Connected Speech in English: Linking and Reductions

Connected speech is the main reason learners who read English well still cannot follow native speakers. In fast, natural English, words run together, sounds disappear, and small words shrink almost to nothing. "What do you want to do?" becomes something like "Whaddaya wanna do?". Nothing about the grammar has changed; the sound has.

This guide explains the main features of connected speech in English with examples, shows how to practise them, and gives teachers a short lesson sequence. It covers the same ground as the opening units of our Natural English Speaking course.

Why connected speech matters

Learners often say "they speak too fast". Speed is part of it, but the bigger problem is that the words do not sound the way learners expect. If you have only heard "going to" as two clear words, "gonna" is a new word. Teaching connected speech is mostly about listening: helping learners recognise these patterns so the words they already know become audible. Producing them is a bonus, and optional for many learners.

1. Linking: consonant to vowel

When a word ends in a consonant sound and the next starts with a vowel sound, the consonant moves across:

- "turn off" sounds like "tur-noff"

- "pick it up" sounds like "pi-ki-tup"

- "an apple" sounds like "a-napple"

Drill: Write phrases on the board with a linking mark between the words. Learners say them slowly with the link, then at normal speed.

2. Linking: vowel to vowel

Between two vowel sounds, English often slips in a small /w/ or /j/ (y) sound:

- "go on" sounds like "go-won"

- "I agree" sounds like "I-yagree"

- "do it" sounds like "do-wit"

Learners who pause between vowels sound choppy. Hearing the glide helps them understand and sound smoother.

3. Reductions

Common combinations shrink into new forms in casual speech:

- "going to" → "gonna" (only when it is the future: "I'm gonna call", not "I'm going to Paris")

- "want to" → "wanna"

- "have to" → "hafta"

- "got to" → "gotta"

- "kind of" → "kinda"

- "let me" → "lemme"

- "don't know" → "dunno"

- "did you" → "didja"

Learners need to understand all of these. Whether they use them depends on context; they are normal in conversation and out of place in a formal presentation.

Drill: Play or read sentences with reductions and ask learners to write the full form. "I dunno, I kinda wanna stay" → "I don't know, I kind of want to stay."

4. Dropped sounds (elision)

Some sounds disappear, especially /t/ and /d/ between other consonants:

- "next week" sounds like "nex week"

- "last night" sounds like "las night"

- "friendship" sounds like "frienship"

- "handbag" sounds like "hanbag"

This is why learners hear "I mus go" and think the speaker made a mistake.

5. Sounds that change (assimilation)

Sounds change to become more like their neighbours:

- "don't you" sounds like "donchu"

- "would you" sounds like "wouldju"

- "ten pounds" can sound like "tem pounds"

- "good boy" can sound like "goob boy"

Teachers do not need to teach every rule. Show the "-t you" and "-d you" patterns first, because they are everywhere.

6. Weak forms of small words

Function words (articles, auxiliaries, prepositions, pronouns) are usually unstressed and reduced to a schwa sound:

- "to" in "I want to go" sounds like "tuh"

- "and" in "fish and chips" sounds like "n"

- "can" in "I can swim" sounds like "kn", while stressed "can't" stays clear

- "of" in "a cup of tea" sounds like "uh"

This is a big source of misunderstanding. In American English especially, learners mishear "can" and "can't", because the difference is mostly in stress and vowel quality.

7. Sentence stress and rhythm

English puts strong stress on content words (nouns, main verbs, adjectives, adverbs) and squeezes the rest. "I WANT to GO to the SHOP" has three beats, however many small words sit between them. Teaching learners to hear the beats is often the fastest way into connected speech, because it tells them which words they can safely miss.

Drill: Clap the beats. Then add words without adding beats: "CATS EAT FISH", "The CATS will EAT the FISH", "The CATS have been EATing the FISH". Each should take roughly the same time.

British and American differences

Connected speech happens in every variety of English, but some features differ:

The flap T (mostly American). Between vowels, /t/ often sounds like a quick /d/: "water" sounds like "wader", "better" like "bedder", "get it" like "geddit".

The glottal stop (common in many British accents). /t/ can be replaced by a catch in the throat: "better" sounds like "be'er", "not now" like "no' now".

Linking and intrusive R (many British accents). Speakers often link a final R sound to a following vowel ("far away" as "fa-raway") and sometimes add an R where there is none ("law and order" as "law-r-and order").

Learners do not need to copy a particular accent. They do need to recognise these features, because they will meet all of them in films, podcasts and international workplaces.

A quick listening test

Read these aloud at natural speed, or ask a fluent speaker to, and ask learners to write the full sentence:

1. "Whaddaya wanna do tonight?" (What do you want to do tonight?)

2. "I dunno, I'm kinda tired." (I don't know, I'm kind of tired.)

3. "We hafta leave at five." (We have to leave at five.)

4. "Didja see the game?" (Did you see the game?)

5. "Lemme check." (Let me check.)

Learners who can decode all five are ready for real audio; learners who cannot should spend more time on reductions and weak forms first.

A five-lesson sequence for teachers

1. Rhythm and stress. Clapping drills and content versus function words.

2. Weak forms. "to", "and", "of", "can", with listening discrimination tasks.

3. Linking. Consonant-vowel and vowel-vowel, with marked-up scripts.

4. Reductions. "Gonna", "wanna", "dunno" and others, listening first, then optional production.

5. Real audio. A short clip from a podcast or TV scene. Learners transcribe a few lines, then compare with the script. Our guide to teaching with TV clips has practical tips.

How learners can practise alone

Shadowing. Play a short clip, pause, and repeat it with the same rhythm.

Dictation of short lines. Write what you hear, then check the transcript and notice what you missed.

Slow-down tools. Many players can slow audio without changing pitch; listen slowly, then at normal speed.

Record yourself. Compare your version of a line with the original.

For individual sounds rather than connected speech, see our list of minimal pairs for ESL.

Common mistakes

- Teaching reductions as "slang" or "lazy". They are normal features of fluent speech.

- Asking learners to produce reductions before they can hear them.

- Using only scripted textbook audio, which is often slower and clearer than real speech.

- Ignoring rhythm. Without stress-timing, linking and reductions do not make sense.

Go further

Our Natural English Speaking course covers reductions, linking, dropped sounds, stress, intonation, idioms and register across 27 lessons, and the first lessons are free to preview. For sound-by-sound work, see the Pronunciation course.