The statement may not be nonsense, but it's certainly not at a level of detail required to be genuinely true. Variables do always have the same value and never

The statement may not be nonsense, but it's certainly not at a level of detail required to be genuinely true. Variables do always have the same value and never change in Haskell. They represent constant abstracted unknowns. That said, a lambda binder which introduces a variable can be applied repeatedly and therefore each independent time it's applied the body of it, where the abstraction occurs, is now understood to operate with a new value for that variable. That's very different from change, though. The semantics of beta reduction clearly don't allow for change---merely substitution. The statics of it clearly demonstrate that even substitution isn't doing anything new---merely replacing like for like. Even when your Haskell program responds to outside interaction---clearly, this is dynamic information, right?---it's still the same static artifact which is doing it. The variables at this point are all erased. It's merely the semantics of your program under outside intervention.