I am interested in checking continuity property of the Poincare group action on the Klein-Gordon quantum field theory defined over the Minkowski spacetime. Maybe the simplest example of QFT out there.

The setup I am working on is the C*-algebraic one, following Haag's local quantum physics (qft = local net of operator algebras). Let me explain briefly the setup. Take $\mathbf{R} = \mathbf{R}^{1+d}$ the Minkowski spacetime, $P = \partial_t^2 - \Delta + m^2$ the Klein-Gordon differential operator and $G = G^+ - G^-$ the advanced minus retarded propagator. This generates a sympletic vector space $V = C^\infty_c(\mathbf{M}) / P C^\infty_c(\mathbf{M})$ with sympletic form $\sigma([f],[g]) = \int_\mathbf{M} (f Gg) dx$. Now the local net of algebras is the CCR-algebra generated by the Weyl unitaries $W(f)$ for $[f]\in V$, with the rule $W(f)W(g)=e^{-i\sigma(f,g)}W(f+g)$, restricting the support of the test functions to the domain of your local algebra.

We have two translation actions on the test functions, which we can call $\alpha$ and $\beta$, such that $\alpha_a f(x) = f(x+a)$ and $\beta_a f(x) = f(x-a)$. Now, we can define the translation action on the C*-algebras by

$\alpha_aW(f) = W(\beta_af)$.

My question: Is this the right action for the theory? Is this strongly continuous? I think it is not, since for the CCR algebra, it is known that $\|W(f)-W(g)\| = 2$ whenever $[f] \ne [g]$. However, Haag's book he writes "it is possible and warranted to choose the algebras so that the action of the translation automorphisms on the elements is continuous in the norm topology". What is the problem here?