4 ms·

> The social hierarchy of hyenas is even more rigid and predictable, with a depth of more than 100 layers I would like to read more about this. The largest kno

> The social hierarchy of hyenas is even more rigid and predictable, with a depth of more than 100 layers I would like to read more about this. The largest known hyena pack [0] is about 130, with most being smaller. So we can say this claim is almost equivalent to saying there's a total ordering of the hyena pack, where between any two animals the seniority is clear. Reading the article supports this interpretation. Sorting 100 things is a difficult feat of computation! How does every hyena in the pack remember its relative place in the hierarchy? How do updates to the hierarchy (presumably it's not strict unchanging seniority order) happen, and how do they get propagated through the pack? If hyena A sees a lower-ranked hyena B exert dominance over a higher one C, how does it now update its place in the pecking order vs B or C? Perhaps this is out of scope for the statistics paper, but it's one for the people who like to develop theories of computation inspired by the animal world. [0] https://ideas.ted.com/everything-you-know-about-hyenas-is-wrong-these-animals-are-fierce-social-and-incredibly-smart/ https://ideas.ted.com/everything-you-know-about-hyenas-is-wr...

wasabi991011 2y ago
You seem to believe this total ordering is accessible to hyenas, is there a reason you suspect this? In my mind just because these scientist found an algorithm to model this total ordering of the pack, does not imply that there must be an algorithm actually in use in the pack. In fact, I would suspect that if the hyenas had total knowledge of the hierarchy, there would be less competition between distantly ranked hyenas, and that the model would actually calculate a shallower competition. But I don't know anything about hyenas, and I may be misunderstanding the paper.
- dmurray 2y ago
  If I'm reading it right, there are very few "upsets", where a hyena the scientists model as lower-ranked exerts dominance over a higher one. That suggests that even if every hyena doesn't know the total ordering of the pack, he knows whether each other hyena ranks higher or lower than him. Maintaining that knowledge doesn't absolutely require knowing the total ordering, but it's an interesting computer science problem to maintain it without.

wasabi991011 2y ago

You seem to believe this total ordering is accessible to hyenas, is there a reason you suspect this? In my mind just because these scientist found an algorithm to model this total ordering of the pack, does not imply that there must be an algorithm actually in use in the pack. In fact, I would suspect that if the hyenas had total knowledge of the hierarchy, there would be less competition between distantly ranked hyenas, and that the model would actually calculate a shallower competition. But I don't know anything about hyenas, and I may be misunderstanding the paper.

dmurray 2y ago
If I'm reading it right, there are very few "upsets", where a hyena the scientists model as lower-ranked exerts dominance over a higher one. That suggests that even if every hyena doesn't know the total ordering of the pack, he knows whether each other hyena ranks higher or lower than him. Maintaining that knowledge doesn't absolutely require knowing the total ordering, but it's an interesting computer science problem to maintain it without.