For a very long time people tried their best to prove the controversial Sapir-Whorf hypothesis at any cost. Just because there's something attractive to the idea.

Lots of great arguments from many many many languages had been put forward, all of which would look very convincing to someone who doesn't know better.
(Neo-Whorfianists are at least proven though.)

But this begs the question. In what other areas have scientists been full of it? For one, what about language history? If they can't not make biased random beautiful-looking conclusions about languages from today, what chances are there that they can be objective about long dead languages?