org/html9/">
<html rdf:about="http://example.org/hello">
<head rdf:parseType="Resource">
<title>A Page</title>
</head>
<body rdf:parseType="Resource">
<h1>My Page</h1>
<p>Hello World!</p>
</body>
</html>
</rdf:RDF>
Hmm, actually it seems quite sensible at this level of nesting, not all that far from Reto's DiscoBits idea. In fact those bnodes could usefully be swapped for # URIs. But I'd prefer not to think how it gets with e.g. a load of nested <div>s.
Dunno, I could imagine an advanced (RDF-friendly!) Wiki syntax looking something like that Turtle.