With my professional hat on, I’m currently musing about how one would match a lost item to a found item, given that the two would have been described by separate people. The crux of the matter is down to distinguishing between categories and picking up on similarities in attribute values that are more readily apparent. For example, I’m not at all clear how one makes a distinction between a handbag and a shoulder bag as I tend to call every bag a handbag and be done with it. But I’m a lot clearer about reading a brand label, or defining it as being green.
It strikes me that categories are based on what one knows about the specific type of item e.g. coleopterists can discern between many types of beetle, but they are all mostly ‘beetle’ to me as I don’t know enough to distinguish. But I can tell a green beetle from a black beetle with relative ease, and that’s my hook into telling them apart.
There are, of course ambiguities with attribute values as well (I’ve certainly been party to many arguments about whether something is blue or green or grey) and for our particular matching problem here, we’re coping with that by saying that green can be a certain degree similar to blue. The fact that we’re having to do the same thing with categories just proves how difficult categorisation is – once you start creating really specific categories, you start requiring more specialised knowledge to tell them apart.