How to read the IPA

The International Phonetic Alphabet has one rule: one symbol, one sound, always.

That is the whole point of it. English spelling gives you though, through, tough, thought, thorough — five pronunciations of ough. IPA gives you /ðəʊ/, /θruː/, /tʌf/, /θɔːt/, /ˈθʌrə/, and every symbol in those means exactly what it means everywhere else, in every language.

You do not need to learn all 107 letters plus diacritics. You need roughly twenty symbols and the logic of the layout. After that, anything unfamiliar is a lookup — and every symbol in our IPA reference has a page explaining how to make it.

Start with what does not surprise you

Most consonant symbols are exactly what a Latin-alphabet reader expects:

/p b t d k ɡ m n f v s z l w h/

These mean what you think. That is fifteen symbols for free.

The symbols that do surprise you

This is the list worth actually memorising — the places where IPA departs from the letter you would have guessed.

Symbol	Sound	As in
/ʃ/	sh	shoe
/ʒ/	zh	vision, French rouge
/tʃ/	ch	church
/dʒ/	j	judge
/θ/	voiceless th	think
/ð/	voiced th	this
/ŋ/	ng	sing
/j/	y	yes — not the English letter J
/ɲ/	ny	Spanish niño, French agneau
/x/	ch as in Bach	German Bach, Spanish hoja
/ʁ/	French/German r	French rouge
/r/	trilled r	Spanish perro — not English r
/ɾ/	tapped r	Spanish pero, American better
/ɹ/	English r	red

Two of these cause most of the confusion in practice.

/j/ is not the English J. It is the y in yes. This follows German, Dutch and the Scandinavian languages, and it trips up English speakers constantly.

/r/ is not the English R. In IPA, /r/ is a trill — Spanish perro, Italian carro. The English r is properly /ɹ/. Most dictionaries cheat and write /r/ for English anyway, because /ɹ/ looks alarming and there is no ambiguity within an English dictionary. Know that the cheat exists.

Vowels: the actual difficulty

Consonants are mostly a lookup problem. Vowels are a conceptual problem, because they are not discrete points — they are positions in a continuous space, and the symbols are landmarks in it.

The space has two dimensions: how high the tongue is, and how far forward.

          FRONT        CENTRAL        BACK
        ┌──────────────────────────────────┐
  CLOSE │  i  y                     ɯ  u   │   see, tu (Fr), you
        │    ɪ  ʏ                 ʊ        │   sit, put
  MID   │  e  ø                     ɤ  o   │   Fr été, Fr peu, Fr eau
        │        ə                         │   about  ← schwa
        │    ɛ  œ                 ʌ  ɔ     │   bed, Fr sœur, but, thought
  OPEN  │      æ                  ɑ  ɒ     │   cat, father, lot
        │        a                         │   Fr patte
        └──────────────────────────────────┘

Where the pairs appear together (i/y, e/ø, o/ɤ), the difference is lip rounding: the second is the same tongue position with rounded lips. This is why French tu /ty/ is hard for English speakers — the tongue position is that of see, but the lips are those of ooh, and English never combines them.

The single most important vowel is /ə/, the schwa, sitting dead centre. It is the sound of the tongue at rest and the most common vowel in spoken English, appearing in every unstressed syllable: about /əˈbaʊt/, sofa /ˈsəʊfə/, the /ðə/. Failing to reduce unstressed syllables to schwa is what makes English sound robotic; over-applying it is what makes an English speaker’s Spanish and Italian sound foreign, because those languages keep every vowel full. Full detail on the schwa page.

Length, stress and the marks around symbols

Mark	Meaning	Example
ː	long	/iː/ in sheep, Finnish tuuli /ˈtuːli/
ˈ	primary stress on the next syllable	/ˈwɔːtə/
ˌ	secondary stress	/ˌɪntəˈneɪʃən/
.	syllable boundary	/ˈhæ.pi/
̃	nasalised	French bon /bɔ̃/
ʰ	aspirated	English pin [pʰɪn]
̥	devoiced	[d̥]

The stress mark goes before the stressed syllable, not on the vowel. This is the most frequently misread convention in the entire system. /ˈrekɔːd/ is the noun; /rɪˈkɔːd/ is the verb.

Slashes vs brackets

/slashes/ mean phonemic: the abstract sound categories of the language, with predictable detail stripped out.

[brackets] mean phonetic: what was actually produced, including detail the language treats as irrelevant.

English pin is /pɪn/ phonemically and [pʰɪn] phonetically — the aspiration is real but automatic, so a phonemic transcription omits it. Both are correct at their own level. The distinction is phonemes vs allophones, and it matters more than it sounds like it should: the things your language treats as “the same sound” are exactly the things you will be unable to hear in another language.

As a learner, you almost always want slashes.

The consonant grid

Consonants are classified on three axes, and the chart is just those three axes drawn out.

Place — where the constriction is, front of the mouth to back:

lips   teeth   ridge   hard palate   soft palate   uvula   throat
 p b    f v     t d s      ʃ ʒ j        k ɡ x        ʁ q      h

Manner — what kind of constriction:

• Stop — airflow blocked completely, then released: /p b t d k ɡ/
• Fricative — narrowed enough to make turbulence: /f v s z ʃ ʒ θ ð x/
• Nasal — blocked at the mouth, air out the nose: /m n ŋ ɲ/
• Approximant — narrowed, but not enough for turbulence: /l ɹ j w/
• Trill / tap — vibrating or single contact: /r ɾ/
• Affricate — stop then fricative: /tʃ dʒ/

Voicing — vocal folds vibrating or not. /s/ vs /z/, /f/ vs /v/, /p/ vs /b/. Put two fingers on your throat and alternate sss and zzz: the buzz is voicing.

Every consonant symbol is one point in that three-dimensional space. Once you can read the axes, an unfamiliar symbol tells you how to make the sound — “voiceless uvular fricative” is a set of instructions, not a name.

The twenty-minute version

If you want the useful subset and nothing else:

1. The fifteen consonants that mean what you expect — free.
2. The fourteen surprising consonants in the table above.
3. Schwa /ə/, and the four or five vowels of your target language that your own language lacks.
4. The length mark ː and the stress mark ˈ.
5. That /j/ is y and /r/ is a trill.

That is enough to read a dictionary transcription in any language and to write a note to yourself that means something specific six days later.

Why bother

Because “your R is a bit off” is not actionable and “/ʁ/, not /ɹ/ — uvular, back of the throat, no tongue movement” is.

The value is not scholarly. It is that a shared, precise vocabulary lets a teacher name a problem exactly and lets a student practise the right thing alone, a week later, without guessing. It converts pronunciation feedback from an impression into an instruction.

Every symbol in the reference has its own page with the articulation, the languages that use it, and the sounds it is most often confused with — which is where to go when you meet one you do not know.