In this essay I have discussed an assumption of semantic externalist theories which I called the coordination principle. This is the idea that natural language kinds and scientific kinds line up or can be mapped onto one another one-to-one. A closer look at water shows that there is not this type of simple one-to-one match between chemical and ordinary language kinds. In fact, the use of kind terms in chemistry is often context sensitive and in cases where chemists want to ensure no ambiguity, they use a very complex and nuanced set of kind terms, none of which could be reasonably associated with the ordinary language kind term "water" alone. Since we cannot just turn to chemistry to find a single chemical kind that can be used to determine the extension of "water," there is not any strict sense in which water is H₂O, because exactly what water is depends on the context in which "water" is uttered.