A few months ago I found a video talking about the etymology of the word "bear." The video itself was interesting, but it was also the first place that I heard of the concept of Proto-Indo-European, which is beyond cool.
Proto-Indo-European (PIE) is a language that linguists believe was spoken around 6,000 years ago in the area of the Pontic Steppe. It spread throughout Europe from there, and is credited with being the ancestor of nearly every European and Slavic language family, including the Celtic, Germanic, Italic, and Balto-Slavic families. None of this, however, is really the coolest part about PIE.
The coolest thing about this language is that in the 18th century, when linguists first began to theorize of its existence, there was nothing left. There were no written records, no speakers, nothing. The only things to even suggest that it had ever existed were subtle similarities in seemingly unrelated languages. Philologists and linguists looked at individual words in all of these languages and, through the similarities, pieced together what their original form must have sounded like. They kept doing this until they had recreated an entire language. From nothing, from a hunch of a few very bright scientists, we now know what the world sounded like in 3700 BC. Our understanding of PIE has come so far that now, for the first time in 6,000 years, you can actually hear it spoken.
The whole idea of PIE is really a great example of what etymology and linguistics are capable of. Linking so many ancient languages also links their people's cultures, and shows us how the world must have interacted at those times. PIE's influence ranges from India to England, before anything like the Silk Road had existed to connect those places. Basically, PIE shows us one of the many ways that languages can shape the world.