Faragiri

Learn Consonant sounds Pronunciation

English keeps its grammar on word endings, so a dropped final consonant removes a tense or a plural. That costs more than any sound that is hard to make.

5Sections
5Min read
4Depth

The consonants worth your hours are not the ones that are hard to make. They are the ones at the ends of words, because English keeps its grammar there: the difference between walk and walked, between book and books, is a consonant hanging off the end of a syllable.

Drop it and you have not produced an accent. You have produced a sentence in the wrong tense, and the listener has to repair a grammatical error rather than a sound, which is slower and much more visible.

The eight consonants that cause the trouble

Ranked by cost

Three consonant contrasts worth the hours, in order

Ranked by how often the contrast holds two English words apart, which is not the same as how hard the sound is to make.

  1. 03th voiced and unvoicedAbsent in most languages and the clearest accent marker there is.think and this
  2. 02r and l distinctionThe English r touches nothing; l and r are merged in Japanese and Korean.light and right
  3. 01v and b distinctionOne sound in Spanish and two here, separating real words.van and ban

The hardest sound to produce is not at the top of this ladder, and its absence is the finding. Ranking by difficulty would put it first and send a year to the cheapest error in the language.

  • the th sounds do not exist in most languages, which is why they are substituted so consistently and why the substitution is the clearest accent marker there is.
  • think and this. Two different th sounds, one voiced and one not, and both are made with the tongue against the teeth rather than behind them.
  • van and ban. V and b are one sound in Spanish and are two here, and the pair separates real words.
  • light and right. The English r is made without the tongue touching anything, which is unusual, and l and r are one sound in Japanese and Korean.
  • sing without the g. The ng at the end of a word is a single sound, and adding a hard g after it is an error most Italian and Indian speakers make.
  • pin and spin. The p is aspirated in the first and not in the second, automatically, and speakers of Romance languages produce the unaspirated one everywhere.
  • judge, which begins and ends with the same sound, and it does not exist in French or German.
  • measure, containing the voiced sh sound, which is common in French and rare in English and appears in very few words.

English piles consonants up, and most languages do not

English allows three consonants before a vowel and up to four after it, which is unusual. Strengths, texts, asked, twelfths: these are ordinary words that a great many languages have no syllable shape for at all.

Syllable architecture

Up to seven consonants around a single vowel

English allows three consonants before a vowel and up to four after it, creating syllable shapes most languages forbid.

  1. Onset cluster (up to three consonants)Clusters before the vowel, frequently broken up by inserting an illegal vowelstreet
  2. Vowel nucleusThe single vowel sound carrying the syllable peak and timingcore
  3. Coda cluster (up to four consonants)Clusters after the vowel like asked and texts packing grammatical tenses onto the endstrengths

Faced with an unfamiliar cluster, speakers automatically insert a vowel or delete a consonant to restore an acceptable shape.

Faced with a shape their language does not permit, speakers do one of two things automatically. They insert a vowel to break the cluster up, so street becomes something closer to istreet or sitreet. Or they delete a consonant, so asked becomes ass and texts becomes tex. Both repairs are invisible to the person making them, because both produce a syllable shape that feels correct.

The insertion is the more damaging of the two, and not for the reason people expect. Adding a vowel adds a syllable, and a listener uses syllable count to find word boundaries in a stream of speech, so a two-syllable version of a one-syllable word is not a mispronounced word but an unfound one.

Where it stops: the ending disappears under speed

Here is the stall, and it is specific. A learner can produce final consonants perfectly in single words and loses them the moment the speech runs at conversational pace, because in a real sentence the ending has to be released into the next word rather than into silence. The drill worked, the drill was in isolation, and nothing in the drill resembled the position where the sound actually lives.

The isolation illusion

Single words succeed while connected sentences lose the ending

Learners drill final consonants in isolation and lose them the moment speech reaches conversational tempo.

  • Clean isolated pronunciationFinal consonants produced cleanly when released into silence during drills
  • Recoverable meaningListeners understand sentences like yesterday I walk because context fills the gap
What isolated word drills show vs connected speech
  • Release into following soundsIn continuous speech the consonant must be released into the next sound rather than silence
  • Lost grammatical tensesPast tense and plural markers vanish under conversational speed
  • Uncorrected error habitsListeners rarely correct dropped endings when meaning is clear, reinforcing the error

Drilling words in isolation releases the sound into silence, which bears no resemblance to connected conversational pace.

Nothing reveals it, because the grammar it removes is usually recoverable. Somebody who says yesterday I walk to the office is understood immediately, since yesterday has already done the work, and no listener will pause to mention that the tense marking went missing. So the error is carried for years by people whose consonants are otherwise excellent.

Watching the mouth rather than reading about it

BBC Learning English publishes free pronunciation videos that show the sounds being made from the front and from the side, in close-up and at reduced speed. That is the format this needs, because a consonant is a movement rather than a position, and it is the movement into the next sound that fails in connected speech.

Static versus dynamic

Replacing textbook head diagrams with close-up video

A consonant is a movement into the next sound rather than a frozen static position.

Textbook cross-sectionA frozen diagram of a head showing the tongue at a single instant
  1. 01Dynamic movementShows the articulation moving toward and away from the articulators
  2. 02Dual perspectivesCaptures lip shape from the front and tongue contact from the side
  3. 03Reduced playback speedSlows the transition into the following sound without altering the mechanical action
Close-up articulation videoFront and profile video showing tongue and lip movement at reduced speed

BBC Learning English videos demonstrate sounds from the front and side, replacing frozen textbook cross-sections.

It replaced the cross-section diagram of a head in a textbook, which shows a tongue frozen at one instant and cannot show you what happens next. The limitation worth knowing is that these are single sounds and short phrases demonstrated carefully, so they will not show you the ending being released into the following word, which is the position that actually breaks.

What it looks like when it goes wrong

The audible failures are a short list. Ass for asked, where the cluster was resolved by deletion. Istreet for street, where it was resolved by insertion. He walk to work yesterday, where the tense went with the consonant.

Two automatic repairs

How the brain unconsciously alters an impossible consonant cluster

Speakers confronted with unfamiliar consonant clusters either delete a consonant or insert an extra vowel.

Which automatic repair mechanism does the speaker apply to the consonant cluster?
  • Deleting a consonant from the cluster
    Consonant deletion (ass for asked)Simplifies the cluster by dropping a consonant, stripping past tenses and plural markers
    • Drops grammatical endings
    • Sounds like a grammatical mistake to listeners
  • Inserting a vowel into the cluster
    Vowel insertion (istreet for street)Breaks up the cluster by adding an extra vowel, altering the syllable count of the word
    • Adds an extra syllable
    • Derails listener word boundary detection in the speech stream

Vowel insertion is more damaging because listeners rely on syllable counts to locate word boundaries in spoken English.

The quiet one is a cluster simplified so consistently that it stops being an error and becomes your version of the word. Strengths has three consonants at the end of it and almost nobody, including native speakers, produces all of them; the difference is that a native speaker drops the one everybody drops, and a learner drops a different one.

Which of these sounds your first language already keeps apart, and which it merges, is settled by ship and sheep: training your ear on one sound at a time, and what happens to an ending when the next word starts with a consonant is why spoken English runs words together, and how to hear it anyway. Both sit under being understood the first time you say it.

The question this opens is a grammatical one wearing a pronunciation coat. If a tense can be lost by dropping a sound, then some of what looks like a grammar problem in your speech is not one at all, and the two have to be told apart before either can be fixed: the errors that survive years of study, and why they stick.

Alongside this

Above this