Sounds of the IPA

John Wells and Jill House

Department of Phonetics and Linguistics, University College
London



W:	We've made this recording in order to illustrate the sounds
associated with the phonetic symbols shown on the 1993 version
of the Chart of the International Phonetic Alphabet.

H:	This is a copyright recording, and may not be reproduced
without permission. It was recorded on the tenth of March 1995,
at University College London. My name is Jill House.

W:	And mine is John Wells. The recording was monitored by
Michael Ashby and John Baldwin, and the sound engineers were
Steve Nevard and David Cushing. Copies of the recording are
available from the Department of Phonetics and Linguistics,
University College London, Gower Street, London WC1E 6BT, United
Kingdom. 

H:	Before we start, we'd like to call attention to two points
that should always be kept in mind. The first is that any
phonetic symbol _ for vowel, consonant, or prosodic feature _
can be applied to a range of sound-types. We have aimed at what
we take to be a neutrally typical version of each. 

W:	The symbol for 'aspirated', for example, does not specify the
degree of aspiration _ how many milliseconds of delay between
plosive release and onset of voicing, delay during which
turbulence may be audibly present _ and in fact there is a
continuous range of possible degrees of aspiration.

H:	The second, and related, point is that when IPA symbols are
used to transcribe the sounds of a particular language, the
neutrally typical versions we try to offer may not correspond
exactly to the language-specific sounds. 

W:	The same letter 'ell' (l) is used in the transcription of
English, Spanish, Zulu, and Korean; in each case not only will
it cover a range of allophonic variants, but also the default
realization will vary somewhat from language to language. So it
is inevitable that our version of the voiced alveolar lateral,
[lA], may not quite correspond to the typical sound of the
language you may be interested in.

H:	We start with the table at the top of the Chart. It
illustrates consonants produced with a pulmonic egressive
air-stream mechanism. We begin with the plosives: sounds made
with a complete obstruction of the air flow. 

W:	They come in pairs, voiceless and voiced. At each place of
articulation you will hear first the voiceless plosive, then the
voiced one. We will make the voiceless plosives with a slight
aspiration, and the voiced ones fully voiced. Each consonant
will be said twice: first followed by a vowel, of the type [A:],
and then with a vowel both before and after.

H:	First, then, the bilabials, made by the two lips: 	[pA	bA]

W:	[ApA 	AbA]. Then the alveolars, made by the tongue tip
against the alveolar ridge: [tA 	dA].

H:	[AtA	AdA]. The retroflex plosives, made by curling the tongue
tip back and articulating against the rear slope of the alveolar
ridge: [ęA	żA].

W:	[AęA	AżA]. The palatals, made by the front of the tongue
against the hard palate: [cA	SA].

H:	[AcA	ASA]. The velars, made by the back of the tongue against
the velum, the soft palate: [kA	gA].

W:	[AkA	AgA]. The uvulars, made by the extreme back of the
tongue against the uvula: [qA	OA].

H:	[AqA	AOA].  Then the glottal plosive, made by bringing the
vocal folds together. Here of course voicing is not possible:
[?A].

W:	[A?A].  



H:	Next we take the nasals. Their articulation is like that of
the plosives, except that the soft palate is lowered, allowing
the airstream to escape through the nasal cavity. We pronounce
all the nasals voiced. First, a bilabial nasal : [mA].

W:	[AmA]. Then the labiodental nasal, made by articulating with
the lower lip against the upper teeth: [FA].

H:	[AFA]. Alveolar: [nA].

W:	[AnA] A retroflex nasal: [cA].

H:	[AcA]. And a palatal: [JA].

W:	[AJA]. Velar: [NA].

H:	[ANA]. A uvular nasal: [żA].

W:	[AżA].



H:	Next are the trills, involving the repeated vibration of one
articulator against another. There are only three of these, and
we pronounce them all voiced. First, a bilabial trill: [A].

W:	[AA]. Then an alveolar trill: [rA].  

H:	[ArA]. 

W:	And a uvular trill: [3A]. 	[A3A]. 



H:	Like a trill, but involving just a single touch, is a tap or
flap. The alveolar tap: [ųA].

W:	[AųA]. And a retroflex flap: [}A].

H:	[A}A].



W:	Now come the fricatives. They involve forcing the airstream
through a narrow gap, which causes it to be turbulent: the
result is heard as friction. The fricatives come in pairs, with
first a voiceless and then a voiced one at each place of
articulation. We begin with the bilabial fricatives: [-A	BA].

H:	[A-A	ABA]. Then the labiodentals: [fA	vA].

W:	[AfA	AvA]. Dental fricatives: [TA	    DA].

H:	[ATA	ADA].  Alveolars: [sA     zA].

W:	[AsA	AzA]. And postalveolars, sometimes known as
palatoalveolars: [SA  ZA].

H:	[ASA	AZA]. The retroflex fricatives are ['A	AA].

W:	[A'A	AAA]. Then the palatals: [CA    sA].

H:	[ACA	AsA]. Velars: [xA	GA].

W:	[AxA	AGA]. And uvular fricatives: [XA	RA].

H:	[AXA	ARA]. Then we take the pharyngeal fricatives, made by
constricting the pharynx: [`A	    A].

W:	[A`A	AA]. And last along the fricative row, the so-called
glottal fricatives: [hA	A].

H:	[AhA	AA]. And then there are also the lateral fricatives, at
the alveolar place of articulation: [KA	­A].

W:	[AKA	A­A].



H:	Approximants are similar in their manner of articulation to
fricatives, but are produced without friction, without making
the airstream turbulent. They are all pronounced with voicing.
First, the labiodental approximant: [PA].

W:	[APA]. Next, an alveolar or postalveolar approximant: [œA].

H:	[AœA]. And a retroflex: [OA].

W:	[AOA]. A palatal approximant or semivowel: [jA].

H:	[AjA]. And a velar one: [ŻA].

W:	[AŻA].



H:	The lateral approximants are articulated with the air
escaping around a central obstruction. We pronounce them voiced.
The alveolar lateral is [lA].

W:	[AlA]. Retroflex: [YA].

H:	[AYA]. The palatal lateral: [LA].

W:	[ALA]. And a velar lateral: [›A].

H:	[A›A]. That brings us to the end of the consonants shown in
the main grid of the chart.



W:	Next we come on to the non-pulmonic consonants. Their
initiation originates somewhere other than in the lungs. First
we take the clicks, made with an oral or velaric ingressive
airstream mechanism. We pronounce them without any specific
simultaneous articulation, and they are accordingly voiceless.
First, the bilabial click: [®A].

H:	Between vowels, that's [A®A]. The dental click is [A] (old
symbol [,], Zulu c, affricated). You will notice that it has an
affricated release.

W:	[AA]. An alveolar or postalveolar click, often called
retroflex,  is [!A] (old [Ÿ], Zulu q; 'hollow' sound).

H:	[A!A]. A so-called palatoalveolar click is [’A]; it has
alveolar contact and an abrupt release ('flat' sound,
dental/alveolar, no affrication). 

W:	[A’A]. And lastly the alveolar lateral click is [€A] (old
[Ž], Zulu x; (post-)alveolar, affricated).

H:	[A€A]. 



W:	The voiced implosives are made with a pharyngeal or 
glottalic ingressive airstream mechanism. The bilabial implosive
stop is [€A].

H:	[A€A]. Dental or alveolar, [_A].

W:	[A_A]. And a palatal implosive: [>A].

H:	[A>A]. Velar: [<A].

W:	[A<A]. Then the uvular: [A].

H:	[AA]. The ejectives also have a pharyngeal or glottalic
airstream mechanism, but an egressive one. They are voiceless.
First, the bilabial ejective stop: [pA].

W:	[ApA]. Dental or alveolar (this one is alveolar): [tA].

H:	[AtA]. Velar: [kA].

W:	[AkA]. And lastly, an alveolar fricative ejective: [sA].

H:	[AsA].

W:	Now we come to the vowels. We'll take these in the order in
which they're displayed on the chart, reading across. We'll say
each vowel first on a falling tone, and then on a mid level
tone. We'll make each vowel fairly long.

H:	We start with the close vowels, which can also be termed
high. Close front unrounded, cardinal one, (falling) [i].

W:	(level) [i]. Close front rounded, cardinal nine, (falling)
[y].

H:	(level) [y]. Close central unrounded, cardinal seventeen,
barred-I, (falling) [1].

W:	(level) [1]. Close central rounded, cardinal eighteen,
barred-U, (falling) [}].

H:	(level) [}]. Close back unrounded, cardinal sixteen,
(falling) [M].

W:	(level) [M]. Close back rounded, cardinal eight, (falling)
[u].

H:	(level) [u]. Next the chart shows three vowels that are lax
or mid-centralized counterparts of some we've just had. Fairly
front, fairly close unrounded, laxed cardinal one, small cap I:
(falling) [I].

W:	(level) [I]. Then the corresponding rounded vowel, laxed
cardinal nine, small cap Y, (falling) [Y].

H:	(level) [Y]. And then fairly back, fairly close rounded,
laxed cardinal eight: (falling) [U].

W:	(level) [U]. Now we turn to the close-mid vowels, also termed
half-close or mid-high. Close-mid front, unrounded, cardinal
two, (falling) [e].

H:	(level) [e]. Close-mid front, rounded, cardinal ten,
(falling) [2].

W:	(level) [2]. Close-mid central, unrounded, a closish schwa,
reversed E, (falling) [ˆ].

H:	(level) [ˆ]. The corresponding rounded vowel, close-mid
central, barred O, (falling) [7].

W:	(level) [7]. Then a close-mid back unrounded vowel, cardinal
seventeen, ram's-horn, (falling) [4].

H:	(level) [4]. Close-mid back rounded, cardinal seven,
(falling) [o].

W:	(level) [o]. Then the schwa, which is quite loosely defined:
a mid central vowel, normally unrounded, (falling) [@].

H:	(level) [@]. Next we take the open-mid vowels, also termed
half-open or mid-low. Open-mid front, unrounded, cardinal three,
(falling) [E].

W:	(level) [E]. Open-mid front, rounded, cardinal eleven, O-E
digraph (falling) [9].

H:	(level) [9]. Open-mid central, unrounded, an openish schwa,
reversed epsilon, (falling) [3].

W:	(level) [3]. The corresponding rounded vowel,  open-mid
central, (falling) [%].

H:	(level) [%]. And the open-mid back unrounded vowel, cardinal
fourteen, inverted V: (falling) [V].

W:	(level) [V]. Open-mid back rounded, cardinal six, (falling)
[O].

H:	(level) [O]. Then a front unrounded vowel between open-mid
and open, A-E digraph: (falling) [{].

W:	(level) [{]. And a central unrounded vowel at the same
height, inverted A: (falling) [6].

H:	(level) [6]. Lastly among the vowels, we come to the open
vowels, also termed low. An open front unrounded vowel, cardinal
four, (falling) [a].

W:	(level) [a]. The corresponding rounded vowel, open front
rounded, cardinal twelve, cap O-E digraph, (falling) [&].

H:	(level) [&]. Then an open back  unrounded vowel, cardinal
five, (falling) [A].

W:	(level) [A]. And an open back rounded vowel, cardinal
thirteen, (falling) [Q].

H:	(level) [Q]. And that completes the vowels.



W:	Now beneath the vowels on the chart is a section labelled
"other symbols". These are an assortment of consonants. First
comes the voiceless labial-velar fricative, inverted W, [WA].

H:	[AWA]. Then the voiced labial-velar approximant, [wA].

W:	[AwA]. And the voiced labial-palatal approximant, [HA].

H:	[AHA]. Then the epiglottal fricatives, articulated by making
a constriction between the epiglottis and the wall of the
pharynx: first the voiceless epiglottal fricative, ['A].

W:	[A'A]. Then the voiced epiglottal fricative, [AA].

H:	[AAA]. If we obstruct the airstream completely at this place,
we get an epiglottal plosive, [AA].

W:	[AAA]. At the top of the righthand column under Other Symbols
we have the alveolo-palatal fricatives, voiceless and voiced,
[,A	ØA].

H:	[A,A	AØA]. Then we have an alveolar lateral flap, [A].

W:	[AA]. And then the famous sound of Swedish identified here
as consisting of simultaneous [S] and [x], namely [A].

H:	[AA]. Beneath this symbol the chart mentions the possibility
of using a tie bar to join two symbols for affricates and double
articulations. The illustration of a double articulation is the
voiceless labial-velar plosive [k_pA].

W:	[Ak_pA]. And the affricate is a voiceless alveolar affricate,
[t_sA].

H:	[At_sA]. Another use for the tie bar might be to show
nasalized clicks, where the click articulation occurs
simultaneously with a voiced velar nasal that uses a pulmonic
air-stream, thus a nasalized postalveolar (retroflex) click
[N_!A].

W:	[AN_!A].



W:	Next we turn to the suprasegmentals. The English word
[%foUn@"tIS@n], so pronounced, illustrates secondary stress on
the first syllable and primary stress on the penultimate.

H:	[%foUn@"tIS@n]. We shan't demonstrate the symbols for degrees
of length, for syllable break, for group boundaries, or for
linking, because these can't easily be demonstrated with
segments in isolation and out of context.



W:	Turning to tones and word accents, we start with level tones.
An extra high tone is [e†] ([¢e]). A high tone is [e‘] ([“e]). A
mid tone, [e‡] ([oe]). A low tone, [eŠ] ([”e]). And an extra low
tone, [e‚] ([öe]). 

H:	To demonstrate downstep and upstep, we can take an English
sentence, He's de"termined to    take  —charge. That was an
unmarked intonation pattern.

W:	He's de"termined to   take —charge. With a downstep on take,
it becomes He's de"termined to Etake —charge.

H:	He's de"termined to Etake —charge. And with an upstep on
take, it becomes He's de"termined to take —charge.

W:	He's de"termined to take —charge. Looking now at the contour
tones, we have first a rising tone, [e<] ([oe]).

H:	[e<] ([oe]). And then a falling tone, [e] ([—e]).

W:	[e] ([—e]). A high rising tone, [e”] ([£e]); and a low
rising tone, [eŒ] ([–e]).

H:	A rising-falling tone is [e¤] ([e]).

W:	An English sentence with a global rise is EWHAT did you say
you wanted?

H:	And a global fall is heard in EWhat did you say you WANTed?



W:	Lastly we come to the section of the chart labelled
diacritics. The left-hand column begins with the diacritic to
show that a segment is voiceless. A voiceless alveolar nasal is
[nįA].

H:	[AnįA]. And a voiceless lenis alveolar plosive, [dįA].

W:	[AdįA]. At the top of the grid here you will also notice the
symbol for a voiceless velar nasal, with the diacritic placed
above rather than below the rest of the symbol: [N A]. 

H:	[AN A]. Next, the diacritic to show that a segment is voiced.
A voiced fortis alveolar fricative, [s‰A].

W:	[As‰A]. Aspiration can be shown by a small raised H:
aspirated voiceless alveolar plosive, [tŽA].

H:	[AtŽA]. And an aspirated voiced alveolar plosive, [dŽA] (i.e.
[dDAD]).

W:	[AdDAD]. 

H:	We shan't illustrate the remaining diacritics in this column
until we come to the one at the bottom, which shows rhoticity. A
rhotacized schwa is (falling) [@8].

W:	(level) [@8]. In the central column of diacritics, we start
with the diacritic for breathy voicing. A breathy voiced
bilabial plosive and open vowel are [bDAD].

H:	[AbDAD]. Creaky voicing, on the other hand, sounds like this:
the same sequence creaky voiced is [b„A„].

W:	[Ab„A„]. Linguolabial consonants are articulated with the
tongue tip against the upper lip. Voiceless and voiced
linguolabial plosives are [tOA	  dOA].

H:	[AtOA	  AdOA].  Next we have a number of symbols to show
secondary articulations. First, labialized segments: [t«A	d«A].

W:	[At«A	Ad«A]. And palatalized: [t™A	d™A].

H:	[At™A	Ad™A]. Velarized: [t¬A	d¬A].

W:	[At¬A	Ad¬A]. And lastly pharyngealized: [tAA	dAA].

H:	[AtAA	AdAA]. The tilde-through can be used to show either
velarization or pharyngealization, as in dark L, [5A].

W:	[A5A]. 



H:	In the third column of diacritics we start with symbols
specifying with greater precision the place of articulation of
sounds  involving the tip or blade of the tongue as primary
articulator. First, voiceless and voiced dental plosives,
articulated against the teeth: [tIA	dIA].

W:	[AtIA	AdIA]. Then, apical alveolar plosives, made with the
tip of the tongue: [txA	dxA].

H:	[AtxA	AdxA]. And then laminal alveolar plosives, made with
the blade of the tongue: [tOA	dOA].

W:	[AtOA	AdOA]. The symbol for nasalization is a tilde over a
symbol: a nasalized front close-mid unrounded vowel, (falling)
[e~].

H:	(level) [e~]. The next three symbols show different varieties
of plosive release. We hear nasal release in the sequence [d"nA].

W:	[Ad"nA]. And we hear lateral release in the sequence [džlA].

H:	[AdžlA]. There is no audible release of the plosive in the
sequence [AdŖ].

W:	The same is true of the first plosive in the sequence [AdŖbA].



H:	Towards the bottom of this grid you see the symbols to show
raised and lowered varieties. A raised close-mid front unrounded
vowel is (falling) [eO].

W:	(level) [eO]. And a lowered close-mid front unrounded vowel
is (falling) [eO].

H:	(level) [eO]. Where a symbol might be interpreted as denoting
either a fricative or an approximant, the raising symbol makes
it clear we mean the fricative, as in the voiced apico-alveolar
fricative, ["OA].

W:	[A"OA]. And the lowering symbol makes it clear we mean the
approximant, as in the voiced bilabial approximant, [BOA].

H:	[ABOA]. The last two symbols at the bottom right of the chart
are used to show advanced tongue root and retracted tongue root.
Cardinal two, a close-mid front unrounded vowel, usually has a
relatively advanced tongue root, giving a wide pharynx and tense
quality: (falling) [eO].

W:	(level) [eO]. With a retracted tongue root we get a narrower
pharynx and lax quality: (falling) [e™] (= [I]).

H:	(level) [e™]. And with that we come to the end of this
recording. 

W:	We hope you've found it useful.