Reply to Touretzky and Pomerleau: Reconstructing Physical Symbol Systems