 have probably heard of the phenomenon of tonal languages. Languages where you need to say things with just the right tone, otherwise you'll end up saying something completely different. The most famous tonal language is Chinese, but this category also includes plenty of other languages, especially in Southeast Asia and Africa. In these languages, controlling the pitch of your voice while you're saying something is just as important as getting the consonants and vowels right. For instance, in Chinese, ma means horse, while ma means mother. Compare that to English, where horse, horse, horse, horse, all of these just mean horse. So there you go. If a language uses pitch, it's tonal, and if it doesn't, it's not. Super simple. Until you start to think about it for even like a couple minutes, at which point everything breaks. Because here's the thing, tone might not change the meaning of this particular word, but we do use tone to communicate meaning. In fact, I can think of at least three different ways to communicate different things in English using tone. Firstly, you can use pitch to emphasize certain words. He didn't steal the money, he didn't steal the money, and he didn't steal the money. These three senses communicate completely different ideas, and the only difference between them is the tone I set them in. Secondly, you can use tone to signify questions. Compare you stole the book to you stole the book. The first might not technically be a question, but at the very least, it does make it sound like I'm expecting an answer, whereas the second one doesn't. Lastly, it seems like defenses in tone is the main way we figure out which syllable the stress of a word falls on, and we use stress to distinguish between different words. Permit is a noun, while permit is a verb, and you can actually experimentally show that the main way we tell the difference is by looking at which syllable has the higher pitch. Situations like these are rare in English, but there are more, like incline versus incline, or intern versus in turn. If you're anything like me, you might easily look at all of this and get the impression that English is actually a tonal language, or at least that this whole thing is a lot more complicated than people generally let on. Well, I'm here to tell you that it's the second one, but people don't usually like to go into the details because it gets really confusing and difficult to talk about really quickly. So, on that note, let's dive in! Buckle up, everyone, because we are about to get technical. Let's start by defining some terms, because I've been using the word tone kinda sloppily this whole video. On a piano, which key you press determines the pitch of a sound, while how hard you push it controls the volume. Pitch is basically the frequency of the sound wave while volume is the amplitude. Now, you might notice you can play a note on a piano and then play the same note at the same volume on a violin, but they'll still sound different. This is because of timbre, which for our purposes is basically the texture of the sound. Pause the screen now if you want to get into the physics of it, but for the rest of you, we can just move on. Now, almost all of the time in language, it's the timbre of the sound that's communicating the information, which is why you can take an audio recording and raise the pitch or make it louder or quieter, and most of the time it'll mean exactly the same thing as it did before. But as we noted earlier, with these three things, pitch also sometimes communicates information. So if your definition of a tonal language is just a language that incorporates pitch somehow, then yes, English would be a tonal language, along with probably every language ever. But usually when linguists say that a language is tonal, they mean that it uses pitch to communicate lexical information. All that means is that pitch is used to distinguish one word from another. Pink and fluffy might mean something different from pink and fluffy, but those differences are post-lexical, above the level of the word. The pitch changes things like which word is more important and whether the phrase as a whole is a question or not. It doesn't change what words there are or what order they're in, which lets us deal with these two, but there's still a case to be made that English is tonal because of accent. What I mean by accent is that within a word, some syllables feel more prominent or important than others. In permit, the accent is on the first syllable, while in permit, it's on the second one. It looks a lot like the main way we tell which syllable has the accent is the fact that the accented syllable has a higher pitch, so English uses pitch to distinguish between words, so that must make it tonal. However, there are two huge differences between what English does and what true tonal languages do. The first is that pitch isn't always the main way we tell where the accent is in a word. Thing is, in English, when a word is the focus of a whole sentence, we usually give the accented syllable of that word higher pitch, and when we say permit and permit, all on their own like that, we're kind of saying them as if they're making up a whole sentence all on their own, so of course they're going to have the focus, so of course we're going to make the accented syllable higher pitched. But listen to the way I say them in context when they don't have the sentence's focus. I didn't permit that. I didn't give them the permit. And if we cut those words out of the surrounding recording and play them back, permit, permit, suddenly it doesn't sound like pitch is the main thing we're using to tell them apart. In this context, you can show that it's actually the relative length of the two syllables, or which of them you spent more time pronouncing, that's doing most of the work of getting across which syllable has the accent. That length difference is there whether or not the word is the focus of the sentence, which makes it seem like it's actually the main signifier of stress and not pitch. The second reason English accent doesn't count as tone is that English accent is syntigmatic rather than paradigmatic. I told you we were gonna get technical. All right, let me try to explain. If English really had tones, then it would be natural to ask, okay, how many tones does it have? And the answer would seem to be two, one for when a syllable is accented and one for when it's unaccented. But that would imply that a two-syllable word could have one of four different tone patterns, one where the first one's accented, one where the second one's accented, one where both are and one where neither are. But that's not allowed. Each word needs to have exactly one accented syllable, no more, no less. Permit and permit are both valid words, but permit and permit aren't. This makes it look like English accent isn't really an innate quality of the syllable, but rather which syllable is accented is a property of the word. If the accentedness of each syllable varied independently, then we could call English accent paradigmatic, but the fact that they can't and that each word has exactly one accent means that it's syntigmatic. With this in mind, we can say that a tonal language is a language that makes paradigmatic, lexical distinctions based primarily on pitch. And English pretty clearly fails that definition. In fact, there are cases of languages that are even closer to being tonal, but still don't meet this bar. Japanese has a syntigmatic accent like English, but unlike English, this accent is primarily distinguished using pitch. Languages like these are called pitch accent languages, because, well, they mark accents with pitch. So I hope that clears up why English isn't tonal, but for me at least, all this kind of raises more questions than it answers. Like, if English uses pitch to communicate post-lexical information, like which word is most important or whether the sentence is a question, then can tonal languages not do that? Like, if Chinese is already using pitch so much to distinguish between different words, it sounds like they wouldn't be able to use pitch to emphasize words or mark questions like English speakers do. But that's not actually true. It works differently than in English, but Chinese uses pitch to communicate both lexical and post-lexical information. One way that linguists used to think about this was basically in terms of adding functions. You start with a function for the overall pitch of the whole sentence, which communicates post-lexical information, and then you get another function that describes the tone of each word, and then you just add the two functions together to get the way pitch will vary over the course of a sentence. This is called the overlay model, because you're sort of overlaying different things that affect pitch. And it's pretty similar to the way a lot of people subjectively experience how pitch works in tonal languages. I think it's also a good way to introduce someone to the idea that in tonal languages, you can use pitch both to distinguish between words and to communicate all the stuff the rest of us use it for. Only problem is, it's kind of wrong. Thing is, if you actually use the overlay model to make predictions about how the pitch of someone's voice will change over the course of a sentence, and then you go out and test it, the results aren't great. It took linguists a long time to figure that out, in large part, because it hasn't been very long since the equipment necessary to measure this kind of thing objectively was invented. But once it was invented, it wasn't long before there were a lot of tech companies who wanted to make computers that could communicate with humans with normal spoken language. And these tech companies suddenly got extremely interested in getting this kind of thing right. So all of a sudden, there was both the means and the pressure to do some actual science to this area of linguistics. And the main result of that was the auto-segmental metrical theory. Whereas the overlay model makes it sound like the speech centers of our brains are generating two or more functions than adding them together to produce the pitch we want to make, the AN theory says that it's much more useful to think of a chunk of speech as containing a linear string of tonal events. Each language will have a limited number of possible tone events, rules for how tone events are actually realized on the pitch of someone's voice, and rules for which tone events happen when, depending on what words we're saying, the syntactic structure that they're in, and whatever post-flexical information we want to communicate. These rules might be very complicated, and they're going to be different in different languages, but this basic model has proven to be very useful for modeling how different languages deal with pitch differently. For instance, with Mandarin Chinese, some words in a sentence will be more important than others, and a lot of the time this is marked by an exaggeration of the tone it would have otherwise. Low tones get lower and high tones get higher. This can be described within the AM framework reasonably easily. You can just say that the string of tone events is different depending on what lexical tones there are, and also depending on which word has the focus, and that when the tone events mark a word as having the focus, it winds up getting realized as an exaggerated form of the normal tone. I'm not entirely sure how one would explain this with an overlay model. There's a lot more I could get into. Each language has its own unique system for combining lexical and post-lexical information to create variations in pitch and volume and timing, but I hope now the idea of tonal languages at least makes a bit more sense to you. See you soon for more linguistics videos!