This is an introduction to the International Phonetic Alphabet (IPA) for English-speaking Wikipedians. Its purpose is to explain the IPA's basic principles to English speakers. IPA clearly and unambiguously indicates how a word or name actually sounds with one letter for each sound. Wikipedia uses IPA because it's the global standard used by professionals and the only system used in most schools in the world.
IPA's most daunting feature is that it has discrete letters for almost all of the distinctive sounds found in the world's languages. (See International Phonetic Alphabet#Letters.) Fortunately, using the IPA for English requires learning only the following small subset of them:
- Vowels: English orthography uses 6 vowel letters (a, e, i, o, u, y) to represent some 15 vowel sounds. While the English system is compact, it is also ambiguous. The IPA is unambiguous, representing each vowel sound with a unique letter or sequence. (See the vowel audio chart). Note that most of what in English are called "long vowels", A, E, I, O, U, are in fact combinations of two sounds, which is why they are transcribed in the IPA with two letters apiece: /eɪ/, /iː/, /aɪ/, /oʊ/, and /juː/.
- Consonants: IPA consonants are mostly intuitive to an English speaker, with the same letter used for the same sound. Thus you already know /b, d, f, ɡ, h, k, l, m, n, p, r, s, t, v, w, z/, as long as you remember that these each have a single sound. For example, /ɡ/ always represents the sound of get, never of gem, and /s/ always the sound of so, never of rose. The letter which most confuses people is /j/, which has its Central-European values, a y sound as in English hallelujah. Two English consonant sounds, ch in chair and j in jump, are transcribed with two IPA letters apiece, /tʃ/ and /dʒ/. The English digraphs ch, ng, qu, sh, th are not used. See and hear also consonants audio chart.
The first principle is to not use English alphaphonemic pronunciations, as if you were reading the English alphabet. In the words below, the vowel letters are pronounced as in the English alphabet, but this is not a system found in any other language:
- A: make, angel
- E or EE: meet, delete
- I: rice
- O: note
- U: use
The English digraphs ee, oo, au, ei, ai, ou, ie, eu, etc. are not used.
Several of these sounds are actually two vowel sounds combined, rather than pure vowel sounds as they are in Spanish or Italian: The letter A is pronounced /eɪ/, E, EE is /iː/, I is /aɪ/, O is /oʊ/, and U is /juː/. In the IPA, the letter /j/ is used for the English Y sound, thus you and ewe are transcribed /juː/. (See below.) While transcribing in the IPA, you can write English alphaphonemic vowels as capitals: [rAk], [sEEm], [rIs], [dOt], [Uz], etc., and then convert from the conventions above:
- A: /eɪ/ rake /reɪk/ (not /raɪk/, which would be Germanic reich)
- E: /iː/ seem /siːm/
- I: /aɪ/ rice /raɪs/ (not /reɪs/, which would be race)
- O: /oʊ/ dote /doʊt/ (not /daʊt/, which would be doubt)
- U: /juː/ use /juːz/
Notes: English commonly requires ea or ee to write the /iː/ sound: read, reed.
A w-like sound can be heard at the end of O in words like echoing (say: echo-echo-echoing, and it may come out like echo-wecho-wecho-wing) and after the co- in cooperate; that is what the /ʊ/ in the transcription /oʊ/ captures.
There are a couple other long vowels and diphthongs in English: OO sound in food (but not good) is written /uː/: /fuːd/. That is, it is written like the vowel of use without the initial y sound /j/. As noted above, the OW sound of doubt or cow is written /aʊ/. There is also the OY sound /ɔɪ/ of joy, /dʒɔɪ/.
English short vowels are all transcribed by a single letter in the IPA.
Because English short vowels a e i o u are closer to the Classical pronunciation (still found in Spanish and Italian) than the long vowels are, it is the short vowels which are transcribed with IPA letters which resemble the English letters a e i o u. However, they are modified to show that they aren't exactly the Classical sounds. For the a sound of cat, the Old English letter ash, ‹æ›, was resurrected: /kæt/. The e i u sounds of pet, pit, put (not putt) were originally written as capital letters, and that is sometimes still done with manual typewriters. However, small caps looked better, so they were for a time written E I U. These took more cursive forms over time, and are today written /ɛ ɪ ʊ/: pet /pɛt/, pit /pɪt/, put /pʊt/. The latter, of course, is also the short oo sound of good /ɡʊd/. The u vowel of putt or cut, is written as an upturned letter "v" e.g. cut /kʌt/. The exact pronunciation of this sound varies considerably among English dialects.
The a sound in bra is written with a cursive a. Because it's long in many dialects, it's /ɑː/ in the IPA: /brɑː/. Likewise, the aw sound of law is long in many dialects, but different than the bra sound. It's written with an "open" o (just as /ɛ/ looks like an open e, since a small cap o looks just like a regular oː law /lɔː/. (Many of you might not make this distinction, in which case you can think of these vowel letters as being the same when reading the IPA.) For those of you who distinguish it, there is a third similar sound, the o of mop. This is written with the bra vowel letter rotated 180°: mop /mɒp/. (A rather unusual IPA letter, as that's an unusual vowel, not found in many languages). The vowel sound in bird is written as an upturned /ɛ/], therefore it is written as /bɜrd/.
Finally, there's the slurred schwa sound found in many unstressed syllables, as at the end of sofa. This is written /ə/, a symbol used in many US dictionaries. The stressed syllable is marked with a tick: sofa /ˈsoʊfə/. Note that the letter /ə/ is never used for a stressed vowel; for words like cut, we use /ʌ/: butter /ˈbʌtər/, cuppa /ˈkʌpə/.
While most IPA consonants are intuitive for English speakers, there are some caveats:
- The sound of the consonant Y is /j/, as in yes /ˈjɛs/ and yellow /ˈjɛloʊ/.
- (This is the value the letter J has in central European languages like German and Polish. The IPA letter /y/ is used for a non-English vowel, the French u, German ü, and Swedish y sound.)
- The NG sound of sing is written by combining the letter n with the tail of the g, /ŋ/, as in sing /ˈsɪŋ/. This is not the same as the sound in finger, which has an extra g sound: /ˈfɪŋɡər/. This sound also appears when n comes before a k, such as in sink /ˈsɪŋk/.
- The digraph TH is used for two sounds in English. Since the IPA uses a single letter for each sound, two new letters are required for these two sounds:
- The sound of the digraph SH is transcribed with the 'stretched' S seen in old books. It's used in its cursive form, /ʃ/, to make it easier to read, as in push /ˈpʊʃ/ and shelf /ˈʃɛlf/.
- There is a sound with no letter or digraph in English, though sometimes written ZH in foreign words. It's usually written si, as in vision. In the IPA, it's written with a 'stretched' Z, /ʒ/: vision /ˈvɪʒən/.
- As noted above, the digraph CH is a sequence of sounds, T plus SH. This may be hard for an English speaker to hear, but is obvious to a French speaker, which is why we get spellings like Tchaikovsky but also catch in English. (Adding a t to ch doesn't make any difference, because the ch already has a t sound within it.) The IPA uses the same stretched S for this sound here as anywhere else: itch /ˈɪtʃ/.
- Similarly, the English consonant J is a sequence with a d sound in it. For instance, in judge, adding the d doesn't affect the consonant sound, just the vowel. In the IPA, this is transcribed /dʒ/: jump /ˈdʒʌmp/, judge /ˈdʒʌdʒ/, or Jesus /ˈdʒiːzəs/.
- Finally, the IPA letter [r] is officially a trill, as in Italian and Spanish. The rather unusual English R sound is transcribed with a turned r, [ɹ]. However, since this makes no difference within English, and not all English dialects actually use the [ɹ] sound, it's very common to see English R transcribed with a plain /r/, and that's the convention used on Wikipedia.
- English is divided into rhotic and non-rhotic accents. Non-rhotic accents such as Received Pronunciation and Australian English do not pronounce [ɹ] at the end of a syllable. However, Wikipedia convention writes in a way that recognizes the rhotic pronunciation, even for places or words normally pronounced with a non-rhotic accent. For example, the pronunciation of the British town of Guildford is written as /ˈɡɪlfərd/, though the local pronunciation is /ˈɡɪlfəd/. Wikipedia does not follow the usual approach of many United Kingdom dictionaries which place the final r in parentheses.
The English digraphs ch, ng, qu, sh, th are not used.
IPA's purpose and Wikipedia's use of IPA
IPA's purposes are to 1) represent the phonetics of words (how they sound) and 2) to give samples of the phonology of a language (how the language as a whole sounds). The second purpose concerns only linguists. The first purpose concerns any interested reader, but only to a limited degree, as transcribing words into IPA does not need to be perfect or overly precise (something for fluent IPA users to consider). The word "transcribe" is used to distinguish this from normal writing or spelling, which has other purposes (such as preserving word etymologies and meaning).
IPA is complex enough to represent nearly anything, but high-fidelity transcriptions will use glyphs that are unfamiliar to English readers and unpracticed in English phonology. For example a transcription of something like the Icelandic name Eyjafjallajökull is pronounced [ˈeiːjaˌfjatl̥aˌjœːkʏtl̥] (listen), meaning island-mountain glacier, may approximate Icelandic phonology, but such information will likely be too much for English readers, who may need to reference the name using what is at best an approximate pronunciation anyway. (Often an English version of a foreign name will try to employ translation in combination with partial transcription, but this often stays unnecessarily close to the original spelling and therefore prevents English speakers from using sounds they can easily produce. For example Eyja-fjalla glacier (['eija-f'jala] glacier) is a sufficiently close approximation, but Eyja-fjatla glacier (['eija-f'jatla] glacier) would be closer and still easy to pronounce.)
- The English digraphs ee, oo, au, ei, ai, ou, ie, eu, etc. are not used at all in the IPA, or similar combinations of two letters are used to logically represent two sounds, for example /eɪ/ for the two vowel sounds in "may", not the single vowel sound at the end of "receive ".