The previous post made an anecdotal case for multilingualism being within the capability of ordinary people. Hyperglots are people who are endowed with or who've exploited that common capability to an uncommon extent. (For a video of an Italian polyglot, Luca, who speaks eight languages, click here and skip the opening ad.) We're all born with the ability to learn not just one language, not even two, but an unknown multiplicity of languages if we want to.
Why so and how? Wouldn't one language have been enough for the minds of individuals? (Societal multilingualism is another matter.) But nature abounds in redundancies. Perhaps they've evolved to provide backup if one ability fails. If we go blind and can't read or write, we can dictate and – these days – listen to audio descriptions; and if sight and sound both fail, there's still communication by touch.
How do we deal with many languages? Not only learn them but also store and recall them and use them properly? The model in favour is that languages are coded as networks of neurons; but there's disagreement over whether a new network forms for each language, or a single network is augmented and the new language 'tagged' for activation and suppression. Either way, our brains come provided with so many billions of neurons that there are plenty for another language and in practice there's no physical limit. Studies of interaction between the languages tend to focus on the leakage or interference between them; but the interferences are relatively minor compared with the vast amount of language that we successfully differentiate and use separately; that's another miracle.
Of course there's the flip side: the billions of people who don't know more than one language. But it's not because they can't, though they often come to think they can't. The most prevalent reason is that they don't need it because they live out their lives in one sufficient community with one dominant language. There's also cultural superiority: the pretension that other people should learn your language and not you theirs; common among native English speakers, including English Canadians.
As for translation, logically the Natural Translation Hypothesis would predict that multilinguals and hyperglots can translate between all possible pairings of their languages. Somebody test it, please! That doesn't mean they can translate equally well or with equal facility in all the combinations; because their translating ability is limited by the extent of their knowledge of the languages and by other cognitive factors. But NTH says they can always do some translating.
Meanwhile, two languages are a sufficient minimum for translating, so it's understandable that most studies of individual translators, including interpreters, have looked at them as bilinguals when many of them were in fact multilingual. At the Expert level, you have a hard time getting a translation position in the European Commission or the United Nations unless you know three or four languages.
Bilingualism studies of individuals have also concentrated on what the bi in the name implies: two languages. This gives the impression that multilingualism is an extension of bilingualism and that it's less normal. But from the point of view of human mental capacity and structuring, multilingualism is not an extension of bilingualism. Rather, bilingualism is a special, minimal case of multilingualism.
Michel Paradis. A Neurolinguistic Theory of Bilingualism. Amsterdam: Benjamins, 2006. There's a paperback edition.
This is for the different hypotheses regarding language acquisition, storage and organisation, especially Paradis' own. Much of it is accessible here and here.
Audio description. Wikipedia. Click here.