Pronunciation: word stress, rhythm & intonation¶
At a glance
Priority: P0 · Complexity: ⅘ · Est. time: 3 h · Phase: 1 · Prereqs: none. Pairs with Pace, pauses, filler words & clarity and Shadowing & self-review; drilled in Week 2, Week 3, Week 9 and Week 17. You're done when: you can say the 40 stress-pattern words in the reference table with the stress on the right syllable without checking, and a 60-second recording of a design-review summary shows a clear nuclear stress (the "peak" word) in every sentence when you play it back.
Why it matters¶
Your goal is intelligibility, not a different accent. Indian English is a legitimate, well-described variety with millions of speakers; you do not need to sound American or British to lead a global team. What you do need is for a listener in Copenhagen, Rotterdam, Houston or Shenzhen to understand you the first time, on a compressed video call, without spending effort decoding you.
What research on intelligibility says, in brief (hedged, because findings vary by listener and context):
- Accentedness, comprehensibility and intelligibility are partly independent (Munro & Derwing's work on second-language speech). A speaker can sound clearly "accented" and still be fully understood, or sound less accented and be harder to follow. Chasing "accent reduction" is the wrong target.
- Prosody carries a large share of comprehensibility. Word stress, which word in a phrase gets the peak, and rhythm affect processing effort for listeners more than most single-sound differences.
- Jenkins' Lingua Franca Core (English as a lingua franca research) proposes which features matter most between non-native speakers: most consonant contrasts, vowel length contrasts, consonant clusters at the start of words, and nuclear stress placement. She treats things like the "th" sounds and weak forms as lower priority for ELF-to-ELF talk. Note the limit: many of your listeners are native speakers of US/UK English, and native listeners are more sensitive to word stress than ELF research suggests. Word stress therefore stays high priority for you.
So the priority order in this page is: (1) word stress, (2) sentence stress and rhythm, (3) intonation for questions and emphasis, (4) a small number of sound contrasts that genuinely cause misunderstanding, (5) everything else (including retroflex t/d) as optional polish.
A note on generalisations
India has 20+ major languages and dozens of accents. What follows describes tendencies that are commonly reported for many, not all, Indian-English speakers. A Tamil-background speaker and a Punjabi-background speaker differ. Record yourself and let your own data decide what to work on.
Core concepts¶
1. Word stress: the highest-value skill¶
English words of 2+ syllables have one primary stress: that syllable is longer, louder, higher in pitch and has a full vowel. The other syllables are shorter and often reduced to a schwa /ə/ (the "uh" in about). Get the primary stress wrong and the listener may not recognise the word at all, even if every sound is "correct".
Tendencies that make English stress hard for speakers of syllable-timed languages: stress is not marked in spelling, it moves when a word changes form (PHOtograph, phoTOGraphy), and unstressed vowels collapse.
Common patterns (rules of thumb, with exceptions):
| Pattern | Rule of thumb | Examples |
|---|---|---|
Suffix -tion, -sion, -ic(s), -ical, -ity, -ify, -ial |
stress the syllable before the suffix | in-for-MA-tion, an-a-LYT-ics, ar-chi-TEC-tur-al, sca-la-BIL-i-ty, ver-i-fi-CA-tion |
| Noun/adjective vs verb, 2 syllables | nouns/adjectives: first; verbs: second | REcord (n) / reCORD (v); OBject / obJECT; CONflict / conFLICT |
| Compound nouns | stress the first element | DATAbase, BACKup, ROLLback, DEADline, WORKflow |
| Phrasal verbs | stress the particle | set UP, roll BACK, back UP, sign OFF |
Prefix re-, de-, pre- in verbs |
stress usually the root | re-FAC-tor, de-PLOY, de-PRE-cate (adj. DEP-re-cat-ed) |
Exceptions
These rules hold for most words but not all (for example -ify verbs such as JUS-ti-fy keep stress early). Check a dictionary; do not memorise rules blindly.
Stress patterns of common tech words (American English; UK differences noted)¶
| Word | Stress | Notes / common slip |
|---|---|---|
| database | DA-ta-base | US "DAY-ta" or "DAT-a"; compound noun, first-syllable stress. Not da-ta-BASE |
| architecture | AR-ki-tec-ture | first syllable; not ar-chi-TEC-ture |
| architectural | ar-ki-TEC-tu-ral | moves! |
| deploy | de-PLOY | verb, second syllable; deployment = de-PLOY-ment |
| cache | one syllable, sounds like "cash" | not "ca-shay" |
| analytics | an-a-LYT-ics | stress moves from analyze (AN-a-lyze) and analysis (a-NAL-y-sis) |
| latency | LAY-ten-see | first syllable; latent LAY-tent |
| infrastructure | IN-fra-struc-ture | first syllable |
| algorithm | AL-go-rith-um | first syllable; last syllable is a weak "um" |
| asynchronous | ay-SIN-kro-nus | 2nd syllable; often shortened to "async" (AY-sink) |
| concurrency | con-CUR-ren-cy | second syllable |
| redundancy | re-DUN-dan-cy | second syllable |
| authentication | aw-THEN-ti-KAY-shun | primary on -CA-; secondary on -THEN- |
| authorization | AW-thor-i-ZAY-shun | primary on -ZA- |
| documentation | DOC-u-men-TAY-shun | primary on -TA-; secondary on first |
| priority | pry-OR-i-tee | second syllable; not PRI-o-ri-ty |
| parameter | pa-RAM-e-ter | second syllable |
| hierarchy | HI-er-ar-kee | first syllable; the "ch" is /k/ |
| scalability | SKAY-la-BIL-i-tee | primary on -BIL- |
| repository | re-POZ-i-tor-ee | second syllable; often "repo" (REE-po) |
| idempotent | eye-DEM-po-tent | stress on -DEM-; the 'i' is "eye" |
| kubernetes | koo-ber-NET-eez | 3rd syllable (Greek origin, "helmsman") |
| migration | my-GRAY-shun | second syllable |
| mitigate | MIT-i-gate | first syllable; mitigation mit-i-GAY-shun |
| escalate | ES-ka-late | first syllable; escalation es-ka-LAY-shun |
| stakeholder | STAKE-hold-er | compound: first element |
| incident | IN-si-dent | first syllable |
| environment | en-VY-run-ment | second syllable; the 2nd "n" often silent-ish in US speech |
| development | de-VEL-op-ment | second syllable; not DE-vel-op-ment |
| strategy / strategic | STRAT-e-jee / stra-TEE-jik | stress moves |
| diagnose / diagnosis | DY-ag-nose / dy-ag-NO-sis | stress moves |
| protocol | PRO-to-col | first syllable |
| metadata | MET-a-DAY-ta | first syllable, secondary on DA |
| throughput | THROO-put | compound: first element; not through-PUT |
| sequence | SEE-kwence | first syllable |
| schedule | SKED-jool (US) / SHED-yool (UK) | both correct; pick one |
| recursion | re-CUR-zhun | second syllable |
| orchestration | or-kes-TRAY-shun | primary on -TRA- |
| refactor | re-FAK-ter | verb; second syllable |
| default | de-FAWLT | second syllable (noun and verb) |
| update | up-DATE (v) / UP-date (n) | both stress patterns exist for the noun; the verb is up-DATE |
2. Sentence stress and rhythm¶
English is usually described as stress-timed: stressed syllables occur at roughly regular intervals, and the unstressed syllables in between are squashed to fit. Many Indian languages are described as more syllable-timed (each syllable gets similar length), and Indian English often carries that, giving an even, staccato rhythm. To a US/UK listener, that evenness can make it hard to hear which words matter, because everything sounds equally important. Note that stress-timing is a tendency, not a physical metronome.
The rule: stress content words (nouns, main verbs, adjectives, adverbs, negatives, question words). Reduce function words (articles, prepositions, auxiliaries, pronouns, conjunctions) to weak forms.
| Written | Spoken (weak forms in lowercase) |
|---|---|
| We can deploy it to the staging environment | we kən de-PLOY it tə the STAG-ing en-VY-run-ment. Stress: deploy, staging, environment |
| We can't deploy it | we CAN'T de-PLOY it: can't has the full vowel /æ/ and is stressed; this is the key difference from can /kən/ |
| The service was down for two hours | the SER-vice wəz DOWN fər TWO HOURS |
| What's the root cause? | WHAT'S the ROOT CAUSE |
The can/can't trap
In American English can before a verb is /kən/ (weak, nearly "kn") and can't is /kænt/ (full vowel, often with a glottal stop or a very short t). Listeners rely on the vowel and stress, not the final t. If you say "we CAN deploy" with full stress and full vowel, a US listener may hear "can't". For a positive, use the weak form; for a negative, use the strong vowel and add stress. In UK English can't is /kɑːnt/ (long vowel), which is easier.
Nuclear stress (focus): in each phrase, one syllable is the peak: the syllable that carries the pitch movement. By default it is the last content word; you can move it to signal contrast:
- "The API is slow." (neutral)
- "The API is slow, but it's not down." (contrast between slow and down)
- "I said the API is slow, not the database." (contrast the subject)
Getting nuclear stress right is the single most powerful prosody habit for a senior speaker: it tells the listener what is new, what is contrasted, and where to look.
3. Intonation: which way the pitch moves¶
| Pattern | Use | Example |
|---|---|---|
| Fall | statements, wh-questions, commands, finality | "We're rolling back. ↘" "Who owns this service? ↘" |
| Rise | yes/no questions, checking, "please continue" | "Can we ship Friday? ↗" |
| Fall-rise (↘↗) | reservation, "but...", partial agreement, polite hedging | "That could work ↘↗ (but...)" |
| Rise-rise-fall | lists | "Latency ↗, errors ↗, and cost ↘" |
| Level/high pitch, then step down | mid-sentence pause with more to come | "Given the deadline, ↗ we'll defer the migration ↘" |
Common problems reported by listeners (again, not universal):
- A rising tone at the end of statements ("uptalk") makes a recommendation sound like a question. Falling on the final word signals authority: "My recommendation is to defer. ↘"
- A monotone or narrow pitch range signals low engagement on calls. Widen your range by about 30-50% on key words; it will feel exaggerated to you and sound natural to them.
- A fall on the first item in a list makes listeners think the list is finished.
- Rise on the wh-question ("Why did it fail? ↗") can sound uncertain or hostile. Fall is the default.
4. Linking, reductions and connected speech¶
Native speakers do not say words separately. Learn to listen to these before trying to produce them:
- Consonant to vowel linking: "check it out" -> "che-ki-tout"; "an API" -> "a-nAPI".
- Flap t (US): t between vowels becomes a quick /d/-like flap: "latency" -> "LAY-den-see", "data" -> "DAY-da", "better", "meeting". It is not an error to keep a clear /t/ in these words, but you will hear the flap constantly and should recognise it.
- Dropping the middle consonant in clusters: "next week" -> "nex-week", "most people" -> "mos-people". Do not add vowels between consonants.
- Weak forms: to /tə/, for /fər/, and /ənd/ or /n/, of /əv/ or /ə/, have /əv/, "want to" -> "wanna" (informal, avoid in writing).
5. Sound contrasts that matter (and ones that do not)¶
Rank by risk of misunderstanding, not by how "accented" they sound.
Higher priority
- Initial consonant clusters (s + consonant): some Indian-English speakers add a vowel in front: "school" -> "is-kool", "station" -> "is-tation", "small" -> "is-mall". This changes syllable count and is a Lingua Franca Core item. Practise s-clusters with the first sound held: sss-tation.
- Aspiration of p, t, k at the start of stressed syllables. In American/British English
p, t, kcome with a puff of air; in many Indian languages the unaspirated versions are used, and to US ears an unaspirated /p/ can sound like /b/: "pill" heard as "bill". Hold a tissue in front of your mouth: it should move on pill, time, cap (initial position). - v/w (see below), where the merger can make "wine/vine" and "worse/verse" ambiguous.
- Vowel length pairs: bit/beat, full/fool, pull/pool, cot/caught, sit/seat. Keep long vowels long. Use the phrase to disambiguate when unsure: "Please sit down" vs "the seat is taken".
- -ed and -s endings: past tense
-edis /t/ after voiceless sounds (checked, deployed is /d/), /d/ after voiced (deployed), and /ɪd/ after t/d (wanted, updated). Plural -s: /s/ (caches: /ɪz/ after sibilants). Dropping these endings changes meaning ("we deploy" vs "we deployed").
Medium priority
- th /θ/ and /ð/: many speakers use dental stops [t̪, d̪] ("tink" for think). It rarely blocks understanding in context, but a few words become ambiguous: thin/tin, three/tree, thank/tank, then/den, they/day. Tip: gently place the tongue tip against the back of the upper teeth (touching or just between) and push air out; do not bite.
- v/w: the merger to a single sound [ʋ], between /v/ and /w/, is widely reported. For /v/: upper teeth touch the lower lip (feel the buzz). For /w/: lips rounded, no teeth contact.
Low priority (accent, not intelligibility)
- Retroflex t/d: many Indian-English speakers use a retroflex (tongue curled back) or a very hard t/d, whereas American English uses an alveolar (tongue tip on the ridge behind the upper teeth). Because English does not contrast retroflex and alveolar sounds, this is a matter of accent and rarely affects intelligibility. If you want to soften it for your own preference: aim the tip of the tongue at the ridge just behind the top teeth (say "t" as in "tea"). Optional.
- Tapped or trilled r; variations of /ɔ/ and /æ/; "w" and "v" in fast speech: leave until the higher items are stable.
Minimal-pair lists (say each pair aloud; record and compare)¶
| Contrast | Pairs |
|---|---|
| th / t / d | thin-tin, three-tree, thank-tank, thought-tot, both-boat, bath-bat, then-den, they-day, though-dough, breathe-breed |
| v / w | vine-wine, vest-west, veil-whale, verse-worse, vary-wary, vow-wow, veal-wheel, vet-wet |
| p / b (aspiration) | pill-bill, pack-back, pat-bat, pay-bay, cap-cab, pit-bit |
| t / d (aspiration) | time-dime, tab-dab, ten-den, tie-dye, tear-dear |
| k / g (aspiration) | coat-goat, cap-gap, cold-gold, class-glass, back-bag |
| vowel length | bit-beat, sit-seat, full-fool, pull-pool, ship-sheep, live-leave |
| cot / caught (US may merge) | cot-caught, don-dawn, stock-stalk |
| s-clusters | school (not is-kool), station, small, speak, status, spec, store |
| cache / cash / catch | cache-cash-catch: "cache" = "cash" |
| -ed | checked /t/, deployed /d/, wanted /ɪd/, updated /ɪd/, logged /d/ |
Resources¶
| Resource | Type | Why this one | Level | Cost |
|---|---|---|---|---|
| Rachel's English | video/site | Best free American-English stress, rhythm and reduction training; her "Mastering English Stress" and intonation videos are exact fits | intermediate-advanced | free/freemium |
| YouGlish | interactive | Type a word (e.g. "latency"); hear it in real talks by real speakers; test US/UK/AU variants and verify stress | all | free |
Sounds of Speech (Univ. of Iowa)  |
interactive | Animated mouth cross-sections for each English sound; shows the tongue position for th, v/w, t/d clearly | intermediate | free |
| Praat |
tool | Free phonetics software; record yourself and view pitch contour and duration, so you can see whether the nuclear stress rises | advanced | free |
| Cambridge Dictionary | reference | UK and US audio for every word, stress in the IPA; check every doubtful word | all | free |
| Merriam-Webster | reference | American pronunciation with stress marks and audio | all | free |
| BBC Learning English: "The English We Speak" and the pronunciation series (site) | video/audio | Short, clear lessons on sentence stress and connected speech (UK model) | intermediate | free |
| Better English Pronunciation with Paul (YouTube channel) | video | Careful explanations of weak forms, linking and word stress from an experienced teacher | intermediate | free |
| Hancock, English Pronunciation in Use (Intermediate/Advanced, Cambridge) | book | Systematic word stress and intonation units with audio and exercises | intermediate-advanced | paid |
| Jenkins, The Phonology of English as an International Language (OUP, 2000) |
book | The Lingua Franca Core research; explains why intelligibility, not accent, should be the target | advanced | paid |
Hands-on lab (90 min)¶
Goal: find your own top three stress and rhythm issues with evidence, and fix one.
- Record (10 min). Read this aloud at normal speed, then record a 60-second free summary of a design you recently reviewed: > "The migration to the new event-driven architecture is scheduled for the third quarter. Our analytics team needs lower latency, and the deployment pipeline needs better redundancy. We should prioritize authentication and documentation before rollout."
- Word-stress audit (20 min). Look up each multi-syllable word in the Cambridge or Merriam-Webster audio (or YouGlish). Mark your version vs the reference. Copy every mismatch into a table: word, my stress, target stress.
- Sentence-stress audit (15 min). Play the 60-second summary. For each sentence, write the one word that carries the nuclear stress. If you cannot identify one in most sentences, your rhythm is too even.
- Intonation audit (10 min). Write ↘ or ↗ at the end of each sentence you said. Statements should mostly be ↘.
- Drill (25 min). Pick your top 10 mismatched words. Say each 5 times in a phrase ("the deploy pipeline", "the latency budget"), then in a sentence. Shadow (see Shadowing & self-review) one minute of a talk with clear stress.
- Re-record (10 min). Same prompt. Compare.
Expected output: a personal list of ~10 stress-pattern corrections, one intonation habit to change, and a before/after recording pair saved in docs/log/.
Questions¶
L1 — Recall¶
Q1. Which syllable takes primary stress in analytics, analyze and analysis, and what does the pattern show?
Name the pattern.
Answer
an-a-LYT-ics, AN-a-lyze, a-NAL-y-sis. The stress moves when the word form (suffix) changes. -ics pulls stress to the syllable before it; the verb keeps stress early. Stress cannot be transferred from one word to its relatives; learn each form.
Q2. In "We can deploy", is can stressed? What is different in "We can't deploy"?
Answer
Positive can before a verb is weak: /kən/, not stressed. Can't is stressed with a full vowel /kænt/ (UK /kɑːnt/). Listeners rely on vowel quality and stress, not the final t. Saying a fully stressed "CAN" can be misheard as "can't" by US listeners.
Q3. Which words are usually stressed and which are reduced in a sentence?
Answer
Stressed: nouns, main verbs, adjectives, adverbs, negatives, question words. Reduced: articles, prepositions, auxiliaries, pronouns, conjunctions. Reduction is the point: it makes the stressed words audible.
Q4. What does the falling tone signal on a statement, and what does a rising tone signal on a statement?
Answer
Fall: completion, certainty, authority. Rise on a statement: uncertainty or a request for confirmation ("uptalk"); on a recommendation it undermines authority.
Q5. Why is the Lingua Franca Core relevant to how you set priorities?
Answer
It is research-based on which features actually cause misunderstanding in international communication: consonant contrasts, vowel length, initial clusters and nuclear stress ranked higher; th and weak forms ranked lower. It supports "intelligibility first, not accent removal". Caveat: native-speaker listeners are also sensitive to word stress, so keep that high.
L2 — Apply¶
Q6. Mark the stress and say it aloud: database, deploy, latency, infrastructure, authentication, idempotent, throughput, kubernetes.
Answer
DA-ta-base, de-PLOY, LAY-ten-see, IN-fra-struc-ture, aw-THEN-ti-KAY-shun (primary on KAY), eye-DEM-po-tent, THROO-put, koo-ber-NET-eez. Compound nouns (database, throughput) stress the first element; verbs like deploy stress the second.
Q7. Underline the words to stress: "We need to roll back the release because the migration failed in staging."
Answer
We NEED to ROLL BACK the RE-lease because the mi-GRAY-tion FAILED in STAG-ing. Roll back is a phrasal verb: stress the particle (BACK) if the noun object follows the particle; in "roll back the release" both roll and back are stressed with BACK stronger. Release, migration, failed, staging are content words. The rest reduce.
Q8. You say "We are deploying to production on Friday? ↗" as a decision. What is the problem and the fix?
Answer
The rise turns the decision into a question and signals uncertainty. Use a falling tone on the last stressed syllable: "We're deploying to production on FRIday ↘." If you actually want confirmation, keep the rise and add the words: "Are we still good to deploy on Friday? ↗"
Q9. Choose the sound contrasts to drill for this list: west, vest, wine, vine, pill, bill. What is the drill?
Answer
v/w and p/b (aspiration). For v/w: /v/ upper teeth on lower lip, buzz; /w/ rounded lips, no teeth. For p/b: put a hand or tissue in front of your mouth; pill should push air out, bill should not (voiced). Say the pairs in a carrier sentence: "I said vest, not west."
Q10. Put the stress in the noun and the verb: "Please record (v) the record (n) of the meeting.".
Answer
Please re-CORD the REC-ord of the meeting. Noun first syllable, verb second: the two-syllable noun/verb rule. Also update, export, import, object, permit, project, increase.
L3 — Judge and choose¶
Q11. A colleague says you should "lose your accent". Another says clarity is what matters. How do you decide what to work on?
Answer
Choose by evidence: record a work-relevant talk and ask two non-Indian colleagues "which words did you have to guess?" Fix what causes guesses, which usually means word stress, nuclear stress, s-clusters, aspiration and v/w. Leave retroflex t/d, the exact th realisation and rhythm nuance until later, if at all. Accent is identity; intelligibility is your responsibility to your listener. Munro and Derwing's work supports treating them separately.
Q12. Should you copy a native American speaker's flap t in latency?
Answer
Not required. A clear /t/ is fully intelligible and is standard in many varieties. The important thing is to recognise the flap when you hear it ("LAY-den-see"). Copying it can help if you find you over-articulate and sound stiff, but it is style, not correctness.
Q13. Two sentences: "I didn't say he stole the data" and "I didn't say he stole the data". Which implies that someone else did, and how does it differ from stress on the first "I"?
Answer
Stress on he implies someone else did it ("I said she did"). Stress on data implies he stole something else. Stress on I implies someone else said it. Stress is the carrier of meaning; that is why nuclear stress placement is a priority.
Q14. You have 3 weeks and 15 minutes a day before a global town hall. Order these: th sounds, word stress on 30 key terms, uptalk, retroflex t/d, filler words.
Answer
(1) Word stress on your 30 key terms (highest intelligibility gain per minute), (2) uptalk on statements and the nuclear stress on key sentences, (3) filler words (see Pace, pauses, filler words & clarity), (4) th, (5) retroflex t/d (accent, not intelligibility). Defend on effect per minute.
L4 — Staff-level¶
Q15. A director on your team, also Indian-English-background, is rated "hard to understand on calls" by a VP in another region. You are asked to give feedback. How?
Answer
Do not say "your accent is a problem". Use SBI: Situation ("in yesterday's architecture review"), Behaviour ("the words 'architecture' and 'latency' came through with stress on a different syllable, and nearly every sentence had the same volume and pitch"), Impact ("the VP had to ask twice, and the decision was delayed"). Offer a concrete, low-effort remedy: slow down 10%, pause between sections, stress the key number or verb, put a slide summary in chat. Offer to share tools (YouGlish, recording review). Keep the focus on the listener's experience and shared success. See Feedback, disagreement & conflict.
Q16. You are hosting a 40-person call where many people are non-native speakers and audio is poor. What pronunciation-level tactics improve intelligibility for everyone?
Answer
Speak 10-15% slower; pause at the ends of sentences; stress numbers and negatives; spell unusual names or acronyms once ("A-P-I, A-P-I gateway"); give the key point in chat; avoid "wanna/gonna"; repeat decisions using a fall; avoid idioms where possible. Those help native speakers too.
Q17. Your CTO says, kindly, that your accent makes you less "executive". How do you respond?
Answer
Reframe to what is actionable: "I want to be as easy to follow as possible. Can you give me two examples where you had to work to follow me?" Then work on those (stress, pacing, signposting). Do not accept that accent equals seniority: many respected leaders are accented. If the comment pattern is persistent or biased, discuss it with your manager or HR, since accent-based bias is a real fairness issue.
Real-world use cases¶
- Incident bridge (P1). A responder says "we can deploy the fix" vs "we can't". A missed negative costs minutes. Stress on the negative and the verb, and confirm in chat: "To confirm: we CAN'T deploy before 14:00."
- Design review with a US principal. Saying "da-ta-BASE" or "ar-chi-TEC-ture" makes it hard to follow at speed; stress-first-syllable patterns for tech nouns stop the listener re-parsing.
- Vendor call with a Rotterdam and a Houston lead. Minimal pairs (vine/wine, three/tree) matter when you say "version three" vs "version tree". Say it with a carrier: "version, three".
- Interview loop. Panels have very short attention for "who is this speaker"; clear stress and falling declaratives add perceived confidence. See Executive presence & handling Q&A.
- Voice AI product work. Understanding stress and intonation also helps you evaluate ASR/TTS output (why a TTS voice sounds "flat").
Pitfalls & anti-patterns¶
- Treating "accent reduction" as the goal instead of intelligibility.
- Memorising IPA for every sound instead of fixing the few contrasts that cause misunderstanding.
- Over-articulating each syllable equally, which makes the rhythm worse.
- Adding a vowel to s-clusters ("is-tation") or dropping -ed/-s endings.
- Rising at the end of every sentence in an attempt to sound polite.
- Practising isolated words but never in phrases; stress lives in sentences.
- Copying an actor's American accent without checking with a dictionary: you may learn the wrong stress.
- Judging yourself only by how it feels; you must listen to recordings.
Checklist¶
- I can say the 40 stress-pattern words in the table without notes.
- I recorded myself, marked the nuclear stress in each sentence and re-recorded after drilling.
- I can explain why
canandcan'tdiffer in sound and stress. - I know which of th, v/w, aspiration and s-clusters I actually need to work on, from evidence.
- I have verified 20 words in YouGlish or a dictionary and corrected my stress.
- I answered all L3 questions out loud in < 3 min each.