How to Read English Words: Guess Pronunciation from Spelling

Vietnamese is written how it is read. The word "chuyên" has only one pronunciation, and you never have to look up a dictionary to know how a new Vietnamese word sounds. That habit follows you into English, then you hit colonel and break.

The most common story goes like this. You read a word in a document, understand the meaning, use it when writing, but in your head it has a pronunciation you built yourself from the spelling. When you hear someone say that exact word, you do not recognize it. You have the vocabulary, but your ear does not.

English spelling is irregular, but it is not random. There is a layer of rules stable enough to teach, and knowing them means when you meet an unfamiliar word you can guess a reasonable pronunciation instead of freezing or reading it the Vietnamese way. This article gathers those rules, with real examples and IPA transcription, and says straight out where the rules stop working.

• How predictable English spelling is
• Final e lengthens the vowel before it
• The letters c and g: soft or hard depends on what comes after
• Suffixes that shift stress elsewhere
• Suffixes that lock in pronunciation
• Vowel pairs, where the rules are loosest
• The ough group, where you have to memorize
• The endings -ed and -s follow the sound before them
• Words whose spelling deceives
• After guessing you still have to listen to know you got it right
• Frequently asked questions

How predictable English spelling is

English spelling has been more or less frozen since the printing press, while pronunciation kept moving for another few hundred years. That is the root of the mismatch between letters and sound, and the article on silent letters in English has the full history if you are curious.

The rest still has a system. The rules in this article do not work at the level of individual letters, but at the level of position and suffix: what letter c stands before, where e sits in the word, what cluster the word ends with. Learning by position means one rule covers hundreds of words at once. In exchange, none of the rules here is absolutely true, so treat each one as a hypothesis you need to check again.

Transcription in this article defaults to British English; where American English differs clearly, there is a note. If IPA symbols are still unfamiliar, the English pronunciation article has a full IPA chart and a list of sounds Vietnamese speakers often get wrong.

Final e lengthens the vowel before it

This rule is worth learning first because it hits the most basic words. Final e is not pronounced as its own sound, but it changes the vowel before it from short to long or to a diphthong.

The reverse mistake is common too: you see e and try to pronounce it, and a one-syllable word gets stretched into two. The word made has exactly one syllable, /meɪd/.

Then the rule runs out. A very common group of words has final e while the vowel before it stays short.

The group ending in -ve has a clear reason and works as a sub-rule: English does not write a word ending in v, so the e there is only padding. The come, some, done, gone group you have to remember separately; on the upside, those are words you meet every day.

The letters c and g: soft or hard depends on what comes after

The letters c and g each have two pronunciations, and the letter right after decides which one. Before e, i, y they are soft. Before a, o, u or a consonant they are hard.

With c, this rule has very few exceptions and you can treat it as a law. With g it fails more often, and the annoying part is the failures are all common words. The practical approach is to treat c as a rule, and g as a rule plus a short exception list you need to memorize.

The last row of the table is often skipped: two c's together where the second stands before e or i, you pronounce both, the first hard and the second soft. That is why accept has both /k/ and /s/.

Suffixes that shift stress elsewhere

Vietnamese pronounces syllables almost evenly; each syllable carries its own tone and none gets swallowed. English works differently: one syllable is stressed clearly, the rest shrink and their vowels often drop to /ə/. Vietnamese speakers used to reading evenly do not recognize the spelling they know when they listen.

Stress does not stay put within a word family either. Add a suffix and it moves, and the whole sound shape of the word changes with it. photograph is /ˈfəʊtəɡrɑːf/, stress on the first syllable. photographer is /fəˈtɒɡrəfə/, stress on the second, and the first vowel drops to a faint /ə/. photographic is /ˌfəʊtəˈɡræfɪk/, stress on the third. Three words from one root, three stress placements, and when you hear them they sound like three unrelated words.

The rule in one line: the suffixes -ic, -ical, -ity, -tion, -sion, -ial, -ian, -graphy, -logy pull stress to the syllable right before them.

A few exceptions worth remembering with -ic are Arabic, politics, arithmetic and lunatic. These words keep stress at the front instead of moving to right before the suffix.

Stress has one more kind of movement, between noun and verb spelled exactly the same, like record or export. That pattern has its own table in the English pronunciation article. How far unstressed syllables shrink when people speak fast belongs in connected speech.

Suffixes that lock in pronunciation

Some word endings sound the same in every word that carries them. Learn the seven endings below and you can handle the tail end of a lot of long words, and long words are exactly where learners hesitate most.

The ending -sion has two pronunciations you can tell apart by the letter right before it: after a vowel it is /ʒən/, after a consonant it is /ʃən/. The pair -sure splits the same way.

The ending -ous reads light and faint as /əs/, never /aʊs/. This mistake is very common and easy to understand, because ou elsewhere really is /aʊ/.

The ending -ate is different: you have to look at word class to read it right. Same spelling, noun and adjective read /ət/, verb reads /eɪt/.

The sentence "We need to separate the two files" and the sentence "They are separate files" spell separate the same way but read it differently. The reverse helps your ear too: hear /ˈsepəreɪt/ and it is the verb, hear /ˈseprət/ and it is the adjective.

Vowel pairs, where the rules are loosest

Two vowels side by side is where English is most chaotic. One cluster can yield three or four different sounds and the spelling gives no clue which to pick. The practical way to learn is to memorize the list of possibilities, then attach each possibility to a few words you already know.

The cluster ee is much more stable than the rest, reading /iː/ in most common words. The cluster ea has three main readings, and the pair read present /riːd/ with read past /red/ fits neatly into the first two.

The cluster oo has a small trick worth trusting: right next to k it reads short /ʊ/, like book, look, cook, took, shook, hook. Outside that case you have to remember word by word.

The cluster ow has one usable anchor: at the end of a two-syllable word it is usually /əʊ/, like window, yellow, follow, tomorrow, borrow. In a one-syllable word it splits: now and how on one side, know and show on the other.

The pair ai and ay divide by position: ai sits inside the word, ay at the end, both read /eɪ/. Two exceptions worth remembering because you use them daily are said /sed/ and says /sez/.

The ough group, where you have to memorize

The cluster ough has no rule. It has six or more readings and the spelling reveals nothing to help you choose, so the only way is to learn each word by heart.

The only comfort is this list is short and closed; no new words will join it. If you want to know why four letters can yield so many sounds, the history is in the article on silent letters.

The endings -ed and -s follow the sound before them

These two endings get overlooked because they do not change word meaning. They are where English hides tense and plural, and Vietnamese speakers often swallow both.

The most important thing about -ed: pronunciation depends on the sound before it, not the letter. Only when the verb ends in /t/ or /d/ does -ed become its own syllable /ɪd/. In every other case it is just a consonant tacked onto the end of the word.

Spelling deceives a lot here. laughed ends in the letters ed but the sound before it is /f/, so it is /lɑːft/. liked has e in the middle but no sound there, /laɪkt/, one syllable. The other way, the habit of adding /ɪd/ to every verb makes worked come out as two syllables, and the listener has to stop and guess what you just said.

The ending -s splits three ways by the same logic.

In listening, this is where the most information gets lost. "He works here" and "He worked here" differ by exactly one final consonant, /s/ versus /t/, and both are very light. An ear not used to catching final consonants loses the tense signal for the whole sentence, and you understand the content but get the time wrong.

Words whose spelling deceives

There is still a group of words no rule can save, because spelling and pronunciation have drifted too far apart. The bad luck is many words in this group are everyday words.

Most share one mechanism rather than being fully random: an unstressed syllable in the middle of the word drops entirely when spoken. comfortable drops the second syllable, vegetable the third, chocolate and interesting do the same.

Wednesday, colonel and iron are truly history's leftovers; no mechanism lets you guess them. If your job sometimes requires reading names or email addresses over the phone, the name for each letter is a different story altogether, and the English alphabet article covers that plus the NATO table.

After guessing you still have to listen to know you got it right

Every rule above gives you a hypothesis about how a word sounds, and only the ear can confirm that hypothesis.

The most dangerous spot is words you have read for years but never heard anyone say. In your head they have a pronunciation you built from the spelling when you first learned them, then sat there unchanged. Nobody corrects them because you never say them out loud. When that word shows up in an ordinary sentence, you hear it pass by without recognizing the word you know the meaning of.

Dictation listening finds exactly those words, because it forces you to write down what you think you just heard. The gap between what you wrote and the real transcript is the list of words sitting wrong in your head.

A concrete example. The audio sentence has comfortable, you hear three syllables going very fast and write comftable, or leave it blank because you could not keep up. Open the transcript and see it is comfortable, and right then you know the four-syllable reading you have been using silently for years is something you made up.

Another example from the -ed section. "He asked me last week" spoken naturally has /skt/ glued to /m/ and the ending nearly vanishes, so you write "He ask me last week". Look at the transcript and ed is there, and next time your ear starts waiting for it. Repeat a few times and the reflex shifts, without learning any new rule.

The podcast library on WELE has a transcript for every sentence, so the loop of listen, write, compare runs immediately. If you are not sure what level to start at, the listening level test takes a few minutes, and the five-step process describes how to transcribe a clip for results.

Frequently asked questions

Are reading rules enough for standard pronunciation?

They are not enough. Rules tell you how a word should sound; getting your mouth to produce that sound is a different skill, depending on tongue position, mouth shape and airflow. The two support each other, they do not replace each other, and the part about producing sound is in the English pronunciation article.

Why do I know what a word means but still cannot catch it when I listen?

Usually because the pronunciation stored in your head differs from how the word actually sounds, or because you put stress in the wrong place so the sound shape of the word shifts. Vocabulary learned through the eye is stored as writing; when you listen you need it as sound.

Should I look up IPA or listen to dictionary audio?

Do both in the same lookup. Cambridge Dictionary and Oxford Learner Dictionaries both have transcription plus audio for British and American English. To hear the word in real speech, Youglish and Forvo are more convenient.

Should I follow British or American English?

Pick one accent for speaking practice and stay consistent; for listening you need both because you do not choose who talks to you. The three clearest differences are final r, vowels in class or dance, and t between vowels in water or better.

Do native children learn these rules?

Yes, in primary school they learn phonics, the same rule groups as in this article. The difference is direction: they already have the sounds in their heads and are learning how to write those sounds down, while you go the other way, rebuilding sound from writing. That is why the listening step cannot be skipped.

The most sensible way to use this article is to keep it next to where you read English daily: meet an unfamiliar word, guess by the rules first, then listen once to confirm. Do that for a few weeks and the number of words sitting wrong in your head shrinks. Learning words in context makes that confirmation step happen more naturally than sitting down to look up words one by one.