Thursday, November 27, 2014

JayJayWords, a new language learning site

Welcome to JayJayWords. This project, this website I've been working on for the last 18 months is born out of my interest in languages.

Languages and computer programming go very well together, at least as far as the bit that's analytical, grammatical, and, frankly, robot-like is concerned. Then there's the whole other side of language-learnin' - the one that has nothing to do with that, like when your teacher is asking you something in Chinese and has finished the question, and now its your turn and you have two and a half seconds to say something to quell the growing suspicion that you have failed, once again, to comprehend what was said and you're hectically scanning the context, her expression, your friend's expressions, the stuff you've all just been talking about, to guess, frankly, what the question was, never mind compose a sensible answer. You're on your own there muchacho: computers can't help you with that. But give a computer a chunk of text and watch it roll up its sleeves and get to work. And that's what we have here.

James Joyce/Trieste, Italy

James Joyce/Trieste, Italy, by thom trauner.

It all starts with a piece of text and the burning question: what is its structure? That, to me, is the most interesting question. What the meaning of the text is is obviously important too, but once again, that's hard for computers to work out, whereas structure is a more tractable problem to solve. The meaning can end up becoming (sorta) clear quite quickly a lot of the time anyway, with a few naive word substitutions, especially in pairings like French/English and Spanish/English. If you didn’t speak French I bet you could still work out what a recent Le Monde headline like “Plus de 2000 soldats déployés à Ferguson.” meant. Or here’s one from El País: “El presidente catalán y líder de CiU, Artur Mas, desveló ayer su plan para proclamar la independencia de Cataluña” which if I told you that "desveló" meant "unveiled" would probably be sufficient to make sense of the rest of the sentence.

So to tease out the structure of an article or any chunk of text you might start by breaking it into sentences. Easy, right? Fullstops. Pas si vite! In Greek, a sentence can end with a question mark, naturally, but what is unnatural is that their question marks look like our semicolons. Wtf; Of course in English a semi-colon that only means that the sentence is having a little rest, but otherwise is still alive and kicking, whereas in Greek that sentence is finished, has expired, and is generally considered a dead parrot. So you can’t rely on a question mark to divide stuff up into sentences, not if you're not making any assumptions about the language the text is in. Which I’m not.

Go down to the word level and things must surely be simpler, you'd think. Why's that? Because words are unequivocally, universally, and pancosmically separated by spaces. NotinChinesethey’renot, ah!

So, just to spell it out if it’s not already clear: almost everything you know about how words and sentences go together you've learned by speaking one language, your own. But that won’t do in a noisy multilingual world.

All this programmatic shenanigans in JayJayWords is just my way of helping myself learn languages by taking arbitrary bits of text in different languages and analysing them to reveal the underlying building blocks. Text such as news feeds, emails from language teachers and friends, or class notes from those same teachers that I've otherwise had to keep in Google Docs, which, for all its virtues knows as much about the contents of that text as I know about the history of Mongolia. And obviously what's good for me, pedagogically speaking, is good for you too, because I’m normal, just like you.

Meeting native speakers and having classes, all that, is expensive. You can only do so much of it, and often have to pay for the privilege, even on great sites like italki. Computing power, on the other hand, is cheap. Looking up words one at a time in a dictionary, even with a Smartphone app, is expensive in terms of your time, your patience, your willingness to go on. Having a computer do it and present it to you in a nicely formatted way, inline, alongside the text, is (relatively) cheap.

And that's where you come in, dear reader. You can help me by picking words and telling JayJay the meaning, part of speech (if you know whether it’s a noun or a verb, etc), even exciting stuff like whether it's masculine, feminine, a Traditional Chinese character, or just a plain old ipso facto Latin phrase. Or even if, like the French word 'constitution', we don't need to do anything with it. This is the 'contribute' feature, and it's nearly ready on JayJayWords.

In honour of that great word-messer James Joyce I have named the app JayJayWords. If that also makes it sound like it has something to do with a bird, then I'm happy. Even though birds don't use words (that we recognise anyway, and they certainly don’t write - again, that we know of. No, no, no, that’s just silly, birds don’t write, end of story), far less worry about the structure of Greek sentences, they do plenty of other praiseworthy things in their own peculiar way. But like I say it’s really named after James Joyce, who made up a lot of his own words, and certainly took pleasure in playing around with classical Latin and Greek terms. Plus he was from Dublin, like me, so that’s nice.