-ed has three sounds, and one rule decides which
Ask a learner to read walked, played, wanted aloud and you'll usually hear three even endings: walk-ed, play-ed, want-ed. But natives say walkt (one syllable), playd (one syllable), and only wanted actually earns its extra syllable. The -ed spelling never changes; the sound changes every time — and one completely physical rule decides which of its three sounds you get.
First, feel the difference between voiced and voiceless
Put two fingers flat on your throat and make a long ssssss. Nothing. Now zzzzzz. Buzzing. That buzz is your vocal cords running — the sound is voiced. /s/, /f/, /k/, /p/, /ʃ/ (sh), /tʃ/ (ch) are voiceless: no buzz. /z/, /v/, /g/, /b/, /m/, /n/, /l/, /r/ and every vowel are voiced: buzz.
That's the whole toolkit. The -ed rule is just your mouth being lazy in a very consistent way: the ending copies the voicing of whatever sound came before it.
The rule in three lines
After a voiceless sound → /t/. Walked = walk + /t/. Kissed = kiss + /t/. Laughed, hoped, watched, washed — all end in a crisp /t/, no extra syllable. Your voice was off at the end of walk, so it stays off for the ending.
After a voiced sound → /d/. Played = play + /d/. Opened, loved, called, robbed, studied — all end in /d/, again no extra syllable. The voice was already buzzing, so it keeps buzzing.
After /t/ or /d/ themselves → /ɪd/, a real extra syllable. Wanted, needed, decided, started, ended. Here the mouth has no choice: you can't glue a /t/ onto a /t/ and hear anything, so English inserts a small vowel. This is the ONLY case where -ed adds a syllable.
Why saying it wrong costs more than an accent
Adding the extra syllable everywhere (walk-ed, kiss-ed) doesn't just sound old-fashioned — it can hide your grammar. If walked comes out mushy, the listener can't tell whether you said walk or walked, and suddenly your past tense has disappeared. In the other direction, learners who fear the ending drop it entirely, with the same result.
The endings are small, but they carry the tense of the whole sentence. English listeners are tuned to catch that tiny final /t/ or /d/ — see the /t/ lesson and /d/ lesson for how each one is made and released.
The shortcut: stop thinking about letters
You don't need to memorize lists. When a verb ends, notice whether your throat is buzzing. Buzzing → add buzz (/d/). Silent → add silence (/t/). Ends in t or d already → add the tiny vowel (/ɪd/). Because the rule is physical, your mouth actually wants to follow it — most learners find that once they stop pronouncing the spelling, the right ending appears by itself.
One decoy to watch: adjectives that look like past tenses keep the old /ɪd/ pronunciation — a wicked plan, a naked truth, a beloved teacher, aged cheese. They're leftovers from an older English where every -ed was a syllable; the modern rule simply doesn't apply to them.
Hearing the endings in real speech
Producing the endings is half the job; catching them at native speed is the other half, and it's harder than it sounds. A final /t/ or /d/ before a consonant is often unreleased — in walked back, the /t/ is a tiny silent closure, not a crisp tick. Before a vowel it jumps onto the next word entirely: walked in sounds like walk tin, played it like play dit. And in American English, a /d/ or /t/ between vowels can flap: waited comes out WAY-did, with the quick tap of butter.
So don't listen for a loud ending — listen for the timing and the vowel length. Walk and walked differ in how the word closes, not in volume. A good ear drill: have someone (or a recording) say random pairs like play/played, want/wanted, kiss/kissed and score yourself on which you heard. When your ear reads the closure reliably, your own endings improve almost automatically, because you finally hear what you're aiming at.
A two-minute sorting drill
Take the last email or message you wrote, pull out every past-tense verb, and sort them into the three buckets by saying each one with your fingers on your throat: finished → buzz off at the end of finish → /t/ bucket. Called → buzzing l → /d/ bucket. Started → ends in t → /ɪd/ bucket. Ten verbs, two minutes, and you've practiced with exactly the vocabulary you actually use.
The usual suspects sort like this — /t/: worked, asked, helped, watched, hoped, missed, laughed. /d/: called, played, opened, moved, learned, answered, tried. /ɪd/: wanted, needed, started, decided, waited, visited, ended. Notice how small the third bucket really is: only verbs already ending in /t/ or /d/. If you've been adding the extra syllable everywhere, you've been working three times harder than the language requires.
Check yourself with the tool
Paste a paragraph of past-tense sentences into the pronunciation tool — try I walked to work, opened my laptop, and wanted coffee — and look at the endings in the notation: walked shows a final /t/, opened a final /d/, and only wanted gains a syllable dot. The engine applies the voicing rule to any verb you throw at it, including ones you invent.
Then read the sentence back with the marks, connecting each ending to the next word (walked-to, opened-my) — the linking lesson shows how those final consonants latch on. Five minutes of this with your own writing beats an hour of abstract drills, because these are the sentences you'll actually say.