My training is in language variation, and I specialized in sociophonetics, so I'm not an expert on syntax, but my understanding is that whenever chomsky's theories are empirically tested on real language data, they don't seem to hold water and he has to go back to the drawing board.
> I wonder if there's much history of groups starting out with common languages, then groups separating and the languages evolving separately, either intentionally to demonstrate group membership, or just due to lack of communication between groups, and continuing to differentiate until they're effectively different languages.
>Language doesn't dictate patterns of thought, but it certainly shapes them.
What you're talking about is the Sapir-Whorf hypothesis, and it is not certain at all. It's highly contested. Language may have an effect on some cognitive processes (again, contested) but definitely not all.
However, language does reflect the cumulative experience of a culture, whether modern speakers are aware of that experience or not. Here's a good illustration: http://youtu.be/kIzFz9T5rhI
So language can reflect a culture, but it probably doesn't dictate it. In the case of the Mongolian girl, she's unlikely to be a nomad these days, and what separates an English speaking Mongolian from a non-English speaking one is usually wealth. So, someone from a wealthy country is probably more culturally similar to her than they might expect.
> What you're talking about is the Sapir-Whorf hypothesis, and it is not certain at all.
Yes, I am. The weak variation of the hypothesis. Coming from a computing background, it's difficult to convince myself that it's the case that the structure of a language (whether artificial or natural) doesn't encourage or discourage specific ways to represent meaning.
It's not always that "natural" of a selection. Many native Americans were basically kidnapped and forced to learn English. Governments in the past have intentionally tried to eradicate languages with the mistaken belief that one language is superior and more communicative and another is barbaric or degenerative. This still happens in the form of dialect prejudice, which is institutionalized in the American public school system. People tend to project their prejudice towards a particular group of people onto their language. This has been proven many times with dialect perception experiments. I think we have an obligation to preserve cultures and languages that have been eradicated due to ignorance and prejudice, and we need to learn from the past and stop trying to enforce linguistic standards in school. If a single homogenous language is necessary, it should emerge naturally.
Having said that, languages do die naturally for socioeconomic reasons too, and I agree that it's unnecessary, and probably harmful to try to force people to use a language that they have no socioeconomic incentive to use.
Being kidnapped and forced to do anything is the problem here. Being rejected for being different is too. The "learning english" part is really not a big issue compared to it. Actually I'm pretty sure people would accept much more easily somebody speaking a dialect next to them if they knew in the end, they have at least one language in common.
Babies are born with the ability to distinguish between any two frequencies that are important for a phonetic distinction. Within the first year, they begin to lose that ability and only retain frequency distinctions necessary for whatever language they're exposed to. So, when an adult can't tell the difference between two phonemes or has difficulty articulating it, it's because they weren't exposed to it early enough. So, the difficulty you describe would apply to any lingua franca. In the bullshit, completely impossible scenario where everyone in the world speaks one language, this wouldn't be a problem because there would only be one set of phonemes.
It isn't necessary to have only one language for people to communicate. Humans are perfectly capable of learning multiple languages, and and it may even be good for a child's cognitive development to do so.
It's also unlikely that it is possible to have one global language. People acquire language from the people they grow up around. They get it from their parents, friends, neighbors, etc (note that I omitted teachers). Even if everyone had the possibility to communicate with any other person in the world, from birth, they wouldn't, and so dialects would form and eventually become languages.
Actually, the impact that standardized language education has had on dialect variation isn't quite clear. It appears to have had some impact, but not as much as laypeople usually think. In fact, standard written American-English has undergone quite a few changes over the last 100 years and things that were non-standard have become standard. Likewise with mass media. There is no evidence that mass media has had any effect on the rate of linguistic change and variation in America. The biggest impact has probably been the increase in ease and speed of modern transportation. Regardless, there is likely more dialect variation in America today than there was 1-200 years ago, though geographic variation may be slightly diminishing (at least intraregionally) and socioeconomic variation may be more important.
The biggest factor in dialect development is time, which is why you see a large variety of dialects in England as compared to the USA. England is much smaller, but has had much longer for dialects to develop.
