I feel like you're using the word "fictional" in a sense that renders all models of computation "fictional" to the same extent and hence renders the word "fictional" itself meaningless/uninformative. Personally I feel like there should be some distinction between features depending on how much supporting machinery you need to physically implement them. To me at least, "low-level features" are those that require the smallest amount of physical machinery to implement them. Things like bitwise operations, indirect addressing, arbitrary IP adjustments, etc. Definitely not things like enforcing stack discipline, enforcing block structure and such. (Just so that you know what I mean when I say "low-level". Opinions on that may differ very wildly, of course, and it's perfectly possible that other people use that term differently.)
I'm not saying that "fictional" things are bad. Tail calls themselves are "fictional" to me in the sense that all tail calls are jumps but not all jumps are tail calls, and the distinction between a tail call and a general jump is too difficult to make for a reasonably sized physical machine and best left to the compiler. But that is not me being judgmental against tail calls, of course -- I love tail calls and consider their lack a huge design failure wherever they're absent.
I'm not saying that "fictional" things are bad. Tail calls themselves are "fictional" to me in the sense that all tail calls are jumps but not all jumps are tail calls, and the distinction between a tail call and a general jump is too difficult to make for a reasonably sized physical machine and best left to the compiler. But that is not me being judgmental against tail calls, of course -- I love tail calls and consider their lack a huge design failure wherever they're absent.