Update: 2016-02-03 02:55 AM -0500


The Human Voice


by Joe K. Tun (aka 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.

TIL website main index index.htm | |Top

Contents of this page

Alpha and Beta - alphabets and consonants
Sound producing animals - the talkative ones
Vowel sounds
Sonority (sound)
Spectrum (instrumental sound)
Sonority hierarchy (human voice sound)
Objectivity and Subjectivity

Passages worthy of note:
The smallest degree of pitch discrimination between two pitches depends on their intensity and frequency range (see DIFFERENTIAL THRESHOLD). Under the best conditions, a person with good hearing can discriminate about 1400 different pitches, of which 120 are used in the western scale of equal temperament.

UKT note
Canadian abugidas checked and free vowels
sibilant spirant 
tense and lax vowels
vowel triangle

Contents of this page

α β Alpha and Beta

alphabet n. 1. The letters of a language, arranged in the order fixed by custom. 2. A system of characters or symbols representing sounds or things. 3. The basic or elementary principles; rudiments. [Middle English alphabete from Latin alphabētum from Greek alphabētosalpha alpha; See alpha bēta  -- AHTD

aleph n. 1. The first letter of the Hebrew alphabet. See note at alphabet .
[Hebrew alep from elep ox from Canaanite alp] -- AHTD
beth ( bt) n. 1. The second letter of the Hebrew alphabet. See note at alphabet .
[Hebrew bt from bayit house] - AHTD
Taken together, they mean Ox-House 'stable for cattle'.

{d~da} - n. 1. meaningful speech sound; word; term. 2. grammar - MED517

The word 'alphabet', coming from the first two letters [phonemes - speech sounds, or, graphemes - representing phonemes], stands for a vowel and a consonant. The two terms consonants and vowels have confused me from the time I met them as a child learning to read and write. I had always thought they were grammatical terms. Of course, I didn't know that "grammar" aka {d~da} means 'sound' - what you speak (primarily) and not for what you write. Writing is secondary.

We were taught that there are 5 vowels in the English alphabet: <a e i o u>. Little did we know that since we were being taught to write, <a e i o u> are graphemes. The rest, we were taught, are consonants. We were also taught that in the Myanmar writing system, what we were writing down, are called "akshara". (Bur-Myan does not differentiate singulars and plurals.) We were always under the impression that the "alphabet" and the "akshara" are the same: they are not.

They are {by:} (equivalent of "consonant") and {a.ra.} (equivalent of "vowel"). Little did we realised that they are terms used for "sounds" (spoken words) rather than for "writing" (written words). Now what are consonants and vowels in production of "sound", "voice" or "syllable"?

Skt-Devan: व्यञ्जन vyajana n. consonant (grammar) - SpkSkt
Pali-Latin: व्य [vyajana] / [sara]
Pali-Myan: {byi~za.na.}
Bur-Myan: {by:} / {a.ra.}/{tha.ra.}

Skt-Devan: अनुनासिक -व्यञ्जन  anunāsika-vyajana  m. nasal consonants - SpkSkt
Skt-Devan: हलन्त  halanta  adj.  ending in a consonant - SpkSkt

Skt-Devan: स्वर  svara  m. vowel - SpkSkt
Skt-Devan: स्वराः  svarāḥ  noun  vowels - SpkSkt
Skt-Devan: अजन्त  ajanta  adj.  ending in a vowel - SpkSkt

<vowel killer> [the term used in TIL for {a.t}]
विराम  virāma  m. leave, pause, stop, cessation, end, termination, holiday - SpkSkt

According to Childers p.592, there are 2 variations of spelling for "consonant" vyajanaṁ and byajanaṁ . I have found that Burmese-Myanmar word for consonant corresponds to the second variation. Please remember that I don't speak or write Hindi-Devanagari and Sanskrit-Devanagari, but using the correspondence between the 2 scripts, I am finding that I can transliterate one from the other.  And in the process, using the POA and the sounds in IPA table, I can derive at how the two groups of people, the Indians and Myanmars, articulate the various phonemes.

The distinction between consonants and vowels is made in the following manner:
if the air, once out of the glottis, is allowed to pass freely through the resonators, the sound is a vowel ;
if the air, once out of the glottis, is obstructed, partially or totally, in one or more places, the sound is a consonant

In fact, across all languages, there are about 10 vowel sounds, and about 30 consonant sounds -- I am giving "round" figures. The number of graphemes allotted to these sounds depend on the writing system of the language. Instead of saying that there are 5 "vowels" in English, what we should be saying is that there are 5 graphemes (generally known as "letters") to represent the basic vowel sounds. Since there are more vowel sounds than 5, we use "digraphs" to represent them. How, we pronounce the vowel sounds -- either as monophthongs or diphthongs -- depend on the dialect we are speaking.

It appears that speakers of one linguistic group pronounce their vowels quite differently than from speakers of another group. And even in one group, individual speakers (man, woman, child) tend to pronounce their vowels slightly differently. Therefore when we are comparing languages, or even the way a person speaks, we need only a minimum of three vowels which are not easily affected by variation in articulation (known as quantal vowels) to characterize a language, dialect or idiolect. For Bur-Myan we need to concentrate on either:  {a.}, {i.}, {u.}, {au:} group, or, {a}, {i}, {u}, {au} group. Such a minimum of three characteristic vowels is known as a vowel triangle which is actually a quadrilateral.

quantal vowel n. A vowel whose acoustic qualities are little affected by variation in its articulation and whose perception is little affected by variation in its acoustic quality: one of [i u a]. Stevens (1972). - from A Dictionary of Phonetics and Phonology cited by http://www.bookrags.com/tandf/quantal-vowel-tf/ 100414


Contents of this page

Sound producing animals

Most of the higher animals can produce sounds using their breathing apparatus. Animals in the sea also produce sounds, but since we cannot hear them without instruments, we will leave them out. The most "talkative animals" are humans themselves, and the song birds. Though the common house crow (Corvus splendens) (not raven) , a commonly hated bird known derisively as {kyi:kn:}, is not considered to be a song bird, it can produce quite a variety of "songs" which I love to listen, simply, because they are always present around my house in Yangon, Myanmar. Eventually, I came to admire them because of their intelligence. From that time onwards they are not {kyi:kn:}, they are always {kyi:} or {kyi: ngak} to me. (The suffix {kn:} meaning "blind" is applied to them not because they are blind, but that's how their haters wish to see them: they have perhaps the sharpest of eyes always on the lookout to grab something to eat.)

In these higher animals, sound is produced by air coming out the lungs. Of course, there are other ways to produce sounds. But we will only look at the sounds produced by the egressive (flowing from inside to outside) air flow. The air flow coming out the glottis produces vowels or vowel-like sounds.

Contents of this page

Vowel sounds

Where in the mouth the consonants are produced can be easily seen. These places are known as Points of Articulation (POA). But the vowels are produced deeper in the throat in the larynx. But before we go to study of the larynx where the vowels are produced, let us say something of the vowels sounds.

It is said that the "vowel sound is a free-flowing sound" which can be either laminar flow (voiceless or vl.), or turbulent (voiced or vd.) within one "puff" of air coming out of the glottis. Most of the languages have only voiced vowels, but there are some languages which have voiceless vowels (Wikipedia http://en.wikipedia.org/wiki/Vowel 071230). After leaving the glottis, the signal carrying air puff has to travel upwards along the pharynx which opens into the mouth (oral) cavity and slightly further up into the nose (nasal) cavity. The flow can be modified by placing restrictions to the path, as in the flow of water in a concertina-like plastic-piping system. The restrictions can be "valves" and/or "weirs". The modification is also brought about by changes in the "shape" and/or "length" of the pipes. Never compare the vocal system to a system of rigid metal pipes with trap-door-like valves. Modification can occur before the vowel sound, or after the vowel sound. When the modification occurs before the vowel sound, we say that the consonant is in the onset of the syllable, and if it is after the vowel, we can say that the consonant is in the coda. This definition is important in the transliteration of Burmese-Myanmar into Romabama {ro:ma.ba.ma} 'backbone of Burmese', which is an "almost" one-to-one transliteration of Bur-Myan into Burmese-Latin.

The statement "vowel sound is a free-flowing sound" is somewhat self-contradictory because of the so-called creaky vowels in Burmese. The statement is strictly true for {a} but not for {a.} which is supposed to have a creaky pitch-register usually called tone. The vowels {a.}, {a}, and {a:} are described as the 3 pitch-registers of Burmese. They have been represented as: /a̰/ (as in /kʰa̰/ 'fee'); // (as in /kʰ/ 'shake'); and /kʰ/ 'be bitter' (Wikipedia http://en.wikipedia.org/wiki/Burmese_language 071226). However, I prefer to represent them with suprasegmentals {a.} (1 eye-blink) [ă], {a} 2 blk [a], {a:} 2 blk+emphasis [aː] . {a} [a] is the modal tone.

The question of short and long vowels arise in the case of checked vowels (vowels followed by consonants as in {Daat-hsi} (long vowel of 2 eye-blink duration followed by a killed consonant) vs. {tt-kywm:} (short vowel of 1 eye-blk followed by killed consonant). The rime in the first syllable can be represented by /t/ and in the second /t/. The problem of short and long vowels also appear in the case of free vowels in words denoting possessive pronouns: e.g. {nga} (a word considered to be somewhat rude to denote first personal pronoun "I' vs. {ng.}/{ngaa.} possessive pronoun "my". Please note that {nga.} and {ngaa.} have almost identical pronunciations. We differentiate them from the context. )

In Romabama we are faced with the problem to represent the 3 pitch-registers of Burmese with two tones (short vowels and long vowels) of English, and we run smack into what I have termed the Two-three tone problem.

Before proceeding further, it should be noted that the line between vowels and consonants cannot be clearly drawn; a continuum exists between the two extremes. This is because, the modification can be very slight or very drastic. From the point of view of Bur-Myan, we will come to see that among the well-defined vowels there are two divisions: the free and checked vowels. Vowels can also be divided into tense and lax vowels. Similarly, among the well-defined consonants, there are two major divisions: the obstruents and the sonorants. This division is best shown by sonority scale. See Sonority hierarchy in my notes.

There are also intermediate instances, such as the semi-vowels (e.g. <y> and {ya.}) and the (frictionless) spirants. Now, what are Burmese-Myanmar {sa.}/ {c} (r2c1) and {a.} (r6c5) (usually transcribed as {tha.}/{a.} ? And, Devanagari स [sa] (U0938) ? 
Are they spirants or sibilants, or something else? We will come to this presently.

UKT: The word for <Sunday> given by Wikibook is from  {ra.wi.} , <Sun> रवि - Dictionary of Pali Language, R. C. Childers, 1909 p403

In the production of human speech-sounds, there are two ways in which air can flow, depending on the position of the soft palate (velum). The soft palate acts like a valve which can completely shut off or open the nasal resonators, or stay in half way positions. However, because of the involvement of more than one pair of muscles in lowering and raising the velum, the action is far more complex than opening or closing a mechanical valve.

John Laver (1980, p.70) writes:
  "The physiology of the velopharyngeal system has been the subject of research by many workers, though largely from other disciplines than general phonetics. The facts about the action of the groups of muscles that serve to open and close the velum are reasonably well established. ... We are thus obliged to accept that different speakers may achieve auditorily (and perhaps articulatorily) similar results by physiologically different means. This is very likely to be true not only of the velopharyngeal mechanism, but of the entire speech apparatus. ... It is to be stressed that the velum does not move like a hinged trap door -- as is so often claimed in various books. In reality, the palate represents the anterior portion of the complex velopharyngeal valve, which functions mainly as a circular sphincter." -- The Phonetic Description of Voice Quality by John Laver, Univ. of Edinburgh, Cambridge Univ. Press. 1980, p70-75, part of the book available in http://www.ling.mq.edu.au/ling/units/sph302/papers/laver_1980_nasal.pdf 071029

In producing the nasal sounds, (with nasal consonants /ŋ, ɲ, ɳ, n, m/; <ng, --, --, n, m> or {nga., a., Na., na., ma.}), the soft-palate is lowered opening the nasal resonator. The air then flows through both the nasal channels (nose: there are two) and the oral channel (mouth). In the production of oral sounds (with oral consonants such as /k, t, p/ <k, t, p> or  {ka. ta. pa.} and {hka. hta. hpa.}), the soft-palate is raised shutting off the nasal resonator, and air flows only through the mouth.

It should be noted that as with all valves used in fluid flow, the soft palate can be completely open or tightly shut, or somewhere in between. We should always remember that there is a pronunciation difference between rapid speech (used by common people) as opposed to careful speech (as used by lawyers, politicians and public speakers).  The nasal/oral opposition concerns not only the consonants but vowels as well.

Whenever we come across English-Latin <k, t, p>, we must be careful about the pronunciation of /k/, /t/, and /p/. English pronunciation of the individual grapheme is of two kinds which are called allophones. For example, the letter <p> with the  pronunciation /p/ (broad transcription) has two allophones represented in narrow transcription: [p] in <spin> corresponding to {pa.} sound, and [pʰ] in <pin> corresponding to {hpa.} sound. The allophones [k t p] corresponding to {ka. ta. pa.} are known as tenuis consonants. They are unvoiced and unaspirated. The allophones [kʰ tʰ pʰ] are voiceless and aspirated.
English-Latin has only two nasals, <n> and <m>. Actually, there are 3 if you count in <ng> of <sing>. The IPA representation of <ng> is /ŋ/. In many English words, <g> is not pronounced, in which case <ng> (perhaps best represented as <ng>) behaves exactly like Burmese-Myanmar {nga.}. Remember, <ng> is a digraph, and the "g" is generally not pronounced: <singer> /sɪŋ.əʳ/ (US) /sɪŋ.ɚ / vs. <finger> /fɪŋ.g|əʳ/ (US) /fɪŋ.g|ɚ / -- <ng> is pronounced differently.
IPA arranges the oral plosives in the order p --> k, whereas in the aksharas including Devanagari, the usual way is k -->p. I am following the akshara way.


Contents of this page


Based on: www.sfu.ca/sonic-studio/handbook/Sonority.html 080322
UKT: The terms such as quality and timbre are semi-technical terms and you should be very careful when you see them.
I have left some of the original online links intact because I might have to go into them later.

Sonority or the sonorous quality of a sound is an auditory conceptual quality.

The tonal quality or timbre of a sound. The term is usually used in a subjective, descriptive manner, often with such adjectives as 'full' or 'rich'.
   Timbre, on the other hand, can be accounted for in terms of frequency content (spectrum) and its behaviour in time, and therefore can be regarded analytically as well as descriptively.

Quality (electroacoustics / music): The faithfulness (or fidelity) with which an original sound is reproduced electroacoustically in terms of frequency response and lack of distortion.
   In music the term quality was for many years used to express timbre, but this is now used less frequently.

Timbre (psychoacoustics / Music): Timbre or tone quality is determined by the behaviour in time of the frequency content (or spectrum of a sound [{see next section}], including its transients which are extremely important for the identification of timbre.
   The presence and distribution of these frequency components, whether harmonic or inharmonic [{antonym of harmonic}], and their onset, growth and decay in time (see Fourier Analysis), together with phase relations between them, combine to give every sound its distinctive tonal quality or timbre.
   Often qualities of timbre are described by analogy to colour or texture (e.g. bright, dark, rough, smooth), since timbre is perceived and understood as a 'gestalt' impression reflective of the entire sound, seldom as a function of its analytic components.
   With musical instruments, timbre is a function of the range in which the sound has its pitch (see mass), as well as its loudness, duration, and manner of articulation and performance. The same applies with speech, where timbre is the basic quality which allows one to distinguish between different voices, just as between different instruments or other sounds.
Sound Examples:
Cello note <)) Gamelan instrument <)) School bell <)) Steam whistle <)) 

Pitch (Psychoacoustics / Music): The subjective impression of frequency, in the same sense that loudness is the subjective sense of the intensity or amplitude of a sound. As such, pitch is a psychoacoustic variable, and the degree of sensitivity shown to it varies widely with people. Some individuals have a sense of remembered pitch, that is, a pitch once heard can be remembered and compared to others for some length of time; others have a sense of absolute pitch called perfect pitch .

UKT: The above reminds me "colour" and "taste". Each of us "seems to know" whether two colours, or two tastes are different. However it is an impossible task to come up with numerical values in terms of measurable quantities. It is probably the best to define pitch as a psycho-acoustic variable of sound.

The pitch of a tone or note allows it to be placed in a musical scale ; thus notes of a scale are often called pitches, and given names (A, B, C, C#, doh, re, mi, etc.).
   The smallest degree of pitch discrimination between two pitches depends on their intensity and frequency range (see DIFFERENTIAL THRESHOLD). Under the best conditions, a person with good hearing can discriminate about 1400 different pitches, of which 120 are used in the western scale of equal temperament. The lowest pitch corresponds to the lowest frequency giving a sensation of tone, around 20 to 30 Hz. The highest pitch depends on the highest audible frequency, which varies with age and especially noise exposure, but lies generally in the range of 15 to 20 kHz with younger people.
   The sense of pitch depends on the intensity of the tone, as shown in the graph; below 1000 Hz, pitch tends to drop with increasing loudness, and above 1000 Hz, tends to rise. A tone must have a certain duration for pitch to be ascribed; if not, it is heard as a click. The nature of the spectrum of a complex tone will affect the sense of pitch as well. A note rich in overtones will appear to have a more definite pitch than a sine tone of the same frequency and intensity, for instance. The pitch of the complex tone will correspond to its fundamental frequency. Compare: SIMPLE TONE.

In a very complex inharmonic spectrum, however, a sound may appear to have several pitch components. A sound with a continuously changing pitch is called a glissando. A pitch change caused by a moving sound source or observer is termed DOPPLER SHIFT.
Sound Examples :
Simple tone (sine wave) <))
Complex tone (triangle wave) with the same frequency as the sine tone <))
1 kHz tone with duration of 40 ms being shortened to a 2 ms broad-band click where it loses its sense of pitch. <))
Cello note (harmonic spectrum) <))
Gamelan instrument (inharmonic spectrum) <))

The pitch ascribed to a complex tone or sound may not necessarily correspond to a frequency that is physically present in the sound. For instance, if a spectrum consists of harmonics beginning with the second or higher harmonic, the sound will still be heard as having the pitch of the fundamental, called the periodicity pitch or the missing fundamental (see FUNDAMENTAL for further discussion). In summary, Schouten has stated, "The pitch ascribed to a complex sound is the pitch of that component to which the attention, either by virtue of its loudness or of its contrast with former sounds is strongest drawn. Therefore the pitch of a complex sound may be different depending on the circumstances under which it is heard." See reference under RESIDUE.
   The distance between two pitches is called an interval . However, equal frequency intervals do not always give the same sense of pitch distance, depending on the range in which the interval is situated. For instance, a fifth in a high frequency range may seem to be a smaller pitch distance than a third in a lower range. The MEL scale is an attempt to measure this variation. An alternative theory of pitch perception judges each note in terms of its chroma or distinctive tone colour and its tone-height. In this system, pitches may be arranged in a helix instead of a one-dimensional order, with the recurring loops of the helix at OCTAVE intervals. See: LINEAR.

Contents of this page

Spectrum (sound)

From: http://www.sfu.ca/sonic-studio/handbook/Spectrum.html 080322
UKT: The terms such as quality and timbre are semi-technical terms and you should be very careful when you see them. Please remember that my training was in Chemistry, and I have the ears of the proverbial water-buffalo which cannot appreciate the sweet sounds of the harp. For my education, I have left some of the original online links intact because I intend to go into them later.

Spectrum may refer to a RANGE of frequencies, as in the audio spectrum (see AUDIO FREQUENCY) or the RADIO SPECTRUM. See Fig.2.04.

The first 32 partials of a Harmonic Series shown as a line spectrum is shown in Fig.2.04. The amplitude of each partial is inversely proportional to the partial number.

The frequency content of a sound or audio signal, often displayed as a graphic representation of amplitude (or intensity level) against frequency. Three-dimensional displays of a spectrum add the time variation on the third axis (see below). The spectrum of a sound is a primary determinant of its perceived timbre.
   Compare: MASS, PITCH, VOLUME.

A partial spectrum consists of discrete frequencies known as OVERTONEs, HARMONICs or INHARMONICs. A continuous spectrum consists of noise components. The spectrum of a sound may be determined by a sound analyser or by Fourier analysis   and is distributed over the audible range (20 to 20,000 Hz). A partial spectrum is also known as a line spectrum, where discrete frequencies are present. A continuous spectrum, on the other hand, shows frequencies continuously distributed over the audible range.

Analyzing the spectrum of a sound is a way of understanding its behaviour in the frequency domain, as opposed to its behaviour in the time domain,  according to its waveform (or envelope). The auditory system is designed to balance the simultaneous resolution of detail in both domains, as expressed by the LAW OF UNCERTAINTY [{comparable to Heisenberg Uncertainty Principle of Quantum Physics}].

 The spectral envelope refers to the contour or shape of the spectrum, particularly when it shows the maximum strength of each frequency component during the sound.

The spectrum of a sound may be altered electronically by filtering (or EQUALIZATION).  

Line spectrum of the partials of a viola string, omitting their time variation is shown in Fig.2.05.

Third-octave spectrum of the Salvatore Mundi bell, Salzburg. Peak intensities occur in the frequency bands centred on 200, 315 and 630 Hz, with a fundamental about 80 Hz is shown Fig.2.07.

Sound example of Salvatore Mundi bell, Salzburg, <))

Three-dimensional plot of a trumpet tone showing the amplitude envelope of the first 20 harmonics. From J. Beauchamp & A. Horner, "Synthesis of trumpet tones using a fixed wavetable and a centroid-controlled second order filter," Proceedings of the 1994 International Computer Music Conference, used by permission of the authors.


Contents of this page

Sonority hierarchy (Human voice sound)

sonority n. pl. sonorities 1. The quality or state of being sonorous; resonance. 2. A sound. -- AHTD

UKT: Sonority is a quality of sound: not loudness.
The following is from:
From: Syllable, Stress and Accent, Phone 2, Peter Roach http://www.personal.rdg.ac.uk/~llsroach/phon2/mitko/syllable.htm 080106
The graphical representation on the right is from Peter Roach.

The term sonority is also applied to human voice sounds. The vowels are the most sonorous, and the consonants are the least. It should be remembered that sonority or the sonority scale (also known as the sonority hierarchy) of the sounds of a language is a perceptual auditory quality. For me the sounds of English (especially RP) and Burmese are comparable. (You may take it as an approximation.)

Wikipedia http://en.wikipedia.org/wiki/Sonority_hierarchy download 071006
UKT: The original Wikipedia article did not explicitly give the allophone [tʰ]. It simply gave [t] which I presumed to be [tʰ]. You should always remember that when Western linguists give [k t p] they usually mean /k t p/. They usually do not differentiate between [k t p] and [kʰ tʰ pʰ].
J. Laver Principles of Phonetics, 1994, p504
Sean McLennan, 2002, http://www.shaav.com/professional/linguistics/sororities.html 071113
http://speech.bme.ogi.edu/tutordemos 071230

See also phonotactics in my notes.
On the right is my adaptation of sonority hiearchy:  Note my proposal to place palatal <c> in between plosives and fricatives to solve the problem of POA of {sa.}/ {c}.

A sonority hierarchy or sonority scale is a ranking of speech sounds (or phones) by how much 'sound' they produce. For example, if you say the vowel [a] {a}, you will produce much more sound than if you say the plosive [tʰ] {hta.}. Sonority hierarchies are especially important when analyzing syllable structure; rules about what segments may appear in onsets or codas together, such as SSP (Sonority Sequencing Principle), are formulated in terms of the difference of their sonority values.

Some languages also have assimilation rules based on sonority hierarchy, for example, the Finnish potential mood
(e.g. -tne- → -nne-).

Sonority hierarchies vary somewhat in which sounds are grouped together. The one below is fairly typical (1 indicates lowest sonority).

sonority hierarchy given in J. Laver Principles of Phonetics, 1994, p504
  -- low vowels
  -- mid vowels 
     [{Laver does not identify "mid vowels" but I presume one of the front mid vowels to be {} /e/.
     If he has mentioned cental vowel I would have taken schwa /ə/}]
  -- high vowels
  -- fricatives
  -- affricates
  -- stops

"Approximants are therefore more open than fricatives. This class of sounds includes lateral approximants like [l], as in lip, and approximants like [ j ] and [w] in yes and well which correspond closely to vowels  and semivowels." -- http://en.wikipedia.org/wiki/Approximant_consonant 071230.
"The glides (/j/ and /w/) and the liquids (/9r/ and /l/) in American English can be grouped together in a larger category called the approximants." -- http://speech.bme.ogi.edu/tutordemos/SpectrogramReading/cse551html/cse551/node38.html 071230

UKT: I will have to add more consonants (obstruents) to my adaptation of the sonority hierarchy: Against the sonority scale: evidence from Frankish tones, by Ben Hermans & Marc van Oostendorp, Meertens Institute, Amsterdam www.vanoostendorp.nl/pdf/sonorityscale.pdf 080103. The pdf file set in HTML is in TIL library. A relation from Hermans & Oostendorp is given below:

Different phonemes inherently contribute different amounts of energy to the acoustic signal relative to each other. This measure, often referred to as sonority, although difficult to tie down to an absolute acoustic correlate, allows us to rank phonemes relative to each other on a sonority scale (Ladefoged, 1982). -- Sean McLennan, 2002, http://www.shaav.com/professional/linguistics/sororities.html

Contents of this page

Objectivity and Subjectivity

UKT: Based on:
http://en.wikipedia.org/wiki/Objectivity_(science) download 070906
http://en.wikipedia.org/wiki/Intersubjective_verifiability download 070906

It is relatively easy to pinpoint anatomically where various sounds are produced. But when we want to describe the sound that is produced at a certain POA, it becomes very subjective. For example, where are Bur-Myan {a.} (approximant) and {sa.}/{ca.} (plosive-stop) produced? When I, as a Bur-Myan speaker, articulate {a.}, and ask another Burmese speaker to emulate me, he usually produces [θ]. Ask a Hindi-speaker, and he produces [s]. Ask an English speaker and he usually produces [θ]. Why is it so? When I articulate {a.}, is only one POA involved, or there are more than one? In the majority cases, we simply do not know. What a person hears and articulates depends largely on his first language (L1). And when the second person tries to emulate the first speaker, he usually produces what his L1 would allow him to do.

Objectivity in science is the property of scientific measurement that can be tested independently from the individual scientist (the subject) who proposes them. It is intimately related to the aim of verifiability and reproducibility. To be properly considered objective, the results of measurement must be communicated from person-to-person, and then demonstrated for third parties, as an advance in understanding of the objective world. Such demonstrable knowledge would ordinarily confer demonstrable powers of prediction or technological construction.

Intersubjective verifiability is the capacity of a concept to be readily and accurately communicated between different individuals ("intersubjectively"), and to be reproduced under varying circumstances for the purposes of verification. It is a core principle of empirical, scientific investigation.
   Although there are areas of belief that do not consistently employ inter-subjective verifiability (e.g., many religious claims), intersubjective verifiability is a near-universal way of arbitrating truth claims used by people everywhere. In its basic form, it can be found in colloquial expressions, e.g., "I'm from Missouri. Show me!" or "Seeing is believing." The scientific principle of replication of findings by investigators other than those that first reported the phenomenon is simply a more highly structured form of the universal principle of intersubjective verifiability.
   Each individual is a subject, and must subjectively experience the physical world. Each subject has a different perspective and point of view on various aspects of the world. However, by sharing their comparable experiences intersubjectively, individuals may gain an increasingly accurate understanding of the world. In this way, many different subjective experiences can come together to form intersubjective ones that are less likely to be prone to individual bias or gaps in knowledge.
   While specific internal experiences are not intersubjectively verifiable, the existence of thematic patterns of internal experience can be intersubjectively verified. For example, whether or not people are telling what they believe to be the truth when they make claims can only be known by the claimants. However, we can intersubjectively verify that people almost universally experience discomfort (hunger) when they haven't had enough to eat. We generally have only a crude ability to compare (measure) internal experiences.  -- http://en.wikipedia.org/wiki/Intersubjective_verifiability

Much has been said about vowels: how they are produced and where. Vowels are commonly described based on the Daniels Jone's vowel quadrilateral. The problem I am running into in transliteration of Burmese to English, and vice-versa is in describing the Burmese-Myanmar vowels and consonants. Undoubtedly, much of it is due to my lack of precise understanding of the descriptions given by the Western linguists and phoneticians. That can be mitigated by sustained effort: not only by the study of phonetics and linguistics, but in other areas as well. But I would have say that much of the problem depends on the impreciseness of what the phoneticians and linguists are saying. And also on the very nature of vowel perception and production. Please refer to Rosner & Pickering, in Vowel Perception and Production, Oxford Univ. Press, 1997. On p.1-2, they write:
   "Vowel perception entails two processes. ... categorization [{and perception}] ... [{First}] An English listener hears two different vowels [{ for the " i sound" in }] ... <bead>  and <bid> [{ /biːd/ and /bɪd/}]. The different meanings ... depend on the English phonemic distinction between the vowels /i/ and /ɪ/. ...
   "The second process ... perception is ...  when a given vowel is produced in the same context and with the same stress by different speakers, such as man and a woman. ... English listeners hear identical vowels in ... <bid> and <lid>. Rhyming, amongst other phenomena, depends on this fact. ...
   "All these acoustic variations in different productions of a given vowel raise a second question. How do listeners hear the same vowel despite the vicissitudes of the physical stimulus itself? In psychological terms, this issue is a constancy problem. ... Can a single type of representation, physical or psychological, cope with the effects of the major sources of variation vowels: speaker identity; consonantal environment; stress; and momentary speaker characteristics?"

My question is: how much can we rely on the IPA transcriptions?

You have to be aware that Bur-Myanis an Abugida-Akshara (a phonemic script -- see Canadian abugidas in my notes.), and English-Latin is an Alphabet-Letter (a non-phonetic script). And, as a university lecturer, whenever I speak, I speak clearly and precisely without slurring. Though, I came from a bilingual family, in Burmese and English), I am more attuned to Burmese than to English, especially because I was born in Myanmar, and because I had received my early education in Myanmar. My English teachers were Anglo-Indians, Anglo-Burmans, Indians and Burmese. My early English pronunciation was that of well educated Myanmars, some of whom had been trained in Britain and US. I had early contacts with the Americans -- my father's co-workers, but not much with the British. I have spent considerable time in the US, in Australia, and now nearly 30 years in Canada (I am now a naturalized Canadian citizen), and my pronunciation has changed, and is still changing, especially when I got involved in linguistics. And because of my life-long interest in languages, I always try to note how a person speaks: his or her vowels and consonants, and grammar.

The Daniel Jone's vowel quadrilateral is a subjective description. Though it is useful to me who has taken up linguistics and phonetics rather late in life, I feel it is inadequate. Please remember that I have no formal training in those subjects. And, remember that I am chemist, one of those are said to be "never satisfied unless they can isolate a substance, purify it, and put it in a bottle, and, most importantly, PUT a label on it". I am interested in an objective way to place the vowels, monophthongs and diphthongs, both of Burmese and English, in the vowel diagram. One way to do it, is to look into the "human-voice" production by machines, and to "measure" the vowels and consonants in terms of formants primarily as F1 and F2. Now, what are the formants? See:
How sound is produced and heard [former hv6.htm] - snd-hear.htm - (link chk 160203)

Contents of this page

UKT notes

Canadian abugidas

From: Wikipedia http://en.wikipedia.org/wiki/Canadian_Aboriginal_Syllabics download 070908

Canadian Aboriginal syllabic writing (often "syllabics" for short) is a family of abugidas used to write a number of Aboriginal Canadian languages of the Algonquian, Athabaskan, and Inuit language families.

Canadian syllabics are presently used to write all of the Cree dialects from Naskapi (spoken in Quebec) to the Rocky Mountains, including Eastern Cree, James Bay Cree, Swampy Cree and Plains Cree. Syllabics are also used to write other Algonquian languages, such as the major Ojibwe dialects in Western Canada and Blackfoot, and to write Inuktitut in the eastern Canadian Arctic. Among Athabaskan languages, syllabics have been used to write Dakelh (Carrier), Chipewyan, Slavey, Tli Cho (Dogrib), Tasttine (Beaver). These native nations are found across Canada, in British Columbia, Alberta, Saskatchewan, Manitoba, the Northwest Territories, Ontario, Quebec and Nunavut. Syllabics are occasionally used in the United States by Native communities that straddle the border, but are principally a Canadian phenomenon.

The bulk of the characters (graphemes) used by this writing system are encoded into the Unified Canadian Aboriginal Syllabics character table as part of the Unicode standard. This page will use Unicode characters that may only be readable with an appropriate font.

Go back cana-abugi-note-b

Contents of this page

checked and free vowels

UKT: More on this in intro-vow.htm

Bur-Myan system of writing is an Abugida-Aksara system and is totally different from that of Eng-Latin which uses an Alphabet-Letter system. In the Abugida-Akshara the syllables are pronounceable, whereas the Letters are mute.

In both systems, the vowels are pronounceable.

Each Akshara-consonant has an intrinsic vowel is pronounceable, e.g. {ta.} /ta/. The intrinsic vowel can be killed with a Virama {a.t} 'vowel-killer', resulting in {t} which is no longer pronounceable.

The Alphabet-Letter t is mute. The opposite of {a.t} 'vowel-killer' is logically the 'vowel-giver' {a.} in Burmese, a in English, ა /a/ in Georgian, and अ a. Alphabet-Letter and Abugida-Akshara systems can be equated. As an example, take the case of Dental plosive-stop of the same POA (Point of Articulation) and pronunciation /ta/ :

Bur-Myan: {ta.}
Eng-Lat:    ta 
Georgian:  თა
Skt-Dev:    त ta

The general form of the syllable is CV where is a killed consonant. Free vowels are those found in syllables of the form CV. Even in syllables of the form CV that end in "killed" {a.t} consonants, (V) is a "vowel". Here the "killed" aksharas are those of the {wag}-division excluding the nasals, and {a.wag}-consonants.

Vowels of CV syllables are free vowels, and those of C(V) are checked vowels. Free vowels can have 3 registers (usually described as tones) of the form {a. a a:} which can be differentiated from each other primarily by their lengths. Burmese is one of the rare languages that have very short duration or register #1 [ă]. Most of the Indic languages, have only short and long vowels. Register #1 is commonly described as "creaky tone". Middle register, #2, is described as "modal". See The Analysis of Voice Quality in Speech Processing, by Eric Keller, Informatique et mthodes mathmatiques (IMM), Facult des Lettres, Unversit de Lausanne, Switzerland, eric.keller@unil.ch www.unil.ch/webdav/site/imm/users/ekeller/public/keller/Keller_04_VoiceQuality.pdf 071101. You will see my HTML version, vq-keller-analysis.htm, in my library. Register #3 {a:} may be viewed as made up of two segments [a] and [ː] or [aː]. (Note in [ă] [ː] [aː], I have used the suprasegmentals).

The following is from:
Wikipedia http://en.wikipedia.org/wiki/Checked_and_free_vowels 071006

Remember that syllables are parts of a word. In a monosyllabic word, there is only one syllable, whereas in a polysyllabic word there is more than one syllable, and the most prominent syllable is known as the stressed syllable. In phonetics and phonology, checked vowels are those that usually must be followed by a consonant in a stressed syllable, while free vowels are those that may stand in a stressed open syllable with no following consonant.

Usage of the terms:

The terms checked vowel and free vowel originated in English phonetics and phonology. They are seldom used for the description of other languages, even though a distinction between vowels that usually have to be followed by a consonant and those that do not have to is common in most Germanic languages.

The terms checked vowel and free vowel correspond closely to the terms lax vowel and tense vowel respectively, but many linguists prefer to use the terms checked and free as there is no clear cut phonetic definition of vowel tenseness, and since by most attempted definitions of tenseness /ɔ/ and /ɑ/ are considered lax, even though they behave in American English as free vowels.

Checked vowels is also used to refer to a kind of very short glottalized vowels found in some Zapotecan languages that contrast with laryngealized vowels. the term checked vowel is also used to refer to a short vowel followed by a glottal stop in Mixe, where there is a distinction between two kinds of glottalized syllable nuclei: checked ones, with the glottal stop after a short vowel, and nuclei with rearticulated vowels (a long vowel with a glottal stop in the middle).

The schwa /ə/ and rhotacized schwa /ɚ/ are usually considered neither free nor checked, since they cannot stand in stressed syllables at all.

Go back check-free-vow-wiki-b

Contents of this page


From Wikipedia http://en.wikipedia.org/wiki/Sibilant_consonant 070918

UKT: From the following, I have to conclude that Burmese-Myanmar {sa.} and {za.} are sibilants. The inference is {a.}/{tha.} the r6c5 character (grapheme), is not to be confused with /s/. And Pali-Myanmar {a.}/{tha.} is not to be pronounced as /s/.

A sibilant is a type of fricative or affricate consonant, made by directing a jet of air through a narrow channel in the vocal tract towards the sharp edge of the teeth.

UKT: "A sibilant is a type of fricative ... " does not mean that all fricatives are sibilants. If you go on reading this downloaded file you will come across a non-sibilant [θ] voiceless alveolar. [θ] is sometimes called a thibilant.
See Wikipedia http://en.wikipedia.org/wiki/Thibilant 070918

sibilant Linguistics adj. 1. Of, characterized by, or producing a hissing sound like that of (s) or (sh): the sibilant consonants; a sibilant bird call. n. 1. A sibilant speech sound, such as English (s), (sh), (z), or (zh). [Latin sībilāns sībilant-,present participle of sībilāreto hiss] -- AHTD

Sibilants are louder than their non-sibilant counterparts, and most of their acoustic energy occurs at higher frequencies than non-sibilant fricatives. [s] has the most acoustic strength at around 8,000 Hz, but can reach as high as 10,000 Hz. [ ʃ ] has the bulk of its acoustic energy at around 4,000 Hz, but can extend up to around 8,000 Hz.

The spin-off terms shibilant, and rarely thibilant, are used to describe particular kinds of sibilant.

Symbols : Of the sibilants, the following have IPA symbols of their own:
Alveolar: /s/ , /z/ (either apical or laminal)
  - / ʃ / , /ʒ/ (Palato-alveolar: that is, "domed" (part ially palatalized) postalveolar, either laminal or apical)
  - / ɕ / , / ʑ / (Alveolo-palatal: that is, laminal palatalized postalveolar; these are equivalent to ʃʲ, ʒʲ)
  - / ʂ / , / ʐ /: (Retroflex, which can mean one of three things: (a) non-palatalized apical postalveolar, (b) sub-apical postalveolar or pre-palatal, or (c) non-palatalized laminal ("flat") postalveolar, sometimes transcribed / s̠ z̠ / or / ʂ̻ ʐ̻ /..

Othr definitions of sibilants: Some authors, as for instance Chomsky & Halle (1964), group [⁠ f] and [⁠v] as sibilants. However, they do not have the grooved articulation and high frequencies of other sibilants, and most phoneticians (for instance by Ladefoged & Maddieson 1996), continue to group them together with the bilabial fricatives [⁠ɸ, β] as non-sibilant anterior fricatives. Some researchers judge [⁠f] to be strident in one language, e.g. the African language Ewe, as determined by experimental measurements of amplitude, but as non-strident in English.

The nature of sibilants as so-called 'obstacle fricatives' is complicated - there is a continuum of possibilities relating to the angle at which the jet of air may strike an obstacle. The grooving often considered necessary for classification as a sibilant has been observed in ultrasound studies of the tongue for supposedly non-sibilant [θ] voiceless alveolar non-sibilant fricative (Stone and Lundberg, 1996, Journal of the Acoustical Society of America, vol. 99: 3728-3737). More research on the phonetic bases of the terms sibilance and stridency, and their interrelationship, is required.

Go back sibilant-note-b

Contents of this page


spirant Linguistics n. 1. See fricative . adj. 1. Fricative. [Latin spīrāns spīrant-,present participle of spīrāreto breathe] -- AHTD

sibilant Linguistics adj. 1. Of, characterized by, or producing a hissing sound like that of (s) or (sh): the sibilant consonants; a sibilant bird call. n. 1. A sibilant speech sound, such as English (s), (sh), (z), or (zh). [Latin sībilāns sībilant-,present participle of sībilāreto hiss] -- AHTD

From: http://cancerweb.ncl.ac.uk/cgi-bin/omd?spirant 070920
UKT: Going online to search using the string "spirant" does not give a Wikipedia article. The above was one of the articles that were available.

A term used differently by different authorities;
by some as equivalent to fricative, that is, as including all the continuous consonants, except the nasals m, n, ng;
with the further exception, by others, of the liquids r, l, and the semivowels w, y;
by others limited to f, v, th surd and sonant, and the sound of German ch, thus excluding the sibilants, as well as the nasals, liquids, and semivowels.

The above paragraph was written as one sentence. I have inserted three "bullets" for easy reading. I have also put in bullets into the second sentence of the excerpt below.
Because of "... by some as equivalent to fricative", I used the string "fricative" in my internet (Google) search. The following is what I got: http://en.wikipedia.org/wiki/Fricative_consonant 071218. An excerpt from the Wikipedia file is given below.

Fricatives (or spirants) are consonants produced by forcing air through a narrow channel made by placing two articulators close together. These are:
the lower lip against the upper teeth in the case of [f], or
the back of the tongue against the soft palate in the case of German [x], the final consonant of Bach.
This turbulent airflow is called frication. A particular subset of fricatives are the sibilants (sometimes referred to as stridents). When forming a sibilant, one still is forcing air through a narrow channel, but in addition the tongue is curled lengthwise to direct the air over the edge of the teeth. English [s], [z], [ʃ], and [ʒ] are examples of this.
... ... ...
Central non-sibilant fricatives
[f], [v], [
θ], []

Go back spirant-note-b

Contents of this page

tense and lax vowels

Excerpt from: Wikipedia http://en.wikipedia.org/wiki/Tenseness 071009

In general, tense vowels are more close (and correspondingly have lower first formants) than their lax counterparts. Tense vowels are sometimes claimed to be articulated with a more advanced tongue root than lax vowels, but this varies, and in some languages it is the lax vowels that are more advanced, or a single language may be inconsistent between front and back or high and mid vowels (Ladefoged and Maddieson 1996, 3024). The traditional definition, that tense vowels are produced with more "muscular tension" than lax vowels, has not been confirmed by phonetic experiments. Another hypothesis is that lax vowels are more centralized than tense vowels. There are also linguists who believe that there is no phonetic correlation to the tense-lax opposition.

In many Germanic languages, such as RP English, standard German, and Dutch, tense vowels are longer in duration than lax vowels; but in other languages, such as Scots, Scottish English, and Icelandic, there is no such correlation.

Since in Germanic languages, lax vowels generally only occur in closed syllables, they are also called checked vowels, whereas the tense vowels are called free vowels as they can occur at the end of a syllable.

Go back tense-lax-vow-wiki-b

Contents of this page

vowel triangle

From Wikipedia: http://en.wikipedia.org/wiki/Vowel_triangle 071222

A vowel diagram or vowel chart is a schematic arrangement of the vowels. Depending from the particular language being discussed, it can take the form of a triangle or a quadrilateral. Vertical position on the diagram denotes the vowel closeness [{height of the tongue measured by the first formant F1}], with close vowels [e.g. { i }] at the top of the diagram, and horizontal position denotes the vowel backness [{measured by second formant F2}], with front vowels at the left of the diagram.

UKT: For comparing two languages a vowel triangle is enough, and it is usual to compare [i , a , u] of individual speakers. However, since occurrence of [a] is rare in English, I have to take [i , ɑ , u] which correspond to Burmese-Myanmar {i , au , u}. Note that Pali-Latin <o> is not the IPA [o] but [ɑ] equivalent to Pali-Myanmar {au:}. The Pali-Latin transcription <o> is the source of confusion in transcription of Burmese.

The vowel systems of most languages can be represented by such a chart. Usually they are evenly distributed on the chart, a phenomenon that is known as vowel dispersion. For most languages, the vowel system is triangular. Only 10% of languages, including the English language, have a vowel diagram that is quadrilateral. Such diagrams are termed vowel trapeziums or vowel quadrilaterals. German phologists know these as, respectively, a Vokalviereck and a Vokaltrapez.

The IPA vowel chart comprises the cardinal vowels, and has the form of a trapezium. By definition, no real vowel sound can be plotted outside of the IPA trapezium, because its four corners represent the extreme points of articulation. The vowel diagrams of most real languages are not so extreme. In English, for example, high vowels are not as high as the corners of the IPA trapezium, nor are front vowels as front.

Go back vow-triangle-note-b

Contents of this page

End of TIL file.