Are errors in false belief tasks symptomatic of a broader difficulty with counterfactuality?