Okay, so the worryingly missing proof of injectivity of a certain translation in my late-2003 notes may be feasible after all. I was panicking pretty hardcore there for a second.