Recognition models reevaluated: A response to Smith