It all started when I went searching the web for action figure vs. doll. You see, I have two articulated (and therefore posable) plastic figures of adult humans with removable and exchangeable clothing (and accessories) — yes, I will (sort of) explain, eventually, how a man of my age and station comes to have such things — and objects of this sort are called either dolls (like Barbie) or action figures (like G.I. Joe), according to whether females (mostly girls) or males (mostly boys) play with them, respectively (which means that since the figures function as, among other things, potent models for adult gender roles, dolls of this sort are themselves mostly female, while the corresponding action figures are mostly male).
The crucial factor in the distinction is then not objectively in the figures, but in the function they serve in our culture — a state of affairs that you can find all over the place, as scholars of categorization and the vocabulary that accompanies it have pointed out many times. Objective properties of referents are far from irrelevant, but the functions of those referents play a major role as well, sometimes a deciding role.
Back to the figures. The neat “dolls are for girls, action figures are for boys” split turns out to fail in both directions. The figures I have are meant for use by males (and are themselves male) but are called dolls, while (I discovered in my web searches) there’s a whole universe of figures that are meant for use by females (and are themselves largely female) but are called action figures. On the one hand we have Billy dolls, on the other figma figures. (I didn’t expect these objects or their names to be familiar to you, because each is embedded in a (sub)culture that is not mainstream North American/Western European.)