Update:
2008-04-01 09:32 AM -0500
TIL
Properties of Consonants
prop-con.htm
by U Kyaw Tun, M.S. (I.P.S.T., U.S.A.). Not for sale.
Prepared for students of TIL Computing and Language Center, Yangon, MYANMAR.
UKT: Based on
• Properties of Consonants and Vowels,
Kevin Russell, Linguistics Department, University of
Manitoba, Winnipeg, Manitoba, R3T 5V5, CANADA
http://www.umanitoba.ca/faculties/arts/linguistics/russell/138/notes.htm.
• Online Phonetics Course (UNIL), Department of Linguistics, University of
Lausanne, Switzerland.
(This source was downloaded in 2000 or a few years later, and instead of the original
links, you can still get to them from:
http://www.unil.ch/ling/page30184_fr.html -- UKT: 070823)
RBM4M |Top
BG4M-indx
Contents of this page
Voice segments
Consonant parameters
Place of Articulation (POA)
Consonant charts
• English consonants
• IPA consonants
• Devanagari consonants
UKT notes
, prop-con11 /
prop-con12,
mostly from Wikipedia for updating my knowledge
in separate file prop-con1.htm
Contents of this page
One of the problems I (UKT) have met in taking material
from more than one source is the inconsistencies
in terms used by various authors. What is the difference
in "voice" in "Voice Segments"
and "sound" in "Sound Production"?
They might be talking about the same thing.
What about the "syllable"? Moreover,
what is a "segment"? See
segment and
syllable in my notes.
UKT based on Wikipedia: In analysing a syllable,
it is convenient to idealized it to be made up of
segments. For example even in a vowel like
{a:} /a/, it is sometimes convenient to look it
as made up of two segments: {a} /a/ and {:} /ː/.
However, in another analysis, it may be convenient
to look at it as a single segment /aː/.
-- Wikipedia
http://en.wikipedia.org/wiki/Chroneme download 071011
When you speak, there are several different things
going on in your mouth at the same time. As an illustration,
pronounce the last part of my name, Tun. In Burmese-Myanmar it is
,
which I have spelled in Romabama as {htwan:} with
the "tone" symbol (:) to show that
the tone is #3 of the three tones or registers
of Burmese (short-light {a.}, light {a} and heavy {a:}).
I have taken the inherent vowel /a/ of the Burmese-Myanmar
akshara system, and have used <a> to represent
the vowel of the syllable without taking the
"phonetics" into consideration. My name
would be pronounced as if the vowel is <a> in
Yaw and in Rakhine, but as <u> in the Irrawaddy Delta.
Please note that you are NOT saying it as {twän} with a
{thé:thé:ting}.
Now leave the <t> out and say
{wan} / {wun},
{wam} / {wum}, and
{wän}. If you were born and raised in mainland Myanmar,
you will be pronouncing all the syllables exactly the same.
(I still have to check how ethnic Burmese groups
which have but scant relationships with the mainland,
such as the peoples of {ming:doan:} (Pyu), {ywa-ngän} (Danu),
{ing:lé:} (Inntha), etc., pronounce the rimes with
killed syllables.) That does not matter for what
I am going to say presently. If you were born and
have grown up in Canada just say [un]. You will
notice that your tongue moves only slightly, but will
never come up to touch the "roof" of your mouth.
Now say my name
{htwan} / {htun} again. You will notice that
at least the following are happening:
• Your tongue tip raises from a relatively neutral position
till it touches your alveolar ridge and cuts off airflow through the mouth
• Your tongue body starts in a high position and
drops to a relatively neutral position
• Your lips start out rounded because of
{wa. hswè:} and become less rounded.
• Your soft palate (velum) starts out raised, blocking off
airflow through the nose, but becomes lower and lower,
allowing more and more air out through the nose
• Your vocal cords are vibrating
The parts of the vocal tract are (mostly) independent
of each other. Several parts are changing, but they aren't
necessarily changing all at the same time.
Now to be on safer grounds, let's take what Kevin Russell
has given. Unfortunately, the example he has chosen was
the pronunciation of <spoon>, the <sp>
of which many Myanmar will say in two parts the <s>
and the <p>, something like: /sə.puːn/.
<spoon> /spuːn/ -- DJPD16 p502
UKT: Remember, there's a problem with English <p>
from the point of view of a Myanmar. <p> at
the beginning of a syllable onset is pronounced as {hpa.}.
If <p> follows <s> it is pronounced
like {pa.}. However, you will not see this difference
in DJPD16 because, it gives "broad transcriptions"
instead of narrow transcriptions".
See
•
allophone
•
allophone problem
•
complementary distribution
•
phonetic transcription in my notes.
The parts of the vocal tract are (mostly) independent
of each other. Several parts are changing, but they are
not necessarily changing all at the same time.
For example, for most English speakers saying [un],
the soft palate lowers and lets air out through the nose
long before the tongue tip raises and blocks off
airflow through the mouth. The whole set of movements
by four parts during the word <spoon> might look
like fig.1 but not fig.2 shown below:
Fig.1 
Fig.2.
On a practical level, it's almost impossible to
start work using the first kind of picture, even if
it is more realistic. To begin studying phonetics,
we have to idealize the real world.
We have to act as if speech events look like
the second picture rather than the first. We can
divide the idealized picture up into segments
which have well defined properties. Fig.2 has four segments:
1. a segment with un-rounded lips,
the tongue tip near but not touching the alveolar ridge,
neutral tongue body; raised soft palate, and
spread vocal cords -- we can write this as [s];
2. a segment with un-rounded lips,
neutral tongue tip, neutral tongue body, raised soft palate,
and spread vocal cords -- we can write this as [p];
3. a segment with rounded lips,
neutral tongue tip, high tongue body, raised soft palate,
and vibrating vocal cords -- we can write this as [u]; and
4. a segment with un-rounded lips,
the tongue tip against the alveolar ridge, neutral tongue body,
lowered soft palate, and vibrating vocal cords --
we can write this as [n].
Fig.3
When phoneticians and phonologists talk about a sound,
they are usually referring to one of these idealized segments.
Ultimately, a theory of phonetics and phonology
has to explain realistic pictures like Fig.1 rather than
simplified pictures like Fig.2. But for many practical purposes,
it's good enough if we use the idealization that any
stream of speech can be broken up into a series of segments
and each segment given a phonetic symbol.
For example, for most English speakers saying [un],
the soft palate lowers and lets air out through
the nose long before the tongue tip raises
and blocks off airflow through the mouth.
Taking the cue from what has been said above, my name
{htwan:} should be pronounced as [tʰʊːn],
and the usual English-way spelling < Tun >
is quite close. However, Romabama is based on
Burmese-Myanmar orthography and the Romabama
spelling has to be {htwan:}. I have to stick to {a}
because of the assumption that the inherent vowel
of the akshara is the English <a>.
Contents of this page
UKT:
•
Consonant means "corresponding or alike
in sound". Because "consonant"
is derived from "sonant", it means
that it is pronounceable. Thus <k>
is just a letter because it cannot be pronounced,
whereas {ka.} being pronounceable is a consonant.
• In voice segments, above,
I have explained how my last name, Tun, is to be pronounced.
I have even attempted to spell it in IPA: /tʰuːn/.
I still have to explain how the first part of my
Burmese-Myanmar name, Kyaw, is to be pronounced.
(See pronouncing the {ya.ping.} sound later chapters.)
I have been trying to do it since I came out the West,
to US, in 1957. So far, I have been unsuccessful,
and my American friends "christened" me
"Joe", which has been made official in Canada.
And so my name is officially, Joe Kyaw Tun in Canada,
and I found that not to get floundered on my name,
the best policy is to introduce myself as Joe Tun. However
in Myanmar, the land of ancestors and where I was born,
I am still U Kyaw Tun -- the name my Burmese-Myanmar parents
have given me, with the honourific U in front.
What makes one consonant different from another?
But, first let's see what a
consonant is, in my notes.
UKT: Remember,
consonant is a sound, NOT a letter of the alphabet.
Thus, <k> is a letter of the English-Latin alphabet.
Similarly, /k/ is a symbol of the International Phonetic Alphabet.
Both <k> and /k/ cannot be pronounced,
whereas /ka/ is pronounceable because it has the vowel /a/.
Burmese-Myanmar akshara {ka.} is a consonant
because it can be pronounced. The sounds of
{ka.} and {hka.} can be represented by IPA /ka/
as in broad transcription. In narrow transcription:
{ka.} is represented by [ka] and {hka.} by [kʰa].
"Consonants involve constrictions, or gestures that
narrow the vocal tract at a particular point.
Most consonants use only one constriction, but some have
more than one. In order to fully describe a consonant sound,
we need to describe each constriction as well as
some other properties of the vocal tract." --
http://www.umanitoba.ca/faculties/arts/linguistics/russell/138/sec3/params.htm
UKT:
Before we proceed further, let's remind ourselves
how to distinguish consonants from the vowels:
• if the air, once out of the glottis, is
allowed to pass freely through the resonators,
the sound is a vowel (not necessarily "a e i o u").;
• if the air, once out of the glottis, is
obstructed, partially or totally, in one or more places,
the sound is a consonant (not necessarily "b c d f g ...)
Now, let's see how a consonantal sound is made.
As an easy illustration, take the case of pronouncing
<p> /p/. We immediately run into trouble
because /p/ must be supplied with a vowel before
it can be pronounced. This is not the case with {pa.}
which because of the inherent vowel can be pronounced.
To
pronounce <p>, let's supply it with a vowel
<a>. This gives us /pa/ which can be pronounced.
In articulating this sound both lips are pressed together,
and then they are suddenly opened which produced
a "popping" sound. That is the sound of <pa>.
There is a time lag between the opening of the lips and
the passage of a puff of air coming out of the glottis.
This time lag is known as the Voice Onset Time (VOT).
See VOT
in my notes.
According to Kevin Russell, a consonant can be completely
described by its 7 properties:
1. active articulator, 2. passive articulator,
3. constriction degree, 4. state of glottis,
5. nasality, 6. laterality, 7. airstream mechanism.
See
Describing consonant segments for consonant /k/.
Caveat: Burmese-Myanmars should take care
whenever they come across <p> used by
Western linguists. Their <pa> sounds more
like {hpa.}, and they take that there are
two allophones for /p/ such as (in narrow transcription)
[p] and [pʰ].
English /k t p/ have two sets of allophones:
[k t p] equivalent to sounds of
{ka. ta. pa.}; and
[kʰ tʰ pʰ] equivalent to sounds of
{hka. hta. hpa.}.
The {wag}-aksharas (excluding nasals)
are best described as:
• tenuis consonants:
{ka. ta. pa.} - [k t p]
• voiceless consonants:
{hka. hta. hpa.} - [kʰ tʰ pʰ]
• voiced consonants:
{ga.
da. ba.} - [g d b]
The easiest way to demonstrate the voiceless voiced contrast
is to place a finger lightly over the Adam's apple
to feel the vibrations of your vocal folds inside
your throat. And, say <k t p> or
{hka. hta. hpa.}. Next, say <g d b> or
{ga. da. ba.}. Though, your voice seems to be generated
in the same place, when you say <p>, you feel
only slight (or no) vibration. With <b>
you can feel some vibration.
I feel very uneasy to "equate" <k t p> to
{hka. hta. hpa.}. In English, unless
these are preceded by <s>, the sound
produced is always "aspirated".
Since, Burmese-Myanmars, unless trained
cannot pronounce <ska sta spa>,
I have no choice but to represent <ka ta pa> as
{hka. hta. hpa.}.
Now, repeat with <k, g>. You find
a difference in the strength of the vibrations.
Similarly, you will find differences in the strength
of vibrations for pairs, <t, d> and <p, b>
We therefore describe <k, t, p> as voiceless
and <g, d, b> as voiced.
Consonants such as <p, t>
are sometimes described as "fortis",
and <b, d> as "lenis". See
fortis-lenis contrast in my notes. In his dissertation,
Wouter Jansen, contrasted voiceless aspirated fortis
plosives (e.g. [pʰ, tʰ, cʰ, kʰ, qʰ])
and plain voiceless fortis plosives
([p, t, c, k, q]). This paper promised to be
a source of information for Burmese-Myanmar consonants,
and I will have to go into it in the future.
What W. Jansen has described as plain voiceless fortis plosives
[p, t, , k] are known as tenuis consonants in Wikipedia
http://en.wikipedia.org/wiki/Tenuis_consonant download 071001.
They have a VOT close to zero. Though these
are not found in English they are found in languages
descended from PIE (Proto-Indo-European) language,
such as ancient Greek and Sanskrit. See
Glottalic theory and
Laryngeal theory in my notes.
We will soon be talking about the parts of
the vocal tract in describing the consonants.
I have described the vocal track before in
Sound Production (in intro-voc.htm).
However, to save you the trouble of going to it, and getting lost,
I am giving the speech organs on the right.
Parameters of a constriction gesture
1. active articulator:
What articulator moves to make the constriction?
2. passive articulator:
What articulator does the active articulator touch or approach?
3. degree of constriction:
How close do the active and passive articulators get to each other?
4. laterality:
How is the tongue body shaped?
These should be specified for every constriction
involved in a consonant. The following properties
don't apply to any particular constriction,
but to the consonant as a whole:
Other parameters
1. state of the glottis: What are the vocal cords doing?
2. nasality: Is air escaping through the nose?
3. airstream mechanism: How is the flow of air being produced?
Answering each of these questions about the state of the vocal tract
is enough to uniquely identify any consonant. If two consonants are different,
they must differ in their answers for at least one of these questions.
UKT: I will have to disagree with Kevin Russell that the above 7 parameters are
enough to uniquely identify any consonants. It may be true for English, but for
Burmese-Myanmar, they are not sufficient to differentiate {pa.} [p] and {hpa.}
[pʰ] and other c1 and c2 {wag}-aksharas. This is because, [{p}] and [{hp}]
are not allophones -- they are separate phonemes.
Together, the active and passive articulator are often referred to as
the place of articulation (POA). The other parameters are often
lumped together and referred to as manner of articulation.
Contents of this page
UKT:
Myanmar akshara,
and in fact all the aksharas of the East, are founded on
solid phonemic principles long known in the East. See the works of
Panini, the
Sanskrit (Magadha? Pali?) grammarian of the 5th century BC,
http://en.wikipedia.org/wiki/P%C4%81%E1%B9%87ini download 071003. Panini
पाणिनि
himself might have been preceded by the oral-grammarians of the Harappan
civilization of the Saraswati River, quite a few cities of which are being
unearthed on the Indian subcontinent (now in Pakistan).
The akshara system preceded the IPA by thousands of years,
and orthography is quite phonetic. I can pronounce
my Burmese-Myanmar as it is written, and I would be
perfectly understood. However, the common people being
the common people, will pronounce the syllables in more
than one way, fully admitting that "what is written is correct,
but what they say is just sound":
{ré:tau. a.mhan/ hpat-tau. a.thän}.
A hallmark of the akshara system is the way
in which the consonants are presented:
{wag}-aksharas, nasal-aksharas, and {a.wag}-aksharas.
The presentation is a matrix of 7 rows X 5 columns.
Since both Devanagari and Myanmar aksharas are
presented in identical ways, I can easily transliterate
from Devanagari into Myanmar and back. In that way,
I could easily derive how a Devanagari word
should be spelled in Myanmar, or a Myanmar word
in Devanagari. Thus:
Transliteration of Panini's name in Devanagari
to Myanmar: पा णि नि -->
{pa-Ni-ni}
After doing this transliteration,
I checked the meaning of "pani"
with an Indian friend of mine. She confirmed that
it means "water". The name Panini probably
means someone living along a river, or one immersed
in the study of human voice. The production of
human voice is due to the air flowing out of the lungs
along the vocal tract similar to the flow of water
along a stream.
POA is pretty well described in Burmese-Myanmar grammar
or more precisely in Pali-Myanmar, which, I am ashamed to admit,
I had not known before: until my good friend U Tun Tint
of MLC pointed out that to me.
POA in Pali-Myanmar is
{Htaan}, and there are six: -- Pali-Myan-Gramm. p.005.
1. velar, 2. palatal, 3. retroflex, 4. dental, 5. labial, and, 6. nasal.
Since IPA listed 11, some would have to be regrouped
to be equal to 6 of Pali-Myanmar {Htaan}. See
Table of
IPA Consonants.
The active articulator usually moves in order to make the constriction.
The passive articulator usually just sits there and gets approached.
UKT:
The tongue is the most active articulator. All three parts of it,
the tip (apex), the middle part (dorsum), and the back (root)
can all take part in sound production. See
Coronal and Dorsal consonants /
Radical consonants in my notes.
Of the three parts of the tongue, the tip is the most active. It plays a major
part in production of "hissing" sounds, or
sibilants.
There are other articulators beside the tongue
giving rise to consonants. The following is from Wikipedia,
http://en.wikipedia.org/wiki/Epiglotto-pharyngeal_consonant download 070912.
An epiglotto-pharyngeal consonant
is a newly reported type of consonant, articulated with
the epiglottis against the back wall of the pharynx.
This contrasts with the pharyngeal consonants, where
the root of the tongue contacts the back wall of the pharynx,
and prototypical epiglottal consonants, where
the aryepiglottic folds contact the epiglottis.
Epiglotto-pharyngeal consonants
have been reported (and videotaped) in one language,
the Formosan language Ami of Taiwan,
which has an aspirated stop and, apparently, a
fricative as phrase-final allophones
of its (ary)epiglottal consonants. The IPA
does not have diacritics to distinguish these sounds
from the epiglottals; the discoverers used
the ad hoc and somewhat misleading transcriptions
[ʕ͡ʡ] and [ʜ͡ħ].
They are also said to occur in the Tsez language
of southwestern Dagestan (in Russian Federation).
A sound's place of articulation is usually named by using
the Latin adjective for the active articulator
(ending with an "o") followed by the Latin adjective
for the passive articulator. For example, a sound where
the tongue tip (the "apex") approaches or
touches the upper teeth is called an "apico-dental".
Most of the common combinations of active and passive articulator
have abbreviated names (usually leaving out the active half).
UKT: In describing the POA, IPA starts from the lips
into the interior of the mouth, whereas in
the akshara system, the description is from the interior
towards the lips. I am following the direction used
by the akshara system, and will describe the consonants
starting from "velar consonants". The POA
is the <velum> /viː.ləm/ or
the nearest in Burmese-Myanmar
{bhi-lam}.
The Velum. The palate is generally defined as the roof of the oral cavity
and separates the nasal and oral cavities from one another. It is divided into a
region with underlying bone called the hard palate and a region made up of
connective tissue and muscle called soft palate, or velum. (The terms ‘soft
palate ’and ‘velum' as used by linguists and phoneticians are largely
interchangeable. We will continue to use the term ‘soft palate,’ since the term
"velum" is often used to refer to other membranous tissues in the body.) The
hard palate comprises about two thirds of the palate; the soft palate makes up
the posterior third of the palate. The hard palate is fixed and immovable
because it is made of bone; the soft palate is fleshy and moveable because it is
made of muscle. The opening between the oropharynx and the nasopharynx is
regulated by movement of the soft palate. This action controls the degree of
nasality speech sounds. Additionally, the soft palate raises so the food will
pass down into the esophagus and not up into the nasal cavity. -- from:
http://www.jladefoged.com/pdf/chapter9.pdf
DJPD16 gives only 8 POA for English, however, IPA gives 11:
English : bilabial, labio-dental, dental, alveolar, post-alveolar, palatal,
velar, glottal
IPA: bilabial, labio-dental, dental, alveolar, post-alveolar, retroflex,
palatal, velar,
uvular, pharyngeal, glottal.
See also the diagram showing the POA for English in my notes.
Contents of this page
Fully describing a consonant involves answering
each of the seven questions discussed earlier, that is,
specifying the consonant for each of the parameters:
1. active articulator
2. passive articulator
3. constriction degree
4. state of glottis
5. nasality
6. laterality
7. airstream mechanism
UKT: We still have to discuss more fully on
Nasality and
Laterality,
which we will do in later chapters. And, we will also have to know about
the Manner of articulation which too will be discussed in later chapters.
Until we have done so, I will just give the consonant charts.
Some charts have a row for affricates (e.g., the chart below from DJPD16).
The IPA consonant chart doesn't -- you can always get the affricate symbol
by putting together the plosive and the
fricative symbols in the relevant column.
Contents of this page
UKT:
• The English digraph <th> has two pronunciations,
[θ] in <this> and [ð] in <that>.
Though it is usual to transcribe Burmese-Myanmar
{tha.}
as English <th>, I have found that it becomes
very misleading when affricates are considered. Thus,
I have no choice but to use {þa.} for transcription
whenever it is necessary. It is to be noted
that <þ> (known as the letter 'thorn') was used
in Old English in place of the digraph <th>.
• I have also found that it would be necessary
to introduce palatal /ç/ (Alt0231 - ASCII character),
to explain the confusion over
{sa.}. Please note that the pronunciation of {sa.}
is similar to that of the
French 'c with cedilla' [ç].
In the Burmese consonant system there are
four consonants for each of the following pairs.
According to Rev. U Kawwida of Toronto Myanmar Buddhist monastery,
the Burmese consonants {ga.} and {Ga.} sounds almost
the same to layman Myanmars. However, the Myanmar monks are
taught to pronounce them differently when reciting Pali.
Of the two {Ga.} has an {ha.} sound at the end. {Ga.}
is the Burmese-Myanmar equivalent of Hindi-Devanagari घ
(U0918 Devanagari letter Gha.)
Contents of this page
Note:
• Is it possible that Burmese-Myanmar
does not have alveolar fricative [s]?
To be unequivocal I would have to call [s] a "hisser" using the term I found in
UNIL. That Burmese
{sa.} is not
the hisser [s] can be seen in when it is used as a "killed" akshara in the word
{kis}
(meaning: to pilfer) transcribed as /[ki']/ (or IPA /kɪʔ/) by MLC
on MEDict018.
• In English <c> can have 3 pronunciations:
<cell> /sel/, <cat> /kæt/, and <cello> /ʧel/,
but the IPA [c] seems to be similar to [ç] with a pronunciation
similar to <s>. French [ç] in French word
[caleçon]
sounded like {sa.} or {hsa.} in my ears. The [c, j] pair has a
counter-part in [ç, ʝ] in the
fricatives.
• Burmese-Myanmar {a.wag}-consonants, {ya. ra. la. wa. ha.}
were medial-formers at one time. However, today, only
{ya. ra. wa. ha.} are the only allowable medial formers.
They form {ya.ping.} {ra.ris} {wa.hswè:} and {ha.hto:}.
My problem is where to put {wa.} and {ha.} in the IPA table.
From the position of English <w>,
{wa.} can be said to be a bilabial
approximant. Since, no fricatives can form
medials, {ha.} can only be a pharyngeal approximant.
I am expecting comments from my peers -- UKT 070915.
• Note that /w/ is not included in the IPA table of pulmonic consonants.
• I have placed the four conjunct formers, in red. You will note that they are
all approximants.
• Now, what is [ʔ] the
glottal stop? Why is there no letter of alphabet
in English-Latin to represent this sound in the regular writing system? Similarly,
why is there no akshara (character, or grapheme) to
represent this sound in Myanmar and Devanagari in the respective writing systems? What do
Western linguists do to represent what they think this sound is,
when they are trying "invent" a writing system
for a Tibeto-Burman language? See (online) "Recent Adaptations of the
Devanagari script for the Tibeto-Burman languages of Nepal",
by Michael Noonan, Univ. of Wisconsin-Milwaukee,
www.uwm.edu/~noonan/Recent-Adaptions-of-the-Devanagari-Script-for-the.pdf
which I have download on 071012, and set in HTML. My question is this.
Do we really need a grapheme for this sound?
My opinion to this date (071016) is this. It is just
a pure invention not needed in any system of writing.
The glottal stop was invented to explain the involvement
of <h> sound at the end of some syllables. See
glottalization in my notes.
• Each Burmese-Myanmar akshara represents three registers or tones,
e.g. /a/ {a. a a:}, /k/ {ka. ka ka:}, etc., which
the IPA represents with suprasegmentals.
• We sometimes see two terms
liquids and glides. They are nothing but {ya. ra. la. wa.}.
Go back to tab-IPA-con-b
Contents of this page
From:
Indian Script Code for Information Interchange, ISCII
http://varamozhi.sourceforge.net/iscii91.pdf
UKT: The Latin characters in Varg 4 are wrong.
They are: [ta tha da
dha na]
Contents of this page
End of TIL file.
