Reinterpreting 'the Company a Word Keeps': Towards Explainable and Ontologically Grounded Language Models