I had a look at the source code but can't actually find where the limit of 3 characters if defined. Anyone happens to know ?
Once we know where it is, two situations might be found: either it's already extendable, in which case a third-party extension could implement the required additional logic. Or other option, it's not, and we should then open an issue to make it more extendable.
Unless there's a very simple way to tell which "characters" (in Unicode sense) should be counted as a full word, I think the implementation is best left to third-party extensions. If it's simple, then we could consider adding it to core directly.
Simply allowing search with one or two characters, including English ones, would most likely result in database requests overwhelming the server with the enormous amount of results whenever just one or two English letter are typed.
Also if anyone happens to know if/how other forum/search software handle that, that would be great insight.