It might prove helpful to look at humans etc. to understand the things that trigger the topic of beauty, in the sense that you might learn interesting related ideas in greater detail by studying these things. But the detailed conditions of triggering the topic are not necessarily among them, so "What makes humans perceive X as beautiful?" may be a less useful question than "What are some representative examples of things that are perceived by humans as beautiful?". The world gives you detailed data for investigation, but you don't necessarily care about the data, the ideas it suggests might make the original data irrelevant at some point.