In the case of water, it follows a priori that if something has the micro-level properties attributed to H2O by chemistry, then that thing would possess waterish macro-level properties
The characteristics of something that you can actually see, touch, or measure with your senses—the big-picture features we experience directly.
micro-level properties(as used in philosophy of science)
The characteristics of something at the smallest scales, like the arrangement of atoms and molecules that make it up.
waterish macro-level properties(as used in philosophy of mind and science)
The observable qualities we associate with water, like being liquid, clear, and able to quench thirst—the things that make water 'feel' like water to us.
The only general argument that seem to be available for this would be the principle that, for any two levels of discourse, A and B, they are more-than-causally connected only if one entails the other a priori. And the argument for accepting this principle would be that the relatively uncontroversial cases of a posteriori necessary connections are in fact cases in which one can argue a priori from facts about the microstructure to the manifest facts. In the case of water, for example, it would be