Alternative: publish a DR that clarifies that "atomic_notify_*" functions are well defined and do not cause a race even if the object no longer exists at the pointed address. I think this will work for recent Windows STL, libc++, Glibc, and MUSL, at least for most types. I haven't checked other implementations.
What if the object has not only been destroyed, but deallocated, and the memory is no longer part of the address space? Should it still work?
I don't see how you can start the discussion talking about what's possible in the abstract machine, and then proceed to "and you should just be able to perform operations on objects outside their lifetime".