/s/as in sun
You probably make a clear s at the start of words like 'sun', but at the end of words — especially in plurals like 'books' or clusters like 'texts' — you may drop the s, weaken it, or add a small vowel after it, since…

What you probably do
You probably make a clear s at the start of words like 'sun', but at the end of words — especially in plurals like 'books' or clusters like 'texts' — you may drop the s, weaken it, or add a small vowel after it, since Mandarin never closes a syllable this way.
How natives do it
Americans keep the tongue's narrow groove firm all the way through the end of the word, finishing with a crisp, complete hiss and no extra vowel, even when another consonant comes right before it as in 'texts'.
Why it matters
Because final s often signals plural, possessive, or present-tense meaning in English, dropping it can make 'cat' and 'cats' sound the same to a listener, creating real misunderstandings rather than just an accent.
Hear the difference
Words this touches
How the mouth differs
Mandarin has s only in syllable onsets; no native syllable ends in a fricative or any consonant cluster, so English final s (plurals, possessives, third-person verbs) and s-clusters ('texts', 'books') fall outside Mandarin phonotactics. Speakers often weaken, drop, or add a vowel after the final s, and may simplify clusters like st or ts.
Listen for the pattern
Because final s in English frequently carries grammatical information (plural, possessive, present tense), dropping it doesn't just sound accented — it can erase meaning, making 'book' and 'books' sound identical to a listener.
Practise it
Exaggerate the Hiss: No Vowel Before 's'
sourced- Say 'state' and deliberately hold the very first 's' sound for two to three seconds, like a long hiss, before moving into the rest of the word.
- Notice that during that long hiss, no vowel sound ever appears — your mouth stays in the 's' position the whole time.
- Repeat this exaggerated, long-hiss version three times for 'state', 'spy', and 'strange'.
- Gradually shorten the hiss by about half, still keeping zero vowel before it.
- Shorten it once more until the 's' duration feels natural but you still clearly remember there's no vowel before it.
- Record yourself saying all three words at this natural duration.
- Play back the recording and confirm each word has exactly one syllable with no vowel before the 's'.
Success check: Your final recordings should show no vowel before the 's' in any of the three words, and each word should contain exactly one syllable at natural speaking speed.

Why this works. Deliberately over-prolonging the 's' hiss at the very start of the word, with an exaggeratedly abrupt, vowel-free onset, then gradually shortening it to normal duration, gives learners a clear perceptual and motor landmark for 'no vowel before the cluster,' consistent with exaggerated-contrast training improving categorical learning before fading to natural production.
Sources (2)
- Distinguishing universal and language-dependent levels of speech perception: Evidence from Japanese listeners' perception of English “l” and “r” — Virginia A. Mann, 1986
- Japanese listeners successfully discriminate the acoustic properties of English /l/ and /r/ when presented with non-English contrasts, ruling out a universal hearing deficit.
- The primary source of error is language-dependent categorization, where listeners map English sounds onto their existing L1 phonological categories (e.g., classifying both as liquids).
- Performance improves significantly when the acoustic contrast between /l/ and /r/ is exaggerated or presented in contexts that highlight their distinctiveness.
- The Role of Temporal Acoustic Exaggeration in High Variability Phonetic Training: A Behavioral and ERP Study — Bing Cheng, Xiaojuan Zhang, Siying Fan et al., 2019
- The HVPT-E group showed greater improvement in natural word identification performance compared to the standard HVPT group.
- Training with temporal acoustic exaggeration induced native-like categorical perception based on spectral cues.
- MMN responses demonstrated training-induced changes at pre-attentive neural levels, suggesting enhanced brain plasticity.
Hear It: state vs estate
sourced- Listen to a recording of the pair 'state' and 'estate' without looking at the spelling.
- Count how many syllables you hear in each word.
- Write down which one you think has the extra vowel at the start.
- Replay both words slowly, listening specifically for any sound before the 's'.
- Repeat the exercise with 'spy/espy' and 'strange/estrange'.
- Check your answers and replay any pair you got wrong three times in a row.
- Finish by listening to all three pairs mixed in random order and identifying each correctly.
Success check: You should correctly identify whether each word has an extra vowel before the 's' cluster on at least 5 of 6 attempts, clearly hearing the syllable-count difference.


Why this works. Training learners to detect the presence or absence of an epenthetic vowel before word-initial s-clusters sharpens awareness of a phonotactic habit that is otherwise automatic and hard to self-monitor; discrimination training on minimal contrasts strengthens the perceptual category needed before production changes.
Sources (1)
- Distinguishing universal and language-dependent levels of speech perception: Evidence from Japanese listeners' perception of English “l” and “r” — Virginia A. Mann, 1986
- Japanese listeners successfully discriminate the acoustic properties of English /l/ and /r/ when presented with non-English contrasts, ruling out a universal hearing deficit.
- The primary source of error is language-dependent categorization, where listeners map English sounds onto their existing L1 phonological categories (e.g., classifying both as liquids).
- Performance improves significantly when the acoustic contrast between /l/ and /r/ is exaggerated or presented in contexts that highlight their distinctiveness.
Drill Sentence: Students in Spain Study Statistics
consensus- Read this sentence silently: 'Students in Spain study strange state statistics at school.'
- Underline every word that starts with 's' plus another consonant: students, Spain, study, strange, state, statistics, school.
- Say the sentence very slowly, starting each underlined word directly on the 's' with no vowel before it.
- Record yourself reading the sentence at this slow pace.
- Play it back and count the syllables in each underlined word, confirming none have an extra one.
- Repeat the sentence three more times, gradually increasing your speed toward normal conversational pace.
- Do a final recording at natural speed and verify every underlined word still begins directly on 's'.
Success check: In your final recording, every underlined word should begin directly on the 's' sound with no extra syllable, even at natural conversational speed.
Why this works. Reading a connected sentence dense with word-initial s-clusters under natural speech rate trains the learner to suppress the epenthetic vowel consistently across multiple words in a row, which is more demanding and more representative of real speech than isolated word practice.
Sources (1)
- Wells, J.C. (1982). Accents of English.
Say It Right: state, not estate
consensus- Say the habitual version first: 'estate', noticing the little vowel before the 's'.
- Now try to start the word with your lips and tongue already in position for 's', before any sound comes out.
- Push air through that 's' position immediately, moving straight into 'state' with no vowel in front.
- Record yourself saying 'state' and play it back to check for one syllable, not two.
- Repeat the same process with 'spy' and 'strange', starting directly on the cluster each time.
- Say all three words five times each, always beginning the mouth already in the consonant position.
- Have a listener (or app) confirm your words sound like one-syllable 'state', 'spy', 'strange', not two-syllable versions.
Success check: Your recordings should contain exactly one syllable for 'state', 'spy', and 'strange', with no audible vowel sound before the initial 's'.


Why this works. Producing the contrastive pairs forces the learner to inhibit the automatic epenthetic-vowel reflex transferred from Spanish phonotactics and start articulation directly on the fricative, consolidating the new onset-cluster motor plan through immediate self-monitoring via recording.
Sources (1)
- Wells, J.C. (1982). Accents of English.
Borrow the Linking Trick
consensus- Say the short phrase 'is school' at normal speed and notice how the 's' from 'is' flows directly into 'school' with no gap.
- Notice that your mouth never opens for a new vowel between 'is' and 'school' — the 's' sounds just glide together.
- Now say 'I need school' and mentally treat the 's' of 'school' as if it were attached to the end of the previous word, just like 'is school'.
- Practice saying 'a state', 'the spy', 'my strange story', pretending the 's' belongs to the word before it.
- Record yourself saying these short phrases and check that no vowel appears before the 's' cluster.
- Now try the target words alone — 'state', 'spy', 'strange' — using that same glide-straight-in feeling.
- Confirm each isolated word still comes out as one syllable, just like it did inside the linked phrase.
Success check: The isolated words should sound just as gap-free as the 's' did in phrases like 'is school', with no vowel inserted and exactly one syllable each.


Why this works. Linking the word-initial 's' to the final vowel of the preceding word (e.g., saying 'is state' with the 's' gliding straight across the word boundary) borrows the motor routine learners already use for native-like consonant-to-vowel resyllabification, giving the s-cluster a familiar coda-like onset environment that naturally suppresses the epenthetic vowel without requiring a brand-new articulatory plan.
Sources (1)
- Wells, J.C. (1982). Accents of English.
Freeze Into 's': No Vowel Allowed
consensus- Say 'state' extremely slowly and pause right at the very beginning, before any sound starts.
- Check in the mirror that your jaw has not dropped open — your mouth should already be in the 's' shape.
- Slowly let air hiss through for the 's' sound, holding it for two full seconds without any vowel sneaking in before it.
- Move directly from the held 's' into the 't' sound, freezing again briefly at that transition.
- Release into the rest of the word 'state' at a slow, controlled pace.
- Repeat the whole frozen sequence three times, then gradually speed it up.
- Say 'state' at normal speed and confirm it still has no vowel before the 's'.
Success check: At each freeze point your mouth should already be in a consonant shape, never an open vowel shape, and the final normal-speed word should contain exactly one syllable.



Why this works. Freezing the articulators in the s-cluster's pre-vocalic configuration before any airflow begins lets learners physically confirm that no vowel-opening gesture is present, breaking the automatic epenthesis habit step by step before reassembling the cluster at normal speed.
Sources (1)
- Wells, J.C. (1982). Accents of English.
Reader ratings and feedback are coming soon.
Sources (2)
- Common Accent Features of Mandarin Chinese Speakers in English — 2026
- Mandarin speakers frequently soften or drop final consonants (e.g., 'seat' becoming 'see') because Mandarin syllables rarely end in stops.
- An extra schwa vowel (/ə/) often appears at the end of English words, creating an unnatural two-syllable rhythm where English requires a single stressed syllable.
- Consonant clusters common in English (e.g., 'texts', 'school') are simplified or omitted by Mandarin speakers due to the language's avoidance of stacked consonants.
- Pronunciation Challenges Chinese Speakers Face in English — 2020
- Chinese speakers tend to break English consonant clusters by inserting a vowel or dropping one or more consonants because Chinese lacks final clusters and limits initial clusters to two consonants (one being Y/W).
- The dark L /ɫ/ does not exist in Mandarin, leading speakers to replace it with a schwa [ə] or unrounded u sound, often resulting in words like 'people' sounding like 'pipo'.
- Speakers frequently nasalize vowels at the end of syllables instead of producing the alveolar /n/, causing confusion between /n/ and /ŋ/ (e.g., 'sin' vs. 'sing').
Loads videos from YouTube via YouGlish; YouTube may set cookies.
bus in YouGlish examples