PART I Sounds1 Introduction With the exception of the Si un Lang ves used by the deaf, and written languages. the es wih which most of us are Funiiar 1 Sign Languages are exts spoke la ly on the mediuan of sound nely interesting, exhibiting all the complexities of es, but their serious study requizes the introduction of a consides able amount of specialised terminology for which we do not have space im an introductory book of this kind, As for written languages, they too have many fascinating features, Dut they’are regarded as secondary to sposen Tanguaes fora numer oF reasons, For instance, children are explicitly taught to read and write sometime after they acquire a spoken language, and many cultures have never employed writing systems. Thus, a focus on sounds is entirely appropriate, and this part of the book is devoted to discussion of the way in whieh the sound systems of languages are organised and the role of such systems in the acquisition and processing of la ystems difler from one dialect or variety of @ cs. We will also consider the ways in which sound ven language to another an the that we can identify in the sound system ofa given language over time chang Bolore we can discuss any aspect of the sound system of a langust indtvanseribing we need a systematic way of deserihi speech sounds, and in section 2 we introduce a standard transerption system, while explaint « how the more impor tant speech sounds are produced. It is important to be clear that the purpase of this on is to introduce éermnology that enables us to tall about speech sounds, ‘with some precision, this being a prerequisite to our discussing any of the issues raised in our main introduction, Onee our transeription sysiem is in place, the most si ehtforward way to put ito use is sm conneetion with sociolinguistic issues Therefore. in section 3, we fncus on the ways that sound systems vary across ups ete. We shall soe that one dialet differs from another in dialects, soc ie. that so-called ‘substandard deviations’ are quite 1, contextual and linguistic principles. Seetion 4 examines how sound systems change over time to give rise to new dialects and ultimately new lan: Once more, we shall see thi to ‘Sloppiness* on the part of sp principles. Moreover, we shall di vemnedby soe’ sare neither random nor due fare subject to coherent ‘over that there isa close zelationship between ic al any point in time and historical change variation ina go In section $, we begin to introduce some of the more abstract concepts that are important in understanding the phonological component of a grammar. Amon these concepts is that of the phoneme. a unit of phonological analysis, and we will26 SOUNDS also touch upon the structure ofthe syllable, a particulary important unit in sound systems. Phonological processes have already received a brief introduction pp. 41), and in this section we shall consid intsodueing the important concept of alternation, such as we ean observe in connection with the'a’ vowels in Japanand Japanese. The word Japanese clearly consists ofapar followed by the ending se, and native speakers of English will read thatthe twota’ vowels of Jupaaare different; the firsts Hke the ‘wat ‘about whereas the second is like the “a? af pam. However, in the word Japanese ech of the toa" your's has the opposite quality and we say tht they alfermate— it scoms as if the addition of-ese causes a change in the vowels of apa, This ditforonce is a systematic property of the language and, unlike the examples mentioned inthe main introduction, it dacs not depend on whether we are speak fully or not; much of this section is devoted to such phenomena, and we vill show how they ean be described in terms of processes. In the last two seetions of this part of the book, we ramine some of the developmental and psycholinguistic issues that arise in connection with sound systems, Seotian 6 discusses how phonology can throw light on the acquisition of pronunciation patterns by children leaming their frst language. Additionally. it illustrates the interaction between approaches alluded to in the main introduction, im that we will see that aspects of child phonolog which also find @ role inthe formulation of adult wwe will consider selected aspects of specck perception, along with common ‘everyday errors in speech production (so-called slips of the tongue). This section concludes with a brief discussion of the role of phonology in. understandin certain aspeets of poetic systems and the way that writing systems have devel oped, Overall, the section secks to establish the importance of some of the roguire theoretical notions -ammars, Finally, in sztion 7. theoretical notions introduced in section 5 for the under with which some readers wall already be familiar janding of phenomena2 Sounds and suprasegmentals How many sou are there in English? This seems like a reasonable enough question, but in fet its diffieult to answer. for several reasons. A major problem is thatthe spelling system of English (its orthography) is izewular and doesn't represent sounds in a completely consistent way. Sometimes one sound can be {sound of Katy (ors tC), but worse spelled in several ways as with the fi ‘we Find that some sounds just aren't given their own symbol at all, There is a difference between the fist sounds of sockand sock, but the first of these sounds is represented by two symbols sand fi, each af which corresponds to a sound that is lfferent ta the first sound of shock Moreover, although most speakers of English will distinguish the midalle sounds in put ‘to place’ and puct ‘to strike « golt ball while its on the seven’, ths distinction is never made inthe writing system, Wealso need to he carefi! about what we mean by “English, as pronunciation differs from one dialect to another. ln the Nor of Enwkand, for instanes, both put and puttare often pronounced like put and dialects inthe United States lle as to ‘which (funy of the sounds in bold face in the words mer), mupyand Mary they distinguish, Those are systematic diff aprice on the part of tion 3. In the present speakers, an issue that will be discussed in more detail ins contest, however, such observations indicate a clear need for some way of writing down sounds which bypasses traditional orthography Moving aay from English, as noted alzoady. there are a ureal many languages which have never had a writing system of their own and which until recently have never been written down {hitherto undiscovered la still encountered in some pars ofthe world). For such eases ts essential that knguists can rely on a system af writing which can be applied tony huswa la completely unknown to the in estigator For these reasons, linguists have developed systems ofphon in which each sound is r transcription presented by just one symbol and each symbol represents just one sound, Unfortunately, there are several such systems ia use. Ln this book, we will use the anseription system of the International Phoneties Association, which is generally referred to as the IPA. This system, commonly used in Britain, derives fiom one developed in the 1920s by Daniel Jones and his colle ws at London University. one of whose aims was to provide writing systems for the Unvritten Janguawes of Afiiea and elsewhere One advantage of the TPA is that its accompanied by a wellefined method of dleseribing sounds in terms of the way in which they are produced. An understanding28 SOUNDS sole palate (velum nasal eavity cof how speect sounds are produced is a prerequisite for being able to tsanseribe them, so our introduction of the various symbols employee! in the [PA will be accompanied by an account of the mechanisms of speech praduetion Any sound is a series of vibrations moving thr fur, water oF some other material. To ereate these vibrations, asound souree is need snd these come in various types. On a guitar, for instance, the sound source is the strings, which vibrate when plucked. By themselves, these produce relatively litle noise, but the body of the instrument is basieally a wooden box which amplities the sounds by ip their vibrations and resonating. that is. vibrating in the same way, but more loudly. ITyou strum more than one string ona guitar, the patter of resonance becomes very complex, with several sets of vibrations resonating at once. Speech sounds are produced in basically the same way. with bands ¢ vocal cords ar vocal folds coresponding tothe guitar str in the larynx or voice box. a structure in the throat (see figure 4). When air is forced aut of the lungs, it causes the vocal cords fo vibrate. Comesponading to the body of a guitar and functioning as a resonating chamber is the mouth and cavity above the laryns. Taken together. all these structures are called the tract. The major dilference bobween «guitar and the vocal tract is that we ean mae different sounds by chaning the shape of the latter. by moving the tongue: issue called the s, These are situated the fips and even the larynx Consonants Given the apparatus described above, there are several ways of produ cing speech sounds, Fustly, we can simply set the vocal cords vibrating and rmasntain a steady sound such as Sass’ or Sooooh.* Or we an produce a very short-lived explosive sound sueh asp" or and another important type of sound is illustrated by “P ors’, when we farce air through a namoyy opening to cause aSounds and suprasegmentals 2 Table | JP4 transcription for the English consonants pay TH fa Ti in I) vie I 18) tH I : f (a) rl mm 0) re 5 re Ig) In| her i yo in fon rH fo fl you bl hang Tn] hissing sound, Sounds such as ‘p's "0, aid *s? are called consonants, while those Tike Sawa? ar Soonol? are vowels. The basie list (or inventory) of con sonants in English is wiven in table 1 Lnall eases except for {nin dug and [5] pleasure, the consonant isa the beginning of the accompanying word —[n] and [5] do not occur wortinitilly in English, As will be apparent, in many eases the TPA symbol, written between square brackets, is identical to the ordinary printed symbol. ‘The reasons for laying out the table in this manner will become clear from the subsequent discussion, Lot’ begin by considering the sounds [p] and {f] ‘These differ from each oser in their manner of articutation. The [p) sound is produced in taree phases. Firstly ‘we shut off the vocal tract complotely by’ closing the lips. Then, we try to Foree air out of the lungs, However, this air is prevented from escapmy hecause of the closure and this causes a build up of pressure inside the mouth, ‘Then, we suddenly open the lips releasing this pressure, and the result isan explosive sound that lasts for a very shoot time, Such sounds are called plosives. und the English plosives are [p bed ‘The production of [1] is quite diferent Here we allow a sal wap between the top teeth and the bottom bp and then foree air through this gap. When air at high pressure is foreed through a nacrow opening, itsets up fiition which causes a noise Sounds proxtuced im this way are therefore called frieatives. The English fricatives ate [FV O087.[ 5h], The initial consonants of chuirsand jd which begin as plosives and end as iricatives. They ane known asaffrieat esznd the IPA symbols [tf] sid [eb] snake their comples el clear The remaining sounds im table 1 fall anto two groups. Filly, consider the sounds [mv an]. These are produced by allowing the nasal cavity to resonate Nonnally, the nasal passages are separated from the mouth und throat by a small somiplex sound, 20 SOUNDS piece of flesh, the velum (also sometimes called the soft pate), which is the backward continuation of the roof ofthe mouth (sce igure 4). When the velun is lowered, air an puss through the nose. For instance, we close the bps as if te produce a [b] and then lower te velum, the air fiom the lungs will no Toner be trapped but will passthrough the nose and sot up vibrations there. This is how (im) is produced. and sounds such as [m np] are called masals. The othe: remaining group of sounds is [1 j] and we shall describe how they are produced after we have looked at the other sounds in more deta Consonants are distinguished by more than just heir manner of acticulation, The sounds represented by [p tk] are all plosives, but these symbols represent different sounds. ‘To understand the relevant distinctions here, we need 10 kno somethin about the intemal shape of the vocal tract, and figure $ contains a cross-sectional view showing the way in which [mi] is produced for [p.b]. the yelum would be raised. Phe three sounds [p. b. m] are all formed by bringing the lips together. and they are referred to as bilabial sounds. By contrast, th (td n) are made by plac upper teat this rid Diab etosure Figure 6 Conse-seeton ofthe vaSounds and suprasegmentals Figure 7 Cossssection of the va jes (e.g. French, Spanish, Russian) use sounds which are sibtly different to the [td n) We find in English, Speakers ofthese languages place the tip af the tongue against the upper tet themselves rather than the alveolar ridge and this produces adentalsound, IP we need to distinguish dentals from alveolar, we ean tse special IPA symbols [ql] to refer to the dentals. Different azain ave [k #9] To produce these, we use « diflerent pact of the tongue, the body or dorsum, hich is brought against the velum as ilustrated for (nin figure P. These sounds are known as velars and the descriptions we have introduced here give us the place of articulation of the sound A pce of articulation usually involves tov types of articulator: One isa passive structure such as the alveolar ridge or the tevth; the oer is the aetive articulator which is moved, For the alveotar, dental and velar sounds deseribed above, the active articulator is part of the tongue, For bilabial sounds, we have an odd situation in which both lips can he regarded as simultancous'y the wetive and passive aticulators So far, m our discussion of place of articulation, we have mentioned only plosives, Turning now to ficaives, [8 ) have the sime place of articulation as {tals thus, [s] isan aly colar fricative, whereas [f] isan alveolar plosive, The sounds, [00] are made by bringing the blade of Ue tongue wguins he upper tech or even between the teeth (so thatthe tongue tip prot slighty). These sounds are therefore dental thous they are sometimes aso calledinterdentals (figure 8). As already noted, the production off] (and [y} involves moving the lower ip into close proximity with the upper teeth. These are therefore known as labiodental sounds igure 9) Before considering [f 3], lets bret wwe setaside above, The production of this sound invol towards the roof of the mouth (although not far enough to peoduce friction, see below), The zo of the mouth is eallod the palate (sometimes hard palate). and for this reason [js called a palatalsound (figure 10). Now for [Ps]. we bring the tongue blade forward trom the palate hut not as far forward as for an alveolar sound. The place of articulation for [f 3] is midway between the places of stieulation for palatal and alveolar al for this reson [f 5] are roferred to as2 SOUNDS Lbiadental closure Figue 10 Gross section of th rattan the article lato-alvcolar or alveopalatal ricatives. The affricates fF cb] are made in the nie Enghsh fricative with which we have not yet deal, [h], Formation at is made simply by of this sound docs not involve the tongue or kps: raSounds and suprasegmentals i ) Figure 11 Chassseetion of the vocal tact, dlustating the w tewlaian of passing air through the vocal cords. The past ofthe larynx containing the vocal cords is called the glottis, so we often refer to [b] as» glottal fricative, Equally singe it is made in the laryns, we may call it alarymgeal iFcati We ean now retum to [1.1 j). Above, we have noted that while [7] i palatal, its articulation does not involve moving the blade af the tongue sufficiently elose to the hard palate to produce friction, Therefore, it is not a iicative, and it is necessary 1 recognise another manner of articulation, For each of the sounds in the set [Iw j], the distance between the active and passive articu insufficient to cause fretion, and such sounds are referred fo as approximants, Thus, we can rofer to [j) asa palatal approximant, Next, cansider (3), Production of this sound involves bringing the lips together, but again not close enough to cause complete closure or friction: it is @ bilabial approximant, With the tw remaining sounds, there are adltional factors to take into acount, although it remains convenient ta continue to approximants. Take [J] frst This is produced by placing the tongue tip against the alveolar ridge. However Unlike in the case of [], we do not create a complete obstruction: rather, we give the air an escape hatch by allowing il to pass around one side of the tongue, For this reason, [!] is called alateral sound, The [1] sound is produced by curling the tip of the w for sometimes as far back as the hard palate), but frictional airflow, Sounds made by curling the tongue tip in this way are retrofley, In fet, there is considerable variation in te way that" type sounds are pronounced in English fas ia mnany other languages). Thus. in many dialects we ue towards the alveolar ri cain without gett close enough to cause an abstruction orc and is caused to fap rapidly a centre of the mouth, Traditionally, the sounds (11) are often referred to as liquids, ied glides. We will see an interesting connection between ainst it several times by air passing with fw 5] being ca There is one final distinction we need before our description off English con sonants is complete, We need to understand what distinguishes [p] from [b]. t]SOUNDS from [a], [5] fom (71 [0] from [a], ete. Taking [p] and [b], we have seen that both of these ate bilabial plosives, but they aze different sound, So, what is dhe nate ofthe difference between them? Theianswer to this question ss most easily grasped fora pair of fhiatiyes suc as 2]. Try saying these sounds one afer the other and you will notice that dhe difference between tem is that for [s] dhe voeal cords ate not vibrating (Ue of if you pur your fingers in yous eas), In other words, [s] doesn’t seem to require any sound Source. This may seem rather odd until we realise that, a. ticative, fs] produces its own ftitional nose, To produce [2J. however, vocal cord vibration is also necessary. Ths gives tse to a dilferene invoicing, with sounds such as [by 07] beingvoieed winle [p Vs] areunvoiced, All the English nasals and approximants are normally voice. The three ateibutes of voicing, pave of articulation anal manner of setcultion ee-term deseription for many sounds. Thus, tb] is a voieed palatal colar ulecate, [] i a voiceless labiodental featve, [isa voiced ‘clar nasal and soon, However, for(i], wensed a slighly more detailed description {I} isa voiced alveorfateral approsimant and [isa voiced alveolar non-lateral or retroflex approximant All hese sounds anda number af abers are shown inthe TPA, chat reproduced in appendix. 1 is also convenient fo use more general rms for some groupings of sounds, Thus, dhe bilabial an lbiodcmtal sounes all involve the lips, So Uiese are called kabials ‘The dentals, alveolar. palato-alveolars and palatal all involve the tip ore blade of the rongue tie, the front part of the tongue whieh eseludes the dorsum), These sounds are all earonals, while the sounds that involve the dersisn ane dorsals In addition, otis usolul distinguish the is stro provide a convenient th plosives. aflticates and frcatives, which usually come in voiced'voiccless pairs from the nasals and approximants, which ane intrinsically voiced. ‘The former ane called obstruents (because their production obstruds the airflow) and the latter are called sonorants (because they involve ‘Wine the sounds in table T are standardly regarded as the English consonants, there are a number of other consonantal sounds that are smportant in understanc ing the way English is pronounced. Consider the final sound ofcarwhen the word is spoken in a relaxed and unemphatic manner. In many dialects, this is px sreater degree of resonance) 2, and comes out as a ‘catch’ in the nounced without any intervention ofthe t laryny. This is formed by bringing together the vocal cords, building up pressure behind them as for a plosive and then seleasing the vocal cords. The result is, in fact. plosive but one produced at the glotis, hence its name glottal plosive (or more common. glottal stop) [7]. This sound is a very common replacement to certain oceurrences of [1] in many British dialects, most famously ia London Cockney. where catand /yater would be pronounced [kat] and [ba come to the vowel sounds appearing here short ‘The [t] in words such as brdteris, in fac, subject to fester variation, For instance. im many varieties of American English, it is pronounced a bit like a ‘d?, More precisely. the sound in question is Tite shonter than (dana is produced by very uickly Mapping th tip of the tongue against the alveatar ridge (or the front of the hard palate), Such a sound ss called a flap (or a tap) and its IPA symbol is [7]. we shallSounds and suprasegmentals Finally. we must mention an important aspect of English promunesation that is quite hard o discern, Ifyou listen carefilly tothe pronunciation ofp" in pitand ;pit you should beable to hear thatthe ‘p*afpitis followed by a putYof breath that is absent in spit This pulf of breath is calledaspiration, and you can detect it by holding your hand in front of your mouth as you say the words, The same difference is observed in the °C of tartar andthe *K’ oF carseat and “Kare aspirated in faand earbut not in stand scar. We transcribe aspiration by means of a raised ‘i: [p* € KJ. If we wish to make it clear that a given sound is unaspirated, we use 4 raised equals’ sign, as in [fT IT]. thous no possibility of confusion, itis customary to omit this, Transeriptions for pitand spirineluding this difference in aspiration are thus [pn and [spt]. In transerip- tions, additional symbols such as the raised *h” oF * when there is ‘equals’, auled 10 a basic ‘ymbol to ereate another symbol fora related sound are called deri are good many diacritics used by phonetics (see the IPA chart on p. ALL for aadctional examples) So far, we have restricted our attention to English consonants, but ef course other languages use additional consonantal sounds, In table 2, we see the Engl'sh consonants fram table 1 along with various other IPA symbols for sounds which ‘occur in other a As we can See, itis possible to fll a good many of the cells in table 2 with symbols representing sounds in the world languages. Without spocial training. you won't beable to pronounce many of these sounds. but you should have some idea of how they are produced. For stance, a wetoflex [1] ss made sn the same place as the English rotrafles fi] but wit the lateral manner of articulation characteristic of [I] Retoflex sounds are Found én a large number of k of the Indian subcontinent and in Austrslia amongst other places. Uvular and pharyngeal sounds are made wath places of articulation not found sn English Uvular sounds are ike velars, except that the tongue hody moves further back smel a Title lower to articulate against the uvula, Pharyngeal sounds are common in Arabic (although they are encountered in languages throughout the world). They are made by bri e root hack towards the back ofthe throat, often Fe 2and 3) with constriction of the throat texercis Table 2 Consonantal sounds arranged hy place and manner of artic PLACE Tabios inter palato- MANNER bilabial ental dental alvwolar alveolar pal sal remofles velar avul_pharyn, 4G freatve ] are mid open vowels. Allematively, linguists ooien refer to [¢ ] as tense (mid) vowels and fe 3) as lax (mid) vowels, Thus, we ccan contrast the set of tense vowels [3 ew 0] with the set of lax vowels [10 € 3] We can represent the position of these new vowels inthe quadrilateral m figure 14 note that we do not represent vowel length in this gudntatera Where a vowel consists of two components, as in the examples in (19). st is called a diphthong (i:0m the Greek meaning ‘two sounds”). The single, pure vowels in {17} and (18) are then called monophthongs. Some varieties of English are particularly rich in diphthongs, and diphthongs are also very common in totally unrelated languages such as Cambodian and Estonian, However, some languages lack true diphthongs altogether (c.g. Russian, Hungarian, Japanese)Sounds and suprasegmentals Fisury 1S Finally, we come w another sot of English diphthongs, mainly found in non-thotie dialects. They are illustrated by the words in (20): 20) peer poor [no] us] For many spe that in (18) we have regarded mare as contaising a pure vowel transcribed [pea] or [pes] nd [mea] ar [me], respectively. In ig shown the “trajectory? involved in the formation of eaeh of the diphthongs we have introduced, swords such as pearspairand mare would belong here— note an would be re 1S we have The description of vowels we have offered so far is sufficient for many varieties of English. However, some dialects use different vowel sounds, For instance, in conservative RP. you might hear go pronounced as [go and for many US speakers some of these diphthongs are lonz monophthongs (e.x. day [ded should be noted that kp rounding, which was observed above as a feature of the English vowels [t 0] isalso a characteristic af [uo]. The vowel quadsilaterals we have examined do not explicitly indicate whether a vowel is accompanied by rounding or not lish vowel sounds worth There is one final feature of the transeription of Ei mentioning here. As already observed, unlike many varieties of British English, most dialects of American English have vowels with an *T-colouring to them, as in fired. ears card, more.air.snunder:Ihis produced by retracting the tongue as if proxluce the sound [1] as in rt during the production of the vowel sound. Where vater accuracy isn’t essential, st is offen transcribed by just adding [>] afier the vowel. es. murder [anordor). However, where we nved more precise transcrip tions, we use special symbols such as [> €°}. Thus, we ean transeribe murder as [mad] and air as for}. The little hook on [>] and fe] can be thought af as a iaerive We conclude this survey of baste sounds by briefly look which do not occur i standard varieties of English, Focus! there isa strong tendency in the warld’s la low to be rounded and for font vowels and low vowels to be unrounded, at vowel sound 1 on lip roundin, ss for back vowels whieh are not However, we do find vowels which are exceptions o this tendency, and some of the amore common correspondences are showa in 2 040 SOUNDS en tiont unsound 1 y € e 2 A D 0 ‘Thus, [y ¥ ex] sound Hike [it © €], except that in producing them, the lips are rounded, On the ather hand, the sounds [tt ¥] correspond to [uo] but are produced with spread lips With two exceptions, all the vowels discussed so far have been placed elose to therig dye of the vowel quadrilateral, and generally with a litle practice ‘we can feel confident about locating such vowels. However, we observed that the sound schwa [a] and the vowel [a] eceupy a central position on the front‘back axis, and vowels such as these are generally less easy tobe sure about. From this, it does not follow that such vowels do not exist, and a number of central vowels are shown in figure 16:long with the rounded and unrounded vowels From The four new vowels in figure 16 [i #3 ware all unrounded except for (x) central high rounded vowel. Finally it should he noted that the ‘1°-coouring of Amer above is not the only sort af colouring that that vowels offen receive is masalisat 2a vowels mentioned coloucin n. This isthe result of allowing ais to pass through the nasal passage, as though for nasal consonant suck as [1], while sill the air flow through the mouth, A nasal vowel is indicated by a diveritic symbol placed over the vow J- In languages such as French, Polish Yoruba (one of the main fangu: igeria) and many others, nasal vowels ply 2 important sole. Here are some words of Yoruba in transcription onal sowels nasal vowels fka]—‘wbeplaced on? — [ki] “to touch? fu] “to reaine {Ki} top paint fil ‘aut [s]twascompa font contial back Figure |Sounds and suprasegmentals Nasal vowels are also heard in many varieties of English. A typical pronunciation of the word can in American English especially, is, in fact, (kita), with the sequence [2en] being replaced by a Tong nasalised vowel texercises 4, Sand 6}, Suprasegmentals So far in this section, we have examined segments, that is individeal sounds and their pronunciations, However, pronunciation involves far more than just stringing together individual sounds, We shall now examine the level of ‘organisation thal exists above the level of the segment, thesuprasegmental level, All words eam be divided into one or more syllables, Although most of us ean syllables (ncluding somal children, see section it i rather {Eve a strict definition of the term, One way of determining the numb: of syllables in a word is to tr singing it each spllabe is sting on a separate note (hough not necessarily on a different pitch, of course). We shall be considering the structure of syllables m detail in section & here we will just consider their basi shape \ syllable typivally contains a consonant or set of consonants followed by a vowel followed hy another consonant or set of eansonanis, eg. eat [et] or springs [spltiy]. A sirmg of more than one consonant such as [spe] oF fv] is called a cluster (or, more precise # consonant cluster}, However, either set of consonants may be missing from a syllable as, For example, in spray [spret] (nc final consonant), iayps [tps] (no inital consonants) orere[a] (na consonants at all), Words with one syllable typriagss cat) are monosyltabie, while words with more than one syliahle are polysyllabie. From this, we might conclude that the only obligatory part ofa syllable isthe vowel, but this isnot quite correct, What a lable must have is a nueleus or peak, snd characteristically this is a vowel However. in restricted eases, itis possible for the nucleus of a syllable to be a onunt, For instance, the word table is disyllabie (nas ovo syllables), contain ing the syllables fit] and [hj]. These is no vowel in the second syllable, and its nucleus is the consonant [J] asyllabie consonant In transcription we sepresent a sllabie consonant by a mark placed beneath it, In English {in n) eam also be [bot] and fueron [hate is sometimes usefil to mark the division het een syllables in transcription. This is done by placing a dot between ie [pO ata Ik Nest, we consider the devices involving changes in loudness or the patel of sounds tht syllabic, as in spllables, ¢ suse to convey meaning, These are stress, tone and intona called prosodie phenomena, We b, I we compare the words transport in means of transport and to transport goods, we ean hear an important difference in pronunciation, In sreany of tras portthe frst sylable.t70 tion, which collectively an with stress, cater emphasis than the seoond,-sport, while in to transport goods i’ the second syllable which gets the gecater emphasis. This «emphasis is called stress. and we say that in means of TRANspart the fist syllable al2 SOUNDS bears stress, while in to tronSPORT the second syllable ss stressed. ‘The other syllable remains unstressed. Physically. a stressed spllable tends to be Louder and oofien a fitle longer than an unstressed one. In the offical IPA system, stoss is indicated by means of the sign ' placed before the stressed syllable: [traaspost] TRANsport (noun) v. [ran!spox] teaxSPORT (verb). However, anasy Hnguists prefer to indicate main stress by means of an acut [tedinspo2t] (noun) v. [tsp] (verb) Some syllables have a degree of stress intermed siress. Consider the word phofographie. The main stress falls on the thied syllable in [fou.ta.g0a.ftk]. The second and fourth syllables are unstressed. However, the first syllable has some stress, though not as much as the third. This is called ondary stress. In IPA il is transcribed with the mark , » [ fouto'graftk]. An sto indicate secondary stress by a grave accent placed over the vowel rit) The type f stress which distin ent over the stressed vowel e between full stress and no guises words sul as ['teansport] Rom [tas' pot] is known astexieal stress or word stress, There is another type of stress in which certain words within phrases are given more emphasis than others, Consider 23) ey Town builds houses Jn a neutral pronunciation, each word receives an even amount of emphasis though slightly more falls on the stressed syllable of fuses: Tare fui DU Ses. This is a natural answer to a question such as “What does Tom do? or "What does Tom build!” However, if we put more emphasis on huildsto get Tam BUILDS houses, then this can only be a natural answer to a question like What does Tom do with houses”. or more likely a correction to someane who thinks that Tom repairs houses or sells them, Fully, in TOM bualds houses we hawe a reply to the question “Who builds houses?” This, phrasal stress. (Many Inguists also ref iype of stress is offen called this musta’t be » as accent, thous confused with the term “accent” meaning the particular type of pronunciation associated with given dialect) It can ofien be important in disambiguating sentences which are ambiguous in the purely written form, Turning to our second prosodic plenomenon, the piteh of the voice is very important i la and all Languages make use of it for some purpose. [n some languages different words are distinguished fiom each other by means of piteh Here are some more Yoruba words ey high tone mid tone low tone te that which” pr fi opus? fe istic. fe to do" Fe woften dks hoe” ko husband eS camon ‘The word:with the mark ‘ aver the vowel is pronounced ata higher piteh than the word 4 whieh in tum is pronounced at u higher pitch than 4, These different patches are called tones We say that fits high tone, 7 has mid tone and 4 has low tone. Notice that one of the systems for wansersbing siress uses te sameSounds and suprasegmentals symbols for primary and secondary stress as ure used here for high an low tone. In most eases, this doesn’t cause aay confusion, though languages do esist which hhave bot independent tone and independent stress. In such eases, we can use the IPA symbols for stress and use the grave and acute accents for tone. Some languages distinguish only 10 levels of tone, while others distinguish up to fourlevels, When a language distinguishes words from each other using this way we say that it hastevieal tome ‘The words stvari “things? and shri ‘Gin) a thing” in Serhian-Croatian are distinguished by tone, though in a different way from the Yoruba examples we hhave just described. In the word meaning “things” the pitch falls from higth to low during the course of the vowel fa}. while in the ward meaning tin) a thing’ the pitch rises from low to high on that vowel. Tones of this sort, where the pitch changes dur the tones of Yoruba, which are called fevel ones. some lan complex contour tones in which the tone first rises then falls or vice ve the course of the sy lable are called contour tones. as apposed t ct more a. The classic example is Mandarin Chinese. In (25) we see four words which are distinguished solely by their tones, with the broken Fines indicating pitch and the unbroken lines boing reference pitches (the words appear in the standard Paiyin canseription, the olliial romanisation of the language in the Peoplc’s Republi of China, and correspond to IPA [i]: ( won cain tone o on = hugh evel yi ‘already fall ise vl ideal tains Both level tones and contour tones qualily 4 language as having lexical tone i.e. as being atone language. English is nota tone language, but, like all spoken The uses of piteh with which we are familiar ish are uses of our final prosodic phenomenon, intonation. Consider the iL uses pitch extensivel instances af the word me in (26), where the pit is represented graphically 26) abe J . VIV=-V44 SOUNDS The pitches applied to these words are very similar to the contour tones of lan: Jike Chinese. Howeves, in E donot produce completely different meaning speaker, but by changing the ‘tone’ over the word the speaker chanwes the avtitde heorshe is expressing, Thus, we move from a simple statement (26a) to a question 260). toa strong assertion (26c). toa matter of fact assertion (26d) and in 26e). to an expression of disbelief. Unlike in Chinese, however, these tones cannot be ded as an inherent past of a single word, IPLie utterance consists of more than one syllable, as in (27), then we find the tone is spread over the whole of that utterance and gives rise to the same range of altitudes as we savy in (26) lish such cha = each af (26a) to ¢ to the 4 a b. e en eal peas gut peas’? eat VJ VT - Vv As observed ulreidy, all spoken languages mae use of intonation (including those like Chinese, Serbo-Croatian or Yoruba that have lexical tone), thou lille greatly fom ane langue to snotherand from one dialect to another Knowing intonation pattems isan important though ofien neglected part of speaking a fore language, and many intonation patterns which sound polite in one language or disle sound rude or funny in another. Wis said thst die British regard Americans as ruce-and pushy sn part because neutral, polite Ameean intonation sounds peremptory 10 a British speaker. while Americans often fool tt Britons are osenweening or avin because what is neutral for British intonation sounds over the top to the Anne? This section has provided a basie description af the sounds of lan next section we'll see how different varieties afone and thesame lang distinguished by the types of sousds they use and the ways in whch they use thet (exercises 7.8 and 9). se Inthe canbe Exercises 1 Using the IPA chart, giveu phonetic characterisation of the following consonants (a)3.(b) te bd) Model answer for (12) acs As regards manner, [3] is fticative ay its production does not involve complete closure af the voeal tract, but the articulators da come closely enough together to produce friction. Ia terms af its place of saticulation, this consonant requires the blade of the tongue to approach an area betweon the hard palate and the alveolar ride — it iva palato-alveolar. As a palato-alveolar fricative, is paired with [f] but whereas the latter is unvoiced, [5] is voiced, Thus, it isa voiced6. Sounds and suprasegmentals Iato-alveolar fricative. I occurs in English in such words as fein ure and some pronunciations of ganige Using the IPA chart, give a description of the fallowing sounds, n (a) yb) (eth. (DZ, (0% t 2)N. (hye. fi Give the IPA symbol for euch of the following consonants: (a) voiced uvular nasal stop (b) alveolar implosive stop (6) voiced reteofle lateral approximant (4) voiceless palatal alfricate (¢) voiced lubiodental nasal stop Give a phonetic characterisation of the following vowels: (a) 1,1) 3, fe} 0, (dD TL as fi) However, the distinction between these twa vowels is not usual Model Answer for (4 In terms of height, [fs « high vowel, although not as hig described in terms of height, Instead, [i] is charuetersed as tense. whereas [it] is lax. This distinetion is analogous to that between [a] (lense) and [0], but [i] and {t] are front vowels, whereas [u] and fo] are back vowels. Furthermore, [i] and ff] common with most front vowels, are unrounded, whereas fu} and [0] are rounded. Finally, ft] ian oral vowe! and does not exhibit nasalisation, Thus, we have the conclusion that [1] ih such English words as bid and pit 1 ont unrounded las onal vowel. It occurs in Using the IPA chart, give a description of the following vowels: (ayer, HY, (e) A, 9.) PDA Le) ytd HE. GE COI Give the IPA symbol for each of the following vowels (a) high tense buck rounded (by open (lax) mid front rounded () central mid unrounded (€) central low unrounded (¢) igh tense front rounded (0 iat lax back rounded This is a (ext in IPA transcription of a short passage as it would he spaken by a sneaker with a British accent, Rewrite this in ordinary onthography46 SOUNDS danas ind and do san wa disajuetay al) ay Ao suNgEA weND ALD ker aloo aupt m oman Souk, dst aaaizd dot dome hur Sst soksixdad ra incikin dora tetk It Klouk of fla br kanstdsd snqgs dan OF a. den o> nog) tod blu: sy hod 9y hit hex, ba do mos br lus domo: klouslt ki O> trav fould hy Slous oatund ham ond okt dona zor pds nema den do sin fw aut wal, sod anotlnd tkaylo Uk of hy Klouk, ond Sou Sonn and worablaicgd tokens doc 98 m Way dostieNzA! AO TU. The following is 4 text transcribed as it might be read by a British speaker and an American speaker. Rewrite the text in orthography and then comment on the differences inthe two accents British version jouhan sabastton botk skstizn ent faty to Sevptizn HT Kewn fiom 9 alt lull piscegnst oud patna prow Fa) eajuzefgy bevks aeepue war tmens Kip anolt oe oy meids> ena mk] sceuicw 96 hae (919 eambo mpsks sikestil swants p kyffedouz pissy 1> bo:psikod ql sian ond on nsamos omaunt ov kacial muzik t> O9 uxt! diz ounlt temp 96 WOK hi didgl Kompour wo7 HpIO dou sum eA! Sc1OAL hI MMpeseSK SCH 96 DOSE adhus pag 7m fakt won ay e9 gterust ma. sia2s97 av 92hopa.ank nat American version joubaen sobsestion Fak stkstin ete at 9 Sevptiy #1 keto os feel aul pecust oso rary prot fy mua fz he bs expt soM MENS riot 3:19 meds mem SiMIoW Oy hNEIS — GeImho MBS aKESTOL 28 wkryferanr pis #9 hopsikA gd 9499 dnd! 30 ragMos ara? av Kora mugik 00 tbo Ox ounl rags a kampavif fr dt rat apa, ou sam wee Se1doe hz mpcgesttkseramay 8999 mah wef 7 to et wana Oo grerrast mastopBSS7o¥ oH apoaRrik Keath Transeribe the text below into IPA following your native accent as closely as you can, indicating jexical stress on polysyllabic items. Note that in some eases there m pronouncing nt be several alternative ways of given sound or sound sequence. For some, Britain and the United States are two counties divided by a common language, and the same could be said of other places where English is spoken, such as Canada, Australia, New Zealand South Africa, Nonetheless, on the whole English speakers tend communicate with each other somehow. Nor should we jump t6 the conclusion that i just across ational boundaries that accent and dialect differences occur. The differences in the speech of Americans from New England and those fiom the Deep South ean be at Teast as seat asthe differences between New Englanders and British speakers, for between Australians and New Zealanders3 Sound variation In our main introduction, we observed that language varies across both time and space. If we compare the English spoken in the cities of Perth, Pittsburgh, Post Elizabeth and Plymouth, we can point not only to differences between these four cities, but also 10 historical differences which dist tod is these varieties fom those spoken in these locations 150 years ago. This important study of historical and go iphical variation bas been a preoccupation of linguists for well over a century now, and continues to be a sttong focus of research in dialectology and historical linguistics. [tis only in recent times, however, that ate linguistic variation with communities. The ni from thal spoken in Montreal, but within these cities? Does everyone in Montreal linguists have begun to invest French spoken in Marseilles may be difle ‘what about the use of languag speak an identical variety of French? Clearly no, we might suppose, but it was not unt Une 1960 that Knnguists began to take this view seriously aml study vacation within vig Tn this section we will examine pho language that alts those features whieh have been introduced in the previous section: sounds, syllables, stressand intonation, Becausear the natureat existing s, towns and cities gical variation — the variability: in research, our discussion will be concerned exclusively with sounds, Linguistic variables and sociological variables that itis the enistence within the sp le community of more than one possible scalisation (or variant) of a particular sound, \ simple introductory Table 3 shows how offen diferent social class groups in Bradford use the twe different possibilities [b] or © (Le. nothing). We can see clearly im this table that there are class diffe (h] the class difference is interesting, but more important is the fact that everybody in nces in the use of someone's social lass, the more likely they are to use [fh]. This this Bradford research used hori forms at least some of the time. Even the lower- working-class speakers occasionally used [b] and the middle middle-class speak crs sometimes omitted it, The variation within this community, then, is relativea8 SOUNDS of occurrences lower working middle working upper working class lower middle class middle middleclass 2 Different groups use different proportionyof the two variants, and this is typical of litle format oft 1, situations where on, variation, bso uses particul that form, occur Tess frequently texercise Ph. roup within the community fime in contrast with other groups which never use In orderto desenbe this quantitative variation, linguists have devised the notion of thelinguistic va people’ use of different variants, A varable is a Imguistic unit whieh has two a ‘more variants that are used! in different proportions either by different sections of jable an analytical construct which enables them to contssst the community or in different linguistic or contextual etreumstances, Variables can cal factors, the topic of this section, and also with ‘word structure, word meaning and syntax, For the example above. we say that the variable () —variables ane normally put in sound brackets — has two variants [fh] anid O, the use of which relates to a person's social class ‘The procedure for analysing the use of variable ina particular community isas be concerned with phonolog follows 1 Recordings are made of conversational speech from people belong to different groups in the community 2 Rescarchien listen to these recordings, noting down the pronunciation of representative number of instances of each variable. Normally they analyse at least thirty examples of each variable for eacia person they record. 3 Exch person's relative use af the different variants is calculated. The results of this are often presented as percentages, showing that a particular speaker used v's of one variant and of the other 4 Itis thon possible to amalgamate these results to produce group scores for example. the researcher m: corking-class speakers and compare this ores of middle-class speakers. or an ave for middle calculate am average of the seores gure with the formiddle-aged ced women, ‘men to compare wth the aver We have seen for the example of (h) in Bradford that there appears 10 be a use, Suel a relationship hhas been 140 relationship between social class and langusy found in many westernised speech communities around the world from Chis ‘Copenhagen, from Brisbane to Berlin, Qutside westem societies. however, the notionSound variation , ° has used education level as a means to measure socio ccortelating language use to s rence oP wowed assim onomie divisions when al structure, An example is provided by the oscur- wy Furst (Persian) speakers in Tehran, Assimilation was, briefly mentioned in the introduction (pp. 4 S),and we can ilustrate is role in Tehran Farsi using the Farsi verb meaning ‘de’, The Standard pronunciation of this verb is {[hekon}, but the vowel in the First syllable may assimilate to the second vowel giving the variant [bokon], Pigure 17 shows that the higher the speakers thetess likely they are to assimilate vow Whether we rely on social class or education, what appears common to all societies is that saefal srucnire is reflected fn linguistic structure in some way We should expect, therefore, that, besides the socioeconomic characteristics of speakers, other social factors will also affect and structure linguistic variability This certainly appears tobe the ducational achievement of we if weconsider theenderof tie speaker. The suage variation and speaker vender is probably the most extensively studied in sociolinguistic research, Ong of the consistent findings is that, all other things be relationship between la equal, women use proportionately more standard Yariants than mien for linguistic variables not undergoing change ples can be found from many very different societies around the world and an illustration, based on the work of Peter Trudssll, appears in figure IS, where we can soe that women in each socal cass group are using more ofthe standard variants [19] as opposed to the non-standard [on]in the British ety of Nonwich (exercise 2 The ethnic group to winch « speaker belongs has also been found tha effect on language variation, In the data trom Wellington, presented in figure 19 and based on the work of Janet Holmes, the ethnic (Maori or Pakeha, i.e. White European) identity of New Zealanders is seen to be relevant to the use of range of different phonological variables 2so SOUNDS SE i ° Wide Mise” ow er Me” pee Waking” We Wong © Class s jal Class of speaker Figune 18 tions of ing) ane speaker sex and sek 35: se of ull onels atteai sZeaarder Figine aviation in New Zealand Eaglish (bused on Holmes 19977 ' The devoieing af to [s} so that “was becomes [wns] instead of fn] . ‘The deaspiration of woxd-initial 1, so that “tip” becomes [ap] instead ‘of [t'2p] (note that ‘tin New Zealand English is pronounced [2]) ' The use of full vowels in unstressed syllables, so that ‘run to school” becomes [ran tur skutl] instead of [1n t9 sku Here, for each variable, itis the indigenous Polynesian Maori community that uses more of the non-standard variants,Sound variation Table 4 (i ab 0 Percentage use of local Percentage use of local Bolfast variant of (th) Belfast variant of (x) Hannah D Paul (th)= deletion of [9] between vowels as in eg. mother G9) =e of [x] in words sth as pal taal four A final example of how social structure has been shown to determine a person's suistic behaviour is of a di mentioned above, Linguists fave established that the quantity and mature of person's social network links within the:r community may be ta important factor im such behaviour, Lesley and James Milroy. who camed out sociolinguiste research in the Norther Irish city of Belfast, measured network strength alon rent nature fbom the speaker-delined categories ‘ovo dimensions: firstly, they assossed the extent to which people had elose social ties with family, Bends and workmates in the neighboushood, and secondly. they looked at the extent to which these ties were multi-functional, ¢.8. ifatie to another network meanber was hased on both friendship and employment or both employ ment and Kinship. as opposed 10 just one of these. People who had many ‘multi-functional social ties were considered to have stong social networks and people who didn't were labelled as hay ing weak networks. [twas hypothesised that strong social networks would aet as nom-enforeing mechanisms, subtly putting pressure on their members to conform to normal local beliour, includin linguistic bebavious. A number of variables which showed an intimate connection hotwoen a person's network strength and their use of local Belfast variants were discovered, and a small sample of the results of this research appears in table 4 This table compares the use of two salient Imguistic variables (th) and (4) by Paul and Hannah, two residents of Belfast. They are both in unskilled jobs, have husbands with unskilled jobs and have a limited educational achievement. Yet their linguistic behas iour is radically different and the explanation for this appears to come from the differing strengths of their social networks, Paula isa member of a strong social network in Bellust she has @ large family liv frequently visits her neighbours, many of whem she works with, and she belongs wa local bi locally, she 10 playing elub. Hanna, however, has fewer loeal es. She has no ‘ups and works with snemors inthe locality, isn'ta member ofany local who do not live in her neighhoutbood cntly, rather than accepting the broad sociolog ernicty and class as universal and sven, soeiohnguists have oupings are actually created at the loca! level and examining the relationsiip between these se ups and ln variability. Linguists such as Penslope Eckert, Miriam Meyerhofi’ and Mary Bucholtz have explored the way in which people actively come together to form example, gend bboen looking at how social2 SOUNDS groups that engaye sn common goal or interest and that, over time, develop practices, including linguistic practices, that are shared and recognised as char acteristic of that group. They label such important advance here lies inthe fact that co maintained and adapted by the very people who exeated them in the fist place. kn oups ‘communities of practice®, The mnmunities of practice are developed. this respect, they differ markedly from the groups studied ia ‘traditinnal” socio~ linguistics. which comprise collections of unattached individuals who happea to slate a certain social characteristic, such as being male, or Asian or middle class A well-known example from the United States demonstrates how suc ‘com munities of practie” develop variable linguistic behaviours that help to doling the group. Penelope Eckert spent several years observi igh School. She observed where different groups con during breaktimes, how they walked, the width of their ans, how much they smoked, were they ate, where they hung out and what they did after school. and, later, how they spoke, In this way. she was able to draw at hig oth enaigers ina Dettait srewated around the school shly det fed picture rou! two pol J in the school and how these clices, There Burnouts — and a large, less eleurly po! oupings that naturally em their everyday ps ced, between? group. Jocks were more likely to buy into the ethos of the schoo! as a stepping-stone into higher education and participate in many of the extracurricular activities which centre around the school, such as sports, the school newspaper, cheerleading and the seloo! council. Burnouts, on the other hand, were much less lik sought to place upon them. Given that they aimed for local vocational employ ment, they did not feel that the schoo! offered them the sort of training and ly wo accept the “corporate culture” ofthe school and resented the restrictions i idance that would help them and so felt less inclined to participate in the extensive extracurricular activities which were dominated by Jocks. The social ‘world of the Bumouts beyond schoo! hours was directed towards the employment and entertainment offered by the local urban neighbourhood Intriguingly, Fekert found that these two polarised zroupings also spoke differently: ‘The difference inthe linguistic behaviour ofthe Jocks and Burnouts is demonstrated by the way they pronounced iv the vowel in “cup! anxl‘cut’), Eckert highlighted ome tendeney in her data for ia’ to be pronounced near the back of the mouth (ith realisations such as [3] or (0). Figuse 20shows her results for i! backing Clearly x backing is characteristic of Bamouts, AS noted above, Fekert® work is important because she demonstrated dhe power of observin ind sele-definin sell-formi groups of people, sather dhan simply assigning people to w egories and observing variation within them fexercise 3) -snowa global social In summary, we number ofsociol ave painted a picture of an intimate relationship between a sal variables social class, educational achievement, sender cibnicity, social network and community of practice — and a range of linguistic variables, 1 seems quite clear that our position in society ean shape certain aspects of our linguistic behaviour, Linguistic variability is not divorced trom social conditioning. ‘We now tum 10 a diflerent type of variationSound variation ur nout gis ur out boy oc gis boys Se and social categor y Figure 20 Degree af hacking of? / anne, students ata Detroit high schoo hased on Fekert 200 HS) Adapted aad yop Hs permission af Wiley lacked! Publishing Stylistic variation We are all prohably conscious that we speak differently to a teacher than to our friends over a coffee, We tend on the whole to speak using a more standard dialect with the teacher. and use more non standard or informal language when having a chat, Similarly, we may find that we speak in a more standard wa uisties— than when di when discussing some topies say. polities or lin thers yesterday's baseball game, or your neighbour’ latest anties, Linguistic variability that is dependent on the social context we find ourselves in ar the topic ofthe conversation weare engaged in is usually termed stylistic variation. Atlin Bell a linguist from New Zealand, developed a model for the analysis of stykstic variation known as audience design. He claimed that sn designing our style of speech at any particular time, we assess the sociolinguistic characteristics of our addressees and adapt the way we speak to conform to these characteristics, Let’s loo’ at an example. Nik Coupland investizated the extent to which an eoney in Cardiff, Wales, shilied her speech to match that of the sovial class of her clients. One ofthe vanables he studied was the Napping of) i.e the use of [bra] instead of [hata andthe results ofthis part af his study appear in gure 21. Those results show how the assistant altered her use of this variable quite radically m The model of audience design helps to explain why people seem, to a none stay an A British or Noth assistant sna travel rent social classes (exercise 4). when speaking to clients of native ee, 10 ‘pick up? the aezent of places th American English speaker spending a couple of years in Australia would have a predominantly: Australian Ei and would accommodate 10 that yartety so often when conversing thal, to non-Austraians, they may ‘Sounda SOUNDS a per Miele Figure21 like an Aussie’. What this indicates, then, is that variation in language is com strained not only by the social characteristies of the speaker, but also by those of eractionally determined theaddressee in any conversation: Variation is also Linguistically determined variation We would be wrong to go on from the above to claim hat i ison social factors that determine the stueture of variation within a speceh community Linguistic factors, too, play a considerable role in determining the relative use of itfetent variants of a variable. One variable which appears to beliave in a similar speaking worlds so-called consonant cluster deletion or more specifically --d deletion. This involves the variable deletion of word-final [t] or [d] when it follows another consonant, So we find examples such as those in way nerass the Fs (28), where the candidate for deletion appears in bold and the phonetie transcrip. tions give variant pronunciations depending on whether [1] or [d] delete: es Data set best fren » flsest fen] [les fiend] cold weathe + Pacold weds] - [cul we5] Data set Jie stuf she wrkey —> [hi sist do whit] [la st 09 tks] she seemed funy —> [fissiamd P nit]—Lfitsian fini) Data set} most of th = [imoxat ov 09 farm] mous 9.9 tan] ground aitack > [graund atk] — [gran 20] Data set he seemedndd ——> [it stot od] —[b stm 0a] she passed test — [fixpost test] [fit pos 9 test]Sound variation (Note that in these examples, we transcribe‘ sounds as [x]. common practice Unless more precision is needed.) As you read this set of data, you will probably Feel that the futier you g thesets theless ikely you would be to hear the secondexample in each phonetically transersbed par, that i the example in whch [tor fd] is deleted. his is because in 1 set of data the word final [t] and [dl] are in different Fnnguistic contexts and iis e whether or not deletion of [) or {d] seems likely is Fand 4 they are followed by vowels. Research hus shown tht deletion is fess likey before vowels than before consonants. In data sots 2 and, [1] snd [d] ane the realisation of the past tense ending -e¢, whereas they don’t have this function insets [and 3, ‘We would expect, based on evidence from many English-speaking communities around the world, to find less deletion in the=ed examples, si Hehe [and [A] are the only indication of the fense of the verb. This means that Hinguistc Factors (oshether the candidate for deletion precedes a vowel or a consonant and seth: it cenicodes past tense or not} predict most deletion i data set and least in data set 4 Table 5 provides evidence trom a number of dialests of English to support this prediction. It is important to note that the ordering predicted on the basis of the linguistic factors isthe same in each of the dialects investigated, despite the fact that there ae quite considerable differences inthe actual faures with the Puerto Rican speakers generally deloting final [t] and [d] much more frequently than speavers of Standard American English, What these dill is that social actors as wel as im ‘The pattem that we see intable 5 Mlustrates what is known as animplicational scale This notion is exemplified in a hypothetical ease in table 6, Here “1 signifies that a particular deletion always takes place and ‘= that st never takes place. Thus, in Dialect A, final [t] and [ai] are always deleted, imwspective of linguistic context and in Dialect D they are always del 0 down these contexts which are afect Indlatasots 1 and 2, t] and [4] are followed by consonants, whereas in rences show, of course suistie factors are playing a part an this variation xd when followed by a encade tense otherwise they are never deleted intermediate hetween A and D. Now, we eam consonant so long as they do not in Dialect D, Dialects B and C a look at table Gand formulate the smplicational statement in (29 Table 5 Deletion of {i} and fal} in English Followed hy a consonant Followed by a vowel lsfetion ims deletion in °% deletion in letiom Language varie hon ] than older: [F] was more likely to be used in the emphatic second reply: and [7] discussion, Labov found that the Table 7 Social, contextual and linguistic variables fiom Labovs New York of in charac Tho shop ansista store (upper-middle-class, lower-m job within stare (floorwalker, tll operator, shelf fille, ete.) floor within store (higher floors sell mote expensive produets Sex cethnivity Idle-clas, working-class) first eply given versus emphatic eply given after Labov had pretended not to heat (1) before a consonant versus (1 in wordefinal position58 SOUNDS epariment store Figure Lav 9 ia Pros old rid young fee Saks dopa was more likely in the word floor than in fourth, Particularly important for ou er. is the fact that discussion of the role of variation in language change, every stage in the advancing change ta [r] could be found in the the assistants, Some used viral wed [1]all the time, but most used i¢ some ofthe time pecch af some of no [F] ab all, others — who were obviously further ahead inthe chang but not on every occasion. The study thus provided Labow with a convenient sitapshot of the progress of this change through the speech af individuals, part: cular groups and the whole New York speceh community tevercise 5)Sound variation ofr fourth T Jang EMA TIC pronunciations Figue sphatic promenciaions of fr] in New York ey 66). Reprinted with permission of th Uni Exer If-we are able to shift our speech so readily, why do you think that poople continue to speak dialects with a low ps Design a small linguistic survey appropriate for your own town, ety or rural area similar to William Lahow’s Department Store research, Which variable would you study and why’? What question(s) could you ask to ensure that you xot a reply that contained your variable? Which groups in your local speech community would you study? Think about the school you went to and how teenagers atthe school formed peer etoups, Were there groupings like the Jocks and Burnouts in Detsoit or dla different system of grouping prevail? What were the characteristics of each group? Did the different groups speak dite eaitly? How? In onder to demonstrate the effects of audience design, alee rocorded in large lectures, small seminars and in one-to-one meetings with students, Four linguistic variables were analysed: (T). examining levels of lt lottalisation: (L), focusing on il) vocalisation; (1D), look ing at whether ‘by was dropped: and (A), investigating whether the a in words such as *bath’ and ‘glass’ was tronter [2g] or backer faz]. The results are displayed in How would you explain the findings? Are they what you would expect?ov SOUNDS 2 [= wocatsaionof 1 = opengl a trong [> ttaisation oft 138 of non-standard phonolog seminar Seting of speeer analy Mountain agrcultre ~ ovland dairy farming Indust GGendet and employment type Figure 24 The promwiciation of fo) in Cteda Spanish by spel (fiom Hol Jonathan Holmquist examined the pronunciation of Spanish (0) én esearch showed that some speak (Ucieda in the Spanish Pyrenees. His crs pronounced this sound as [u] as opposed to the Standard Castilian Spanish [0]. When he examined the occupations af different people im the village and thesr use of (oe found the results in Figure 24, How would you explain the differences in the use of (0) by the workers from different employment types? And why do you think there is a gender difference among the agricultural worke:s and not saniong the industrial workers?4 — Sound change Linguistic change isa process which pervades all human languages. ‘The extent of this change can be so radical that the intelliaiil «can be jeopardised, The language of Shakespeare causes some problems for the ly twenty-first-century reade se not insurmountable, However it we go further back to the writings of Chaucer, we a alien, ess easily recognised form of English, If we observe languay a much smaller timescale, say that of the average life span of a human bein comprehension difficulties such as those conionting the reader of Chaucer do not arise, Languages actually change quite slowly. and hence the ability to commu with all generations of speakers of our own language is maintained. In this section, we will !ook al how the sounds of lan change overtime, bath trom a diaehronie and synchronic perspective, Diacaronic research on sound change has enabled us to chart ehanges that have taken place im earlier historical periods, while synchronic approaches allow us to observe language changes in progress today. In addition, we wall examine sound chang from the perspective of one of the principal problems of angus the transition prob) of Former states af the lan but these od with a much more nicate suecesstl riety change, namely ‘what is the route by which sounds cha Consonant change In section 2, we saw thal consonants can be largely classified accord ing to a simple three mn description i voicing: do the vocal eords vibrate? tb) place of articulation: where is the flow of air obstructed? i manner of acticulation: how is the flow of air obstructed! ‘Consonant changes ofen involve a shift in one or more of these terms, One example of a consonant changing from voiveless to voieed is the so-called flapping mentioned n section 2 (p. 34) as common in the English spoken in North America it also occurs frequently in Australasia. 1 will be recalled that a flap involves tapping the tip of the tangue quickly against the alveolar rida and it occurs when the sound 3s surrounded by two vowels. From our point of view, the important thing is that a lap is voived, whores [t) 8 unvoiced, so here ‘we have an instance where a voiceless sound has changed into a voived sound,o SOUNDS i.e. change with respect to (a) above. Some examples from Australian English appear in (0) eb ver (ms} ire hier foro] > [bir] getagt [gewt] > foeeot (Note: [o] is an unrounded central low vowel, somewhat lower than [a], ef. figure 16) There are a number of place of articulation changes currently under way’ in southem British En ish, Each of these isa change wi respect to (hb). One well known example is dhe change trom ft] to 2] as illustrated in 82); (2) inter [b 9] [ox?s] pilot [ple] [pio?] In this example, both the old and the new sounds are voiceless and have the same manner of articulation (they are both plosives). The place of articulation, however has changed from being alveolar to glota A second example is af [a] when it occurs prevocalically. In these contexts, we offen hear [0] as in the examples in (3) G3) ct [10d] >for] trove [basa] [ee re, both the old and new sounds are voiced approximants. They differ in that the alder [a] 5s retrofiex whereas the newer [U}s lsbiodental; that i, the new form has the ste place oF articulation as (v], but the nraner of articulation of [wv] A final example illustrating a change i place oP articulation concems the loss in certain environments, of the snterdental ficatives AF and 0), which ane merging with the labiodental fricatives ff and /¥/ respectively, Examples illus- uppear in (34) and (35). The change in G5) applies only to rnon-initial oy stunts Warm] [4 soaking [nye] [oH] 6s) thor [bx] [ova] veut: [ord] — [bri] in. there is no change in voici []ane both voiced and 0 ace ficatives. What has cha laiodenta [0] and [1] are both voiceless. while 0] and n manner of articulation old and new sounds ed is the place of articulation, fom interdental to Its also possible to identify changes in manner of urticulation, Included in this ategory is the proess of spiranti Cspirant? was the nineteenth-century term for “ficative’, whieh today survives only in the form ‘spirantisation’, showing that even linguistic jargon und historical change!). A classic example of spiruntisation can be found an the uccent sation — a change from plosive to FeativeSound change Table 8 Spirantisation bilabial alveolar vel voiceless pepper h ‘her Ppepeltretay feta) fhesoy nko} — [cy voiced aby sted fhagetle [beibi] > [beui] [ste |S Istezil Pha] — (hah of the English city of Liverpool, whe ss stops [p tk] have become the voiceless fricatves fb s x] respectively. and the voiced stops [b dg] hive become the voiced frcatives [fy] espoctively. in non word-snitial positions In cach ease, the new sound retams its original place of articulation and its voicing characteristics, bul by turning. from a stop into a frvative, # has under- ne a change in manner of articulation, ic. i illustrates a change in ( erm description af consonants. Table 8 includes examples af e in our ofthe six the language having fewer or more sounds. However, the change exemplified in (34) does have this consequence, since [0] is being replaced by [i] in all linguistic contexts = word initial ree, dink}, word medial ferher) and word final fmoth, ‘onelusion af this process will be a vaciely of English which lacks [0] entirely. ometimes changes can involve consonants being completely lost rather than replaced by others. We ean point to examples such as the loss of [hi] in words such as those in G6) Go [hand] Fad] [hus] — [ous] Hayy: (haul Faun) In twentieth-centuzy Britain, this change appeared to be spreading, but recently ‘evidence has suggested it may well be ne in some parts of the country Ithas certainly been receding in Australasia and is not known in North America Another example is the loss of the glide [j] before [ut] in words such as tue, duke, nex: enthysiasm, reste, solution, ofe..1.change commonly known as your dropping So, in some varieties of American English we find chy those in G7): 2 [njuwislond] + [rvezizlondl] [siuzdoat] > [seurdone] vente [evonje] > avon] ‘Some dialects for example those spoken in eastern Ei than oilers in this change, dropping and cute [ku?]. mid —have yone further he [] in words such as heaueifd [but ?atat]ot SOUNDS It is also possible for a consonant to be snserted where one previously didn’t exist A well-known example of this is provided by the dialects which have inserted [p] in the emphatic forms of the words yes and nt 8) veut [je] —+ [ep] yep" nos [nav] —>Trvup ‘nop Also familiar from some British ancl Australasian accents is the certion of [fk] afier-iug in the words nothing and something oy rothing —[oxtt]—>[ratnk] ‘svetir] fs rttk] \ final example trom the history of English involves the insertion of the bilabial stops (p b] in such Middle English words as.shanife and Old English bremet resulting in theit contemporary forms [femb]] shame and [brembl] Vowel change What about vowel changes? Section 2 showed that vowels are usually classified with respect to fa) height; (b) frontibackness:(c} lip rounding or spread- ing. As with consonants, changes ean alfeet sowels along each of hese dimen- sions, Some examples appear in table 9. In addition, itis possible for monophile from Austratian English appears in 0}: 40) [i] Eisen (Or, in the US city of Philadelphia, we find the change in (41) sto become diphthon, An example the pears is pronounced fomdopomnss] an [is] [oso]: mad, bad and glad are respectively pronounced as [me]. [bszad] an’ falessa) The converse process of diphthongs (und triphthongs complex vowels which exhibit three distinet qualities) becom: 1 monophthongs is also attested. The Table 9 Hanvel changes in ¢ change which dialect chy from change to example English? Iheight (raisin fe] Fe ad Southey [bed] [bed] Hemisphere frontback (backing) — fe} fal hell Nonwich. E [bet] [bal] lip position (rounding) fax] nurse New Zealand [ns:s] [neSound change examples in (42) sre from Bast Anutian English, with the last three mvolving wiphthong (2) sire os) > ss praver [ple] —> [ket fire. [aie] > [] fosver. fais] — fro] {We saws above that for consonants i is possible fora sound ehange to result in the loss ofa particularsound when it's systematically replaced by another which already esists in the language. Sinnlur sstuations ean be identified for vowels (vowel mergers al distinct sounds (vowel splits). Figure taking place in London round about 1350 and its consequences for the 5 with the opposite process where a vowvel splits into bw allustrates an example of the latter hot contemporary Londoners, bhi + bl > rt What we see here isa situation where the Ia back vowel [u]split. In L400, all the words pus bursts pull cup, fuck and mid included the vowel [0]. By about 1580, the vowel in exp, fied and matdlad lowered to [X], but put bustiand pul rotained [0 Later, in some dialects (most notably in South East England and Australasia) the lowered vowel incu fuck and niimoved through a numberof stays to the font, so as to become [a] in some contemporary dialects, This split occurred both in southern England snd Scotland and is ound in all the English varieties of North America and the Southern Hemisphere. It did aot occur in northern England, whieh retains [0] m such words as cup fackand rad. There is that some of the present-day U)class words are unrounding in many varieties af English, so books being pronounced by some as [bak], f] being a centralised unrounded high vawe Mergers are far more common than splis, and examples are easy to find from sound the English speaking world, One instance which was noted in seetion 2 is the identical pronunciation (as (meri) of the words merry; marry and Mary én paris of the westem snd central United States. Similar example some dialeets, af fo] and [oe], so both sure and siore become [lot]. and the smrger ina few rural castern English dialects of [au] and [e3] wih the result that cow andcare are pronounced identically as [ke]66 SOUNDS A slightly more complex case can he identified in New Zealand, the Caribbean and Norfolk, where the diphthongs [ia] and (eo) have merged. Interestin however, whereas in Norfolk the merger has resulted in [€9] taking whee [ia] was previously found, sa bath New Zealand and the Casibbewn, a ew ipinhiong [eo] has replaced both of the osiginal sounds. Thus, whereas bot andAcerhave come to be pronounced Eker be pronounced as [bea] inthe othe: two locations Finally, we can note an example of the rucal dialect of Norfolk or underzoin a merges that has affected most other English varieties. This isthe merger of the diphthongs in toe and tow: which were distinel in Middle English, They began to merge in the seventzemth century, but as the esamples from Norfolk English in (43) show this dalet has not been affected by this process over in words cari Noxfolh, they have both come 10 3) root) tow ft ul) roseEna7] —rows[19u7] moan finan] — mone [mau] far, we have looked at a number of essentially independent sound changes, Inthe ease of many vowels, however, linguists have noticed that a change to one san have a knock-on elfeet for othcrs in the neighbouring area of phonetic we understand this notion in terms ofthe vowel quadrilateral fom 2. Sometimes eases arise in which one vowel will change and leave a Jim which a second vowel moves, Its not uncommon far Several vowels to be linked togetbor in this ways a series of changes known as wehain shift As we sanv brielly in the main introduction, while our knowledge of the linguistic changes that have occurred over time isl researeh—a detailed analysis of the gradual historical development of a particu linguistic feature— methods which can accurately chart language changes as they ly based on diachronic take place within a commuaity of speakers have recently been introduced. These alled apparent-time methods involve the simulation of a historscal dimension within a synchronic study, and apparent-time researchers collect recordings of the language varieties used within a particular community and compace the speech of people bom at different times. By comparing the speech af those boon ia 1920 with that of those bom in 1970, itis claimed, we are compasin the language acquired by children at two distinet points in the history af the The language ofthe older speakers should therefore reflect an earlie in the development of the language than the varieties spoken by the ups Appazeat-time studies have enabled linguists to observe some quite complex of chain shilling in progress. For example, William Labov and his colleagues have carried out extensive research on a series of vowel shilis, known asthe Nosthem Cities Chain Shift, whieh is undor way in American cities suel as Chicago, Dettoit and Buifilo, Some shifis in the chain are almost complete and others are in their infancy, but overall the chain forms a complete “loop” in phonetic space, The oldest change in the chain is the raising of [x] in wordsSound change 1 such as fiat pack, fast bath and man. In these words, the vowel is shifting from fe] © [eT oF [P] tthe raised [°| indicates a very weak second component 10 a diphthong). The space left by (ea low front youel, has been filled by a fronting of [a] (in words such as gor, wor and pop) to [ae Similarly, the space vacated by fa. a low back vowel, has been filled by the lowering of [3] t0 fu] m words such as caneght, talk and taught. We thus see a sequence of changes wath vowels taking over the “space” vacated hy other vowels, Furthermore, something like the converse of what we have just deseribed has also occurred as part ofthe Northern Cites Chain Stat. Specifically. the change of [2] 0 [2] or [P] produced a sted” arca of mid closed high front vowels. As a result, these have alse sno move. In particular [1] Gin words such as pis in and si is moving from 1] to (e}. and [e] in words such as per, end and spel) is moving back to the position of [4], Finally, [x] —in cup, butters fuck, ete. —is moving slightly further back and rounding, to fill the position vacated earlier in the process by 3 From the above description and figure 26, it should be clear that the chain involves a series of cha Now notice that 8 which constitute a closed ‘loop? in phonetic space. ome of the chan, vowel moving and pring other vowels behind it This 1s the ease with the [se] in this chain have been caused by one {a} [2] churn: [ee] moved first and the others followed". Such chain shifts are called drag-ehains. Soonetimes, however, a vowel may anove towards the posi- tion of another vowel, causing that vowel to move itself. This is the ease with the {1]—[e]=[a] part of the chain: [] lowered to the position of fe], which backed into the position af [4] which, consequently, had to move back itself: This sort of shift is called a push-ehain (exercise Ne The transition problem: regular sound change versus lexical diffusion Having observed a number of different types of sound change tum to the question of how, more precisely. these cha which they occur. Does a sound change affect every sound at the same tine, or are some words afl es affect the words in rd which contains that ied before wiiers? Are vowel changes phonetically gradual, taking small steps in phonetic space on their routeos SOUNDS tothe new vowel, rare they abrupt, Sumping? from one vowel to another without ing through intermediate phonetic stages? Trwo hypotheses have heen put forward to account forthe way sounds chang ‘The first was initially proposed in the nineteenth century by the Neogrammarian 1up of historical linguists and it regards sous chat Two important ples underlie this hypothesis. The first of these is that ifa sound change takes pri place, it will take place in aff words with similar environments at te ‘There will be no exceptions. The outcome of this s that sound changes must be phonetically gradual, but lexically abrupt A vowel shitl, adbering to this principle, would move through phonetic space towards its new destination in ‘would apply to every ward in sinall steps, rather thaa in one step, and the el which that vowel occurred. If, for instance, we take the change from [e] to [e] in the Southem Hemisphere varieties of English of Australia, New Zealand and South Aftica, we would expect to find (a) small phonetic changes to gradually shift fe] to (of: and ¢b} every word which contained fe] to move to [e}. In the case of Sout African English, this appears to be correct with all words with [e] passing through a stage where they had a vawel intermediate between (e] and (e] The second Neogrammarian principle cliborates on the notion of ‘Similar environment? which appears in the frst principle, Specifically. it states that if a sound change takes place, the ony factors shat can affect that change in any way are phonetic ones, such ss the phonetic characteristics of the segments which s. then, may be phoneti fords may sift faster than surround the feature undergoin cally coneltioned: the cha in others because it is surrounded by a phonotic environment whieh particulary favours the change. Conversely. in some words the phonetic environment may However, agcording to the Neos is impossible for a Sound change to operate, say, in nouns but not in verbs, sin this would be an example of a change being subject to non: phonetic conditioning (grammatical category membership}. An example whieh appears to be con sistent with this emphasis on phonetic envionment appears in Labov's studies of change. These chan; ‘ound in some of the hinder and slow down the chan: smnmarians, the Northern Cities Chain Shift, wich we have just described. He found thatthe fiom [e] to [P] was most favoured when the vowel preceded a nasal consonant, asin ctl but hindered when the vowel preceded a velar consonant, such a inh Despite the predictive succes saummasian principles in some eases, a number of historica! linguists, particulary those working on dialects of Chinese. larity. This was because they discov contorm to the expected neat and regular patterns, Instead, they found instances cof what has come to be known as diffusion. Taking its name from such instances, the lexical diffusion fypothesis also depends on two principles, which are directly opposed to the principles of the Neogrammarians. This hypathesis maintains that (a} rather than bemg phonetically gradual, sound changes areSound change Table 10 fag and fae} i Standand British En Tollowing phonetic environment RP fa] RP fae) 7 Tay, ath al afte. fa na “ie crattafter. shat dait fated 2» pail bath rua). Care “st las past sy ast aster =P sk asker, mascot sl tassek hassle ns ont Sntnp mpl nd demands remad sian rar, pnd (indicates a word boundary and C any consonant in the top two coities of the leit hand column in this tale; the crucial vowel isin bold throughout) phonetically discrete, ‘jumping’ from the old sound to the mew one without passing through any intermediate phonetic stages: and (b) rather than the whole the sound change a the same time, individual words change fom the old form to the new one ina manner which is not phonetically predictable ina neat way ‘One aflen-cited example of lexical diffusion in English is a sound split which took place in southern British English and is sometimes known as the TRAP BATH split, In the latter part of the seventeenth century, the [se] in some but not all words which contained it began to fengten, and then move back, ultimately to [az], Currently, in Standard British English we have the pattem in table 10 (remember that RP is Received Pronunciation, a rather conservative variety of British F Notice thatthe: ze charted intable 10 isnot altogether phonetically re some tendencies: most words with following ave undery change—there are only a few rarely occurring exceptions, Overall, however, fram a phonetic perspective, we have a picture ofa rather messy and isregular change. Since it bas not taken place in a phonetically regular way but has seen ind idual words change independently ofany precise phonetic conditioning it provides an example of lexical diffusion (exercise 2) The change from [ee] t [az] appears to be most advan English and other southete British E ed in Standard British lish dialects but has most notably nortaken place in nosthern England. Between the north and the south we have a mixed picture, and we ean search for more evidence of the lesieal nature of the shift by o0 SOUNDS looking at a dialect which has not yet advanced quite as far as Standard British English inthe reallocation of words trom [2] to [a7]. Such a dialect is that of the small urban centre af Wisbech, a tow located between those areas of Enishi where the shift has or has not taken place, that is roughly the south and the north, ‘There are two findings about the Wisbech dialoct that are notsble here. Fisstly ited the not totaly younger residents ofthe wn are mone likely to have acquired oraliost a Standasd British English system than the older ones—a good, thou reliable indication that change is still under way. Secondly, there does not seem to be a ‘common rote? through the change that all speakers a the commustty follow. In other words, while some speakers will have, for example, (lest) [plonts) and [kes!], but (gla:saz) and (pact, others, with very similar social backgrounds, will have [glaesay] and (plient], but (la:st), [bast] and [pad] Research by William Labor comparing examples of regular sound change with lexical diffusion suggests that regular sound change is most common sn youve! shifis (ronting, raising, backing, ete.) and lexical diffusion most widespread in eases of vowel lengthening (such as the TRAP-BATIT split) and shortening, ft appears to be the ase, thea, that rather than one af our hypotheses heing the universally correct c fexercise 3} ‘one, eae seems to apply to different sorts of chan Suprasegmental change As well as affecting vowels and consonants, change may also occ among supra al phenomena such as stress and intonation, An example of such a suprasegmental change is the shifting of sess in dxyllabie words from the second to the first syllable. Particularly interesting are some noun-verb pairs in which the verb is becoming indistinguishable fiom the noun because of this process, Ie will His sfx bberecalled fiom section 2 (pp. 412) that the standard pattern in Modern E disyllabic verbs to be stessed on the second syllable, whereas comesponding nouns ace stressed on the first syllable, Thus. we have such pairs as 44 and 45): aa) They won the [‘epovest] easily (noun) ‘She wanted co [kv ‘est the ease in court (ver) (45), She hired an ['eskoa] (noun) b, The bouncer needed to [2s’koa] the drunkard thom the elub (verb) An exception to this patter is provided by adhess in most varieties of British English, which is stressed on its final syllable, irrespective of whether itis a noun ora verb (46) a, Give me your fades) (UK, noun) b, She demanded the sight to [ad's] the auience (UK, verb} Now, at the beginning of the seventeenth century, many words whieh could function as either nouns or verbs behaved likeaddress, So, for example. increase,Sound change protestandrecond carried stress on their Final syllables even when they functioned as nouns, We thus see that there has been a process of shitting stress from the Final to the initial syllable in such words when they are used as nouns, a process which has not (yet) taken place in the case of address in British English address has undergone this stress shit in American English 7a, Give me your [edse5] (USA, noun) b. She demanded the right to (dees) the audience (USA, verb) Furthermore, there is evidence that the stress shift is extending to the verbal use of some wonls in varieties of British English, as illustrated by the examples in (48) and 49} (48).a, There way a steep [rts ination last month (nun) b. The goverment was forced so [mkrits) [ayers] interest ratey yesterday barb) (4914, Bob's Pens to the personnel depastment was proving difficult (noun) b. She went tothe hank to tkensFa} [ues fae] some money (verb) What we have, therefore, is situation whi nerally no stress-hused distinction be some 400) years ago there wa en our paits of nouns and verbs, Such a distinetion hhas been sntracuced in the intervening period, with actress exceptionally maintain- ing its original propertis in British English, And now, unde stress to shifl forward from the final syllable, the distinetion is b cnoral tendency for stinning to be lost again, even though the pronunciations af both nouns and verbs are different 16 were 400 years ago. The word eavy offers a final perspective on this process. In 1600), it alre ly exhibited the ‘modem’ stress-based contrast between its uses as a noun and a verb. However, siess-shift has applied to the verb in the intervening period with the result that today we have only the single pronuncia- tion [‘envi). The examples in 48) and (49) suggest that increase and trans embarking on the route which emy has already completed We conclude this section with an example of intonational change which is affecting the varieties of English spoken in Australia, New Zealand and North America. In these localities. some people are acquiting a rising, question-like intonation contour in dee ‘Consider the small dil utterances, ves a young New Zealander ve (2c, nom-questioni ogue in (50), whied inve recounting an experience on a Paci cruise italics mark the clauses with rising intonation, (50) ras: ‘Those guys liner were in fly cheap sore a eabin all they had \was a porhole an! [looked out ofthis porthole ust. [laughs] niecn: [aw TRANK intonation are found most Reseateh has shown that these pattems of sing ries and giving explanations frequently. asin the example above, when tellingSOUNDS and descriptions, and are found rarely inthe expressing of opinions, The change appears to have begun in Australasia just afier the Second World War and is nom emg heard in parts of the UK (exercises dand $). Exercises 1 Consider the data in table 11 fom a dialect of English. The table shows the pronunciations of a number of changing, vowels and pro- vides representative examples of words in which these vowels occur What can you conclude abot the initial stages ofthe changes that took place? How are they related to each other? What happened subse quently? You may need to look at a yowel chart to help you answer these questions, Table 11 Howef changes in an English dialect Pronunciation oF Pronunciation of Pronunciation the vowel before the vowel during of the vowel the cha, ange tonlay fal] fe) [orl [ou] [us [se] fea feu Are the following examples of sound changes. discussed in this see eulur sound change" or of ‘lexical diffusion"? How do tion, cases 6 you know (a) the UA split? (h) the shift to sylluble-initial stress 3 Inmany varietcs of English [tJ s changing into a glottal stop [2 The linguistic contests in whieh glotatisation ean oecur differs from place tw place, and nowhere bas [1] been completely replaced by [2]. Below are some data illustrating the extent of lotlalisation in ane variety of English, Try to describe phonologically the contexts in which wlotta lisation ean and cannot occur “lottalisation possible ion mot possible data deter Peter pesterSound change let me left me lets us bet best coll tomorow call Tony salt soft want washed ‘wn ft button setun center ‘wrapped ‘nept bortle act As well as being spoken in the Netherlands, varieties of Dutch are also used in northern Belgium (wheze they are often called Fle ian and Dutch linguists have been researching the extent to which thestondord varieties of Dutch inthe Netherlands and in Belgium are becoming moze similar or more different, Figure 27 (hasedl an the work of van de Velde, van Hout and Gerritsen), shows the results of an analysis of radio commentaries on royal and sporting events in Belgium and the Netherlinds at regular periods between 1935 and 1993, The feature investigated here is the devoieing of 'w! to [f) in words such as those immediately helo (a) vuor [sy] [ft] fre (by lever flewar] > fle'far] liver (c) aanval [anval] [enfal] attack What has happened to ‘v/ over the past seventy years? How might we unt forthe patierns found? = «tl 1235 7980) i Figur devin oft fin Net 1935 ad 1998 fe Bei dati P95 ad 398 How aud Gerritsen 1996: 161)SOUNDS Collecting data on variation and change in language invelves under stanafing the way the speech community is structured socially as well as linguistically, I! you were to conduct research in your own neisgh bourhood, what sociological factors do you think you would need to take into account and why?5 Phonemes, syllables and phonological Processes We beyansection 2by asking how many sounds there re in English, butwe found there were satious practical difficulties in responding to this question and never sorived alan answer. There is a furthce reason why the question can’t be answered st this is our fist concera in this section, In forwardly, and understand fet. sp i. fashion. We can see this particularly easily in the vowel system, One of the main differences between the [é] of read [tizd] and the [f] of rid [aki] is length But just how fong isa long vowel? An emphatie pronunciation of read, say in a alone~ Pmtrying to READ’, has a much longer vowel than a from cach other ina nondiserete, continuous ech sounds can di plaintive ‘Leaver non-emphatic pronunciation, The precise length of any vowel will depend! on the rate of speaking, degree of emphasis und so on, A similar case is presented by the aspirated plosives. In any dialect, a [p'] sound, as in the word pit will be aspirated ww om the d cater or lesser extent depending fire, that there is a sense in which sounds form a continuum: fom this perspective. ov of emphasis. We see, there there is an iafinite number of specch sounds in any language Phonemes Fortunately, there is another perspective from which sounds are discrete units or Segments, and we can come to terms with this by asking is the difference between the words pit and fie? From section 2, we pitstarts with a voiceless bilabial plosive and hit starts with a voiced bilabial plosive. Otherwise. the words are identical. A pair of this kind, in which eves thing escept the portion under consideration is identical s called a minimal pair, This pair shows that voicing can distinguish one word from another, and that the pair of sounds [p b] can dlistingnish types of [p]. with different deg different picture. There are no words in English which differ solely in whether rds, However. when we consider different cvs of aspiration or no aspiration at all, we get a they contain an unaspirated or an aspirated plosive. That is, English does not hhave distinet words live, say [p't] and [prt] fact, [pr] with totally unaspirated [pl is unpronounceable without explicit taining for most English speakers Conversely, we could never find pairs such as [sptt] and [sp'tt] im English follow ng initial [s]. the onf'¢p" sound we find isthe unaspirated [p} The same is true of (Ft) and [kM], as the pairs of words stan rav-and scar, ea: In other1% SOUNDS words, the distribut n of the sounds [p p!] is governed by a rule or prineiple according to which we never find [p] in the positions reserved for [p"] and we never find [pin the positions reserved far [p]. This type of patterning is called complementary distribution (the positions in which we Find the «wo sounds complement each other) Things needn't be this way: There are languages in which [p] and [p*] ean be uscd to distinguish words, that is, in some languages [p/p" Ut? Kk] and simi parsare contrastive sounds, [n{ 1) weshow examples from Bengali (or Bangla) spoken in Bangladesh, in which (p] and {p!. [1] and [t"] and (k] and [&"] contrast {and there is also a contrast between ff] and ft") 61) aspirated unaspirated (ea ‘canal? (kal) stime" ya) ‘ashes’ Lai] ALwant™ aka) storemain® flaca) to see fata) ‘head? {imana] to be enthusiastic” {oul} ‘lower ful bid Retuming to English, we can simplify our description of the sound inventory by thinking of [p tk) and [p" t! k"] as variants of the *p’. *U’ and *k* sounds. Thus, ‘we can say that thors are just the three voiceless plosives, but they have slighty different pronunciations depending on their postion sm the word. kgporing ther Positions, word-initally we get the aspirated variant and fier [s} we get the unaspirated variety, Thus, we could transcribe the words pits, easter, ca sears [pit]. [A0ystaz, [ka sker] on the understanding that a general cule will tell us exactly how to pronounce the plosive. It is no accident, then, that this dlistinetion between aspirated and unaspirated sounds is never marked in ordinary English orthography (though it is masked in the spelling system of Bangla). fn fact, native speakers of English who have not had some kind of phonetic a linguistic taining are usually completely unaware of the distinction Fram the above. it follows that we nced to he abe to talk about sounds at t0 fevels. Atone level we must be able to deseribe the fact that Eng sh has aspicated aswell as unaspizated plosie simply to capture an important difference botoeen the plosive system of English and those of lang French, Spanish, Russian, Samoan, Inuit and many athers in which plosives are never aspirated. On the other hand, we also need to be able to capture the idea that in English [p] nd [p"] are variants of the saone sound?” But shat sound? To answer this question, we need another less concrete, concept ofsound?, We vill call these moze abstract sounds phonemes and write them between slashes: PLL... transcription into such phonemie symbols is called a broad transerip- tion. However, when we want to talk: about the precise, conere bbe doteoted by phonetic analysis, we will speak about phones, These are writen rn square brackets. Thus, [p pt #!& &"] represent six phones but in Entish respond to only three phonemes, /p tk’. A transcription which ineludes ‘tie detail about the promunciation af individual phones, and written in squarePhonemes, syllables and phonological processes brackets, is rofemred to as amarrow transcription, There is always some chorea to exactly how much phonetic detail an analyst might include, so the notion of ‘narrow transcription isa relative one, We will also say thatthe ovo variants [p p*] ofthe phoneme ip’ areallophones of that phoneme, The term ‘allophone’ is based on a Greek expression mean “diferent sound’. The phenomenon of variation in the pronuneiation of phonemes in different positions 's called allophony o* allophonic variation. and we can iMlustrate this diagrammatically for our English voiceless plosives as in (52) 62) ip! av ‘ki phonemes. /\ /\ 7S ipl Ira [eT kh] allophones Note that the transcription at the Jevel af allophones bas to be rather approximate, cn that we can have different d infinite number of distinctions at tis level. Howe (ree) of voiceless plosive phonemes in the languag IP we turn tothe vowel system, we have noted that le permitting any number ofdistinetions, Obviously. this is also the ease forthe front ‘ees of aspiration in principle, there is an there is only a fixed number wth isa continuous quality back and high low ases introduced in section 2 as playing a major role in the categorisation of vowels. However, we can simplify this complesity by takin some devisions as to what features of the pronu be sad to belong to the phoneme, and which are less to do this in diferont ways, and we shall do no more than illustrate the issues that arise here. Consider the pairs of vowels it ut] and [f 0], Members ofthe fist pair are Jonger than members of the second pai, hut there is also a difference in gua (isuz] are ep. 38} Furthermore, the distinction betwen the puss is crucial, since we have such minimal pairs as hewehirand poot fi We will assume that vowel length is the important factor in these distinetions, Thus, we can say that [tut] are she long vowels corresponding to (f). This m that the more fax pronunciation of the short vowels [tis secondary to the Feng distinction. na broad, phonemic trnseription we could thus use just one syrabol forcach, sty 4 w. with an additional indication of lengts, Thus the fo ‘would be pronounced [it] and the short phoneme would be pronounced [i], ancl similasly for ‘wt (pronounced as (ut) and sw (pronounced as fo). Likewise, we might want to say that {4 -p) are shom equivalents af (a: a2) There # some controversy as to whether this gives a satisfactory answer for English, howeve (for reasons which go well beyond the seope of an intgoduction suc as this). fa addition, it is helpiul to get used to the more aecunite marrow tanscriptional system for vowel sounds, since vowels differ so much trom one variety to another ‘Therefore, we will eantinue to make more dBtinctions than may be srietly necessary We can now revast our original question as ‘How many phonemes are there in English", and we jon are crucial andl henee an ial, Different accounts tend nse vowels, whereas [FU] are fas sven in tableSOUNDS Table 12 ‘The English phoneme inventory Conon corona laibio- inter palato- dlossals-gutturals bilabial dental dental alveolar alveolar palatal velar glo Plosives pb i K Frivatives fy 80 87 $3 b Nasal m n n Approximants la i Vowels Shor 1 ‘ © > . a D ix u a Diphihones eratay a10u1a 0 (Note that the term guaturalsis used to refer tothe class of uvular, pharyngeal and glottal consonants, English has only ‘hv in this class continue to use distinet symbols for the Long and short vowels in aeknowle ment ofthe uncertainty to which we have just alluded. This is our fist experience of the importance of distinct levels of analysis in linguistics. an extremely important notion. In the ¢ atively concrete level more closely linked to physical sound and a more ated to the (and ultimately ally, what we ean sugeest is that the phonalo- urzent context, we have a Jbstract level onuunisation of patterny of sounds inthe grammar of the langue in the minds of speakers). Specifi representation, which appears in the lexicon as part of the lexieal entry for a onetic representation, The manner in which a word, is a phonemic and not a p phonemic representation is converted to a phonetic representation is part of the PF component of the grammar (see the Introduction, p. S) and we shall be saying ‘more about this presently texercise 1) syllables When the Japanese borrowed the monosyllabic sporting termspru, it came out as suparingo with four syllables. When an English speaker ties t nerally pronounce the Russian name Mstislav (two syllables in Russian!), it gPhonemes, syllables and phonological processes ” ‘acquires an extra initia) syllable to become [amstislay] or [mistislav]. Speakers of Cantonese Chinese tend to pronounce the words wath, walks and ea dentally, as [a2]. Why 3s this? The answer is that different languages permit different Kinds of syllables, and native speakers of languages bring their Knowledge of syllables and syllable stucture to their attempts to produce words from other languages. To see what kinds of syllables we Find, we need to look at syllable structure more earefully Words like flat, spat and spratare said to shyme, This is because they have identical pronunciations after the fist consonant or consonant cluster We can divide a syllable therefore into two halves, the rhyme (or rime) and the “onset, We have already referred (p. 4110 the vowel in the midalie of the syllable as themucleus (orpeak}. The consonant or consonant cluster after the nucleus will be called thecoda, These terms are illustrated in (63) for the cord qi : ; “ ows ae Nis A kow 1 re The symbol (= Greek letter ‘sigma’) is often used to represent a sylluble The order of the consonants in the onset and the coda is anteresting here because some consonant orders yield impossible words, Thus. compare the con sonant clusters at dhe beginnings and ends ofthe ‘words’ in ($4) and (5). In each «case, the ilieit sequence (marked with *) is intended to be pronounced as a single syllable: G4) nelp nh Hump Sayan hard had 655) play Span prey Sapa quick —*Avktk cue Sik Returning to (53), form such aso walt is fine but © wKtt is an impossible firm in English, There isa systematic reason for this. We distinguished inscction 2 between obstrueats (plosives, affricates and tricatives} and sonorants (nasal and approximants). The reason kt’ makes a aed syllable peshaps has something to