I read something a while ago, who knows where, about evolutionary qualities of languages. The idea is that languages start with a multitude of competing dialects, and gradually the number is reduced as things get consolidated. Dialects may persist, but moreso in isolated areas--the more communication between groups, the more consolidation (the modern extension of this is that many wholly developed languages are being swallowed up). And the way things get consolidated is to some extent a process of natural selection. Some ideas for a language work well, others don't.

One example of something that might get lost in the shuffle for obvious reasons, would be a 15-syllable word to say that a predator is approaching. And we see this in modern times too; it's reasonable to say that in most emergencies there are a few syllables that will get the idea across. And it could be said that while natural predators are more rare these days, effective communication has become necessary for 'survival.' And even not in emergencies, it seems like the evolution of language is one of more and more contractions like 'gonna' and the elimination of things that are hard to say. It starts to look more like laziness, but it's easy to imagine that at one point it was important. Another example given was that words that must be distinguished frequently should not sound very similar, such as 'yes' and 'no.' These words, especially 'no,' are in reality quite universal, those sounds occur in some approximation in many many languages.

When I read this mystery article I had recently started studying Japanese, and I was baffled by how this language managed to evolve yes and no words. As far as I have learned so far there are three levels of such words for different levels of politeness. At the most polite level there's simply 'hai' for 'yes,' which is used for things like responding to requests from superiors. My quarrel is not with that word, which sounded pretty natural to me right from the start. But here are the other two levels:

less polite: 'ee' - yes / 'iie' - no
least polite: 'un' - yes / 'uun' - no

What the hell is this?! Of course it only makes matters worse that there are many other words in the language with similar sounds, like 'ii' - 'good,' 'ie' - 'house' (also 'love' I think), etc. But leaving that aside, I have to wonder how many discussions these words have inconvenienced or worse. As I've spent a bit more time with them, I can see how it wouldn't be so bad in face-to-face conversation. You can sort of shake your head and knit your brow as you say 'iie' or 'uun' and draw out the vowels, although I can't say if a native speaker would do those things, and technically each sound should always have the same length. But how about over the telephone? Well, that's just another reason the telephone sucks as a conversation medium.
Probably the old isolation thing can function as an explanation of how Japanese kept these confusing words. Then again, it is also known for persisting dialects that are incomprehensible outside small areas, but I don't know if any of these have other yes/no words that might have provided competition.
Maybe part of the reason it seems strange to an English speaker is that consonants tend to contribute more to the way we hear an English word than the vowels. Our vowels tend to all get squished toward the schwa, the 'uh' sound, which is the most relaxed position of the mouth. When words get changed, we tend to keep the consonants more than the vowels. For example, 'gonna' keeps most of the consonants in 'going to,' but changes all the vowels to the schwa (at least the way I say it). In Japanese it seems like vowels are perhaps more important to the sound, so while 'ee' and 'iie' might seem quite confusible to us, they're very different to Japanese ears. Then again, I've also found out that Japanese are known for moving their lips very little while speaking, so it's hard to imagine their vowels don't get squashed a bit as well.

P.S. The pronunciation of those vowels in Japanese is like 'tee' for 'i' and like 'eh' for 'e', and somewhere in between 'ewe' and 'uh' for 'u', but more toward 'uh'

P.P.S. This whole thing is actually pretty hypocritical if you consider 'uh huh' and 'uh-uh' in English. Where did we come up with those?