Emma Bellamy-Dodd (Host) · 28:46
That makes me think, what can we do as a society to better encourage linguistic diversity?
Antonella Sorace (Guest) · 28:55
It's interesting, I was at an international meeting yesterday where we were talking about this in many ways. From my point of view, information is essential. So making sure that people have the right kind of information about what it means to be bilingual, multilingual, and getting rid of these misconceptions the old misconceptions, some of the new misconceptions that I haven't mentioned so far, but some people now are thinking that, you know, yes, if you're bilingual, you don't get Alzheimer's, for example, you don't get dementia, which is not true, unfortunately. But research shows that there may be a delay in the first symptoms of dementia. If somebody is genetically predisposed to get this disease, they get it. But bilingualism may be one type of experience that delays the emergence of the first symptoms. Or a child is more intelligent if they have more than one language. This is also not true. We are not talking about intelligence. We are talking about adaptability to change. We're talking about the ability to understand other points of view, the ability to behave in a more effective way, you know, depending on where you are. But no, we're not talking about intelligence per se. So we have to work together to get rid of these misconceptions. And we think that information is essential. As an academic, we have responsibility. We have a special responsibility to make sure that research reaches people also because we benefit from this. Research benefits from this. We understand how bilingualism works much better by going to particular contexts, talking to different people. So it's not just them benefiting from interacting with us. It goes both ways. We do better research by doing public engagement. So information is essential. Another very big issue that is being discussed more and more is artificial intelligence, you know, and the existence of models and systems that can be very useful on the one on the one hand, right? They provide more sources of data or opportunities to hear the language, particularly minority languages. But there are lots of challenges, lots of challenges. And I don't do any work in this particular area, but I have colleagues who do. And benefiting from automatic speech recognition systems, for example, involves having lots of data. Lots and lots of data. Now, those data are there for majority languages, of course, you know, English is a good example, but they're not there for minority languages. And so many minority languages can't be represented that way, can't be helped yet by artificial intelligence because there aren't enough data to build these very large models. And so ultimately, can we stop the disappearance of minority languages? I think the most optimistic view that we can have now is that we can delay it. Whether we can stop the disappearance, I think more and more languages will die. I was at a meeting at UNESCO recently, and they said that by the end of this century, about half of the existing languages may be dead. Because they're not spoken anymore, because they're not transmitted from one generation to the other. And AI can't really help the vast majority of these languages. So the end of this century is 3 generations, 3 generations, and then half of the existing languages in the world will be dead. And that is a very big loss for humanity.