I think we already have this markup language. For example, look at the words we use for different drugs:
Pychedelic
Dissociative
Stimulant
Empathogen
etc.
A description of the experience is already in the names.
Perhaps I'm confused by what a markup language actually is, because language itself seems to be a "markup language" of sorts. For example, I can't describe what "red" looks like, but you know what I mean when I say it. We already use words as variables to simplify the description of incomprehensibly complex states. It is very hard to pin down a word describing certain perceptions common during psychedelic experiences, but we are already trying to do that in the Nexus Wiki, are we not? In essence, a DMT XML would need only consist of a list of words from the Wiki, from which the user must pick Tags that best describe the experience. For an entheogenic markup language, the user could pick Tag words from a list of words like, say, Webster's dictionary.
My understanding of conceptual XML is that it essentially makes use of name/value pairs. So all we need for an Entheogenic Markup Language is a list of attribute names and possible states of each attribute...
Suggested attributes="[list of states]"
actual plateau duration="[number of minutes]"
apparent plateau duration="[number of minutes]"
immersiveness="[number from 1 to 10]"
dimensionality="[integer]"
essence="chaotic","organic","euclidian" (or "gaseous","fluid","crystalline" ?)
primary emotion="euphoria","love","sensibility","grandeur","confusion","fear","hatred"
secondary emotion="euphoria","love","sensibility","grandeur","confusion","fear","hatred"
The problem with creating an Entheogenic Markup Language to quantify what is otherwise expressed in confusingly abstract and esoterically subjective paragraphical descriptions is that the psychedelic experience is indeed confusingly abstract and esoterically subjective. There's no getting around the fact that the insides of the psychedelic experience have never really been seen by human words.
So here's another suggested name/value pair:
ineffable="yes"
or how about:
Inexpressible using XML="yes"