Six English Pronunciation Mistakes You Can Fix Without a Teacher
Most pronunciation advice assumes you have someone listening. A lot of what holds learners back is not a sound you cannot make; it is a rule nobody told you, and rules you can fix on your own. These six are the ones with clear answers.
There are two kinds of pronunciation problem and they need completely different work. One is physical: a sound your language does not have, which takes a mouth, a mirror and usually a person to tell you when you got it. The other is informational: a rule with a definite answer that nobody happened to teach you.
The second kind is worth attacking first, because you can do it alone and because it accounts for more of what makes speech hard to follow than any individual consonant. All six below are rules.
The six at a glance
Each row is the problem, the version learners produce, and what English does instead.
| Problem | Often said | English does | The rule |
|---|---|---|---|
| Word stress | PHOtograph, phoTOgrapher the same | PHOtograph, phoTOgrapher | Stress moves when the suffix changes. |
| -ed endings | want-ed, work-ed, play-ed all alike | wantID, workT, playD | Three pronunciations, decided by the sound before. |
| Plural and third-person s | books and dogs alike | bookS, dogZ, watchIZ | Three pronunciations, same logic as -ed. |
| Silent letters | kuh-nife, comb with a b | nife, coam | Whole letters are simply not pronounced. |
| Weak forms | TO the shop, saying to fully | tuh the shop | Small words reduce to a schwa in normal speech. |
| Sentence stress | Every word given equal weight | Content words stressed, the rest reduced | English is stress-timed, not syllable-timed. |
1. Word stress, and why it matters more than any single sound
English words have one strong syllable, and putting it in the wrong place is the single fastest way to become hard to understand. Listeners identify a word partly by its stress pattern, so a correctly pronounced word with the wrong stress can be genuinely unrecognisable in a way that a slightly wrong vowel is not.
The good news is that a lot of it is predictable from the ending. Words ending in -tion, -sion, -cian and -ity stress the syllable immediately before the ending: inforMAtion, deCIsion, elecTRIcian, posSIbility. Words ending in -ic and -ical do the same: phoTOgraphic, eCOnomic, hisTOrical. Words ending in -ate, -ise and -ify as verbs usually stress the first syllable: EDucate, ORganise, IDentify.
The pattern to watch is a family of related words where the stress moves. PHOtograph, phoTOgrapher, photoGRAPHic are three different patterns from one root, and learning the root does not give you the other two. When you save a word, save which syllable is strong along with it.
2. The three ways to say -ed
Regular past tense has one spelling and three pronunciations, chosen entirely by the sound that comes before it. Not the letter, the sound.
After a t or d sound, -ed is a full extra syllable: wanted, needed, started, decided. After any other voiceless sound, it is just a t: worked, watched, hoped, laughed, passed. After any other voiced sound, including all vowels, it is just a d: played, lived, called, opened, arrived.
The error that matters here is adding a syllable everywhere: work-ed, play-ed, call-ed. It is immediately audible, it is one of the clearest markers of a non-native speaker, and it disappears once you know that only t and d earn the extra beat.
- wanted, needed, started, decided: an extra syllable, because of the t or d before
- worked, watched, hoped, laughed: just a t sound, no extra syllable
- played, lived, called, opened: just a d sound, no extra syllable
3. The same three-way split on plural s
Plurals and third-person verbs work the same way, and once you see the parallel you get both for the price of one. After a voiceless sound the s is a hiss: books, cats, laughs, stops. After a voiced sound it is a buzz, which is really a z: dogs, cars, plays, lives, runs. After a hissing or buzzing sound it becomes a full extra syllable: watches, boxes, buses, changes.
Most learners do this correctly by instinct and then get caught on the third group, because the spelling gives no warning. Watch becomes watch-iz, not watch-s. So do bridges, prizes, wishes and pages.
4. Letters that are simply not there
English spelling records how words sounded several centuries ago, and a number of letters have gone silent since. There is no way to work these out; they are learned one at a time, and they are worth learning because pronouncing them is far more disruptive than an imperfect vowel.
Two caveats, because English pronunciation is rarely universal. Often has two standard pronunciations and you will hear both, so it is not a reliable member of that group. And the wh- words keep a genuine h sound in Scottish, Irish and some American accents, where which and witch are not homophones. Neither is a mistake; both are worth recognising when you hear them.
| Silent letter | Words | Sounds like |
|---|---|---|
| k before n | knife, know, knee, knock | nife, no, nee, nock |
| b after m | comb, climb, thumb, lamb | coam, clime, thum, lam |
| l before certain consonants | walk, talk, half, calm, should | wawk, tawk, haf, cahm, shud |
| t in these | listen, castle, whistle, fasten | lissen, cassel, wissel, fassen |
| h after w (most accents) | what, when, where, white | wot, wen, wair, wite |
| gh | thought, through, daughter, eight | thawt, throo, dawter, ate |
5. Weak forms, which are why you cannot hear the small words
This is the one that most affects listening, and the two problems are the same problem. In normal English speech the small grammatical words are not pronounced the way they look. They collapse to a neutral vowel, the schwa, and they get very short.
To becomes tuh. For becomes fuh. Of becomes uh. And becomes n. Can becomes cun. Have becomes uv. Was becomes wuz. Them becomes uhm. A sentence like "I have to go to the shop and get a few things" contains almost no fully pronounced small words when a native speaker says it at normal speed.
Learners who pronounce every small word fully sound careful and slightly robotic, and, more importantly, they do not recognise the reduced versions when they hear them. That is the real explanation for understanding a textbook recording and understanding nothing when two people talk to each other, which is covered from the listening side in how to learn vocabulary from podcasts.
6. Rhythm: English is stress-timed
Many languages give each syllable roughly the same length. English does not. It gives roughly equal time between stressed syllables and squeezes whatever falls in between, which is why the weak forms above exist at all.
In practice that means the content words carry the beat, and everything else compresses to fit. In "I would have TOLD you if I had KNOWN", two words are strong and the rest is squeezed into the gaps. A learner who gives all nine words equal weight is speaking a different rhythm, and rhythm is a large part of what makes an accent hard to follow, quite separately from individual sounds.
The practical exercise is small: take one sentence from something you are listening to, mark the two or three strong words, and say it several times keeping those beats even while compressing everything else. It is not glamorous, and it transfers faster than almost anything else on this list.
How to work on this alone
You do not need a teacher for any of the six above, but you do need to hear yourself. Recording thirty seconds on your phone and listening back is uncomfortable and it is the whole method: the gap between what you think you said and what you actually said is where the work is.
One thing to avoid: slowing audio down. It makes comprehension easier and pronunciation worse, because the weak forms and the rhythm are the first things to disappear when speech is slowed, and they are exactly what you are trying to learn.
- Record thirty seconds, listen back, pick one thing. Not five.
- Shadow a short clip: play a sentence, say it immediately over the top, repeat until the rhythm matches.
- When you save a new word, save its stress pattern with it.
- Listen at normal speed. Slowing it down removes the thing you are trying to hear.
Where MindDory fits
Stress patterns and silent letters are per-word facts, which means they belong with the word rather than in a separate list you will never reopen. Whatever you use to keep vocabulary, keep the stressed syllable in the same place as the meaning, so that the two come back together. That is what MindDory does with the words you save: the item you review is the word in the context you met it, returned when it is closest to slipping rather than when you last added it.
The free B1 English vocabulary list is a reasonable place to test yourself on this: go down it and mark the words whose stressed syllable you are not sure about. That list is the work.