If we look closely at (1) we see that time and speaker coordinates are not sufficient: to get the full meaning we must know what 'Master of this College' means, and to know that we must know which college is being talked about. It is not clear what sort of contextual feature supplies this information. The postulation of a 'college coordinate' would certainly do it but this would mean that we should have to have such a coordinate in all value assignments and, if so, why not a country, climate, religion, or 'previous drinks' coordinate ?143