True plain text -- as opposed to something like Markdown or another "plain text markup" format -- is hard for a reader mode, because you're going to have to put some effort into deducing the structure of whatever document you're looking at and that could be highly idiosyncratic from author to author. I don't think that's necessarily a reason not to do it, of course, but I can imagine it's why it's not high priority -- reader mode is already having to deal with all the ways modern web sites have found to make HTML highly idiosyncratic.