According to some philosophers, the only way to explain why green is projectible and grue is not is because the former is a natural kind and the second is not. A natural kind is a group of entities such that they share at least one essential property. A common argument for the existence of natural kinds like GOLD, OXYGEN, or BOSON is that they best explain our experimental and predictive success. E.g. if you know that certain properties cluster together then you can predict what properties objects have. What do you think about this defense of natural kinds?
Associate Professor, Department of Philosophy, Lewis & Clark College