Sunday, April 7, 2013

Phonemes and Allophones

Img014-Phonemes and allophones.gif

The difference between phonetics and phonology

The terms phonetics and phonology are both used in discussions of the sounds of English and of English pronunciation in general. There is an important and fundamental distinction, but there is an obvious and equally fundamental connection. Phonology can only be a satisfactorily defined in relation to phonology. Similarly, Phonetics can only be given a satisfactory definition in relation to Phonology.
On the one hand, phonetics is the study of the sounds made by the human vocal device, especially those sounds used in speech. That is, it deals with allspeech sounds, describes how they are made, classifies them and gives some idea of the nature. On the other hand, phonology is the study of the selection that each language make from the vast range of possible speech sounds. It tries to describe how language organizes and uses the selection it makes. Both phonetics and phonology are concerned with the same subject matter which is the speech sounds as produced by the human vocal apparatus, but they look at this subject matter from different points of view.
Phonetics tends to be a more general discipline, concerned with speech sounds without reference to their function or role in any particular language. However, there are many notions and concepts, such as that of the syllable, which it is virtually impossible to discuss without bringing in phonological and other linguistic considerations. Also, phonetics can be more particular when it deals with the sounds of a particular language. This creates a close relationship with phonology as it is difficult to study the sounds of one language without mentioning their relation to each other and the way in which they work together. Likewise, phonology tends to be more particular as it deals with the patterning of sounds in a particular language, so it needs the reference to phonetics. Yet, it can have a more general aspect when it deals with the universal aspects of sounds pattern as many principles governing the way sounds are used apply to all languages although details differ from language to language.
To sum up, phonetics provides the descriptive and classificatory framework for phonology. In other words, phonetics describes and classifies the speech sounds while phonology studies how they work together and how they are used. Thus, phonetic is concerned with what speech sounds are or their nature while phonology is concerned with what they do or their function. The distinction between form and function is to be found in many fields of study. For example, the two English words pin & bin differ in meaning or function due to the difference in sound or nature of /p/ and /b/. It is clear that phonology deals with a system of sounds. An example of a system is the colors of the traffic light. The study of nature of color corresponds to phonetics while the study of systems in which colors are used corresponds to phonology.

Different Branches of Phonetics

Phonetics is the study of sounds, especially speech sounds, made by the human vocal devices. Generally, there are different branches of phonetics.
  • Acoustic phonetics studies the transmission of speech sounds through the air from the speaker to the hearer. That is, it is concerned with measuring and analysing the movement and vibration of the air. This involves investigation within the framework of physics. In other words, an acoustic phonetician deals with speech wave forms and studies their frequency and amplitude in much the same way as a physicist or acoustic engineer.
  • Auditory phonetics is the study of the hearing of speech sounds and deals with such questions as how we perceive and recognize different speech sounds. Such investigations take place largely within the framework of psychology.
  • Articulatory phonetics is the study of the production of speech sounds by the human vocal device. It deals with how the speaker produces them, the sounds he or she uses in speech and how they can be classified and described. This branch of phonetics concerns teachers and learners of language.

Phonemes and allophones

The phoneme is the smallest unit that creates a difference in meaning. A phoneme is not a sound, but it is rather an abstract entity which has particular phonetic forms on particular occasions. In other words, a class of sounds, which are identified by the native speaker as the same sound, is called 'phoneme.' The members of these classes, or the actual phonetic segments poduced by native speaker, are called 'allophones'. For example, [ph]of (phIn) and [p] of [spin] are allophones of the phoneme /p/.
The phoneme is a unit of linguistic structure' which is just as significant to the native speaker as the word or the sentence. Phonemes are psychological units of linguistic structure. Thus, linguists sometimes describe phonemes as the form in which we 'think' of sounds and store them in our memory. For example, the first sound in a word like (path) is pronounced aspirated 'automatically' when the brain sends signals to the articulatory organs to produce a phonetic realization of the phoneme. Since phonemes are psychological concepts, they are not directly observable in stream of speech.
Allophones are the slightly different variants of phonemes which are found in different positions. Thus, the phoneme /p/ has two allophones, one found in (pin) and the other in (spin). The allophones of a phoneme may be siad to differ phonetically although they do not differ phonologically as they belong to the same phonological unit or phoneme. Differences which are significant or distinctive are phonological or phonemic, while differences which are not distinctive are phonetic or allophonic. This underlines the form/function distinction made earlier between phonetics and phonology. However, the allophones in one language may be separate in another. For instance, the difference between the aspirated [ph] and the unaspirated [p=] is allophonic or non-distinctive in English. In Sesotho, the differences between the two is distinctive or phonemic as [phola] means (jump off) and [p=ola] means (thrash). An allophone is, thus, a concrete sound representing an abstract class or group of sounds all having the same function and place in the system or phoneme.

The Characteristics of phonemes and allophones

Because phonemes are important units of linguistic structure, linguists must have a general method for identifying them in all languages. However, this is no possible because the set of phonemes differs from language to language. Thus, linguists have developed an objective procedure to discover the phonemes of a language through examination of a set of words written in phonetic transcription.
At first, phonemes make distinctions in meaning, so minimal pairs can be found with separate phonemes. For example, the minimal pair led and red is evidence that /l/ and /r/ are members of separate phonemes in English. But if two sounds are allophones of the same phoneme,minimal pairs should not exist.
Furthermore, the allophones of a phoneme are not a random collection of sound, but are a set of sounds which have the same psychological function. Accordingly, allophones of the same phoneme should be systematical related to one another. That is, they share many phonetic properties. Therefore, it is possible to predict which allophone will appear in a word on the basis of general phonetic principles.
Moreover, separate phonemes are phonetically dissimilar as each phoneme is a class or family of sounds. Yet, allophones of the same phoneme are phonetically similar. For example, /h/ and /ŋ/ cannot be members of the same phoneme while English has no initial /ŋ/ and no final /h/. This is because the description of the phoneme is different.
Since there are many environments in which a sound appear, it is easier to eliminate all of those except the most plausible by using strategies like the following.
  • First, hypotheses are formulated about the allophones.
    • Voiced nasals and liquids are more common than voiceless ones.
    • Oral vowels are more common than nasal vowels.
    • Consonants of normal duration are more common than long consonants.
    • 'Plain' consonants are more common than those with secondary articulations like velarization,palatalization and labialization.
  • However, some languages exhibit exceptions. For instance, nasal and oral vowels are separate phonemes in French.
  • Second, allophonic variation results from the application of phonological processes. For example, voicing differences in consonants are often caused by voicing assimilation which frequently occurs in consonant clusters. Since /r/ is the second member of all of the clusters, the consonant preceding it constituted the relevant environment.

Distinctive and non-distinctive sounds

In every language, certain sounds are considered to be the 'same' sound, even though they may be phonetically distinct. For example, native speakers of English consider the /l/ in lay to be the same sound as that in play, even though the former is voiced and the latter voiceless. Also, though the the aspirated [ph] of 'pin' and unaspirated [p] of 'spin' are phonetically different sounds, native English speakers overlook this difference.
However, a native speaker of Hindi could not overlook this difference. This is because his language contains many words that are phonetically identical, except that one word has an aspirated stop while the other has an unaspirated stop. In general, speakers tend to notice phonetics differences between two sounds only when the sounds have the ability to change the meaning of a word. Such sounds are said to be distinctive, ad they are determined by identifying a minimal pair.
For example,[tæp] and [thIp] constitute a minimal pair in English, in which [æ]and [I] contrast. However, minimal pairs can never be found for [r] and [l] in Korean because they do not appear in the same positions in words. [r] appears only between two vowels while [l] does not appear in this position. Observations about this sort play an important role in determining which sounds are considered to be the same by a native speaker.

Complementary distribution

There is one pattern for the distribution of sounds that is of practical importance for describing a language and learning. For example, /k/ has three classes of allophones or non-distinctive variants; one before front voweld in /k:/ (key), one before central vowels in /k3:/ (cur) and one before back voewls in /ka:/ (car). The three types of /k/ are in complementary distribution. This means that each one occurs in a fixed set of contexts in which none of the other occur.
Also, within each class there will be a certain amount of random variation. For instance, people asking Where's the car? differ in their individual renderings of /k/. Another example is that a invariably precedes a sequence beginning with a consonant sounds (a large cat) while an precedes a sequence beginning with a vowel sound (an enormous cat). The, distributions in the language are complementary or mentally exclusive. Sequences as well as individual phonemes are in complementary distribution.
The complementary distribution of allophones ensures that they are non-distinctive. For example, the allophone of /k/ in (car) is brought about by the choice of the word as a whole, the distinctive features being the vowels /i;, 3:, a:/ and not the allophones of /k/ which precede them. However, in Arabic, the contrast of what in English are simply front and bad allophones of /k/ is phonemic. Arabic has/kalb/ (dog) with a fronted velar plosive, and /qalb/ (heart) with a back one. Hence, the sounds that in English are allophones of /k/ are in Arabic separate phonemes /k/ and /q/.

Allophonic Variations

An allophone is a non-distinctive variant of a phoneme in a given language. An example of non-distinctive variation is a single phoneme /k/ Three allophones of /k/ are non-distinctive in English and in complementary distribution, each being limited to a particular type of environment.
However, within each of the three types, there are a further range of non-distinctive or free variations. For instance, the phoneme /k/ in a thousand pronunciations of the word car has no two utterances are ever completely identical. For this particular set of variants of /k/, the linguistic context was held constant. Since, they occur in the same word, they are non-distinctive.
Thus, allophones are manifestations of a phoneme which are either in free variation or complementary distribution. Yet, non-distinctive variation has definite limits. For instance, if /d/ is unvoiced, it becomes /t/. If the point of articulation is shifted far enough back, it becomes /g/. If the plosive mode of articulation is shifted to a fricative, it becomes /s/.


Aspiration is the puff of air that blows from the mouth after a consonant under certain conditions. This phoneme must be a voiceless plosive, which /p/, /t/, /k/. Also, voiceless plosive are aspirated when they occur at the beginning of accented syllables followed by a vowel. These aspirated allophones are placed in square brackets []. For example, the allophones of the phoneme /p/ is [ph] as in pin [phIn], and the allophone of the phoneme /k/ is [kh] as in kite [khaIt]. Yet, the aspiration does not occur of the conditions are not applied. For instance, the unaspirated [p=] occur in "spin" [sp=In] because this time it is not at the beginning of the syllable.


VOT is an acronym of "delayed voice on set time." This is a kind of allophone occurring only with vowels. It means that the pronunciation of the vowel is delayed for a very short time measured by millisecond. This allophone occurs when a vowel is preceded by an aspirated plosives. For example, (pin) is pronounced [pIَn] and (kite) is pronounced [kaَit].


Devoicing is a kind of allophones that occurs with consonants. It means means that the voiced phoneme like /n/, /m/ and /b/ is pronounced with less vibration, but it does not become voiceless. These allophones occur in three cases. When a voiced sound occur at the end of a word followed by silence, the voiced phoneme becomes devoiced, such as "bet" /bet/; "bed" [beḍ]. Also, when a voiced sound occurs after a voiceless consonant, it becomes devoiced like "leave" /li:v/ and "sleeve" [sḷi:v]. This allophone is placed between square brackets [] and is marked by (˳) below the affected phoneme.

Dark allophone

Dark Allophones is a kind of allophones that occurs with consonants. For example, the phoneme /l/ has two allophones. One is followed by vowles like /ʌ/, /ɑː/ and /a/ and is called dark /l/. For example, the word "less", pronounced /les/ while "luck" is pronounced [ḷʌk]. Likewise, other phonemes like /s/, /t/, /k/ and /d/ may have dark allophones under the same conditions. For example, "done" is pronounced [ḍʌn], "sigh" is pronounced [ṣaI] and "tough" is pronounced [ṭʌf]. In Arabic, also, the word "Allah" has the dark allophone [ḷ].

Phonological Processes

Phonological processes are the changes which delete, add and modify the sounds present in a careful or 'neutral' pronunciation of a word. The effect of phonological processes is to make the word easier to pronounce. For example, garage is pronounced as [gradʒ] rather than [gəradʒ] because the mouth is required to perform fewer articulatory movements due to the missing of one vowel. Another example is that [litərəčər] is considered to be an easier pronunciation of literature than is [litərəčər] because /d/ spares an articulatory movement of vibrating.
Phonological processes serve also to eliminate phonetic difficulties arising from the listener's perspective. For instance, the word strike may be pronounced as [straik]. This extra syllable containing a tense vowel may be pronounced when someone is forced to shout (strike one!).
These two motivations are opposed to one another. A word or phrase which is made easier to pronounce is also more difficult to perceived. The extra difficulty of understanding the phrase did you eat yet pronounced casually leads to the impression that the speech is 'sloppy' or 'slurred.' But if a word is easy to perceive, it may also be difficult to pronounce. Nevertheless, some compromise must be made between the needs for ease of pronunciation and ease of perception.

Optional Processes

The primary function of the phonological processes is to mark differences in speech styles. Fast, casual or inattentive speech generally contains phonological processes which make sequences of sounds easy to pronounce. Yet, slow, formal or emphatic speech styles contain phonological processes which make sequences of sounds easy to perceive.
The optional phonological processes have identifiable stylistic functions. The changes they induce are important. Phonetic segments may be altered radically, or entire syllables may be added, deleted or moved. Thus native speakers of English are clearly aware of the difference between the casual and formal pronunciations of a word like something, pronounced as [səmθiη] or [səmpm].
The stylistic function of these phonological processes also serves to make them optional. That is, they merely change the pronunciation of a word which is already pronounceable. Though a word like police is easier to say when the first vowel is deleted, the pronunciation containing both vowels is certainly possible.

Changes in Syllabicity

Liquids and nasals are syllabic if they appear after a consonant at the end of a word, as in bottom [badm] or ladle [led!]. Though syllabic nasals and liquids are required in cases like these, syllabicity may sometimes be governed by stylistic considerations. For example, the possessive pronoun mine [main], appearing in phrases like that's mine, is emphatically pronounced as [majn] in some English dialects. Likewise, flour [flawr] and hour [awr] may be pronounced with two distinctive syllables, as [ flawr] and [awr], respectively.

Vowel Deletion

In casual speech, native speakers of English frequently delete sounds, especially unstressed vowels. For example, mathematics is often pronounced as [ maeθmæetiks ] without [ə]. Family and familiar are two related words. In family [fæmili], the unstressed [ə] is deleted in casual speech, yielding [ fæmli]. However, it cannot be deleted in familiar [fəmiliə] because it bears the primary stress.


Dissimilation is a phonological process whereby two sounds, usually adjacent, become less like one another in terms of some articulatory feature. Thus, dissimilations are conceptually similar to assimilations. Though English has relatively few examples of dissimilations, the ones that occur frequently mark differences in style. For example, native English speakers pronounce fifth casually as [fift], or Smithsonian as [smiθtoniən ]. In both cases, sequences of fricatives have been changed to sequences consisting of a fricative followed by a stop. [fθ] changes to [ft] in fifth, and [θs] changes to [θt] in Smithsonian. Dissimilation in these examples serves to make difficult consonant clusters somewhat easier to pronounce.


In casual speech, the [t] and [d] in these pairs of words: (writer-riderlatter-ladder, and matter-madder) may sound the same. Though these words seem to contain a 'd', the sound in pronunciation is actually an alveolar flap, which differs from [d] as it is articulated more quickly. [d] and [t] are pronounced as flaps when they occur between two syllables, and the second must be unstressed. This sound is symbolized by [D].

Vowel insertion

Speakers often insert sounds into words. This may be done to break up consonant clusters that are awkward to pronounce. For example, many native speakers of English say athlete as [æθəlt], inserting a schwa in the difficult [θl] sequence. Moreover, sounds may be inserted for other reasons. In some dialects and casual styles of English, family is pronounced as [fæmbli] with [b] inserted between the [l] and the [m] and with the deletion of the unstressed [ə]. This is because [b] and [m] are identical as bilabial consonants. In contrast to the above examples, insertions may occur in self-consciously articulate speech. For example, words containing initial consonant-liquid clusters such as prayed generally pronounced as a single syllable. But schwa is inserted between the initial consonants, yielding [pəreid] which contains two syllables.


Metathesis is the reordering of sounds. For example, many native speakers of English, especially in casual situations, say [kɅmftrəbl] comfortable, showing metathesis of t and r. Similarly, many people say [pərksripŠn] for prescription. Although metathesis occurs rarely in English, it occurs regularly in a number of American Indian, Polynesian and Philippine languages.

Obligatory Processes

The obligatory phonological processes may be used to signal differences in style, but their application is for the most part not subject to deliberate manipulation. The 'unconscious' nature of these processes is evident in two ways. For one thing, native speakers have tremendous difficulty in perceiving the difference between the sounds of their language which have undergone these processes, and the corresponding sounds which have not been affected. This problem arises because the changes induced by these processes are always minimal. They modify a single feature of a single phonetic segment. Moreover, they are obligatory because the sequences of sounds affected are not pronounceable unless the processes apply first. Therefore, obligatory processes cannot give rise to alternative pronunciations of the same word.


English voiceless stops (p, t, k) are aspirated when they occur at the beginning of a stressed syllable. This means that a puff of air follows the release of the stop and precedes the voicing of the following vowel. The aspiration is indicated by placing a raised 'h' to the right of the voiceless stop. Aspiration makes a sound easier to perceive since it has the effect of increasing the difference between voiced and voiceless stops. For example, pin is pronounced as [phin] while spin is pronounced as [spin] as it is not aspirated.


Assimilation is a phonological process whereby one sound becomes more like an adjacent sound in terms of some articulatory feature, such as voicing, place of articulation or manner of articulation. The general effect of assimilations is to make pronunciation easier by reducing the number of articulatory gestures required for producing a sequence of sounds. Assimilations are extremely common in English and other languages of the world.
Voicing Assimilation
Voicing assimilation occurs when adjacent sounds come to agree with one another in voicing. For example, liquids ([l], [r]) and glides ([w,j]) are usually voiced in English, but are made voiceless when they follow a voiceless stop or fricative in the same syllable. They are said to be devoiced in those phonetic contexts and are marked by a '°' belwo the relevant phonetic symbol. For example, lay is pronounced as [lei], while 'play' is pronounced as [pḷei]. Since voicing assimilation is a general process, it may have the opposite effect. The word pot ends in a voiceless consonant while car ends in voiced consonants. The voiceless [s] is chosen as the plural in the word ending in a voiceless consonant, and voiced [z] is added to the words ending in voice sounds.
Place of Assimilation
Place assimilation occurs when adjacent sounds are made to agree in point of articulation. The nasals in the prefixes of the English words, like imbalance andindefinite, exhibit place assimilation. Those prefixes mean something like 'not'. The point of articulation of the nasal depends on the point of articulation of the consonant immediately following it. Thus, piety and balance begins with bilabial sounds, and the nasal of the prefix is also bilabial [m]. Likewise, /d/ of indefiniteis alveolar like the nasal [n]. Moreover, place assimilation may be applied to adjacent consonant-vowel sequences as in [k] or a [g] before a front vowel like [I, i:, e, æ], the place of articulation of these consonants is moved forward to the palatal region in anticipation of the 'palatal' vowels. [k] and [g] are said to be palatalized, and this is indicated by placing a 'Ʌ' over the affected consonant:
Manner Assimilation
Manner assimilation occurs when adjacent sounds come to agree in manner of articulation. For example, a vowel in English appears before a nasal consonant is nasalized. This is indicated by '-' over the vowel in narrow phonetic transcriptions like [kan] can and [binη] bing.

Vowel length

In English, the articulation of vowels is held a little longer when they occur before voiced consonants. This extra length is indicated by placing a (:) after the vowel in narrow transcription, such as [kæp] cap and [kæ:b] cab. Vowel length may make a sound easier to perceive as it serves to increase the distance between the vowel and the following consonant.

Phoneme and phonological processes

Single phonemes like /p/ have more than one pronunciation like [p] and [ph]. Phonological processes are responsible for the fact that sounds differently from the way they are stored in memory. Thus, they make the substitutions that 'distort' representations of sounds. They do this either to make sequences of sounds easier to pronounce—as when vowels become nasalized before nasalized consonants or to make sounds easier to perceive—as when voiceless stops are aspirated at the beginning of a stressed syllable. The operation of phonological processes on phonemes has three effects.
  • Phonological processes change single phonetic features. For example, voicing assimilation changes voicing, leaving all of the features for manner and place of articulation intact. Likewise, vowel nasalization changes only nasality, and does not affect vowel height, backness and rounding. Because of this, phonemes and their phonetic realizations are assumed to be basically similar. This means that no language have a phoneme /p/ pronounced as /a/ and /s/ because there is no phonological process that would 'convert' /p/ into sounds such different sounds. Also, the variant pronunciations of a phoneme share most phonetic properties.
  • The second consequence of phonological processes acting on phonemes is that the variant pronunciations of a phoneme make phonetic sense. A native speaker of English does not pronounce the phoneme /p/ as [ph] or [p] randomly. The different phonetic forms occur in 'complementary distribution'. That is, different allophones of the same phoneme appear in nonoverlapping phonetic contexts. For example, [æ] occurs only before nasals like in [kæn] a 'can' and [læmb] in 'lamb' as a consequence of vowel nazalization. Complementary distribution results when native speakers must apply a phonological process, so nasalized vowels occur before nasal consonants as it is impossible to pronounce an oral vowel in this phonetic context.
  • Filially, phonological processes may result in variant pronunciations of the same word. For example, pretty can be pronounced either as [priri] or [priDi] with no difference in meaning. When two sounds such as [t] and [D] occur in the same phonetic environment but are not distinctive, they are said to be in free variation. Free variation occurs when a phonological process is optional, whereas complementary distribution is found where the process is obligatory.