Co-evolution in biological systems is based on evolving organisms. There is a broken symmetry here (though personally I find the notion of 'artificial' constraints affecting the space of possible actions rather too dull to get excited about) in that we are considering changes in two categorically distinct mechanisms and organisms.

In fact, imo, what is actually interesting are the specific cases instances of the dull fact, such as of language and thought/expression. And btw, would you consider that specific case "co-evolution"? Is it really a linear history?)