Thank you! Isn’t it amazing how a rigid hierarchical categorization system fails everywhere you actually look into details? See also category theory vs prototype theory.

It's amazing that most people don't realize it, and even in higher education you get people believing in taxonomies and categories as if they were a property of the natural world. There are no categories in the objective reality, rigid or otherwise; there are no metadata tags attached to elementary particles, that say what the arrangement they're part of is, and of what type it is. Whether in biology or in code, taxonomies are arbitrary - they're created by people for some specific purpose, and judged by useful they are in serving that purpose.

You'd think that now that we have LLMs, the actual in-your-face empirical evidence of a system that can effectively navigate the complexities of the real world without being fed, or internally developing, rigid ontologies, that people would finally get the memo - but alas.

Well, no, what we're saying here is that if you use a rigid, hierarchical catergorisation system (cladistics) you can say that there is no such monophyletic grouping as a fish. Ie there is no grouping with a common ancestor that encompasses all the things, and only the things, that we commonly call fish. That system hasn't failed, it's fine, its purpose is to categorise things in terms of evolutionary descent. However, under that system humans are reptiles and trees and fish aren't useful categories. There exist other systems of catergorisation, which are polyphyletic or paraphyletic, which fit better with commonly used language, and we get back fish, trees, non-avian non-mammalian reptiles. Neither of them are wrong, they're just differently used and differently useful. It's like knowing that a tomato is a fruit, but nobody wants it in a fruit salad. People tend to struggle when things exist in multiple naming systems and categories for some reason.

Agree. Latour's got neat arguments too (commenting on Pandora's Hope)