Invariance principle (linguistics)

Transcription

So let’s talk about blueprints. There have to be plans for putting together a system, whether it’s constructing your band or an apartment building. And when we have something as complex as language, with all its delicate nooks and crannies, we all have to be using the same guidelines if we’re all going to end up in the same place. But do we? After all, there’s a great deal of variety to languages. The thing is, it turns out that when you set the right parameters, all languages are principled ones. I’m Moti Lieberman, and this is The Ling Space. If you’ve been watching along with our series so far, you’ve probably noticed that we at the Ling Space are big fans of Universal Grammar, or UG. Under the UG view, we say that all people are born with a set of knowledge about how languages can work. This means that when babies encounter data from any given language, they can build up a linguistic system of their very own. But clearly, this isn’t the simplest point of view to take. It’d be easier to say that we all start off with nothing, and then we pick up how everything works as we go along, like learning the guitar or painting a picture. Back in our first episode, we advanced one argument as to why we support the UG theory: babies just get language too quickly, and make too few mistakes, to be starting off from the infant equivalent of some vast mental wasteland. This time, we want to come back and address another leg of this argument: languages are too similar for them to be growing unconstrained. Now, on the surface, any pair of two random languages, like Swedish and Japanese, don’t really appear to be that similar to each other. But these differences, under the UG view, are just superficial ones. If you peel back that top layer, and look at the wide variety of languages that we have in the world, you actually find a surprisingly small amount of variation. There are some rules that all languages follow. And when things do vary between languages, changing one small thing below can cause a big array of changes on the surface. We don’t need to encode into little baby brains nearly as much as you might think. That’s because there are two sets of things that UG stocks us with from birth. We usually refer to these as Principles and Parameters. Let’s start with the Principles. These are the rules everyone shares, universal properties that are true in every natural language in the world. We’ve already encountered some, like in our video about movement and traces in syntax: the idea that you always leave something behind when you move stuff around is true for all languages. But let’s dig a little bit deeper. Those tree diagrams that capture what we know about how sentences work? There are rules that govern them, no matter what language you’re talking about. Let’s stick with talking about moving around words and morphemes. For a simple sentence, like, say, “Nana heard Ren play the new song”, you could ask a question about almost any part of it. So you could ask, “Who did Nana hear play the new song?” Or “What did Nana hear Ren play?” And every English speaker would be fine with your beautifully formed questions. But we’re not actually free to just take whatever words we feel like and move them around to make those questions. Our system is subject to the principle of Subjacency. By this principle, some words are off-limits, stranded off on syntactic islands, from which they can’t escape. Other syntactic elements just get in the way, locking down that part of the sentence so that no one can get out. So let’s tweak our previous sentence a bit. What if we make it, “Nana heard when Ren played the new song.” Okay, that's simple enough. But can we still ask the question about who was playing, like we did before? Let’s try it: “Who did Nana hear when played the new song?” That’s... not good anymore. The presence of when there just stuck a padlock on who leaving that part of the sentence. There’s no escape. And that’s the same as for any language that moves its question words around. If your language does that, it obeys this rule. So if you put our sentence in German, the question would be just as bad: “Wer hat Nana gehört wann das neue Lied gespielt hat?” Or if you want to put it in French, “Qui a Nana entendu quand a joué la nouvelle chanson?” That's also super terrible. And that’s just one of the many islands Subjacency rules over. We’ve got some more about this in the extra materials back on our website, but for now, you probably get the idea. Our questions never break the principle of Subjacency. Every natural language obeys this rule, because it's part of UG. Now, it’s not just the ways in which languages are the same that are determined by UG, either. The way they differ is just as constrained by what UG has on offer. For all the huge variety of ways that languages look different, a lot of it comes down to making a choice between one thing or another. We call these Parameters. It can be useful to think of Parameters a bit like switches, You can just flip in your head when you’re learning a language. For example, do verbs get put before or after an object? Do sentences need to have an overt subject, or can we understand what we mean without one? Can a syllable have multiple consonants at the beginning, or are you stuck with just one? The switches are all there when you start out - they’re part of the mental framework you’re born with - but how you set them depends on the languages that are around you while you grow up. Let’s look at a couple of different Parameters UG gives us. So, when you’re making your sentences, maybe you want your verb at the beginning of the verb phrase, or maybe you want it at the end. So that’s the difference between “move to Tokyo” and 東京に引っ越す. Languages like English and Malagasy like to put the verb first, whereas language like Japanese and Turkish like to put it at the end. Cool. So there’s this switch that gets set depending on whether you’re learning Malagasy or Turkish, that determines how you treat your verbs. Fantastic. But wait! The amazing thing about parameters is, once you set one thing, a whole bunch of other stuff falls into place. So if a language places its verbs at the beginning of the verb phrase, it should do the same with all its other parts of speech: so nouns at the beginning of the noun phrase, adjectives at the adjective phrase, and etc. So “furniture from the vintage store” vs ヴィンテージの店からの家具.  A syntax tree in Japanese will be pretty much the mirror image of one in English! Or, how about whether a sentence needs a subject? In French, if you want to proclaim your love for a certain auditory experience, you might say “J’aime la musique”. “Aime la musique” doesn’t mean the same thing, right? But in Italian, “Amo la musica” is totally okay. That’s because Italian just lets you drop the subject, whereas French and English don’t. And again, this simple parameter has some interesting consequences! Say you want to tell your friend Yasu that it’s snowing in London. You say, “It is snowing in London”. But if you want to impress Yasu with your rad polyglot skills, and break out your Italian, you might say “Sta nevicando a Londra”. In English, you need that “it” there, because you need subjects every single time! So even if you don’t really need one when you think about it - I mean, there’s no subject doing the snowing, right - your parameter doesn’t make exceptions. So it sticks a word in there that doesn’t even mean anything! It just can’t imagine life without subjects. Now, the Principles and Parameters approach is just one way of framing UG. There are also things like the Minimalist Program, which we’ll talk about in the future. But it’s a really good vantage point to start understanding how amazing human languages are, and how similar they are in spite of their surface differences. We could have way more variation than we do! That we all follow some basic rules, and don’t stray outside the lines, is some strong evidence for the innateness of language. Starting from when we’re little babies, we’re good at following the language blueprints. So we’ve reached the end of the Ling Space for this week. If our Subjacency violations didn’t break your mind, you learned that our linguistic system is defined by a set of principles and parameters; that principles like Subjacency are invariant rules that apply across all natural languages; that parameters are like binary choices about how your language can work; and that a single parameter can be responsible for multiple changes in how a language behaves. The Ling Space is produced by me, Moti Lieberman. It’s directed by Adèle-Elise Prévost, and it’s written by both of us. Our production assistants are Georges Coulombe and Stephan Hurtubise, our music and sound design is by Shane Turner, and our graphics team is atelierMUSE. We’re down in the comments below, or you can bring the discussion back over to our website, where we have some extra materials on this topic. Check us out on Tumblr, Twitter and Facebook, and if you want to keep expanding your own personal Ling Space, please subscribe. And we’ll see you next Wednesday. Vi ses!