A PHONEMIC TRANSCRIPTION KEY

Justin B Rye 13-May-05
(now Unicode-enriched)

FOREWORD

This page is intended to complement my other pages in two ways: for British non-linguists it's a rough guide to the scheme of transcription I'm using, while for non-British linguists it's an introduction to my accent.  If you are already familiar with both Kirshenbaum ASCII IPA and modern RP you're unlikely to find it very useful; and on the other hand if you don't know what I mean either by "<O> as in <bother>" or by "/A./ as in /'bA.D@/" then there's no convenient way I'm going to be able to get the leverage to explain either - or at any rate, not without a set of explanatory sound files tailored to your own particular native dialect, whatever that is (and web-browsing environment, likewise).  It is fairly technical, but that can't be helped: attempts to explain details like this purely in terms of "hard TH" and "long A sounds" just spread confusion.

TERMINOLOGY

Throughout this essay, example spellings and so forth are as usual marked out as follows...

And before I get started I'll need to explain some of the labels I'm sticking on things:

Phonetic

Of or pertaining to the precise articulatory and acoustic properties of speech sounds, irrespective of the way they're used in particular languages; for instance, the <K> in <key> is usually a palatalised and aspirated [k;<h>], phonetically quite different from the <K> in <ask>.

Phonemic

(Here I go again trying to define "phoneme" in a nutshell... compare the attempts on my Espe-Ranto, Spelling Reform and Futurese pages)

Of or pertaining to phonemes, the sounds treated as basic units in a given lingo.  What the <K>s in <key, ask> have in common is that both are forms of one English phoneme, /k/: the phonetic space between them is never used for telling English words apart (so English-speakers learn to disregard it).  The difference between <cod> and <god>, on the other hand, is that the first begins with /k/ and the second with a voiced /g/ - a subtle phonetic distinction that English does recognise as phonemically significant.

IPA

The alphabet of the International Phonetic Association, providing a standard notation of symbols and diacritics for representing the sounds of human languages, either on a phonemic level or in phonetic close-up.  Not to be confused with radiotelephony codes such as the so-called "NATO Phonetic Alphabet".

ASCII

The American Standard Code for Information Interchange is the basic 7-bit set of alphanumerics and punctuation symbols used as a lowest common denominator for text on the net.  Its 128 characters don't stretch to <ç> or <õ>, let alone the weirder squiggles of IPA - hence such makeshifts as...

KIRSHENBAUM ASCII IPA

Evan Kirshenbaum's widely accepted standard for encoding IPA into ASCII - here's his own introduction (for linguists), and a less technical one (for the a.u.e newsgroup).

REAL IPA

I ought to mention that here in the UK, what "I.P.A." most commonly stands for is "India Pale Ale", a style of beer (not to be confused with "IsoPropyl Alcohol").  Real Ale versions are common, and come to that there's also such a thing as a "rye IPA"... so referring to the full-strength, typewriter-unfriendly version of the phonetic symbolset as "Real IPA" is something of a fossilised pun.

RP

"Received Pronunciation", an accent better known perhaps as "BBC English".  Like "General American" it is socially dominant as a standard "educated" accent; but unlike GA, RP is strongly associated with a particular region and social class, that of the southern English (upper-)middle class.

JBR RP

My father being a retired Church of England vicar, I myself am definitively middle-class; I went to a primary school in rural Norfolk, then a public (i.e. private) secondary school in Norwich before coming to Edinburgh.  As a result my accent is fairly "posh", though by no means up to Hollywood Villain standards.  It has a few traceably East Anglian traits, and lacks several features becoming common in more southern areas; see the pronunciations of <insolvencies, Norwich, quarter> listed below.

Standard English

... Has nothing to do with this.  I'm not talking about "correct" grammar, or vocabulary, and certainly not spelling; I'm talking about accent, meaning how things are pronounced - you can quote Bob Marley or PG Wodehouse in any accent!

Voiced/voiceless

Standard labels for consonants, classifying them in terms of whether they are accompanied by vocal-cord buzzing.  If you don't recognise this idea, clutch your throat and go "SSS-ZZZ-SSS-ZZZ"!

Bilabial/dental/velar/etc

Standard labels for consonants, classifying them in terms of the part of the mouth involved (teeth-ridge, soft palate, or whatever).  These you probably can just pick up from the examples.

Stop/fricative/approximant

Standard labels for consonants, classifying them in terms of how thoroughly airflow is blocked.  During a stop like "T", no air can escape the mouth (until it's abruptly released); it scrapes through turbulently in fricatives like "F"; and approximants like "W" hardly impede the airflow at all.

Nasal/affricate/lateral

Three kinds of special exception to the above.  Nasal stops involve air escaping through the nose ("MMM"); affricates turn from stop to fricative ("TCH"); and laterals involve airflow around the sides of a central blockage ("LLL").

Front/central/back

Standard labels for vowels, classifying them in terms of where in the mouth they are articulated; that is, whether the front or back of the tongue is held closer to the palate.  Again, you may never have noticed this; but say "EEE-OOO-EEE-OOO" and pay attention to what your tongue's doing.

Close/half-close/half-open/open

Standard labels for vowels, classifying them in terms of the gap between tongue and palate at its narrowest point - smallest in "close" vowels like "OOO" and greatest in "open" vowels like "AAH".

Rounded/unrounded

Standard labels for vowels produced with or without lip-rounding.  This distinction is obvious in vowels like "OOO-EEE", but trickier in the middle - Americans are likely to have trouble following my "AW sounds".

INVENTORY

JBR RP features the following supply of consonants and vowels, here listed in an order which is more or less "systematic" rather than quasi-alphabetic.  The "Real IPA" equivalents are a new addition, for the convenience of the increasing proportion of visitors I can expect to be using Unicode-capable browsers.

/b/ as in <EBB>, <BIB> (Real IPA /b/)

characteristically a voiced bilabial stop

/d/ as in <ADD>, <DID> (Real IPA /d/)

characteristically a voiced alveolar stop

/dZ/ as in <EDGE>, <JUDGE> (Real IPA /ʤ/)

characteristically a voiced palato-alveolar affricated stop - counted as one item

/g/ as in <EGG>, <GAG> (Real IPA /ɡ/)

characteristically a voiced velar stop

 

/p/ as in <UP>, <PIP> (Real IPA /p/)

characteristically a voiceless bilabial stop

/t/ as in <IT>, <TOT> (Real IPA /t/)

characteristically a voiceless alveolar stop

/tS/ as in <ETCH>, <CHURCH> (Real IPA /ʧ/)

characteristically a voiceless palato-alveolar affricated stop - counted as one item

/k/ as in <OAK>, <KICK> (Real IPA /k/)

characteristically a voiceless velar stop

 

/v/ as in <OF>, <VERVE> (Real IPA /v/)

characteristically a voiced labiodental fricative

/D/ as in <EITHER>, <THE> (Real IPA /ð/)

characteristically a voiced dental fricative - n.b. distinct from /T/

/z/ as in <IS>, <ZOOS> (Real IPA /z/)

characteristically a voiced alveolar fricative

/Z/ as in <SEIZURE>, <VISION> (Real IPA /ʒ/)

characteristically a voiced palato-alveolar fricative

 

/f/ as in <IF>, <FIFE> (Real IPA /f/)

characteristically an unvoiced labiodental fricative

/T/ as in <ETHER>, <THIRTIETH> (Real IPA /θ/)

characteristically an unvoiced dental fricative - n.b. distinct from /D/

/s/ as in <ASS>, <SAUCE> (Real IPA /s/)

characteristically an unvoiced alveolar fricative

/S/ as in <ASH>, <SURE> (Real IPA /ʃ/)

characteristically an unvoiced palato-alveolar fricative

 

/m/ as in <AM>, <MUM> (Real IPA /m/)

characteristically a voiced bilabial nasal stop

/n/ as in <AN>, <NUN> (Real IPA /n/)

characteristically a voiced alveolar nasal stop

/N/ as in <INK>, <HANGING> (Real IPA /ŋ/)

characteristically a voiced velar nasal stop

/l/ as in <ILL>, <LULL> (Real IPA /l/)

characteristically a voiced alveolar lateral approximant

 

/r/ as in <FRY>, <RAH-RAH> (Real IPA /ɹ/)

characteristically a voiced alveolar approximant

(not the US retroflex or the Scots rolled R)

/w/ as in <QUIT>, <WAH-WAH> (Real IPA /w/)

characteristically a voiced labiovelar approximant

/j/ as in <VIEW>, <YOYO> (Real IPA /j/ - n.b. not /y/)

characteristically a voiced palatal approximant

/h/ as in <HUE>, <HA-HA> (Real IPA /h/)

characteristically an unvoiced glottal approximant

 

/I/ as in <ITCH>, <BID> (Real IPA /ɪ/)

characteristically a short close front unrounded vowel

/E/ as in <ETCH>, <BED> (Real IPA /ɛ/)

characteristically a short half-open front unrounded vowel

/@/ as in <UP>, <BUD> (Real IPA schwa, /ə/)

characteristically a short open central unrounded vowel

/&/ as in <AT>, <BAD> (Real IPA /æ/)

characteristically a short open front unrounded vowel

/U/ as in <HOOD>, <BULL> (Real IPA /ʊ/)

characteristically a short close back rounded vowel

/A./ as in <ON>, <BOD> (Real IPA /ɒ/)

characteristically a short open back rounded vowel - the /./ means "rounded"

 

/I@/ as in <HERE>, <BEARD> (Real IPA /ɪə/)

characteristically a glide from [I] towards [@]

/E@/ as in <HAIR>, <BARED> (Real IPA /ɛə/)

characteristically a glide from [E] towards [@], or long hybrid vowel

/V":/ as in <HER>, <BIRD> (Real IPA /ɜː/)

characteristically a long half-open central unrounded vowel - the /"/ means "centralised"

/A:/ as in <ARCH>, <BARD> (Real IPA /ɑː/)

characteristically a long open back unrounded vowel

/U@/ as in <MOOR>, <CURED> (Real IPA /ʊə/) - uncommon, often replaced by /O:/

characteristically a glide from [U] towards [@], or long hybrid vowel

/O:/ as in <HAWK>, <BOARD> (Real IPA /ɔː/)

characteristically a long half-open back rounded vowel

 

/i:/ as in <HEAT>, <BE> (Real IPA /iː/)

characteristically a long close front unrounded vowel

/EI/ as in <HATE>, <BAY> (Real IPA /ɛɪ/)

characteristically a glide from [E] towards [I]

/AI/ as in <HEIGHT>, <BY> (Real IPA /ɑɪ/)

characteristically a glide from [A] towards [I]

/OI/ as in <HOIST>, <BOY> (Real IPA /ɔɪ/)

characteristically a glide from [O] towards [I]

 

/&U/ as in <HOUSE>, <BOUGH> (Real IPA /æʊ/)

characteristically a glide from [&] towards [U]

/u:/ as in <HOOT>, <BOO> (Real IPA /uː/)

characteristically a long close back rounded vowel

/@U/ as in <HOSE>, <BEAU> (Real IPA /əʊ/)

characteristically a glide from [@] towards [U]

For the benefit of visitors who still see some of those Real IPA symbols as "unrecognised character" signs, here's a chart of them in graphical form:

COMPLICATIONS

"Characteristically"

My use of this weasel-word signals that I'm ignoring some phonetic subtleties; for example while everyone more or less agrees on the prototypical phonetic realisation of /t/, there is some disagreement in specific contexts such as <glottal>.  Many US accents prescribe a "flapped D" there, and many UK accents a "glottal stop"; JBR RP has a clearcut "T" in that particular context, but allows a glottalised /t/ in a more limited family of words.

Phonetically Long

When I say something is stressed or long, that doesn't necessarily mean what you think.  Stressed vowels are loud and rather high-pitched; long vowels are characterised by prolonged duration; but the difference between, say, <fur> and <fury> isn't one of stress or vowel length - the two are simply different (/'fV":/, /'fjU@ri:/).

Phonemically Long

That said, the length-marker /:/ that forms part of the symbol for (e.g.) the phoneme /O:/ does not mean that it always takes longer to say than (e.g.) the phoneme /E/; only that length is one of its distinguishing features.  All other things being equal - as in the pair <bought/bet> - I take more time over the former; but other things very rarely are equal.

Complex And Compound Sounds

English has several other kinds of genuinely "long" sounds, including affricated stops like <j> (=/dZ/) and diphthongs like <oy> (=/OI/).  A case could be made for splitting each of these into subcomponents, but at least for my own accent I find such analyses unnatural.

Conspicuously Absent

The typical pronunciations of the letters <X> and <Q> do not correspond to single phonemes (the sounds are transcribed as the sequences /ks/, /kw/).  Likewise, the sound of <U> as in <unit> is just "YOO" - /j/ followed by /u:/ - and needs no separate sign of its own.

Stress-marking

The difference between the adjective <ABstract> and the verb <abSTRACT> is shown in my transcriptions by /'/ marks before stressed syllables: /'&bstr&kt/, /&b'str&kt/.  Secondary stress (not as emphatic as the word's main stress, but stronger than neighbouring syllables) receives a preceding /,/ mark - thus <CHAracteRIStically> is /,k&r@kt@'rIstIkli:/.

Schwa

The symbol /@/ stands for three distinct sounds, each of which occurs in <abundance>, /@'b@nd@ns/.  In phonetic terms, that's [@bVndn-s]; and for many accents it makes sense to consider the three forms as separate entities.  However, they don't seem to be used contrastively as phonemes in my own speech: the pronunciation of /@/ is always predictable from the surrounding context.  Accordingly, I am treating them as one phoneme, and standardising my phonemic notation here on the symbol for the neutral vowel known as the "schwa".

EXAMPLES

Please remember that this is an attempt at a representative selection of accurately reported spoken forms, not a prescriptive guide to how English should be spoken.  These may be what I normally say, but I'm not denying the existence of alternatives (even in BBC English).

SENTENCES

Some full-length case-studies - they may not be phrases that come up frequently in conversation, but all of them have been used before, and "The North Wind and the Sun" is often seen as a sample text in comparative studies.  In each case, I am speaking with normal levels of emphasis at ordinary conversational speeds.

<With tenure, Suzie'd have all the more leisure for yachting, but her publications are no good.>

/wID 'tEnj@ 'su:z,i:d h&v O:l D mO: 'lEZ@ f@ 'jA.tIN b@t hV": 'p@blI,kEIS@nz @ n@U gUd/

For certain US accents and phonological analyses, that's a "quick brown fox" sentence, providing a full inventory of phonemes; for me, even if we ignore the somewhat marginal /U@/, it would require the addition of <air thigh rear chow joy>.  2004 addendum: at last I've found a perfect JBR-RP pangram, with exactly one use of each phoneme in my inventory:

<Are those shy Eurasian footwear, cowboy chaps, or jolly earthmoving headgear?>

/A:  D @U z  S AI  j U@ 'r EI Z @ n  'f U t w E@  'k &U b OI  tS & p s  O:  'dZ A. l i:  'V": T ,m u: v I N  'h E d g I@/

Yes, I know there are far too many <R>s in that for it to work for all you rhotic types!  Honestly, for me there's only one - I've increased the spaces between segments to make it easier if you want to double-check.  Now for some less elaborately contrived example sentences:

<Bother, father caught hot coffee in the carpark>

/'bA.D@ 'fA:D@ kO:t hA.t 'kA.fi: In D@ 'kA:,pA:k/

<I teach Ferdinand the calm cat to fetch cold cups of coffee.  Who knows more about tasting things?  He's used the book.>

/AI ti:tS 'fV":dI,n&nd D@ kA:m k&t t@ fEtS k@Uld k@ps @v 'kA.fi:   hu: n@Uz mO:r @b&Ut 'tEIstIN TINz   hi:z ju:zd D@ bUk/

<The North Wind and the Sun were disputing which was the stronger, when a traveller came along wrapped in a warm cloak.  They agreed that the one who first succeeded in making the traveller take his cloak off should be considered stronger than the other.
Then the North Wind blew as hard as he could, but the more he blew the more closely did the traveller fold his cloak around him, and at last the North Wind gave up the attempt; then the Sun shone out warmly, and immediately the traveller took off his cloak.
And so the North Wind was obliged to confess that the Sun was the stronger of the two.>

/D@ nO:T wInd @n D@ s@n w@ dIs'pju:tIN wItS w@z D@ 'strA.Ng@ wEn @ 'tr&vl@ kEIm @'lA.N r&pt In @ wO:m kl@Uk   DEI @'gri:d D@t D@ w@n hu: f@:s s@k'si:dId In 'mEIkIN D@ 'tr&vl@ tEIk Iz kl@Uk A.f SUb bi: k@n'sId@d 'strA.Ng@ D@n Di: '@D@
DEn D@ nO:T wInd blu: @z hA:d z i: kUd b@t D@ mO:r i: blu: D@ mO: 'kl@Usli: dId D@ 'tr&vl@ f@Uld hIz kl@Uk @'r&Und hIm   @n @t lA:s D@ nOT wInd gEIv @p Dj @'tEmp   DEn D@ s@n SA.n &Ut 'wO:mli: @n I'mi:dZ@tli: D@ 'tr&vl@ tUk A.f Iz kl@Uk
@n s@U D@ nO:T wInd w@z @'blAIdZ t@ k@n'fEs D@t D@ s@n w@z D@ 'strA.Ngr @v D@ tu:/

2005 addendum: here's a token sample of what JBR RP looks like (recycling the "pangram" example sentence) when it's transcribed in the Real IPA characters and displayed via Unicode:

/ɑː ðəʊz ʃɑɪ jʊəˈɹɛɪʒən ˈfʊtwɛə ˈkæʊbɔɪ ʧæps ɔː ˈʤɒliː ˈɜːθˌmuːvɪŋ ˈhɛdɡɪə/

With luck that should look a lot better than the ASCII version, but it's hard to be sure - I've got no basis for assuming you have any particular font installed that handles IPA well, so I'm having to leave it as a default serif typeface.  You should be able to find a more appropriate one by modifying your browser configuration... just don't start mailing me your MS-IE tech-support questions, okay?