10 2.3 Describing Speech Sounds: the IPA

Check Yourself

Video Script

In the first part of this book, we’re concentrating on the sounds of human speech. You might have already noticed that there’s a challenge to talking about speech sounds — English spelling is notoriously messy.

Take a look at these words:

say, weigh, they, rain, flame, lei, café, toupee, ballet

All of them contain the same vowel sound, [e], but the sound is spelled with nine different combinations of letters. Some of them are more common ways than others of spelling the sound [e], but even if we take away the ones that English borrowed from other languages, that still leaves five different ways of spelling one sound. One of the problems is that English has only five letter characters that represent vowels, but more than a dozen different vowel sounds. But it’s not just the vowels that are the problem.

English has the opposite problem as well. Take a look at these words:

cough, tough, bough, through, though

Here we’ve got a sequence of four letters that appear in the same order in the same position in each word, but that sequence of letters is pronounced in five different ways in English. Not only can a single sound be represented by very many different spellings, but even a single spelling is not consistent with the sounds that it represents.

Even one letter can be pronounced in a whole lot of different ways. Look at:

cake, century, ocean, and cello

The letter “c” represents four quite different sounds. Clearly, English spelling is a mess. There are a lot of reasons for why that might be.

The area where English first evolved was first inhabited by people who spoke early forms of Germanic and Celtic dialects. But then Normans invaded and brought all kinds of French and Latin words with their spellings. When the technology to print books was invented, there was influence from Dutch. So even the earliest form of English was influenced by many different languages.

Modern English also borrows words from lots of languages.  When we borrow words like cappuccino or champagne, we adapt the pronunciation to fit into English but we often retain the spelling from the original language.

Another factor is that the English spelling system was standardized hundreds of years ago when it became possible to print books. A lot of our standard spellings became consistent when the Authorized Version of the Bible was published in the year 1611. Spelling hasn’t actually changed much since 1611, but English pronunciation sure has, so the way we produce the sounds of English has diverged from how we write the language.

Furthermore, English is spoken all over the world, with many different regional varieties. British English sounds quite different from Canadian English, which is different from Australian English, and Indian English is quite different again, even though all of these varieties are spelled in nearly the same way.

There’s even variation within each speaker of English, depending on the context:  the way you speak is going to be different depending on if you’re hanging out with your friends or interviewing for a job or talking on the phone to your grandmother.

The important thing to remember for our purposes is that everyone who knows a language can speak and understand it, and children learn to speak and understand spoken language automatically. So in linguistics, we say that speaking and listening are the primary linguistic skills. Not all languages have writing systems, and not everyone who speaks a language can read or write it, so those skills are secondary.

So here’s the problem: as linguists, we’re primarily interested in speech and listening, but our English writing system is notoriously bad at representing speech sounds accurately. We need some way to be able to refer to particular speech sounds, not to English letters. Fortunately, linguists have developed a useful tool for doing exactly that. It’s called the International Phonetic Alphabet, or IPA. The first version of the IPA was created over 100 years ago, in 1888, and it’s been revised many times over the years. The last revision was fairly recent, in 2015. The most useful thing about the IPA is that, unlike English spelling, there’s no ambiguity about which sound a given symbol refers to. Each symbol represents only one sound, and each sound maps onto only one symbol. Linguists use the IPA to transcribe speech sounds from all languages.

When we use this phonetic alphabet, we’re not writing in the normal sense, we’re putting down a visual representation of sounds, so we call it phonetic transcription. That phonetic transcription gives us a written record of the sounds of spoken language. Here are just a few transcriptions of simple words so you can begin to see how the IPA works.

snake [snek]
sugar[ʃʊɡəɹ]
cake[kek]
cell[sɛl]
sell [sɛl]

Notice that some of the IPA symbols look like English letters, and some of them are probably unfamiliar to you. Since some of the IPA symbols look a lot like letters, how can you know if you’re looking at a spelled word or at a phonetic transcription?  The notation gives us a clue: the transcriptions all have square brackets around them.  Whenever we transcribe speech sounds, we use square brackets to indicate that we’re not using ordinary spelling.

You can learn the IPA symbols for representing the sounds of Canadian English in the next unit. For now, I want you to notice the one-to-one correspondence between sounds and symbols. Look at those first two words: snake and sugar. In English spelling, they both begin with the letter “s”.  But in speaking, they begin with two quite different sounds. This IPA symbol [s] always represents the [s] sound, never any other sound, even if those other sounds might be spelled with the letter “s”.  The word sugar is spelled with the letter “s” but it doesn’t begin with the [s] sound so we use a different symbol to transcribe it.

So, one IPA symbol always makes the same sound.

Likewise, one sound is always represented by the same IPA symbol.

Look at the word cake. It’s spelled with “c” at the beginning and “k-e” at the end, but both those spellings make the sound [k] so in its transcription, it begins and ends with the symbol for the [k] sound. Likewise, look at those two different words cell and sell. They’re spelled differently, and we know that they have different meanings, but they’re both pronounced the same way, so they’re transcribed using the same IPA symbols.

The reason the IPA is so useful is that it’s unambiguous: each symbol always represents exactly one sound, and each sound is always represented by exactly one symbol. In the next unit, you’ll start to learn the individual IPA symbols that correspond to the sounds of Canadian English.