Professional Documents
Culture Documents
When linguists are developing a phonemic description of a language or dialect they most often
select the most common or widely distributed allophone of each phoneme as the typical
allophone of that phoneme and use its phonetic symbol to represent the phoneme as a whole.
When a symbol is used to represent an actual sound (allophone) it has an entirely different
meaning to the same symbol when used to represent a phoneme. For this reason we always
enclose transcriptions in /.../ when we are indicating phonemes and in [...] when we are
indicating the actually produced sounds.
For example, /k/ might represent a particular English phoneme, which in this case can be realised
in actual speech as an aspirated, unaspirated or unreleased velar stop, as a more fronted palatal
stop, as a more retracted uvular stop, as a stop produced in some intermediate position between
these extremes and also as fricative variants of these stop sounds. The true identity of /k/ is
possibly better described by a system of phonological features (see the topic on Distinctive
features). [k] on the other hand represents a sound that is an unaspirated velar oral stop.
What symbols should we use for a phonemic transcription of Australian English?
a. Consonants
The symbols for the consonants are not problematic as the phoneme inventories of the
consonants of most English dialects are identical and the consonant symbols used below are the
same consonant symbols used for the consonant phonemes of British, American, New Zealand,
Canadian, South African and other English dialects. There are some differences in the
pronunciation of a small number of phonemes but this has not affected the choice of phonetic
symbol to be used for each consonant phoneme across these dialects.
b. Vowels
There is much greater difficulty in determining the symbols for the vowel phonemes. Firstly,
there is a significant variation between English dialects in the number of vowel phonemes.
Secondly there is considerable variation in vowel phoneme pronunciation between dialects (even
those that share the same vowel phonemes). There are different systems applied to British and
American English and even different systems applied to different dialects within the United
Kingdom and the United States of America.
For many years the system that has most often been applied to Australian English vowels is
known as the Mitchell-Delbridge system. This system evolved from the work of Mitchell and
Delbridge (between the 1940s and the 1960s) but which was mostly based on the system applied
to south eastern British English. The Mitchell-Delbridge system is the one used in the Macquarie
Dictionary and traditionally in speech pathology clinics in Australia. However, it is now
progressively being replaced by the HCE system. Whilst still quite widespread, the MitchellDelbridge system is not a good reflection of actual Australian pronunciation.
In this course we use instead the system devised by Harrington, Cox and Evans (1997). We use
this system because it better represents the average pronunciation of these vowels by speakers of
Australian English. It also fulfils the requirement that the symbols used for the phonemic
transcription of each phoneme is the most common or widely distributed allophone of that
phoneme. For a rationale, see Cox (2012).
For a comparison of several phonemic transcription systems used for the transcription of
Australian English vowels, see the topic entitled Vowel Symbols for Australian English
Phonemes.
For an overview of the Mitchell-Delbridge system for the phonemic transcription of Australian
English, click here.
pat
tin
cap
/b/
/d/
//
bat
din
go
2. Affricates
/t/ or /t/
/d/ or /d/
choose
judge
3. Fricatives
/f/
//
/s/
//
/h/
fan
think
so
she
he
/v/
//
/z/
//
van
these
zoo
beige
my
no
sing
5. Approximants
/w/
we
/j/
/l/
//
you
leaf
run
Note: /w/ and /j/ are also called semi-vowels, because they are very similar acoustically to
vowels.
Vowels
This material should be read in conjunction with the lecture materials on Australian English
Vowels.
1. Monophthongs
/i/
/e/
Long vowels
heed
hair
//
/o/
//
//
hard
hoard, saw
who'd
herd
//
/e/
//
//
//
//
//
Short vowels
hid
head
had
mud
pod
hood
the (spoken quickly)
2. Diphthongs
//
/e/
/o/
//
say
high
toy
cure
//
//
//
so
how
here
Note: These vowel symbols differ significantly from the vowel phoneme symbols used for
Australian English in the Macquarie Dictionary, but more closely represent the way average
speakers of General Australian English actually pronounce these vowels. They also differ from
the symbols used for the same vowel phonemes in British and American English. Note the
difference between these symbols and the ones used by Ladefoged for American English in the
textbook. These differences reflect differences in pronunciation.
//
//
/o/
//
//
//
//
//
/e/
/o/
//
//
//
/e/
// note: There is a great deal of inter-personal variability for this centring diphthong. Some
people use it for some words, other people for other words. Some speakers don't use it at all (or
only rarely) substituting either // (a sequence of 2 phonemes) or /o/ (a long monophthong).
For this reason no examples are given here.
Exercises
1) Find the errors in the consonant phonemes in the following transcription. In each word there is
one error, indicating an impossible variety of English.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
slang
crime
lamb
wishing
knives
these
hijack
swimming
sixty
myths
exclusion
human
/sln/
/cem/
/lmb/
/wsh/
/nevs/
/iz/
/hejk/
/swmm/
/sxti/
/mths/
/kskln/
/hymn/
2) A pair of words which differs in only one phoneme is said to be a minimal pair. Find minimal
pairs for the following. It is acceptable to write the orthographic form for each word. Phonemic
transcription is not required.
initial position
final position
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
/p b/
/t d/
/k /
/f v/
/s z/
/m n/
/n /
/ l/
/t /
/p f/
/d z/
/t /
/h w/
/s /
---------------
--------
Word Stress
Primary Stress
In English, the syllables of words differ in prominence. In polysyllabic words (words of more
than one syllable), one of the syllables always has a greater degree of prominence than any other
syllable. The syllable of greatest prominence is known as the primary stressed syllable, or the
syllable that carries primary stress. The primary stressed syllable can be marked by placing
a mark in the following way:
Primary stress
apart
define
pattern
metal
pamela
impudent
miserable
phonetics
part
fine
pat
met
pam
im
mis
net
In monosyllabic words (words of one syllable), the outcome is unambiguous: the syllable, i.e.
word, has primary stress (eg.
heat,
look,
greet) and does not need to be marked.
Secondary stress
Many words have two stresses, one primary and one secondary. The secondary stress can be
marked by placing a diacritic before the syllable which has secondary stress:
Secondary
ac
pol
ec
ex
lib
vir
gen
nom
rho
academic
apoloetic
economic
explanation
deliberation
environmental
enerosity
phenomenoloy
rhododendron
Primary
dem
get
nom
nat
rat
ment
ros
nol
den
Further comments
(i) words that have secondary stress are very often morphologically related to simpler forms
eg.
apoloetic /
enerosity /
apoloy;
enerous.
or
upstairs
Primary
al
re
sys
an
syll
rhu
diph
po
alternate
realise
systematise
anecdote
syllabub
rhubarb
diphthon
potash
Unreduced
nate
lise
tise
dote
bub
barb
thong
tash
In some cases, the occurrence of an unreduced syllable is predictable (eg. words ending in -ate or
-ise), in others it is not. Note that there is also not necessarily a requirement that an unstressed
syllable intervene between the primary stressed and unreduced syllable.
Unstressed syllables
In almost all cases, syllables other than primary stressed, secondary stressed or unreduced
syllables are unstressed. In General Australian English, the large majority of the vowels of
unstressed syllables can be transcribed as //:
/bdkt/
/dmt/
/kmben/
/knfmn/
/kndn/
/lbt/
/ptl/
abduct
admit
combine
confirmation
condition
elaborate
patrol
There are also some cases when unstressed vowels have a quality other than //. It is not possible
to list them all, but some of these include:
Some words that end in unstressed -ish, -ic, -ism and -ing
/d/
/mjzk/
/kmjnzm/
/n/
radish
music
communism
running
annual
actual
/kntnjs/
/din/
/dni/
/veis/
continuous
guardian
junior (nb. can also be pronounced
various
/dnj/ )
very,
potato,
city,
every
barrow
(not
(not
/nflenl/ )
/pin/ )
influential
appreciation
Exercise
Primary and secondary stress
Mark primary stress using and any secondary stressed or unreduced syllables using . You do
not need to give a phonemic transcription. Mark stress as it occurs in the recording.
Example: economic Answer: economic
(i)
(ii)
(iii)
(iv)
(v)
accommodation
certainty
orthodox
fanatical
greenhouse
(vi)
(vii)
(viii)
methodical
unbelievable
Adelaide
Syllabic consonants
A syllabic consonant occurs when there is no vowel in a syllable. Certain sonorant consonants
(nasals and approximants) can replace a vowel as the nucleus of a syllable. We call such
consonants "syllabic". In English, syllabic consonants occur frequently in some contexts and less
frequently in others. Whilst in certain contexts syllabic consonants occur with very high
frequency, it is not obligatory for this to happen. The frequency of use of syllabic consonants is
speaker-dependent.
Syllabic /n/ and /l/ occur, with very high frequency, after alveolar consonants at the end of words.
Additionally, syllabic /m/ sometimes occurs when there is a preceding /b/ or//. In phonemic
transcriptions we include the preceding (weakened or deleted) schwa (//). In words subject to
this process there is a continuous range of degrees of vowel weakening, from partial weakening
to complete vowel deletion but the syllable still exists, so we indicate this in phonemic
transcription by indicating the vowel. This is speaker dependent and also depends upon rate of
speech. For this course, you will include the schwa in your phonemic transcriptions. Only in
phonetic transcriptions will we delete the schwa (when appropriate) and only in phonetic
transcriptions will we indicate, using a diacritic, that a consonant is syllabic (when appropriate).
In the following examples the vowel has been greatly reduced or deleted and the final consonant
has been judged to be syllabic BUT for phonemic transcriptions of such words you must indicate
the schwa. In other words, for your phonemic transcriptions you don't need to determine whether
or not the vowel has been fully deleted, partially deleted, reduced in duration and strength or
fully maintained. If you hear two syllables, whether or not you can hear the vowel, treat the
vowel as phonologically present and include it in your phonemic transcription.
/btn/
/sdn/
/dzn/
/ktl/
/mdl/
/ksl/
/mzl/
/fenl/
/lbm/
/fm/
button
sadden
dozen
cattle
middle
castle
muzzle
final
album
fathom
When words occur in connected speech, there can be many modifications to their phonemic form
compared with their production in isolation. It is not possible to cover all the different kinds of
modifications that can arise. Some of the more important changes are listed below.
am
an
and
are
as
at
been
but
can
could
do
does
for
from
had
has
have
he
her
him
his
is
must
of
shall
should
some
than
that
Citation
/m/
/n/
/nd/
//
/z/
/t/
/bin/
/bt/
/kn/
/kd/
/d/
/dz/
/fo/
/fm/
/hd/
/hz/
/hv/
/hi/
/h/
/hm/
/hz/
/z/
/mst/
/v/
/l/
/d/
/sm/
/n/
/t/
/d/
/z/
(/d/), /d/
/z/
/v/
//
//,//
(/m/)
(/z/)
/z/
/mst/
//, /v/
/l/
them
there
to
us
was
were
who
will
would
you
/em/
/e/
/t/
/s/
/wz/
/w/
/h/
/wl/
/wd/
/j/
/m/
*/e/
/t/
/s/
/wz/
*/w/, /w/
*/h/, (//)
*/wl/, /wl/
*/wd/, /wd/, /d/
*/j/, /j/
/m/
//
//
/l/
/d/, /d/
Note in the above table that there are occasions where it is common to produce the citation form
in connected speech. Such common cases are repeated in the continuous speech column and
marked with a *. In some cases that form is only normally reduced in rapid speech, whilst in
some cases there is both a reduced and an unreduced connected speech form, which may be
selected between at the whim of the speaker or for semantic reasons. Further, occasionally even
the most commonly reduced words, such as "the" may be pronounced as the citation form in
connected speech when there is a strong semantic need to stress that word:
eg. "This isn't just A book, this is THE book." (referring to a book of particular importance to the
speaker)
You might also note that for some of the pronouns there are some forms surrounded by round
brackets. These forms are the "dropped h" forms which some speakers use in normal connected
speech. Other speakers choose forms with the "h" intact in normal connected speech, but may
"drop the h" in rapid speech.
the car
/ k z n dev/
drawing
Tina Arena
(iii) Assimilation
In continuous speech, alveolar consonants can assimilate to the place of articulation of a
following consonant in certain contexts. This process is certainly not obligatory, but it is
nevertheless sufficiently common that the rules should be known.
a) alveolar consonants can assimilate to a bilabial place of articulation before labials
/meppt/
/mepbe/
/mepmk/
/mepwn/
/bpt/
/bbe/
/bmk/
/bwn/
/simpit/
/simbl/
/simmek/
/simwolt/
might put
might buy
might make
might win
should put
should buy
should make
should win
seen Peter
seen Bill
seen Mike
seen Walter
b) alveolar consonants can assimilate to a velar place of articulation before velar consonants
/mekkm/
/mek/
/km/
//
/sikn/
/sie/
might come
might go
should come
should go
seen Karen
seen Greg
c) alveolar fricatives can assimilate to a palato-alveolar place of articulation before // and /j/
/p/
this shop
/iip/
/j/
/telj/
these sheep
this year
tells you
When the following word is a function word, the /j/ may also be elided:
/knidt/
/hletkm/
e) word-internal assimilation
The same processes of assimilation can also take place word-internally. Often this will be in
compounds or words that are formed from a root+affix eg.:
/bbe/
/tbpl/
/mmvd/
/ho/
good-bye
tadpole
unmoved
horse-shoe
Deleted Stop
/menz/
/hnfl/
/dekli/
/nvesmnt/
amends
handful
directly
investment
/t/ and /d/ can also delete in many cases when the following consonant belongs to a different
word:
/nekst d/
/ld mn/
/neks d/
/l mn/
next day
old man
As you attempt the transcription practice exercises you will notice some discrepancies between
what you might expect and what we have supplied as a correct transcription. These discrepancies
may be due to one of the following reasons:1. The principles or "rules" outlined above are what would be expected to occur most of the time,
not all of the time. These rules are probabilistic. That is, they predict what is likely to occur, not
what must always occur. Even the very common alveolar assimilations don't always occur and a
couple of exceptions can be found in the practice exercises. You must listen and decide for
yourself if a predicted rule has been applied by a speaker. Use the rules as your starting point, but
then listen and determine if they have actually been applied.
2. Sometimes schwa is confused for another vowel. Read "some further comments on schwa", in
the text above.
3. Sometimes we have indicated a voiced oral stop or fricative, but you hear its voiceless
counterpart. This is because of a process known as "devoicing" which occurs for these sounds
when they precede a pause or a voiceless consonant. This is dealt with in the section on phonetic
transcription. For now, and only when the circumstances outlined in the next sentence are true,
indicate the voiced stop or fricative when the orthography indicates the voiced sound and the
voiceless stop or fricative when the orthography indicates the voiceless sound. Only do this,
however, for a stop or fricative:
at the end of a word immediately preceding a pause (eg. before a comma or a full stop)
at the end of a word when the next word starts with a voiceless consonant.
BUT, be very careful of how you transcribe the {-s} and the {-ed} morphemes as these are
allocated "allomorphs" (alternative pronunciations of a morpheme) on the basis of the following
rules:{-s} morpheme (verb 3rd person singular present tense, or noun plural)
{-s}
"catches")
/lsz/
/mzz/
/lz/
/kz/
/z/ following a /s/, /z/, //, //, /t/ or /d/ (e.g. " glasses", "mazes", "lashes",
glasses
mazes
lashes
catches
/z/ following a vowel or a voiced consonant (other than /z/, //, or /d/) (e.g.
"dogs", "pads", "bins")
/dz/
/pdz/
/bnz/
dogs
pads
bins
/s/ following a voiceless consonant (other than /s/, //, or /t/) (e.g. "cats", 'tips",
"packs")
/kts/
/tps/
/pks/
cats
tips
packs
For example in
{-ed} morpheme (verb past tense)
{-ed}
/wntd/
/edd/
wanted
guided
/d/ following a vowel or a voiced consonant (other than /d/) (e.g. "rowed",
"dodged")
/d/
/dd/
rowed
dodged
/t/ following a voiceless consonant (other than /t/) (e.g. "tossed", "walked")
/tst/
/wokt/
tossed
walked
4. Occasionally we make mistakes in our transcriptions. If you think that we have made a
mistake, let us know and we will double check to see if an error has occurred.
References
Cox, F., (2012) Australian English: Pronunciation and Transcription, Cambridge University
Press.
Harrington, J., Cox, F., and Evans, Z. (1997) An acoustic phonetic study of broad, general, and
cultivated Australian English vowels. Australian Journal of Linguistics, 17, 155-184.
Nativlang.com
languages
linguistics
logic
books
videos
6. An extra challenge
7. Answers to the exercises
Introduction
Symbol-to-sound correspondence
Imagine that you wanted to represent the pronunciation of a language as accurately as possible in
writing. This seems like an easy task. After all, isn't that what the invention of the alphabet was
all about?
Well, consider English. We write an "a" in change, ball and hand even though that "a" represents
a different sound in each word. Some simple sounds are represented by two letters (like "th" in
thing) while sometimes a single letter represents two sounds (like "x" in example)!
Maybe a more "phonetic" language, like Spanish, fares better? Initially, it would seem so: "a"
sounds the same in avin, sola and amigas. Yet Spanish spelling masks the difference between
the "g"s in gringo vs. agua vs. gente. What's more, the way a Cuban leaves off the final "s" in
muchos or how a Latin American speaker pronounces "s" and "z" the same account for only a
few of the intricacies that hide behind standard Spanish spelling. English and Spanish spelling
demonstrate how languages can't even use their own alphabets to encode their sounds accurately.
To tackle the problem, we first need to understand a basic concept missing from our discussion
so far: the phoneme. Phonemes are individual units of sound that can be pronounced on their
own and considered "one sound". To represent the pronunciation of a language accurately, we
can break words and phrases into these individual sounds, these phonemes. Then, if we assign a
unique "letter" (or symbol) to every possible phoneme, we could write the pronunciation
accurately, without confusion and contradiction.
Steps toward understanding the IPA
Initially, all that may seem like a trivial thought experiment, but it teaches you the thinking
behind the International Phonetic Alphabet. The IPA is a tool, really just a set of many symbols,
that allows you to display and read any stream of sounds in any natural language. It's a way to
"hear with your eyes" and imitate pronunciation more accurately. It rivals our ability to record
the human voice in its usefulness to help us understand and analyze the pronunciation of the
world's languages.
In this lesson, you will begin to learn about the phonetic concepts that underpin the IPA.
Specifically, you will learn about vowels and consonants and how the human mouth produces
them. You will learn many of the most common IPA symbols relating to vowels and consonants.
Then, you will learn how syllables and larger chunks of speech work, and how to use IPA to
represent those. Along the way, you'll have opportunities to work with the pronunciation of
words in a variety of languages.
close
mid
open
front
i,
, e
central
back
u,
, o
example words
sit
let
math
reduction
father
ought
books
The pure vowels /i/, /e/, /a/, /o/ and /u/ are more difficult for most English speakers. They have to
be extracted from diphthongs where they're pronounced before a y-sound (after front vowels /i/
and /e/) or w-sound (after the back vowels /o/ and /u/). The phoneme /a/ is the first element of the
diphthong /ai/ (in "aisle") and /au/ (as in "out") for many English speakers. It is also found in
Australian English, and is the normative "a" sound in all Romance languages. We'll learn more
about diphthongs later.
IPA symbols
/i/
/e/
/a/
/o/
/u/
example words
seem
say
aisle
low
soon
notes
leave out the final "y" sound
leave out the final "y" sound
leave out the final "y" sound
leave out the final "w" sound
leave out the final "w" sound
By comparison and with a good ear, you can notice the same phonemes in any language you're
learning. In languages as different as Japanese and Hawaiian, "i" is pronounced /i/. What's more,
you'll be able to imitate unfamiliar vowels. Spanish, French and Italian "a" all sound like /a/
(rarely heard in American English) rather than //. French, German and Ancient Greek all have a
rounded close front vowel represented by /y/ (sounds like /i/ but pronounced with your lips
rounded).
Other features that differentiate vowels
We can consider secondary features. Other features include roundness (you round your lips when
you pronounce /u/) and nasalization (French & Portuguese vowels before n/m + consonant are
pronounced through the nose; in rapid speech some English speakers say "hand me that" with a
nasal //).
Keep in mind that an essential function of vowel features is to represent how speakers of a
language distinguish vowels. For instance, if mid-front, mid-central and mid back vowels don't
sound different to speakers of language X, then language X does not distinguish backness as a
feature of mid vowels.
Ultimately, as you learn more about sound systems in languages, you'll see you can define vowel
features more precisely by subtle variations in the quality of the sound produced than by how
they are made in the mouth, as we have defined them.
Practice Exercise
1) Identify every vowel phoneme in the words below. First, list the features of each vowel, then
write each vowel in IPA. Note that silent e's are silent, or zero-phonemes:
change, ball, hand, ring, after, said, look, true
2) Read these four Italian words written in IPA aloud. Pay close attention to the vowels.
/wmini/, /stte/, /porta/, /kultura/
sound. If you let gentle stream of air pass between your tongue and gums, you pronounce an /s/
sound instead. This feature is known as manner of articulation, in other words, how you form the
phoneme.
Whether or not you voice the consonant
The final consonant feature we'll learn about has to do with the vibration of your vocal chords.
Notice that when you hum, your throat vibrates, but when you whisper quietly, it doesn't. Your
"voicebox", the vocal folds on your larynx, vibrate when you make voiced sounds (like /z/, /d/,
and /b/), but stay still when you pronounce voiceless sounds (like /s/, /t/ and /p/). This feature is
called voicing. In fact, it's safe to say that the only difference between /s/ and /z/ is the voicing
of /z/ (the same holds for /t/ & /d/ or /p/ & /b/), making these phonemes voiced-voiceless pairs.
Combining the three features to identify specific consonants
Like with vowels, we can't identify specific consonants based on voicing, place of articulation or
manner of articulation alone. We must treat these features as building blocks of consonants.
Certain combinations of the three features produce specific, identifiable consonants. For
example, instead of speaking about "the consonant in the word thing", we can accurately
describe the "voiceless dental fricative". You can consider the relationship between these features
as a chart with three axes, with place of articulation on the x-axis, manner of articulation on the
y-axis and voicing as the z-axis. In the table below, relevant sounds are given as "voiceless,
voiced" pairs.
labial
nasal
m
plosive
p, b
fricative
f, v
affricate
approximant
lateral
dental
alveolar
n
t, d
s, z
t, d
r
l
palatal
velar
k, g
x
glottal
example words
more
none
singing (no actual /g/ sound)
thing (voiceless)
that (voiced)
said
/z/
//
//
/x/
/h/
/t/
/d/
/j/
prize
sharp
pleasure
Scottish loch ("hard H" sound)
hear
cheek
jaw
yes
Notice that the affricates /t/ and /d/ are combinations of two phonemes heard as a single
sound. The glottal stop is often described as a catch in the throat (the sound found between "uhuh", the negative counterpart to "uh-huh), You can hear this glottal stop "catch" before initial
vowels in English: "every" /vrij/.
English speakers have a tougher time with plosives. Specifically, speakers tend to pronounce the
voiceless plosives with a puff of air, really an /h/ sound: "take" starts with a sequence of
phonemes like /t/ + /h/ rather than just a bare /t/. That aspiration (h- or aitch-sound) is absent
when you pronounce /p/ in speak, /t/ in stay and /k/ in sky, so imitate that sound to pronounce /p/
/t/ /k/ as "pure" voiceless plosives. You shouldn't have any trouble with their voiced
counterparts /b/ (be), /d/ (day) and /g/ (guy).
As you further develop your comparison skills and your good ear, you can notice these
consonant phonemes in foreign languages. Spanish plosives "p" and "t" sound like /p/ and /t/
(without the puff of air as I described above). Italian "c" sounds like /k/ before /a/, /o/, /u/ but /t/
in front of /e/ and /i/. Japanese has the affricates /ts/ and /dz/, which aren't heard as single sounds
in English but can be imitated easily by stringing together /t/ + /s/ and /d/ + /z/. The alveolar
tap // of Spanish, Portuguese or Italian /r/ in "caro" is the same sound of American English
"later" or "stutter".
More precise definitions and other features
Place and manner of articulation can be pinpointed more scientifically and exactly than I have
done. The sounds // and //, which I label as palatal, are actually "postalveolar" (slightly above
and behind the alveolar ridge). You can deepen your understanding by considering which part of
your tongue presses against which part of your mouth - the tip of the tongue is involved in
coronal sounds (including dental and alveolar), while the body and back of the tongue articulates
dorsal phonemes (including velar). Labial sounds emphasize the lips rather than the tongue
(including bilabial sounds like /b/ and labiodental ones like /v/). Radical and glottal sounds are
made with the base of your tongue and the back of your throat.
Practice Exercise
1) Identify every consonant phoneme in the words below. First, list the three features of each
consonant, then write each consonant in IPA. Pay attention to pronunciation over spelling!
Syllables
So far, we've studied vowels and consonants separately. We've even broken them down into their
component parts called "features". Yet speakers don't tend to pronounce consonant or vowel
sounds in isolation, but together. We won't just jump from sounds to words and sentences. We
can first organize speech sounds into beats or units. More specifically, speakers of all languages
put vowels and consonants together into speech units known as syllables.
Structure and types of syllables
Syllables tend to be built around a vowel. That vowel is the heart of the syllable, called its
nucleus. We may abbreviate vowel as V (V stands for any vowel).
Clearly, people don't just speak in vowels. Consonants surround the nucleus. A syllable may have
consonants before the nucleus and consonants after the nucleus. As with vowels, we may
abbreviate consonant as C. Consonants before the vowel are part of the syllable's onset.
Consonants after the vowel form the syllable's coda (Italian for "tail").
The English syllable "dab" (which also happens to be a word) has one consonant sound before
the vowel and one after: /db/. So, the structure of that syllable is CVC. But that's not the only
syllable type:
structure
V
(CC)CV
VC(CC)
(CC)CVC(CC)
English examples
ah!
nah! (for "no")
acts
splints
IPA
//
/n/
/kts/
/splnts/
notes
no consonants in onset or coda
consonants in onset, none in coda
consonants in coda, none in onset
consonants in onset and coda
Now, other languages have different constraints on their syllables. Specifically, English can
"overload" on consonants, while languages often allow only much simpler syllables. In standard
Japanese, there are only four possible syllable types:
structure
V
CV
CV:
Japanese
examples
do
d
IPA
notes
/o:/
/do/
/do:/
CVn
dan
/dan/
notes
/n/ and /t/ are both alveolar
/n/ turns labial before adjacent /b/
This brings up a concept I plan to tackle in the next series (introduction to phonology). In the
examples above, we don't say that the /n/ has split into two new phonemes, one /n/ and the
other /m/. Instead, we point to the concept of an allophone. Allophones account for the various
ways the "same sound" (as heard by speakers) shows up in speech: /n/ is pronounced [n] in the
word "into" but sometimes comes out as [m] in the word "input". In this way, we can propose [n]
and [m] as two allophones of one underlying phoneme /n/.
Features of longer streams of sound
We haven't yet considered the kinds of features that might apply to longer speech - to words,
phrases and utterances. At this point, we can only transcribe strings of individual sounds, but
those would be fairly monotonous if read aloud.
One very important feature is accent. One type of accent, called stress, involves emphasizing one
or more syllables by speaking them louder than surrounding syllables. In many languages, like
English and German, words normally have a fixed stressed syllable (consider "ho-tel" versus
"host-tel"). In IPA, stress can be written with an apostrophe before the stressed syllable:
/how.'thl/ versus /'hs.tl/. But this also holds when we look at longer
utterances: /'wijkhd'rajts'sntnslajk's/.
In Japanese, stress doesn't play this kind of role. Instead, Japanese speakers raise and lower the
pitch, or tone. This is called pitch accent or intonation. "Kyto" /kjo:to/ starts with high, falling
pitch in the first syllable, and ends in a lower pitch in the second syllable. In IPA, pitch may be
indicated by accent marks (rising pitch //, high/stable // or falling //), up/down arrows, sloped
arrows or hooked bars.
Sometimes features don't work on the level of words but on the level of sentences and "complete
thoughts". Features at this level are collectively called prosody. English speakers don't use
intonation to distinguish syllables, but to perform larger functions like distinguishing a question
from an exclamation from a simple declaration: "Are you happy?" /'r juw 'hphij/ versus
"You are happy." /juw 'r 'hphij/. Stress may also play a prosodic function: "I didn't
learn anything" versus "I didn't learn anything".
Practice Exercise
1) Oops! As I transcribed these phrases into IPA, I forgot to assimilate certain consonants. Look
at each phrase, consider how it sounds in flowing speech, and correct expected assimilations.
/'khnkwst/, /s'p/, /'sz'jrz/, /n'phrsn/, /bow'sts/
2) Two English speakers read the same phrase differently. I captured their pronunciation below in
IPA. First, identify and place the missing features of stress accent and intonation. Second, spell
out the original phrase these speakers read in standard written English. Third, based on your
judgment, what kind of speaker said a)? What kind of speaker said b)?
a) /ntnnlfnethkhlfbth/
b) /nrnnlfnkhlfb/
3) Listen carefully to these sound files of Italian words and phrases. Transcribe them as best you
can into IPA.
phrase 1, phrase 2, phrase 3
An extra challenge
Before you run off to brag to your friends that you've completed this course on IPA, I leave you
with a continuing challenge.
On this site, I have a number of "learn to pronounce language X" or "language X pronunciation"
guides with explanations and audio files. I challenge you to complete one of those pages and
transcribe every word you hear into IPA as accurately as possible. If you need help, look up "X
language phonology" in Wikipedia - that resource usually has all the IPA you'll need to start
transcribing a specific language.
But don't stop there. I recommend that you start employing your IPA skills in any situation
relevant to language learning or pronunciation. Using the foundation you have gained here
coupled with curiosity and vigor for spoken language, you will master the phonetic alphabet in
no time.
1) /'khkwst/, /'p/, /'s'rz/, /m'phrsn/, /bows'sts/. (Whether you perform every one
of these assimilations in English will depend on your dialect, but the first one, /nk/ > /k/, is
fairly universal.)
2) First: /nt'nnlf'nethkh'lfbth/ and /nr'nnlf'nkh'lfb/
Second: "International Phonetic Alphabet"
Third: Speaker a) sounds like a British English speaker, while speaker b) sounds like an
American English speaker.
3) /pjuk'kao/, /an'napoli/, /'sketso/