The Indo-European language family includes most of the languages of Europe as well as many languages in Asia. There is a long research tradition that has shown, though careful historical comparison, that languages spanning a huge linguistic and geographical range, from French to Greek to Russian to Hindi to Persian, are all related to each other and sprung from a common source, Proto-Indo-European. One of the techniques for studying the relationship of the different languages to each other is to look at the similarities between individual words and work out the sound changes that led from one language to the next.
This diagram, submitted to Reddit by user IronChestplate1, shows the word for two in various Indo-European languages. (The “proto” versions, marked with an asterisk, are hypothesized forms, built by working backward from historical evidence.) The languages cluster around certain common features, but the words are all strikingly similar, especially when you consider the words for two in languages outside the Indo-European family: iki (Turkish), èjì (Yoruba), ni (Japanese), kaksi (Finnish), etc. There are many possible forms two could take, but in this particular group of languages it is extremely limited. What are the chances of that happening by accident? Once you see it laid out like this, it doesn’t take much to put *dwóh and *dwóh together.