YES # Parallel rule labeling (Zankl et al. 2015). Consider the left-linear TRS R: f(a()) -> b() f(a()) -> f(c()) a() -> d() f(d()) -> b() f(c()) -> b() d() -> c() All parallel critical peaks (except C's) are decreasing wrt rule labeling: phi(f(a()) -> b()) = 3 phi(f(a()) -> f(c())) = 3 phi(a() -> d()) = 3 phi(f(d()) -> b()) = 2 phi(f(c()) -> b()) = 1 phi(d() -> c()) = 2 psi(f(a()) -> b()) = 3 psi(f(a()) -> f(c())) = 3 psi(a() -> d()) = 3 psi(f(d()) -> b()) = 2 psi(f(c()) -> b()) = 1 psi(d() -> c()) = 2