You have seen it a thousand times and slid right past it. You look up a word, and there next to the spelling, tucked inside a pair of little slanted lines, sits a row of symbols: /ˈwɔtər/, /əˈbaʊt/, /ˈvɪʒən/. A few of the letters you recognize. A few of them look like they wandered in from a math textbook. So you do what almost everyone does, which is skip the symbols and guess the pronunciation from the spelling, which is the one thing those symbols were put there to stop you from doing.
That little row is the International Phonetic Alphabet, and it is the only writing system that tells you the truth about how an English word sounds. Every article on this blog leans on it, because there is no honest way to talk about pronunciation in a language whose spelling stopped tracking its sounds five hundred years ago. This is the page that teaches you to read it. The reassuring part, the part nobody mentions when they hand you a wall-sized chart with a hundred and sixty symbols on it, is that you need almost none of that chart. You need the few dozen symbols General American uses, and inside that set, maybe ten that aren’t what they look like.
The IPA is an alphabet with one rule: each symbol stands for exactly one sound, and that sound never changes. That is the whole reason it exists, and the whole reason it can do what English spelling can’t. You don’t need the full international chart. American English uses around forty of its symbols: roughly twenty-four consonants and fifteen or sixteen vowels. Most of the consonants are the letter you’d already guess, so the real work is small. About ten symbols genuinely surprise people: /j/ is the y in yes, not a j; /ʃ/ is sh; and the schwa /ə/, that upside-down e, is the most common sound in the language. Two little marks for stress finish the set. Learn those, and a dictionary entry stops being decoration and starts being instructions.
Why spelling lies, and IPA doesn’t
Start with the word that names this whole company. Water, spelled with a t in the middle and an -er on the end, comes out of an American mouth as waa-der: the t has softened to a quick tap that sounds like a d, and the -er is one r-colored vowel, not a vowel plus a separate r. Nothing in the five letters w-a-t-e-r told you any of that. The spelling is a historical artifact, frozen around the time of the printing press and never updated, and it has been quietly lying to learners ever since.
The lie runs in both directions, which is what makes it so hard to reason around. One spelling gives you many sounds: the letters ough are a different vowel in though, through, tough, thought, and bough, five words built from the same four letters and not one of them rhyming with another. And one sound hides behind many spellings: the long ee in see is also spelled ea in sea, ie in field, i in machine, ey in key, and eo in people. Six costumes, one sound. If you try to read English pronunciation off the letters, you will be wrong most of the time, and the worst part is you won’t know which times.
IPA was built to close exactly this gap. Its founding rule is that the relationship between symbol and sound is fixed: one symbol always writes one sound, and one sound is always written by the same symbol, no matter how the ordinary spelling chooses to dress it up. /i/ is the ee sound whether the word is see or machine or people. That single property is the cheat code. Once you can read it, every dictionary entry, every word in this blog, every transcription you ever meet becomes a precise instruction for your mouth instead of a riddle.
One symbol, one sound — and just the ones you use
The full International Phonetic Alphabet is genuinely enormous, because it was designed to write every sound in every human language: clicks, tones, consonants made deep in the throat that English has never used. That is why the chart on the wall looks intimidating. It is a global tool, and you are learning one accent of one language.
General American draws on roughly forty of those symbols. The exact count is a little soft, for reasons we’ll get to, but the shape is firm: about twenty-four consonants and somewhere around fifteen or sixteen vowels, depending on how you count the ones that merge. That is the entire working set. Set the rest of the chart aside and you are left with something a person can actually learn in an afternoon.
It gets easier still, because the symbols aren’t randomly distributed in difficulty. The large majority of the consonants are the plain Latin letter you would have guessed: /b/ is the b in bed, /p/ the p in pen, /m/ the m in man. Those you already know; you just didn’t know you knew them. The trouble is concentrated in a small, nameable group, a handful of consonant symbols that look like one thing and say another, plus the vowels, where the symbols and the spelling part ways almost entirely. Knowing where the difficulty lives is half the battle. You can stop bracing for forty surprises and brace for ten.
The consonants are mostly the alphabet
Here is the full American consonant set. Read down the last column first, the everyday sound each symbol stands for, and notice how many of them are simply the letter itself. Each symbol links to a fuller lesson on that sound if you want to drill it; for now, the goal is just to recognize the glyph and know what noise it makes.
| Symbol | As in | The everyday sound |
|---|---|---|
| /p/ | pen | p |
| /b/ | bed | b |
| /t/ | ten | t |
| /d/ | den | d |
| /k/ | kit | k (also spelled c, ck, q) |
| /g/ | get | hard g |
| /f/ | fan | f (also ph) |
| /v/ | van | v |
| /θ/ | think | voiceless th |
| /ð/ | this | voiced th |
| /s/ | sun | s |
| /z/ | zoo | z |
| /ʃ/ | ship | sh |
| /ʒ/ | vision | the si in vision |
| /h/ | hat | h |
| /tʃ/ | chip | ch |
| /dʒ/ | job | j (also soft g) |
| /m/ | man | m |
| /n/ | net | n |
| /ŋ/ | sing | ng (one sound) |
| /l/ | let | l |
| /r/ | red | American r |
| /w/ | wet | w |
| /j/ | yes | y |
Twenty-four rows, and you can see at a glance that most of them caused you no trouble at all. Sixteen of these are the plain letter, and you could have produced that table yourself. The work isn’t twenty-four new symbols. It’s the handful in the middle of the chart where the glyph stops being the letter you expect, and those deserve their own section.
A couple of small things to notice before we get there. The symbol /r/ is the standard dictionary spelling of the American r, though some phonetics books write it /ɹ/ to be fussy about the fact that the English r isn’t trilled; you’ll see both, and they mean the same sound. And a few of these symbols pull double duty against spelling: /k/ writes the hard c in cat and the ck in back and the q in quick, because the sound is the same no matter which letter the spelling reached for. That is the rule doing its job. The sound decides the symbol; the spelling gets no vote.
The ten symbols that surprise everyone
This is the short list worth memorizing, the symbols where reading them by instinct will lead you wrong. Ten of them carry almost all the confusion in the whole system.
The single most misread symbol in English is /j/. It is not a j. It is the y sound at the start of yes, year, and yellow. So cute is written /kjut/ and few is /fju/, that little /j/ capturing the tiny y-glide you make before the oo. If your first language is German, Dutch, or a Scandinavian one, this will feel obvious, because your j already says y. If it’s English, Spanish, French, or Portuguese, it’s a trap every single time, and the only cure is to overwrite the instinct on purpose.
Then there is the cluster of h-spelled English sounds that get their own clean single symbols, because IPA refuses to write one sound with two letters. /ʃ/ is sh, the sound in ship, and also the hidden sound inside machine, sure, and every -tion ending. Its voiced partner /ʒ/ is the soft buzz in the middle of vision and measure, and at the end of garage, a sound English never gave a letter of its own. The two ch/j sounds are written as little two-symbol teams that show what they’re made of: /tʃ/ is ch as in chip, literally a /t/ released into a /ʃ/, and /dʒ/ is the j in job and the soft g in gym, a /d/ released into a /ʒ/. This is the one place among the consonants where a single sound gets two symbols, and it’s deliberate: the spelling is telling you the recipe.
The two th sounds are the famous pair, because English writes both of them with the same two letters and IPA splits them apart. /θ/ is the breathy, voiceless th in think and bath; /ð/ is the buzzing, voiced th in this and mother. Put a hand on your throat and the difference is plain, the second one hums and the first one doesn’t, and they’re the reason the TH sounds get a whole article. One more nasal hides a similar trick: /ŋ/, the ng in sing, is a single sound made at the back of the mouth, not an n followed by a g. The full story of why singer has no real g in it lives in the NG article.
That accounts for the consonants. The last two on the memorize-this list aren’t full consonants at all. One is the schwa, /ə/, an upside-down e that writes the small uh in the first syllable of about and the last of sofa; it earns its own section below because it is everywhere. The other is the flap, /ɾ/, the quick tongue-tap that the t in water turns into. You won’t see the flap in most dictionary entries, which keep the tidy /t/, but you will see it on this blog and in careful phonetic transcriptions, and it’s worth knowing on sight because it’s the symbol behind the flap-T and behind the name SayWaader itself, /ˈwɑɾɚ/.
The vowels, where the symbols stop matching the letters
If the consonants came mostly free, the vowels are where the work is. English has five vowel letters and roughly fifteen or sixteen vowel sounds, so the symbols here can’t possibly match the spelling, and they don’t. There are simply more sounds than letters, and IPA needs a distinct glyph for each one. This table shows the fourteen everyday vowels and diphthongs; the schwa from the next section and the stressed r-colored vowel round the count up to fifteen or sixteen. The full chart, with the mouth positions and the tense-versus-lax pairs that trip up most learners, is the job of the American vowels article, and you should read it next if the vowels are your weak spot.
| Symbol | As in | Watch out |
|---|---|---|
| /i/ | see | the long ee; not the letter i |
| /ɪ/ | sit | the short, lax ih; the one ship needs so it isn’t sheep |
| /ɛ/ | bed | the eh in bet |
| /æ/ | cat | the flat American a; many languages don’t have it |
| /ɑ/ | father | the wide-open ah in hot |
| /ɔ/ | saw | the rounded aw; many Americans merge it into ah |
| /ʊ/ | book | the short uu; not the oo of moon |
| /u/ | moon | the long oo |
| /ʌ/ | fun | the stressed uh in cut; the schwa’s louder twin |
| /eɪ/ | day | starts just above eh, glides up to ee |
| /aɪ/ | my | glides from ah up to ee |
| /ɔɪ/ | boy | the oy glide |
| /aʊ/ | now | the ow glide |
| /oʊ/ | go | glides from oh toward oo; not a flat o |
Two things in that table do most of the damage, and both are worth flagging now even though the vowels article handles them in full. The first is that several symbols come in tense-and-lax pairs that English keeps apart and many languages don’t: /i/ against /ɪ/ (sheep against ship), /u/ against /ʊ/ (fool against full). When a dictionary shows you /ɪ/ and not /i/, it is telling you, precisely, to use the short relaxed vowel, and that distinction is invisible in the spelling. The second is that the last five rows are diphthongs, vowels that move while you say them, which is why each gets written with two symbols: a starting point and the direction it glides. The two letters aren’t two sounds. They are one sound in motion.
A note on the r-colored vowels, the ones in car, bird, and the endless -er of mother. Dictionaries write them as a vowel plus /r/, so bird shows up as /bɜrd/ and mother ends in /ər/, that schwa-with-an-r you’ll meet again in a moment. In a rhotic accent like American English the r fuses right into the vowel, and the tongue mechanics are the same problem as the consonant r; the full treatment is in the American R. For reading purposes, just know that an /r/ sitting after a vowel inside a syllable means the vowel is colored by it, not that you pause and add a separate sound.
Stress marks and the schwa’s other job
Two small marks in a transcription do an outsized amount of work, and most learners read straight past them. They’re the stress marks, and English lives or dies on stress.
The first is a raised vertical tick, /ˈ/, and it sits before the syllable that carries the main beat. The second is a lower tick, /ˌ/, marking a weaker secondary stress in longer words. So record the noun is /ˈrɛkərd/, beat on the front, and record the verb is /rɪˈkɔrd/, beat on the back. Moving that one floating tick drives the whole change: it decides which syllable stays full and which one reduces, so the vowels shift right along with the stress. The mark looks like a stray apostrophe, which is exactly why people ignore it, and ignoring it is how you end up stressing the wrong syllable in a word you otherwise pronounce perfectly. The deeper logic of why English does this is the subject of word stress; here, just promise yourself you’ll stop skipping the tick.
Stress is also where the schwa earns its keep, and the schwa is worth a second look because it is, by raw frequency, the most common sound in English. The symbol /ə/ writes a small, neutral uh made with the mouth completely relaxed, and its job is to be the sound a vowel collapses into once it loses its stress. Watch it happen: photograph keeps a full, clear /æ/ in its last syllable, FOH-tuh-graf, but in photography the stress jumps to the second syllable and that same vowel dissolves into a schwa, fuh-TAH-gruh-fee. The letters didn’t change. The stress moved, and the unstressed vowels caved to schwa. This is why a transcription of a long English word is so often a couple of clear vowels surrounded by a crowd of /ə/’s: in real speech, only the stressed syllables keep their full vowel, and everything else reduces. A schwa in a dictionary entry marks a syllable that should come out weak, fast, and swallowed; fighting to give it a full, pure vowel is the surest way to sound non-native. The schwa has its own article, because this one habit, reducing the right syllables, separates fluent rhythm from textbook rhythm more than any single consonant ever will.
Reading a dictionary entry, left to right
Put it together on a real entry. Look up water in a good American dictionary and you’ll see something close to /ˈwɔtər/. Read it the way you’d read a word, left to right, one symbol at a time. The opening /ˈ/ says the stress is on the syllable that follows, so the first syllable is the strong one. /w/ is just w. /ɔ/ is the aw vowel, the rounded one, though plenty of Americans will use the ah of father here instead, and the dictionary is showing you one standard choice. /t/ is a t on the page. Then /ər/, the schwa-with-r, the unstressed r-colored ending. String it together and you get the careful, dictionary-clean pronunciation.
What the dictionary entry doesn’t show you is what your mouth does next, and that gap is worth understanding. Out loud, an American doesn’t say a crisp /t/ there; the t sits between two vowels with the stress in front of it, so it flaps to that quick tap, and the real, running pronunciation is closer to [ˈwɑɾɚ], waa-der. The dictionary gave you the idealized version, the sounds the word is made of; the flap is a predictable thing that happens to those sounds in fast speech. The dictionary isn’t wrong here, just working at a coarser level of detail, and the next section is about the brackets that signal which level you’re looking at.
One more, to show the stress mark and the schwa working together. Banana is /bəˈnænə/: schwa, then the stressed /næ/ with a full bright a, then a final schwa. Three syllables, but only the middle one is loud and clear, and the two on the ends are both that same weak /ə/, even though the spelling writes all three with the very same letter a. Read the transcription and it tells you exactly that, buh-NAN-uh, which is precisely the information the spelling banana withholds. That is the whole skill: the entry is a set of instructions, and now you can follow them.
Slashes, brackets, and why dictionaries disagree
You’ll have noticed the symbols sometimes sit inside slashes, /ˈwɔtər/, and sometimes inside square brackets, [ˈwɑɾɚ]. The difference is real but gentle, and you don’t need to lose sleep over it. Slashes mean a phonemic transcription: the sounds a word is built from, the idealized blueprint, which is what dictionaries print. Square brackets mean a phonetic transcription: the actual, physical sound as it comes out, including the small automatic changes like the flap and the puff of air on a stressed p. Phonemic is the plan; phonetic is the performance. This blog uses slashes for the plan and brackets when it wants to show you the performance, the flap in water being the usual reason.
This also explains something that confuses people the first time they compare two dictionaries: they don’t always use the same symbols for the same word, and neither one is wrong. Some of the disagreement is about genuine variation in how Americans talk. The cot/caught pair is the big one, where many speakers have merged the /ɔ/ of caught into the /ɑ/ of cot, so one dictionary writes thought with /ɔ/ and another with /ɑ/, both describing real American speech; the cot–caught merger covers that one in full. Other disagreements are just house style: one book writes the American r as /r/ and another as /ɹ/; one writes the day vowel as /eɪ/ and another simplifies it to /e/; one marks the stressed -er with the compact symbol /ɝ/ and another spells it out as /ər/. These are the same sounds in slightly different handwriting. Once you know the core set, you can read past the cosmetic differences and see that the dictionaries agree about far more than they differ on.
Read these out loud
Reading IPA only becomes automatic by doing it, and there are two drills worth ten minutes. First, decode single words cold. Cover the right-hand column, read each transcription out loud from the symbols alone, then uncover the answer to check. The goal is to look at /ʒ/ or /j/ and hear it instantly, without translating in your head.
| In IPA | The word |
|---|---|
| /jɛs/ | yes |
| /ˈvɪʒən/ | vision |
| /θɪŋ/ | thing |
| /dʒɑb/ | job |
| /tʃip/ | cheap |
| /ˈsɪŋər/ | singer |
| /kjut/ | cute |
| /əˈbaʊt/ | about |
Then read the same trouble-sounds back inside full sentences. Say each line out loud, lean into the respelling under it, and play the audio to check yourself against a native voice. Listen first if you’d rather hear it before you try.
- Yes, the year is young. YES, the YEER is YUNG.
- Her vision is unusual. Her VIZH-un is un-YOO-zhoo-ul.
- I think this is the thing. I THINK THIS is the THING.
- The singer chose a job. The SING-er CHOHZ a JAHB.
- A photograph, not photography. A FOH-tuh-graf, not fuh-TAH-gruh-fee.
- The water in the bottle. The WAA-der in the BAH-dul.
- A few cute beauties. A FYOO KYOOT BYOO-teez.
- She bought a record to record. She bought a REK-erd to re-KORD.
If the y-symbol /j/ still makes you want to say a j, slow down on the first two lines until the override sticks. That one symbol causes more misreadings than the rest of the chart combined.
FAQ
IPA stands for the International Phonetic Alphabet, a writing system in which each symbol stands for exactly one sound and that sound never changes. Dictionaries use it because ordinary English spelling is an unreliable guide to pronunciation: the same letters spell different sounds (though, through, tough) and the same sound hides behind different letters (the ee in see, sea, machine). IPA closes that gap by giving every sound one fixed symbol, so a transcription like /ˈwɔtər/ tells you precisely how to say the word regardless of how it’s spelled.
Around forty, far fewer than the full international chart of roughly 160. General American uses about twenty-four consonant symbols and fifteen or sixteen vowel symbols. The load is lighter than it sounds, because most of the consonants are the plain letter you’d expect (/b/ is b, /m/ is m). The real work is a group of about ten symbols that don’t match their letters, plus the vowels, where the symbols and the spelling diverge almost completely.
Because IPA assigns each symbol a fixed sound drawn from how it’s used across many languages, and /j/ was assigned the y sound at the start of yes. In German, Dutch, and the Scandinavian languages the letter j already says y, so the choice felt natural to the system’s designers. It catches out anyone whose first language uses j for a different sound: /dʒ/ in English, as in job, /x/ in Spanish, and /ʒ/ in French and Portuguese. So cute is /kjut/ and you is /ju/: that /j/ is always a y.
That symbol is the schwa, /ə/, a short neutral uh made with a fully relaxed mouth. It marks an unstressed syllable whose vowel has reduced, and it is the most common sound in spoken English. When you see it, that syllable should be quick, weak, and swallowed, not pronounced with a full clear vowel. In /əˈbaʊt/ (about) the first syllable is a faint schwa and all the weight falls on the second; trying to give that first syllable a full a is a classic way to sound non-native.
Slashes mark a phonemic transcription, the idealized sounds a word is built from, which is what dictionaries print: /ˈwɔtər/. Square brackets mark a phonetic transcription, the actual sound as it leaves your mouth, including automatic changes like the flapped t: [ˈwɑɾɚ]. The phonemic version is the blueprint a word is built from; the phonetic version is what actually comes out when people talk. For a learner the slashes version is usually what you want, with the bracket version showing you what really happens in fast speech.
Two reasons, and neither means one is wrong. First, real variation in American speech: many speakers merge the vowels in cot and caught, so one dictionary writes thought with /ɔ/ and another with /ɑ/, each describing a real accent. Second, house style: the American r is written /r/ in some books and /ɹ/ in others, and the stressed -er vowel appears as both /ɝ/ and /ər/. These are the same sounds in different handwriting; once you know the core set you can read past the cosmetic differences.
Keep a pronunciation key open the next few times you look something up, and read the symbols before you trust the spelling. The forty-odd shapes stop looking foreign within a week or two of actually using them, and from then on the slanted brackets are the one part of a dictionary entry you can take at face value.
