On the existence of prelinguistic categories: A case study