Linking sounds in English

Spoken English does not respect word boundaries. Within a thought group, the end of one word and the start of the next are welded together, so get out sounds like one word, gettout, and come over sounds like comover. This is linking, and it is the single biggest reason the words of a fluent speaker do not match the words on the page: natives link roughly 80 percent of consonant-to-vowel boundaries in normal speech, so a sentence is not a row of boxes but a stream.

Linking is non-negotiable in a way that other connected-speech features are not. You can drop a /t/ or skip an assimilation and still sound clear. But speakers who do not link — who start every word with its own beginning — sound like a computer reading aloud, and listeners have to work harder to segment what they hear. If you do one thing after reading this guide, make it consonant-to-vowel linking.

When a word ends in a consonant and the next begins with a vowel, the consonant belongs, in sound, to the second word. There is no pause, no restart: the /t/ of get is produced exactly where the /aʊ/ of out begins. Saying get · out with a gap is not merely slower — it is a different prosody, the one foreigners are recognised by.

The trick for learners is anticipation: before you finish the first word, you already know the next one starts with a vowel, so you never close the consonant properly — you leave it hanging to attach to the vowel. Practise by rewriting phrases as a single pseudo-word: gettout, keepit, turnon.

When a word ends in the same consonant that the next word starts with, English does not say the consonant twice. big game is not /bɪg ɡeɪm/ but /bɪːɡeɪm/ — one long /g/ held a little longer. Learners often double the consonant because the spelling shows two letters; natives hold a single one. The length difference is real but subtle, so listen for it in black car before trying to produce it.

Two vowels meeting across a word boundary are harder to join, so English inserts a tiny glide — not a new sound, but the natural movement between the two mouth positions. After a vowel made with rounded lips — /uː/, /oʊ/ — the glide is /w/: do it becomes /duːwɪt/. After a vowel made with spread lips — /iː/, /eɪ/, /aɪ/ — the glide is /j/: see it becomes /siːjɪt/, and the end is pronounced thee·yend.

Learners tend to do the opposite: either a hard restart between the vowels, or an accidental /w/ after every vowel. The bridge is automatic once you keep the air flowing and move the tongue through the shortest path between position A and position B. Say do it and see it slowly and you will feel the glide appear on its own.

In non-rhotic accents — most of England, Australia, New Zealand — the r at the end of a word is silent unless a vowel follows. When one does, the r reappears and links: far away is /fɑːrəweɪ/, for example is /fə rɪɡzɑːmpəl/. American and most Irish and Scottish speakers pronounce that r anyway, so for them linking r is not a separate skill — the r is simply always there.

Note what happens with the spelling when there is no r at all: the same link appears anyway. law and order is /lɔːr ən ˈɔːdə/ in British speech, and the idea of it is /ði aɪdɪər əv ɪt/. That is intrusive r, the subject of its own guide — the identical sound, inserted where no r exists.

Linking is not a rule that applies to every boundary — it applies inside a thought group, the stretch of speech you deliver as one unit between pauses. At a pause — end of a clause, before a parenthetical — the boundary is a real stop and nothing links across it. When a word is contrasted or emphasised, the link weakens too. The skill is not "always link" but "link everything inside the chunk, and stop at the chunk edge". Native speakers do the first effortlessly and the second deliberately.

The fastest way to build the habit is shadowing: play a short sentence, repeat it aloud immediately, and force yourself to start each word the way the speaker ended the previous one. Record yourself against the original — choppy playback means gaps, and gaps are the thing to eliminate. For listening, know that a linked phrase like fill it in (fi·lli·tin) is one of the hardest things for untrained ears to segment, so training production doubles as listening training.

Connected speech overview — the four patterns as a system. Intrusive r — the r that appears where no r is written. Chunking and pausing — where the linking stops. The vs thee — the vowel-bridge rule that governs "the". Shadowing — the practice method that automates linking.

Read slowly first. Pronounce each syllable deliberately before trying to blend them at full speed. Rushing hides the vowel and consonant shapes your ear needs to learn.

Record yourself. Most phones have a voice memo app — record a word, then compare it to the playback here. Hearing the gap between your version and the model accent makes targeted practice far easier.

Focus on one accent. Jumping between American, British and Australian in the same session creates confusion. Pick the accent you encounter most and stick with it until the patterns feel natural.

Every word above is a button. Set the accent and voice you want on the main screen, then come back — the guides use whatever you chose. To hear two accents against each other, use Compare; to test whether you can actually hear a contrast, use the listening quiz.