Content reuse (like in a CCMS) seems to complicate everything pretty nastily. For example, if you transclude one markdown file into another markdown file, are

Content reuse (like in a CCMS) seems to complicate everything pretty nastily. For example, if you transclude one markdown file into another markdown file, are the transcluded file's headings rendered as absolute levels - or as relative levels, based on the indentation level at the point of transclusion ? Of course the same issue arises when transcluding HTML, what with <h1>..<h6>. This issue could arise with (among others?) Lightweight DITA, but I don't think its spec is fully developed enough to even ask the question.