Forthcoming in this series F. AARTS and J. ARTS English Syniactic Structures ‘An introduction to units and structures in contemporary written English IVAN POLDAUF English Word Stress ANALYZING ENGLISH An introduction to descriptive linguistics HOWARD JACKSON City of Birmingham Polytechnic SECOND EDITION © PERGAMON PRESS ‘Oxford - New York » Toronto « Sydney + Paris » Frankfurtux. usa, CANADA AUSTRALIA FRANCE FEDERAL REPUBLIC OF GERMANY Pereamon Press Lt (OXS OW, Fagland Pergamon Pres ne, Maxwell House, Faivew Pas imford, Now York 10523, U.S.A. Pergamon Press Canada Lid, Suite 104, 180 Consumers Road, Willoweele, Ontario MAY 1P9, Canada Pergamon Press (Aust) Pi. Lid, P.O. Box SH, Pts Point, N°.W. 2011; Austin Peigamon Press SARL, 24 rue des Beoes, 75240 Pats, (Codex 05, France Pergamon Press GmbH, 6242. Kronberp-Taunus, ammerwet 6, Federal Republic of Germany Headington Hit Hall, Oxford ‘Copyright © 1982 Pergnon Press Lid AIL righ eserved No part ofthis publication may be dnl edn feral eetinor aed ‘ny form or by any meant: electron, leona imac wpe, chanical phoning, reading or ‘herwse, wou permission in hing ro the pub her Fit eiton 1960 Reprine1982 Sezond eon 1982 LUbrary of Congres Cataloging in Publication Data Jackson, Howard, ‘Analying English, (Language courses) Tes inde 1 English anguape-Grammar-1950 1. Tle, Series. PELIGIE RD 2)B1-23834 AACE rks Library Cataloguing in Publieation Data Jackson, Howard nalyzing English; an introduction to descriptive igus 2nd ed (Language courses) 1 Engi rege Le I fn PEL ISBN 0-08-025567-4 Printed nGreat Briain by A. Wheaton & Co. Lit, Exeter for HilaryContents Acknowledgements Introduction Part One: Sounds Making sounds English consonants English vowels Sounds and their variants Sounds in connected speech Combining sounds intonation Part Two: Structures 8, Grammar and sentences, 9. Word classes 10. Phrases 1—noun phrases 1. Phrases 2—verb phrases 12. Phrases 3—adjective phrases, adverb phrases, prepositional phrases 13, Clauses 1—strueture and types Jd. Clauses 2—dependent clauses \Fext 1—rearranging items fext 2—cohesion Part Three: Words 17. Morphemes 18. English morphology 19. Dictionary information 20. ‘Grammar’ of words 21. Relations between words Conclusion and further reading Key to exercises International Phonetic Alphabet Index 2 2 2 48 35 n 8 83 9 102 109 4 19 2s 136 138 Ist 153Acknowledgements This book came into being over a number of years in the course of teaching the analysis of the modern English language (o students on the English degree at Birmingham Polytechnic, Like many students of English coming to study for a dogree which includes a fair proportion of linguistic work, after the purely literature courses they have pursued at. school they find the ways of thinking and the analytical approach required for this work at first often new and strange. I am {grateful to all those students who bore with my attempts to find the most satisfactory way of introducing the modern linguistic study of English, 1 would also like to thank Vaughan James of the Pergamon Institute of English for his help and advice in the preparation of the manuseript for publication. And owe a debt, asin many other things, to my wife Hy Introduction Many people associate the term ‘linguistics’ with the philosophy of language (Wittgenstcin and the like), with learning foreign languages, or with abstract, ‘mathematical symbolism supposedly charting the language that we speak and write. ‘Now linguisties does have something to do with all these three activities, and the third of them in particular occupies the attention and energies of many scholars ‘working in the field of linguistics. Perhaps the most well-known, but by no means only, scholar falling into this category is Noam Chomsky, who initiated the “transformational generative’ school of linguistics. Such scholars are often referred to as “theoretical linguists, Their aim is to construct a unified model or theory to ‘explain’ the nature of individual languages and of human language in general. Like all theoretical sciences, theoretical linguistics aims to make as explicit as possible its axioms and procedures, so that its hypotheses and theories can be adequately tested. The explicitness required demands a measure of formalism, and for this theoretical linguistics looks, like fther sciences, to mathematics and symbolic logic. This makes many linguisties books daunting to the uninitiated and difficult for the layman to gain access to, ‘Much modern linguistics over the past 25 years has been of this kind, often concentrating on individual problems of description at a quite abstract level, However, not all linguistics is like that. More important, in the opinion of this author, and complementary to the kind of approach outlined in the previous paragraph is a branch of linguistics referred to as ‘descriptive’ linguistics. This ‘approach starts with a language as it is spoken and written, withthe data of actually occurring utterances and sentences, and seeks to analyze, catalogue and describe ‘that data and that language. Descriptive linguistics is about describing languages rather than about constructing theories and models. It must, of course, use categories of description, but these arise inductively from a consideration of the language data rather than deductively from the axioms of a theory. Jn many ways, descriptive linguistics isthe successor to ‘traditional grammas’, as it used to be taught in schools. And it is ironic that, while in recent years great advances have been made in the description of contemporary English, the education system in Britain no longer demands that pupils should be acquainted with the nature and structure of their mother tongue. In the seventies a number of significant reference works have been published, distilling the latest knowledge about the English language, among them the Grammar of contemporary English by R Quitk, § Greenbaum, G Leech and J Svartvik, published by Longman in 1972. ‘This present book is aimed at those wishing to discover the techniques and insights Of linguistics as applied to the description of the contemporary English language. ‘Such persons might be students in further or higher education undertaking a coursexX Introduetion in linguistics oF English language, without any school background in language analysis. Or they might be teachers or advanced learners of English as a Foreign Language, secking the beginnings of a systematic linguistic description of modern English. Or they might be interested lay people wishing to know what this still relatively new science of linguistics has to say about the nature of the language they speak. ‘i ‘The book is divided into three sections, dealing with Sound (ie pronunciation), ‘Structures (je grammar) and Words (ie vocabulary and meaning). The terminology used is, much of it, derived from traditional grammar, and it relates directly to that found in A Grammar of contemporary English and similar works. Each chapter presents a framework of analysis for the point being discussed, and is provided with fn exercise, to give practice in applying the analytical techniques presented in the chapter. Solutions to the exercises are given in a Key at the end. The aim of the ook is 10 show how linguistic analysis and description is done and to lead the reader to make his own analyses. It is intended that the book may be used seltinstructionally, or i could be used by a teacher. And it is hoped that after working through the book the reader will then be able to make practical analyses of any English sentence or text himself—with the aid of the reference works cited in the Conclusion, Part One: Sounds| | 1. Making sounds Speaking and writing ‘We can transmit our language in two ways: either by speaking or by writing. We are all conscious of what writing invalves: the use of a pen or pencil in our hands, the ‘making of special marks (letters) on paper. But we are not at all conscious of what speaking involves, It involves sound coming out of our mouths. But how is that sound produced? And how does it relate to the letters that we write? Afterall itis the same language, whether we speak it or write it ‘The reason we are more conscious of writing is because we had to be taught how to write, how to hold the pencil correctly, how to shape the letters so that they could be recognized hy someone else, But we learned to speak on our mother’s knee. [No one told us how to make the right sounds, what to do with our tongues or our lips. All natural languages have been transmitted by speech, but not all by writing. ‘And all normal healthy people have learned to speak their language in childhood, but not all have Ieamed to write, In this sense speech is prior to writing and characteristic of us as human beings. When we speak we use over half of our bodies to do so, from the diaphragm, situated below the lungs, to the mouth and nose in our faces. Speech is quite simply ‘column of air, that originates in the lungs, and is modified in various ways before its passage through the lips, and s0 out of the mouth and into the air. Like any other sound, speech is the vibration of the the air to make sound waves. In this case the, vibration, and the characteristics of the sound waves are determined by the human vocal organs. ‘What we have just described is the most usual basis of speech, ie egressive lung ‘ir—air originating in the lungs and passing outwards. It is also possible to speak while breathing in, with ingressive lung air. We sometimes speak like this in ‘moments of tension, but it is not a normal way to produce speech sounds. However, ingressive mouth air sounds, produced when drawing air into the mouth but not into the lungs, do occur. For English speakers the tut-tut sound is made like this, and in some languages such sounds are regularly used. Egressive mouth air sounds also occur; for English speakers blowing a raspberry or giving a kiss are sounds of, this kind, But for articulating the sounds of English words egressive lung air is virtually always used in normal English speech. ‘Vocal cords ‘As air is expelled from the lungs, it passes up the windpipe (trachea) and into the larynx (see Figure 1). In the larynx is found the glottis, the passage between the ‘vocal cords (oF vocal folds). Here the air coming from the lungs receives its first4 Analyzing English major modification. The vocal cords, despite their name, are not to be imagined like strings of a guitar or piano. They are a pair of fleshy, lip-like membranes, that are hinged at the front and may be moved together or apart in order to impede or allow the passage of air through the glottis The action of the vocal cords (ie the state of the glottis) determines whether a sound being produced is voiced or voicelesss.When the vocal cords are brought close together, but not shut tightly, the air escaping from the lungs eauses them to Figure 1: The vocal organs vibrate: this is called voicing, and a sound produced while the vocal cords are Vibrating is a voiced sound, When the voeal cords are apart, no vibration occurs, and sounds produced with the glottis in this state are voiceless sounds. This is also the state of the glottis for normal breathing, The vibratory action of the vocal cords ‘may be easily demonstrated by blocking up the ears and producing alternately an s And ax sound. The s isa voiceless sound and the z a voiced sound, The articulation of the latter should induce a buzzing sensation There are two further possible states of the glottis, which are both of marginal interest to the production of speech. One is when the vocal cords are brought tightly together and the glottis is completely closed. We do this when we are lifting, heavy objects, and it occurs in speech as a momentary action, in the production of the ‘glottal stop’, familiar from the Cockney pronunciation of the t in butter. The other state of the glottis is intermediate between vibrating and completely open: the vocal cords are brought together, but not sufficiently for vibration to occur. This is the state of the glottis associated with whispering, ‘Summarizing, there are four possible states of the glottis: 1, Open and relaxed, wsed for normal breathing and making voiceless sounds; 2. Vibrating, used for making voiced sounds; 3. Completely shut, used for making the glottal stop; Making sounds 5 4. Close together but not vibrating, used for whispering. (Only the first two are of importance in the description of speech sounds, giving us the categories of voiceless and voiced sounds. A speech sound must belong to one or the other of these categories, and this is determined by the state ofthe glottis at the time of its production. (Oral and nasal sounds As the column of air passes beyond the glottis, through the pharynx, there are two possible ways of escape from the face: either through the mouth, or through the nose. At the back of the mouth, as an extension of the root, isthe soft palate, oF volum. This is under muscular control and may at any time be in one of two positions: either raised, or lowered, Ifthe soft palate is raised, the passage through the nose is blocked and air may escape only through the mouth. Ifthe soft palate is. lowered, the passage through the nose is free, and air may escape both through the nose and through the mouth, Sounds produced with the soft palate lowered are nnasal or nasalized sounds. Sounds produced with the soft palate raised and air escaping only through the mouth are called oral sounds. Indeed we assume that sounds are oral unless we use the term nasal or nasalized. The difference between oral and nasal sounds may be illustrated from the word sudden. In a normal, fairly rapid pronunciation of this word, the final two sounds are da. The only respect in Which these sounds differ from each other is that d is an oral sound and n is a nasal sound. If you monitor your pronunciation carefully you should be able to feel the action of the soft palate in the transition from d to n, i from a raised to a lowered position ‘The function of the soft palate is, then, either to allow or not to allow the passage of air through the nose. When it is lowered it does not restrict the escape of air through the mouth. However the eseape of air through the mouth may be restricted in other ways, and air may be escaping only through the nose. For example, in the pronunciation of m, air cannot escape from the mouth because the lips are tightly shut, so that the escape of air is only through the nasal cavity. The most complex modifications of the column of air that originated in the lungs take place in the mouth. Here we are concemed with two aspects of articulation: firstly, the manner of articulation; and secondly the place of articulation, The ‘manner of articulation refers to the way in which a sound is made, and the place refers tothe position in the mouth at which the sound is made. A number of sounds made in different ways may be made at the same position, ‘Manner of articulation With manner of articulation we frst of all make a broad distinction between sounds that are produced without any obstruction in the mouth, and those that are made with some kind of obstruction, The former we call vowels and the latter consonants. ‘The differing quality of vowel sounds is determined by the openness of the mouth, the configuration of the tongue and the shape of the lips. The quality of consonant6 Analyzing English sounds is determined by the kind of obstruction to the passage of air (or elasure), 8 well as by the place of articulation. igure 2: Near closure fors Complete closure ‘A umber of kinds of closure may be identified. First there is the complete closure, when a total obstruction is made to the flow of air. When the soft palate is raised and air is prevented from escaping through the nose, air-pressure will build up behind the obstruction, and may then be released with an explosion. Sounds produced in this way are called plosive sounds (or sometimes stops). Such a sound in English isthe intial b in bin, Here the complete closure is made by the two lips. Obviously the closure, the build up of air-pressure, and the plosion are momentary activities; but they are all three stages in the articulation of plosive sounds. If, with a complete closure, the soft palate is lowered, then air may escape through the nose. Sounds made in this way are called nasal sounds, eg m in English man, where the closure is again made by the lips. Figure 3: Complete closure for g Near closure ‘A second kind is a near-closure. Here the air is allowed to escape, but not freely. ‘The escaping air causes friction at the point of near-closure, and the sounds produced in this way are called ftieative sounds. Such a sound in English is the initial v of van, Here the near-closure is between the bottom lip and the top front teeth, with friction resulting. Laterals ‘A third kind is made with the front of the tongue forming a complete closure, but With the air being allowed to escape over the sides of the tongue, Sounds produced in this way are called lateral sounds, since the air escapes laterally. Now, the air Making sounds 7 escape may be completely free, as in the case of the initial 1 of English lid, or the sides of the tongue may be raised, so that air escapes only with friction, as in the initial 11 of Welsh Lian-, This latter sound is called a lateral fricative. Intermittent closure A fourth kind is an intermittent closure, In making sounds with an intermittent closure, « complete closure is made very quickly and repeated several times. Such a sound isthe rolled r, sometimes heard with an emphatic (or Scottish) pronunciation ‘of F in, for example, red. Here the intermittent closure is between the tip of the tongue and the back of the upper front teeth. An intermittent closure may, however, involve only one such closure, in which case itis called a Napped sound, sometimes heard when F occurs between vowels eg in very. Near closures without friction ‘A fifth and final kind is a near-closure, like the second kind, but without friction. The articulators come close together, but not close enough for friction to occur. ‘These sounds are called frietionless continuant sounds. Such a sound isthe intial w in wer. Here the near-closure occurs between the two lips. These sounds could be classed as vowels, since they are made without any obstruction to the air-flow through the mouth, but they are usually regarded as consonants, at least as far as English is concerned, because they act like consonants in all other respects, especially in syllable and word structure. These, then, are the categories that we use to describe the manner of articulation of sounds. There is first ofall the broad distinction between vowels and consonants, ‘and then among consonants we distinguish between plosives, nasals, fricatives, laterals, rolls and flaps, and frictionless continuants. Place of articulation Now let us consider the place of articulation of sounds. From what has been said about the distinction between vowels and consonants, it will be clear that ‘consonants are easier to describe for place of articulation than vowels, since it is possible to feel where the obstruction is taking place in the mouth. Nevertheless, ‘vowel sounds may be described in terms of their place of articulation Vowels ‘wo sets of categories are relevant to this description. The first involves the ‘openness of the mouth and height of the tongue, for which the categories are close, half-close, half-open, open (an alternative corresponding set of categories is also sometimes used: high, mid-high, mid-low, low). The second involves the general configuration the tongue and the area of the mouth in which the sound is made, the categories here are front, eentral, back. A third set of categories i relevant to the description of vowel sounds, involving the shape ofthe lips; the categories are8 Analyzing English ‘rounded and spread (also referred to as unrounded). Ilustrating from English, the vowel sound in beat, represented by the letters ea, isa close front vowel with spread lips; the vowel sound in boot, represented by the letters 09, is a close back vowel with rounded lips Figure 4 Vowel Chart showing and 00 close ST Ty . WG | 5, are involved, In making consonant sounds, two parts of the mouth, or articulators, For the majority, one of these articulator isthe tongue, oF atleast some part of the tongue. Since this is so, the label for tongue, (from Latin lingua), is not usually included in the description of the place of articulation of the sound. Consonants We will begin our review of places of articulation at the front of the mouth. The Figure 6: Labio.dental f Jace involves the two lips as articulators, and sounds made here are called bilabial sounds (from Latin labium). They may be illustrated trom the initial b in bit, which is bilabial plosive, or the initial m in men, which is a bilabial nasal. At the next place of articulation the bottom lip makes a closure with the top front teeth; these sounds are called tabio-dental sounds. In English the inital fof fun isa labio-dental fricaive ‘The other consonant sounds all involve the tongue as one of the articulator. First, consonants are made by the articulation of the tip of the tongue and the back of the "upper front teeth; these are called dental sounds (from Latin dens). In English the final sound in teeth, represented by the letters th, isa dental fricative, Next, a sound ‘Making sounds 9 is made by the articulation of the tongue and the bony ridge just behind the upper front teeth. The part of the tongue used is that just behind the tip, called the blade, Figure 7: Dental Figure &: Alveolar d and the bony ridge is called the alveolar ridge, which ‘made in this position —alveolar sounds. In English the initial d of din is an alveolar plosive. The next area of the roof of the mouth, behind the alveolar ridge, is called the hard palate. An articulation occurs between the hard palate and the part of the Figure 9: Palatal y tongue behind the blade, called the front of the tongue. These sounds are called palatal sounds; in English the intial sound of yes, represented by the letter y, is a palatal frictionless continua. If you run your tongue over the roof of your mouth you will feel the hard palate being replaced by a soft area, called the soft palate or velum. There is an articulation between this area and the back of the tongue, making velar sounds: in English the intial g of ge isa velar plosive. The velum extends into the uvula, the piece of flesh that you can see dangling atthe back of your mouth when you look in, mirror. The uvula articulates with the back of the tongue to produce uvular sounds. In French the sound is often uvular roll, with the uvula making an intermittent elosure against the back of the tongue. Moving further back in the ‘mouth we come to the pharynx; itis possible for an articulation to occur between the back or root of the tongue and the pharynx, giving a pharyngeal sound. No such10 Analyzing English sounds ovcur in English or the European languages; they are typical, however, of ‘Arabic, Proceeding beyond the pharynx we come to the glottis, where the position Figure 10: Velarg of the vocal cords may produce glottal sounds, We have mentioned the production of the glottal stop (the Cockney pronunciation of the tt in butter) by means of a complete closure in the glottis. From this description of the possible places of articulation for speech sounds it will bbe clear that for purposes of description, the tongue is divided into a number of identifiable areas as is also the roof of the mouth. Beginning at the front end, the Figure arts ofthe tongue and root of the mouth Team 7 parts of the tongue that are distinguished are: tip, blade, front, back, root. The areas of the roof of the mouth are alveolar ridge, (hard) palate, yelum (ie soft palate), uvula, Deseribing speech sounds In deseribing speech sounds there is a convention well worth following. When labelling a vowel sound, give first of all the category for height, then the one for the general area of the mouth, and then the one for the position of the lips. Thus, the Vowel of heat is described as a close front spread vowel, Vowels are assumed to be Making sounds 11 voiced, ie articulated with the voeal cords vibrating, since this is predominantly the ‘case. And in English, at any rate there is no regular distinction to be made between voiced and voiceless vowels. In labelling consonant sounds however, voicing is portant and is given frst, then the place of articulation, and finally the manner of articulation. Thus, the b of beat is described asa voiced bilabial plosive, and the fof fun as a voiceless labio-dental fricative. International Phonetic Alphabet It will have become clear that the Roman alphabet, which we use for writing English, is by no means adequate for representing the diversity of speech sounds that we use for speaking English. The inadequacy becomes evident if we reflec that thete ate over 40 distinguishable speech sounds in English, while our alphabet contains only 26 symbols. To make the point more concrete, consider that we have {0 use a pair of letters (th) to represent the single voiceless dental fricative sound at the end of teesh, Moreover, this same pair of letters also serves to represent the voiced dental fricative sound, as found at the beginning of shen. So in representing speech we make use of the International Phonetic Alphabet, developed by the International Phonetic Association. This alphabet is based on the Roman alphabet, but with the addition of symbols from other sources. We shall be introducing the symbols that are necessary for representing English speech sounds in the following chapters, and the complete alphabet is given on page 151. Exercise 1 ‘Make a, description of the following speech sounds in the way recommended eg the b of beat—voiced bilabial plosive; the en of bea—close front spread vowel 1 the € of beat 2. tho v of van 3. the k of kiss 4 the th of thin 5. the mof now 6 7 8 the a of bath the p of pin the €e of seed the 0 of bole 10. the Lof like 11, the g of go 12, the # of 002. English consonants In the previous chapter we defined consonants as those speech sounds which are produced with some kind of closure in the mouth, restricting the escape of air. And ‘we distinguished among consonants according to the place in the mouth that the closure occurs (place of articulation) and according to the kind of closure made (manner of articulation). These factors, along with the state ofthe glottis (voicing), determine the kind of consonant sound that is produced. We shall now take each of these classes of consonants in turn, discover which sounds are used in English, and introduce the phonetic symbol for each sound from the International Phonetic Alphabet. Fricatives The largest group of consonants in English comprises the fricatives, sounds that involve a near-closure with friction resulting between the articulators, In most cases there is a voiced and a voiceless fricative occurring at each place of articulation. Bilabial fricatives, where the friction occurs between the two lips, have symbols taken from the Greek alphabet: for the voiceless bilabial fricative the symbol is, igure 12: Bilabialiative and for the voiced bilabial fricative 8, These sounds do not regularly oceur in hglish, except in certain contexts as variant sounds for labio-dental fricatives (see Chapter 5). ‘The labio-dental fricatives have the familiar symbols f and v for the voiceless and voiced variety respectively. Friction occurs between the lower lip and the upper front teeth. These sounds occur regularly in English words, in all positions, eg initially in feignivain; medially in referee(reverie; finally in lifelive. When the phonetic symbol is a familiar one, as in this ease, it should not be assumed that English consonants 18 when the corresponding letter occurs it always represents the sound that the phonetic symbol stands for. So, in the case of the letter f for example, the word ‘written of is in fact pronounced with a voiced labio-dental fricative as the final sound, ie represented by the phonetic symbol . Figure 13: Dental fricative In producing dental friatives the friction occurs between the tongue tip and the back of the front teeth. For some speakers the tongue tip actually protrudes between top and bottom teeth, and these sounds are sometimes called interdental fricatives. One symbol is taken from the Greek alphabet: for the voiceless dental fricative the symbol is; and for the voiced dental fricative—8, which is a specially invented symbol. These sounds occur in all positions in English words, eg thiefthis; lethaltleather; clothiclothe. Iveolar frgative ‘The alveolar fricatives have familiar symbols, s for the voiceless alveolar fricative, and z for the voiced alveolar fricative. Again, it should be noted that not all ‘occurrences of the letter $ are voiceless alveolar fricaives, ie symbolized phoneti- cally by s; eg rise has a voiced alveolar fricative as its final sound, represented phonetically by 7. With alveolar fricatives the friction is between the tongue tip or blade and the alveolar ridge. But they differ from dental fricatives not only in place Of articulation, but also in the shape that the tongue takes up in the articulation of the sound. With alveolar fricatives in English the tongue is shaped so that there is a groove down the centre line along which the air passes, whereas with dental fricatives the tongue is relatively flat with only a narrow slit. So s and 2 are groove fricatives in English, while @ andO are slit ficatives. The alveolar fricatives occur in14 Analyzing English all positions in English words, although x is not common initially: seal/zeal, racerirazor, lacellaze. Figure 15: Palao-alveolar fictive Further back in the mouth we come to the pair of palato-alveolar frieatives. Friction occurs between the tongue front and the hard palate, but at the same time the tongue tip is raised towards the alveolar ridge. These sounds are symbolized by J for the voiceless palato-alveolar fricative, and 3 for the voiced palato-alveolar fricative. They occur in all postions in English words, eg, ship/gigolo,fssion!vision, rushirouge. But the voiced palato-alveola fricaive is of very restricted occurrence initially and finally, found only in words borrowed from French; and there is a tendency in some people's speech to replace the voiced palato-alveotar frieative in these positions with the voiced affricate (see below). Palatal fricatives do not occur in English. Of the velar frieatives only the voiceless ‘member occurs, and then in only a few words borrowed from other languages. The velar fricative is produced with friction between the tongue back and the soft palate, and the voiceless member is symbolized by x. It occurs in the Scottish pronunciation of lack, and in some German names, eg of the composer Bach, ‘The only other fricative to note in English is the glottal fricative, which occurs only as-a voiceless sound, It is represented by the symbol h. This consonant occurs mainly initially, although it is also sometimes found medially, eg headlahead, huiltuphill. [is sometimes discounted as a consonant by phoneticians of English. In articulating the h the mouth takes up a configuration for the following vowel sound, ‘and for this reason it is considered to be a voiceless onset for the vowel which follows. But for practical recognition and transcription purposes it is useful to regard it as @ separate sound, and we shall call it a voiceless glottal fricative, symbolized by h. Figure 16: Velar fricative English consonants 15 Plosives Prosive consonants in English involve a complete closure in the mouth, a raised velum preventing escape of sir through the nasal cavity, and plosion after the release of the closure. Figure 17: Bilabial plosive Bilabial plosives occur commonly in English and are represented by the familiar symbols p and b, for the voiceless bilabial plosive and the voiced bilabial plosive respectively. They are found in all positions in English words, eg pearibear, chapelicable, ropeirobe. Figure 18 Alveolar plosive Figure 19: Vela plosive ‘The alveolar plosives are made with a closure between the tongue blade, or sometimes tongue tip, and the alveolar ridge. They are represented by the familiar symbols t and d, for voiceless alveolar plosive and voiced alveolar plosive respectively, They occur frequently in all positions in English words, eg trainidrain, writeririder, matlmad. ‘A final pair of plosives in English is made at the velar place of articulation, involving a closure between the tongue back and the solt palate. The voiceless velar plosive is symbolized by k and the voiced velar plosive by g. Like the other plosives they are found commonly in all positions in English words, eg cane/gain, pack- agelbaggage, rickirig.16 Analyzing English In many cases the voiceless plosives in English are articulated with aspiration ‘accompanying the release phase of pronunciation. This means that when plosion ‘occurs, a puff of air accompanies it, So we talk about aspirated and unaspirated ‘voiceless plosives, and aspirated voiceless plosives are symbolized pt, th, KS. The aspiration of voiceless plosives in English isa more-or-less phenomenon rather than an either-or one; that isto say, there are a numbar of degrees of aspiration rather than aspiration versus non-aspiration. Strong aspiration is typical when voiceless plosives occur initially, and a complete absence of aspiration typically occurs when voiceless plosives follow s, cfpin/spin,un(sin,Rilliskill. A lighter aspiration is often present when voiceless plosives occur medially or finally, eg happerthope, latter! le, packeripack. There is one further plosive that we should include—the glottal plosive, or glottal stop, that we mentioned in Chapter 1. [tis not really a plosiv, in that no observable plosion occurs; hence ‘stop’, the alternative name for plosive, seems more “appropriate inthis instance. Only a voiceless glottal plosive occurs; its represented by the symbol 9. Apart from its widespread use in Cockney and some other accents, its oceurrence in English speech is restricted to special uses. It may be heard in the pronunciation of words where two vowels follow each other but belong to separate syllables, eg co-operate, re-action. Here a glottal stop is often used to separate the two vowels, It may also be heard on occasions as a reinforcement of the articulation of a final voiceless plosive, eg in leap ot leak. ASfricates| ‘One pair of sounds in English is related to both plosives and fricatives: the palato-alveolar affricates. The voiceless palato-alveolar aifricate may be illustrated by the inital and final sounds in church, and the voiced by the initial and final sounds in judge. Atfricates involve a complete closure, as for plosives, but the release phase is not with plosion but with friction, In the case of the palato-alveolar affricates the closure is made with the tongue blade and front at the alveolar ridge and hard palate area, and the release is by means of a palato-alveolar fricative. ‘These sounds are symbolized by t{ for the voiceless palato-alveolaraffricate, and by Figure 20; Palato-alveolar afreate 4 for the voiced palato-alveolar allricate. They occur in all positions in English words, eg chinigin, richesiridges, lunchilunge. English consonants 17 Nasal Like plosives, nasals involve a complete closure in the mouth; but unlike for plosives, the soft palate is lowered so that air may escape through the nose. Nasals, Unlike plosives, but like fricatives, are continuant sounds. This means that their articulation may continue for as long as the person speaking can find breath: compare mmm, ff, p. In English, nasal consonants are normally voiced; voiceless nasals do sometimes occur in some contexts, but they are then variants of normal voiced nasals, and may be ignored for our purposes. [English has a bilabial nasal, represented by the familiar symbol m, which occurs in all positions in English words, eg mind, limit, rhyme, ‘The labio-dental nasal, involving @ closure between bottom lip and upper teeth, is represented by the ‘symbol 1). This isnot a regular sound in English, occurring only as a variant of the bilabial m or the alveolar nasal n before labio-dental fricatives, eg in symphony and in infant. And then it occurs really only in fairly rapid speech, It will be mentioned again in Chapter 5. |As already implied, English has an alveolar nasal, represented by the familiar symbol n, which occurs in all positions in English words, eg nigh, lantern, line. Finally, English has a velar nasal, involving a closure between the back of the tongue and the soft palate, It is reprosented by the symbol n, and is of restricted occurrence, never being found initially. It occurs frequently before a velar plosive, and in some accents of English, eg in the Midlands, itis always followed by a velar plosive. In other accents it may occur finally as well as medially, eg linger, fongue, singer. Laterals Like nasals, the lateral consonants are also continuant sounds: there is a complete closure, but air escapes over the side of the tongue. In English laterals are normally voiced, although there is a voiceless variant that regularly occurs in a particular context and which will be discussed later (Chapter 4), The only lateral occurring in English isthe alveolar lateral, represented by the familiar symbol |. Itis found in all positions in English words, eg lav, filler, rail. In fact, there is a marked difference in {quality between the alveolar lateral in inital position and that in final position. The difference arises from the configuration of the body of the tongue in the mouth: for the alveolar lateral in final position, the back of the tongue is raised towards the soft palate, and it is said to be velarized, The velarized alveolar lateralis symbolized by ‘igure 21: Alveolar letral ~ non-velarized and velarized18 Analyzing English compare leafifel, lateiil. The details of the contexts in which the two varieties ‘occur will be discussed in Chapter 4. Frictionless continuants Last in our review of English consonants we.come to the class of frictionless ‘continuants, the sounds involving a near-closure in the mouth but without friction, ‘The normal English r sound is a post-alveolar fritionless continuant: a near-closure is made between the tongue blade and the area just behind the alveolar ridge. It is represented in the Intemational Phonetic Alphabet by the symbol J, but for purposes of broad transcription (see Chapter 4) it may be more convenient to use the usual orthographic symbol. It is of restricted occurrence in English, not normally being found in final position eg run, free, furrow. The orthographic symbol appears frequently in final position in English spelling, but careful ‘observation will reveal that in British English itis not normally pronounced in that position, eg car, more, bear, farm, although it is pronounced here in many ‘American aecents. Other kinds ofr sound are also found in English speech and will be discussed later (Chapter 4). The postalveolar frictionless continuant is reckoned to be the normal English sound. It is usually voiced, although a voiceless variant does occur in some contexts (see Chapter 4), ‘Two further sounds are included among the frictionless continuants, although they are sometimes also called semi-vowels. They have similarity with particular vowel sounds of English, and were it not for the fact that they pattern like consonants in the structure of English words and syllables, they would be counted as these vowels, One is the bilabial frictionless continuant, represented by the familiar symbol , which is similar to the elose back rounded vowel, as in boot. This occurs in initial and medial positions, eg wet, cower. But in final positions its occurrence in spelling is usually considered to be a close back rounded vowel. ‘The other vowel-like consonant sound is the palatal frictionless continuant, where the near-closure is between the tongue front and the hard palate. It is represented by the symbol j, and it is similar to a close front spread vowel as in beat. The ‘occurrence of the palatal frictionless continuant is restricted: it is found in initial position, and in initial consonant combinations after certain consonants, ¢& ye, beauty (bj), queve (kj). Both these frictionless continuants are normally voiced, By way of summary the consonants of English are set out in a chart below. Place of articulation is indicated across the top of the chart, and manner of articulation down the left-hand side. If a particular kind of consonant has a voiceless and a voiced variant, then two symbols appear in the relevant box in the order voiceless followed by voiced. Consonants that occur only as variants of other regularly ‘occurring sounds have been omitted from the chart English consonants 19 ‘igure 22: English consonants ] a) 3 ay | al ails atl y A | 3 | 2 se) s]a}é als eli] alii fv | @6 | sz ss | b pe ta ke Afriate v4 Nal = [el] 3 Tatra [a Frionles | wy 1 fi Continuant | _ _ Exercise? ‘Transcribe the following English words in the symbols of the International Phonetic Alphabet charted in Figure 22 to represent their pronunciation. Since every word has a ‘owel sound, # number of vowel symbols will b needed. The following vowels have been {sed in the words listed: | as peep, asin pit a8 in bet, anda asin the educated southern [English pronunciation of pat cf then Sen; jar dg; lick hk 1, shed 2 teeth 3. chart 4. detest 5. jet 6. guard 7. these 8 barge 9. vase 10, chet 11 peach 12, thieve 1B. yeast 14, green 15, Mars 16. wealth 17. beet 18, wrench 19. crease 20. charge 21, shriek 2. fling 23. stink 24. yield3. English vowels In Chapter 1 we defined vowels as sounds which are made without any kind of closure or impediment to the escape of air through the mouth. Because there is no contact oF near-contact of articulators, vowel sounds are more difficult to deseribe than are consonants. But, as with many consonants, the tongue is crucial in determining the quality of vowel sounds, and it i the differing configurations of the tongue in the mouth that cause the varying qualities, As we mentioned in Chapter 1, there are three parameters by which we deseribe vowel sounds: firstly, the height of the tongue or openness of the mouth; secondly, the area of the mouth having the highest part of the tongue or the general area of the mouth in which the vowel is made; and thirdly, the shape of the lips. The frst of these gives us the categories lose, half-close, half-open and open; the second the ‘categories front, central, and back; and the third the categories rounded and spread. Vowel sounds are, of course, normally voiced, so that voicing is not relevant for their description. Before we look at the English vowels in detail we should mention a broad distinction to be made among vowel sounds between pure vowels and diphthongs. ‘This is particularly relevant for English, since the proportion of diphthongs is unusually large. Pure vowels are made with the mouth taking up a single position during the articulation. In the ease of diphthongs the configuration of the mouth changes in the course of the articulation of the vowel sound. As we shall see, «diphthongs can be deseribed in terms of the pure vowel from which the articulation starts, and the pure vowel in whose direction the articulation moves. Pure vowels First of all, we shall consider the pure vowels. In English, pure vowels are made in all three areas of the mouth—front, central and back Front vowels Inthe front area English has four vowels; all are made with spread lips. Two of the front vowels are in the close area. One is the lose front spread vowel found in beat or sheep and symbolized by i, so these words would be transeribed phonetically as ‘biv and sip. The other is a more open and a more central close front vowel, usually described as a lowered and centralized close front spread vowel. It is represented by the symbol 1, and it occurs in bir /bit) and ship IS English has no vowel at precisely either the half-close or the halfopen position, but ‘one mid way between the two. This is usually described as a mid front spread vowel It is sometimes symbolized by ¢, which is the International Phonetic Alphabet English vowels 21 symbol for a half-close front spread vowel, but itis more usually symbolized by €, the IPA symbol for a half-open front spread vowel, and this is the symbol we shall be using. This vowel occurs in ber /bet! and jeich ite ‘The other English front vowel isin the open area, but it is not completely open: itis ‘somewhere between half-open and open, although itis usually described as an open front spread vowel. However, the IPA symbol for an open front spread vowel (a) is, not used, but rather the symbol z. It occurs in bat bet! and catch Jka} Back vowels In the back area of the mouth we can recognize five vowels in English. Four of them ‘are made with rounded lips and one with spread lips, There are also two vowels in English that fallin the close back area, One is the elose back rounded vowel, as found in boot and too! and symbolized by u, so these words will be transcribed phonetically as /but/ and /tul. The other is a more open and more central variety, usually described as a lowered and centralized close back rounded vowel. It is represented by the symbol v and occurs in (Southern British pronunciations of) ‘book Sook! and put (pov. {AS in the case of front vowels, English has no vowels at precisely half-lose and. hhalf-open back positions. There is one vowel somewhere between the two, probably nearer to half-open thn hall-close. The IPA symbol for a half-close back rounded vowel (0) is sometimes used to symbolize it, but more usually the IPA. symbol for a half-open back rounded vowel is used, 9, It may be described asa mid back rounded vowel and it occurs in boughe lot and law ia! In the open back area two vowels are found in English, one made with rounded lips, the other with spread lips. The open back rounded vowel is, like its front counterpart, not completely open, but somewhere between half-open and open. Nevertheless, the IPA symbol for an open back rounded vowel is used to represent it, p. This vowel occurs in hor /hoV and moss /mos’. The open back spread vowel is completely open but not completely back: it tends towards the central position, The IPA symbol for an open back spread vowel is used to represent it, and it oceurs in farm Vom and cart Skat. Central vowels ‘We come now to the final group of pure vowels, those made in the central area of the mouth. There are three vowels produced in this area in English, all with spread lips. One of these central vowels isin the open area,but like z and o in fact mid-way between half-open and open, although it is described as an open central spread vowel. It is represented by the IPA symbol for a half-open back spread vowel, rnamely a, and it occurs in (the southern British pronunciation of) bur /bat! and some isam, The other two,central yowels in English are both mid central spread vowels, that is mid-way between half-lose and half-open. These sounds may be illustrated by the vowel in bird, represented by the symbol 3, and the final vowel in father,22 Analyzing English Fepresented by the symbol 9. The latter sound is often referred to as the schwa vowel; it occurs in English only in unstressed syllables and has a generally lax articulation. By contrast the 2 vowel occurs in stressed syllables, and is generally {longer in duration than the schwa vowel. They may be fusther illustrated by loser Nusa, ahead fabed/, git igeW, first Nast. This completes our description ofthe tel’ English pure vowels. f you do not speak with a southern Brilsh accent, you may have found that some. the ilstatons given didnot match your ptonenction ofthe word, Indeed sone Midland and Northern speakers nay fad dilfleltt dcovera word Ih ey speech that contains an open central spread vowel (a). The ellerenoos in Eghat Accents ae due tn considerable extent to the vowel sounds, and fn two tee Vowel sounds ifr among accent in their quality, for example, In wnne aceon ‘he mid back rounded vowel is considerably more open atkulation tna described here. in caught ka! And vcondly, vowel sun difersnane accor, in the words in which they occur; for exampley most Northemer sad some Midlanders wil wea lowered and centrlzed clos back rounded vovel Seon, rather than an open central spread vowel (1). Itis impossible to do justice to the variety of British English accents in a work of this kind, not to mention American, Australian and other accents. So, transcriptions will be given as representing Southern British pronunciation, and the reader will be left to determine what the transcription will be for his own accent ‘The vowel chart So far we have characterized English vowel sounds by means of a descriptive label such as ‘close front spread vowel’, ete. Phoneticians also use another means of characterization: the vowel sounds are plotted on a vowel chart. This is a fepresentation of the human mouth (in a squared out form!) and the peripheral Points constitute a set of ideal sounds (called cardinal vowels) against which the sounds actually occurring in a language are compared and their points plotted on the chart. The TPA symbols are those given to the cardinal vowels. The chart of the pure vowels of southern British English is as shown below: Figure nglsh vowels _ V_\ Ha-close iain \2] Open ea English vowols 23 Diphthongs Let us now consider the diphthongs of English, of which cight are usually recognized. They may be divided into three groups according to the vowel towards ‘which the articulation of the diphthong moves. In the case of three diphthongs the articulation moves towards a lowered and centralized close front spread vowel (1); in the case of two it moves towards a lowered and centralized close back rounded ‘vowel (v); and in the case of the remaining three it moves towards the unstressed ‘mid central spread vowel (9). 1 diphthongs ‘Taking each of these groups in turn, we begin with that in which the diphthong tends towards 1 One of the diphthongs in this group begins with a half-close front spread vowel, s0 it is represented by the symbol e1. The beginning point of this diphthong is then more close than the articulation of the mid front spread pure vowel of English (), ‘This diphthong is found in late leit! and blame (ler. AA second diphthong in this group begins with an open front spread vowel and is ‘represented by the symbol a1. The beginning point is more open than the articulation of the English pure vowel described as open front spread (2). It occurs in sight Isatt) and mice may ‘The third diphthong in this group begins with a half-open back rounded vowel; it is sometimes represented by the symbol 21 and sometimes by the symbol ni, since the half-open position is mid way between the English pure vowel symbolized and the vowel symbolized 0, the former being mid way between half-close and half-open, and the latter mid way between halfopen and open. We shall be using the latter symbol, 01. This diphthong occurs in soil soil and boy Show, v diphthongs ‘The second group of diphthongs comprises those tending towards v ‘The first in this group starts with the unstressed mid central spread vowel and is represented by the symbol 9v. This diphthong occurs in home /hovmy/ and in boat Ibavu. ‘The other in this group starts with an open central spread vowel, a vowel which slightly more central than the English pure vowel described as open back spread; and in fact the symbol used for this pure vowel is also used in the symbolization of the diphthong, av. This occurs in house /havs! and ow! fav). {It will be noticéd that in making both the diphthongs of this group, not only is a cchange in the shape and height of the tongue involved, but also a change in the shape of the lips, moving from a spread vowel to a rounded one.24 Analyzing English 2 diphthongs ‘The third group of diphthongs tends towards the mid central spread vowel 0 First there is a diphthong beginning with the lowered and centralized close front spread vowel and represented by the symbol 19. This occurs in fierce ifs! and clear kl. " ‘A second in this group begins with a half-open front spread vowel, that is with a ‘vowel more open than the pure vowel described as mid front spread. The symbol of this pure vowel is used in the diphthong and is the LPA symbol for a half-open front spread vowel. So the diphthong is represented by the symbole and occurs in scarce Iskeos! and fair (feo. ‘The remaining diphthong of this group begins with a lowered and centralized close back rounded vowel and is represented by the symbol v2. It occurs in cruel /krval/ and pure Ij. ‘The eight diphthongs of English may be plotted on a vowel chart: Figure 24: English diphthongs Front Central Back se Hale-close Hattopen Open xereise 3 “Make a transcription ofthe pronunciation ofthe following English words, using the symbols iscussed. The Key will give # representation of a southern British pronunciation; you may ike t0 also consider what the transcription should be in your own accent tg humble tambo, lathe ile, badge Soa! 1. caught 2. another 3. faint 4. sock 5. bag © cover 7 tam 8 fuel 9. count 410. look 11. catching 12. te 413. large 14. bee 415. dozen 16. dirty 17. share 18. goal 19. food 20. flesh 21, fear 22, fall 23, morning 24, pleasee 25. confess 26, moist 27. plain 28. salt 29. lurking 30, climb | oGtnctine Gotures 2 (oo Demers 4 Farmer) English vowels 25 ‘Summary ‘Asa summary of the last chapter and the present one, here is list ofthe 44 sounds ‘of English that we-have. identified and described. Each one is given with its symbol followed by its label. Consonants voiceless labio-dental fricative voiced lahio-dental fricative voiceless dental fricative voiced dental fricative voiceless alveolar fricative voiced alveolar fricative voiceless palato-alveolar fricative voiced palato-alveolarfricative voiceless glottal fricative voiceless bilabial plosive voiced bilabial plosive voiceless alveolar plosive voiced alveolar plosive voiceless velar plosive voiced velar plosive voiceless palato-alveolar affricate voiced palato-alveolar aflricate bilabial nasal alveolar nasal velar nasal alveolar lateral bilabial frictionless continuant post alveolar frictionless continuant palatal frictionless continuant. Seg -segganrateu raSN aces Vowels close front spread vowel lowered and centralised close front spread vowel mid front spread vowel ‘open front spread vowel close back rounded vowel lowered and centralized close back rounded vowel ‘mid back rounded vowel ‘open back rounded vowel ‘open back spread vowel id certral spread vowel ‘unstressed mid central spread vowel ‘open central spread vowel26 Analyzing English 1 diphthong beginning with half-lose front spread vowel, moving towards lowered and centralized close front spread vowel 1 diphthong beginning with open front spread vowel, moving towards lowered and centralized close front spread vowel DI diphthong beginning with half-open back rounded vowel, moving towards lowered and centralized close front sproad vowel 90 diphthong beginning with unstressed mid central spread vowel, moving. towards lowered and centralized close back rounded vowel ‘90 diphthong beginning with open central spread vowel, moving towards lowered and centralized close back rounded vowel 19 diphthong beginning with lowered and centralized close front spread vowel, moving towards unstressed mid central spread vowel 69 dipthong beginning with half-open front spread vowel, moving towards unstressed mid central spread vowel 'v9 diphthong beginning with lowered and centralized close back rounded vowel, ‘moving towards unstressed mid contral spread vowel. xercise 4 Make transcription of the pronunciation ofthe following English words, using the symbols sven, 1. playground 2. Mexible 3. Chinese 4. Snudgery 5: insinuation 6 pleasurable 7. blinkers 8. understate 9. Search-warrant 10. blackmail 11. migration 12 frequency 13, robust 14. forceps 15. paranola 16. birthday 17. catedstor 18. writhing 19, cures 20. overestimate 4, Sounds and their variants — In the preceding chapters we have in general assumed that each sound that we perceive as a separate sound in English is transcribed by just one symbol: p sounds are transcribed by /py, | sounds by /V, and so on, At the same time we have also noticed that some such perceptually different sounds may, in different contexts, have different phonetic values. We have distinguished, for example, between a non-velarized and a velarized alveolar lateral (I and 4); and we have noticed aspirated and unaspirated varieties of voiceless plosives (eg p and p). But in ‘transcription we have not taken account of these finer differences, although we could have done. If we had done so, we should have produced a narrow transcription instead of the broad one we have been making, Broad and narrow transcription A broad transcription is one that only takes account of the sound differences that are important to distinguish words from each other in a language. The distinction between [p4] and [p] does not make a difference between words in English, If we substitute [p] for [p"] in [pln] we produce a peculiar pronunciation of pin but not a new word; similarly, if we substitute [p] for [p] in [spin]. But the substitution of p for t does make a difference of word: pin /piw/ and tin /tn/ are different words in English. A narrow transcription attempts to represent more of less accurately the ‘way in which a particular speaker pronounces his words. Obviously the amount of Phonetic detail that may be represented can vary enormously; that is to say, there are degrees of narrowness of transcription, and a narrow transcription may be made for the pronunciation of a group of speakers, with a particular dialect or accent. ‘This distinction is, in fact, not merely a distinction between different kinds of transcription; its also.a distinction between different ways of looking at the sounds of a language. When the sound system of a language is viewed as a system of units for maintaining distinctions between units on a higher level of linguistic organiza tion (ie words), we call the sounds phonemes. The sounds that people actually make ‘when they speak are called phones. Its part of the task of phonology (or phonetics) to establish and describe the relationship between phones and phonemes in a language, Phonemes and allophones ‘The phonologist starts with phones, the accurate phonetic transcription of the sounds that hé hears speakers making, But it soon becomes clear that not all the distinctions made at this level are relevant for distinguishing words from each other. Or, to put it another way, native speakers do not perceive as distinct sounds all the28 Analyzing Engi phones that the phonologist has recorded. Several phones may, therefore, belong to the same phoneme; or a phoneme may have a number of variants, called allophones. In grouping phones together into phonemes, three criteria are used by the phonologist: complementary distribution, free variation, and phonetic similar- ity, Complementary distribution ‘The criterion of complementary distribution states that if two or more phones occur in non-identical environments, then they may be members of the same phoneme. So, we find that for English the non-velarized alveolar lateral (I) occurs before vowels, while the velarized alveolar lateral [#] occurs before consonants and at the cends of words. Their distribution in English words is complementary; they do not occur in the same environment, Hence these two I-phones belong to the same phoneme. Or, we can say thatthe /V phoneme has wo allophones: (1) [1] occurring before vowels; and (2) *] occurring before consonants and word-tinaly. From this statement you will notice the following convention: a broad transcription ie in terms of phonemes, is enclosed between slashed brackets / /; and a narrow transcription, ie in terms of phones, is enclosed between square brackets }. Taking ‘contrary example, we find that the phone [p] and the phone [0] both oceur in a umber of identical environments eg onipun, shopishor, sippinglsiuing. Their distribution is, therefore, not complementary and they cannot be considered as belonging to the same phoneme. Free variation ‘The criterion of free variation states that if two or more phones occur in the same environment, but without changing the word in which they occur, then they may belong to the same phoneme, So, in English, a word like bid is sometimes articulated with @ fully voiced final alveolar consonant [4]. and sometimes with @ devoiced (voiceless) final consonant, represented by [d]. But whichever phone is used, the word is the same. These two phones, then, belong tothe same phoneme. ‘Or we can say that the phoneme /d! has two allophones when it oceurs word-finaly () [al and (2) {J}, which are in free variation. Iti also the case that a devoiced velar plosive [g] occurs word-finally; so that we eould hypothesize that {d) and (g] were in free variation, But we soon realize that the substitution of one forthe other makes a difference in the word: bid [bid] and big [big] are diferent words. Phonetic similarity ‘These two criteria—complementary distribution and free variation—by themselves Teave a loophole, which has to be filled by the criterion of phonetic similarity. In English, on the basis of the criterion of complementary distribution, the voiceless alottalfricative (h) and the voiced velar nasal (9) would be members of the same phoneme, since [hi] occurs only word-initially and [9] occurs only word-medially and word-finally. But, quite obviously, it is ridiculous to suppose that these two sounds have anything in common: a glance at their descriptive labels reveals that i Ve | | Sounds and their variants 29 they share nothing—not in terms of voicing, nor place of articulation, nor manner of articulation. This insight is formalized in the criterion of phonetic similarity, which states that if two or more phones are.to be members of the same phoneme they must be phonetically similar. In practice, this means that the allophones of a phoneme usually share at least two of voicing, place of articulation and manner of articulation in common. English phonemes and allophones ‘A phoneme may be defined, then, as a class of phonetically similar phones in complementary distribution or in free variation or in both. Remembering that transcription in terms of phones is a narrow transcription and is enclosed in square brackets, and a transcription in terms of phonemes is @ broad transcription and is enclosed in slashed brackets, we shall now turn to a detailed consideration of English sounds (phonemes) and their variants (allophones) Plosives We begin with plosive sounds, which asa class show more variation than any other. ‘We have already noted, in Chapter 2, that voiceless plosives in English (pt, k) may: be either aspirated (accompanied by a puff of air on release) of unaspirated. ’That is to say, the phonemes /p,/¥ and /ki each have an aspirated allophone [p(t] and {ki}, and an unaspirated allophone [p}, {t] and [k]. And these allophones are in ‘complementary distribution: the unaspirated allophone occurs after /s/ and the aspirated allophone occurs in all other positions. As we pointed out in Chapter 2, though, the degree of aspiration varies from environment to environment, There is strong aspiration when the plosive occurs initially in a stressed syllable eg in pin, tn, kin, Between vowels the aspiration tends to be rather weak, and may indeed be absent altogether eg in upper, utter, sucker: we could say that we have @ ease of allophones in free variation in this instance. And in final position the amount of aspiration is variable, depending whether the plosive is released fully or not eg as in sip, sit, sick. The voiced plosives in English, d, g/ may be considered to have two allophones. ‘The main allophone, and the one which occurs in all environments, is the fully voiced [b], [4], [f]. The other allophone occurs only in worde-final position and is & devoiced variant (b), [0 [QI isin free variation with the fully voiced allophone, that is, either of them may occur in this position with no particular factor conditioning the occurrence of one rather than the other, excep forthe speakers whim atthe time eg asin ib, id ig. ‘The difference between /p,t,k/ and /b, d,g/ in English is not merely one of voicing, or even unequivocally one of voicing; that is, the distinetion between voiced and voiceless. It is also a difference of aspiration. The one environment in which unaspirated voiceless plosives regularly occur, ie after /, is the one in which voiced plosives do not occur: there are no English words beginning with sb, /sd/ oF /s ‘There is, therefore, no contrast between voiceless and voiced plosives in English in this environment. In all other environments (initially, finally, intervocalicall) /p,30 Analyzing English W/ and /b, d, g/ are distinguished both by voicing and usually also by aspiration. In fact, aspiration might be said to be more important in distinguishing /p, tk! (or words containing them) from /b, d, g/ (or words containing them), ef pipibib, suid, ickigig. For this and other reasons, the series /p, ,k/ are often referred to not as ‘voiceless’ plosives, but as fortis plosives; and the /b, d, series not as ‘voiced plosives, but as lenis plosives. The terms fortis and lenis refer to the muscular energy used in articulation: fortis consonants are more energetically articulated than lenis conson- ants. These terms are used not only for plosives, but also for fricatives and affricates in English, ‘The variants of plosive that we have dealt with so far involve the whole class of, plosives in English. We now consider variants of individual members of the class. Besides the allophones already mentioned, the alveolar plosives each have a further allophone. In words like widih /wd8/ and eighth fut the articulation ofthe plosive is not alveolar, but dental: the tongue tip makes @ closure with the back of the upper front teeth, rather than with the alveolar ridge. In other words, the articulation of the plosive is the same here as that of the following dental fricative, and this allophone occurs only in this environment. So /d! has an allophone [d] before a dental fricative, and /V has an allophone [}] before a dental fricatve. Velar plosives also have variants in addition to those already mentioned. Compare the articulation ofthe velar plosives in the following words: keepigeese, cupleum, ooplgoose. By careful observation you will feel thatthe contact of the tongue with the roof of the mouth is made in slighty different position for each of the three Pairs of words. The closure ofthe velar in the frst pair is more forward than that of {he second pair, and the closure of the velar in the lat pair is further back than that in the middle pai. The conditioning factor is the nature ofthe vowel that follows the inital velar plosive: a front vowel, asin keep/kip/ or geese /eis! tends to pull the articulation forward, and a back vowel, as in coop /Kup’ and goose igus, tends to push the articulation back. So if we regard the articulation of the velar before central vowels, asin eup /kapy and gum /gatn, as genuinely velar, we can call the articulation before front yowels ‘pre-velar" and the articulation before back vowels ‘post-velar’, So the phoneme /k/ has an allophone {k] or [kt] before front vowels, and an allophone [k] or [k#] before back vowels; and the phoneme /g/ has an allophone [8] before front vowels and an allophone [g) before back vowels. One further variant needs to be mentioned for plosives, When a voiceless (or fortis) alveolar plosive /¥’ occurs in final position in a syllable, it may be articulated as a glottal stop cg in suit So the phoneme // has allophone (?] occuring in syllable final Position, and which isin free variation with [t) Larerals We have already mentioned the fact that the alveolar lateral /V in English has two allophones, a non-velarized allophone [1], occurring before vowels and /j as in leave, early, illuminate, and a yelarized allophone [H], occurring before consonants, Iw! and word-finally (ie before pause), as in altogether, always, pool. These ‘Sounds and their variants 91 allophones are, then, in complementary distribution. There is, in addition, a third allophone which isin complementary distribution with these two: after a forts (ie voiceless) plosive /p, k/ the alveolar lateral in English is devoiced (ie voiceless), eg in play, clean. So the phoneme /V has the allophone {I} after /p, W/. There is yet one more allophone of /V/ in complementary distribution with the others: like the alveolar plosives, the alveolar lateral has dental articulation when it occurs immediately before a dental fricative, as in filth, stealth, although. Since this ‘occurrence is before a consonant, the allophone is also velarized: (I) P sounds In (Chapter 2) we said that a number of ‘sounds occur in English speech. Itis possible to recognize three regularly occurring allophones of the /t/ phoneme in English, The allophone with the widest distribution isthe post-alveolar frictionless. ‘continuant [1]. After fortis consonants pt, k, f, 9, J ete a devoiced (voiceless) allophone occurs [J] og in tain, crane, free, shrug. These two allophones are in complementary distribution. The third possible allophone is a flapped consonant Ir], which may occur intervocalically eg in very, marry. This allophone is in free varition with the voiced post-alveolar frictionless continuant. While other ‘r= sounds do occur in the speech of English native speakers eg a rolled consonant // or ‘ uvular roll /R/, they do not occur regularly enough for them to be included in a general description of English sounds. Nasals ‘Among the nasals we have to consider two variants. Like its alveolar counterparts, among the plosives and laterals, the alveolar nasal has a dental articulation before ental fricatives as in plinth, ninth, So the phoneme /n/ has an allophone (9) occurring before dentals, and an allophone [1] occurring elsewhere; they are in complementary distribution, The other variant occurs as an allophone of both the bilabial nasal and the aveolar nasel. When these occur before a labio-dental {ricative if, v! 9s in symphony, convent they are often articulated as a labio~dental ‘nasal (m)]. This allophone is probably in free variation with the normal allophone [m] and [n]; its occurrence is more likely, the more rapidly a person speaks. So the phoneme /mm/ has an allophone [ny] before labio-dental fricatives, which is in free variation with the normal allophone [m]; and the phoneme /n/ has an allophone [79] before labio-dental fricatives, which is in free variation with the normal allophone (nl. Fricatives Lastly among the consonants we come to the fricatives. Like voiced plosives, voiced fricatives have a devoiced allophone occurring word-finally (before a pause), which is in free variation with the normal fully voiced allophone eg rise [2], live [y], seethe [B]. The palato-alveolar fricatives /f, 5/ present at an interesting case for the application of phoneme theory to English. They could almost be regarded as allophones of the same phoneme. The voiced palato-alveolar fricative (5/ occurs32 Analyzing English only intervocalically, except for a few loan-words from French, where it may occur initially and finally: although, as pointed out earlier, many speakers substitute a palato-alveolar affricate in these positions. Moreover, there is hardly & pair of words in which /f/ and i5/ contrast; the nearest is perhaps fission and vision. However, the fact that /f/ does occur intervocalically eg also in mission, cushion, ‘bushel, means that the two sounds cannot be eonsidered to be in complementary distribution or in free variation, and so must be separate phonemes, Vowels ‘The English vowels can be divided into two groups: a number of the pure vowels and all the diphthongs are in general longer in duration than the remainder of the ure vowels, A comparison of seat and st wll make it plain that the vowel of seat i! hhas a longer duration than the vowel of sit /i/. The set of short vowels is /,¢, 4,5 , a/ and the set of fong vowels /:,u:, a:,9:,2:/ together with the diphthongs. The symbol : denotes length. Some of these vowels may be matched into pairs of complementary long and short vowels eg i,t, /u:, 8 93, D, /:,a/ but it is important to note that the distinguishing factor between these vowels is primarily’ one of quality determined by difference of articulation, and not one of length. ‘The short vowels have no variants. The long vowels and diphthongs, however, vary in Iength according to the environment in which they occur. They ean be considered te have two allophones: (1) a short allophone before a fortis consonant ie /p, t,k, tf, 4,0, 15 (2) along allophone in all other environments. Compare, for example, the length of the vowels in bead/bear, rude/root,cardlcart,cause/course, curdicurt. The {ong allophone may be marked with a length mark eg [i] and the short allophone without eg [i]. In broad (phonemic) transcription, there will be no need to indicate the length of long vowels at all, since it is not a phonemically distinguishing characteristic. But in a narrow (phonetic) transcription the long vowels will need to be marked for length when they occur in environments where they are relatively long, ie not before fortis consonants. Neutralization One further topic remains to be dealt with under the heading of sounds and their variants: neutralization, We have referred to phonemes as contrastive units of sound: the substitution of one phoneme for another will result in & different word being formed, ef pinibintinisinithin, butbetlbatbootlbaitiboatlbutlbitetboughtlbeat etc. Sometimes, however, in a particular environment, the contrast between phonemes in a particular set will be neutralized: the normal contrast will not Operate. We have already mentioned an instance of this in English, though without using the term ‘neutralization’, ‘This was the neutralization of the contrast between fortis (voiceless) and lenis (voiced) plosives after /s/. The plosive after Is) is ‘unaspirated and voiceless: the firs feature it shares with /b, d,g/ the second with /p, 1, W/. The contrast between /p, t, k/ and /b, 4, g/ is neutralized in this environment, since there are no words in English that differ from each other by the fact that one begins with /sp/ and the other with /sb/ (or /st, /sd/ and /sk/, sg/). In fact, the phone | ‘Sounds and their variants 38 is ne of either the fortis ‘occurring in this environment could be considered an allophone of either t set of the lenis set. It is usually allocated to the fortis set, however, and this is reflected in the orthography. ‘Another case of neutralization occurs in Englishi among the nasals /m, n, y/ when they occur before fortis plosives Jp, t, K/. This is a slightly different case from the preceding one. Here the neutralization of contrast between fm, /n/ and /p) arises from the fact that which of these nasals occurs is totally predictable from the following plosive: a bilabial nasal occurs only before a bilabial plosive, an alveolar nasal before an alveolar plosive, and a velar nasal only before a velar plosive, as in limp, lint, link, That is to say, only the combinations mp, /nt/ and /gk/ occur, and not the combinations /mU, /mk/,/np!,/nk/, pl, From what has been said in this chapter it will be clear that the list ofthe 44 sounds fof English at the end of Chapter 3 represents an inventory of the phonemes of English, those units which are regarded as being contrastive, making a difference in ‘words if one is substituted for another. By way of summary for this chapter, there now follows the same list, together with the allophones of each phoneme: Consonants ‘pv [p] after /, [pH] elsewhere ‘bv [p] word-fnally (free variation), [b] normally iy [t) after i, [0] betore 1, 01, (41 word-fnaly (free variation) [1] ekewhere ‘a [g) before 8, 6, [d] word-fnally (fee variation), [) elsewhere ‘ki [before front vowels, (k] before back vowels, [k] after I, {k¥] elsewhere ig! f] before front vowels, [g} before back vowels, [g) word-finally (Free variation), [z] elsewhere wt iv! [y] wordefinally (free variation), [v] normally (0) 8! (] word-finally (free variation), [6] normally iw fs) ‘rd [z] word-fnally (free variation), [2] normally aw 3) tw [h] im [m] before /f, v/ (free variation), {m] elsewhere In! [p] before /®, 6/, [mp] before /f, w/ (free variation), [n] elsewhere ‘a fo) 1 [HJafter/p, Ws {1} before 1, 0, [t] before consonant, pause, [I] elsewhere ‘1! ({)after fortis consonant, [| intervocalially (fee variation) (1] elsewhere iw! [w] wv fil Vowels [i] before fortis consonant, [i:] elsewhere94 Analyzing English wt a! Ie} ‘a! [e) ‘ai [a] before fortis consonant, a: eewhere ‘oo ‘3! [3] before fortis consonant, [>] elewher a [Ol 7 a ‘[u] before forts consonant, u:] elsewhere a) bl ‘a) [3] before fortis consonant, [:] elsewhere tl ‘ aoe ‘feu! [et] before fortis consonant, [et] elsewhere ‘au’ [at] before fortis consonant, [ats] elsewhere You! [o1) before fortis consonant, [o1:) elsewhere ‘av! [90] before fortis consonant, [2v:] elsewhere ‘avi [av] before fortis consonant, [av] elsewhere ‘io! [19] before fortis consonant, [9:) elsewhere ‘ea/_ [va] before fortis consonant, [e9:] elsewhere Jv2/_ [v9] before fortis consonant, [v9:] elsewhere, Exercise § Make a broad (phonemic) and narow (phonetic) transcription ofthe following English pret fre (bed), pune ope [Pap shaving wick SermnscW ex :wpsh) freak ‘wealth 5. kee & plinth 5. Thursday 6 achive 7, forgeous 8 comfortable 5: pereusion 10. fre 11 aleahot 12, thirty reject 14 15, especialy 16. mis 0 1 paramount 9. conversion 20. penne 21 favourite 22, recive tosure 24, ablaze 25. sreech 5. Sounds in connected speech All the variations in pronunciation that we considered in the previous chapter were inthe pronunciation of words a units in isolation, Changes such as /n! to [n] before 10) we accounted for by word-internal factors, or by the assumption of pause at the ‘end of the word, as for example in the conditions for the occurrence of a velarized alveolar lateral [¥], But we rarely speak in single word utterances. And in studying the phoneties of a language we have to take account of the kinds of changes in pronunciation that take place as a result of words being strung together. Most of these changes occur at word boundaries and affect the final end inital sounds of words. Assimilation Some of the allophonic variation which occurs within words, that we discussed in the previous chapter, occurs also at word boundaries. Alveolar consonants at the end of a word have a dental articulation if the intial sound of the following word is 4 dental fricative eg the dental allophone [t] of occurs in no hin (nol Orn). Here there isin fet no aspiration of the alveolar plosive, because itis unreleased, the release occurring through the following fricative. In ten thumps, the dental allopbone [a] of’ occurs: ['e9 Onmpts]}. And in well thought, the dental allophone fH] of /1/ occurs: wet Gat, A second case of allophonic vatiation occurring at a ‘word boundary is the substitution of alabio-dental nasal fora bilabial or alveolar nasal before a labio-dental fictive. In the sequence fen forks, the alveolar nasal may have the labio-dental variant {r) in a fairly rapid pronunciation; and in the Sequence come Jor me the bilabial nasal at the end of come may have the Jabio-dental variant in a rapid delivery. A third case of allophonie variation dceurring at word boundaries as well a8 within words is the incidence of the voiceless (devoiced) allophones of /V and /r ie [I] and J], when these phonemes Cccur after a fortis (voiceless) consonant. So in the sequence af last, the initial alveolar lateral of last is devoiced after the forts alveolar plosive:[3tJast]. And in the sequence a rest, the post-lveolar frictionless continuant is likewise devoiced: ot Jest) ‘There isa further case of variation at word boundaries which could be considered as allophonie variation, but this isnot a kind which occurs within words, This is the case of word-final lubio-dental fricatives having a bilabial articulation when the initial sound ofthe folowing word isa bilabial plosive. This probably only occurs in a fainly rapid delivery. For example, in the sequence rough paich, the final labio-dontal tricative of rough may become [®] in rapid pronunciation: [ra phe]. And in the sequence live bird, the final voiced labio-denalfricative of live ‘may become (Bin apid pronunciation: [Ian b96 Analyzing English ‘The conditions that we gave for the occurrence of one allophone of « phoneme rather than another applied to words spoken in isolation. It happens in a couple of cases that these conditions no Tonger apply ina sequence of words, particularly in relation to wordefinal positon. One of the conditions for the occurrence of velarized alveolar lateral is word-final position; but if an alveolar lateral word- finally is followed by a vowel in intial postion in the subsequent word, then its non-velarized eg in the sequence flit compared with fil: [fi 1, [68]. We said that lenis (voiced) plosives and frcatives tend wo be devoieed in word-fnal position. But it ina sequence, the word following begins witha vowel or voiced consonant, then they will not be devoiced, though they wil be if the inital sound ofthe following ude vote mona Compr ood das [pe et wih ood es iz ‘The influence of one sound on another to become more like itself is called assimilation. Assimilation may be to the place of articulation of a neighbouring sound, as in the dentalizaton of alveolars before dental fricatives, or the substitu- tion of a labio-dental nasal for a bilabial oF alveolar nasal before a labio-dental fricative. Or assimilation may be in voicing, as when voiced plosives and fricatives become devoiced before voiceless sounds or silence (pause), or when /V and iti become voiceless after fortis (voiceless) consonants. All the cases of assimilation that we have considered so far have been cases of allophonic variation: the substitutions have been regarded as allophones of the same phoneme, with the criterion of phonetic similarity being of particular relevance here eg in counting [}] and [Bas allophones of // and !v/ respectively. But there are many cases of assimilation where the substitution is not allophonic but phonemic, the substitution of one phoneme for another. Alveolar sounds in word-final postion are particularly able 10 assimilate to the place of articulation ofthe initial consonant ofa following word; that is, the sounds 1, 5,2. Final is ikely to become pl before bilabial sounds /p,b, mv eg in the sequences shat pen [Dep pPen}, sat boy [Omp bo, that man [Owp msen]. And Final ‘becomes /by before the same sounds eg in good pen [gob plea] ete, Final is likely to become /k/ before velar plosives, g/g, in the sequences that cup [Bzxk i shat gir Ozek go 3]. And likewise ia! becomes before the same sounds eg in good cup [avg k*npH) ete ‘A word-fina alveolar nasal /n/ becomes bilabial /n before bilabial sounds /p,b, and velar iy before velar plosives /k, gl. Examples: en pens [em prenz ten boys {tem bor:2], ten men [Nem menl; ten cups [te keaphs, en girs (te ga Ha). Ira Sequence of alveolar nasal and alveolar plsive occurs worde-finall, then both may assimilate to the place of articulation ofa following consonant, Examples: int in don’ assimilates to the place of articulation of the following iy in don’t be late [ov :mp s lett;/n/in won’ assimilates to/k/ in he won's come {hu wav: kb ms ‘nin ound assimilates tothe bilabial articulation of following ib in he found both {hu fav:mb bv); /nd in kind assimilates othe velar articulation of ina kind gift fo kearing git. Alveolar fricatives /s, 2/ assimilate to a following palatal sound, either palato- ‘Sounds in connected speech 97 alveolar fricative ot palatal frictionless continuant J, j/ and become palato-alveolar fricatives,3/- Examples: this shop [O15 fop) this year [rf jt J; has she [has fi], those young men [80:3 jay men], Assimilations of the kind we have been discussing could, in theory, give rise 0 ambiguous utterances, since in substituting one phoneme for another the new phoneme may already make an acceptable sequence in English. For example, the sequence [lenk*wikt}} could be taken as representing either rang quickly or ran quickly with assimilation of fw to following ‘ki. Or the sequence (woif jo : wert] Could represent either watch your weight or what's your weight with assimilation of Js! to following palatal /j ‘When wordefinal alveolar plosives and fricatives/t, d,s, 2/ are followed by an initial palatal frictionless continuant //, then assimilation is to palato-alveolar place of articulation and the initial /! usually disappears, although for /s, x this is not always the case, as indicated above. The alveolar plosives become palato-alveolar affri- cates tf, d5/ in this instance, and the alveolar fricatives become palato-alveolar fricatives f, 3, as discussed earlier. Examples: woud you [wod3 u: ], what you want [woufu: wont); as yer [eget], in case you need it [iq Kerfus nid 10 Now we come to assimilations involving nasalization. A voiced alveolar plosive /d! ‘may become a nasal when followed by a nasal. Ifthe following nasal is alveolar, /d/ becomes /n, eg in the sequence he wouldn't do it, the Jd! of would becomes an alveolar nasal: [hi wonn(t) du: 10]. There is a tendency for the /V of not to disappear. Ifthe following nasal is velar, then /d! becomes /; eg in the sequence he wouldn't go, the ‘nl assimilate to the following velar /g/, and the /d/ of would then assimilates to the velar nasal following, while the /k/ from /t/ probably disappears {iv wonn(k) gov :]. IF the following nasal is bilabial, then Jd becomes /m/, assimilating again both in place of articulation and in nasalty eg good morning {[gom mo nin]. It is not just word-final alveolar consonants that may be involved in assimilation; for example, the voiced labio-dental fricative /W/ may assimilate in place and nasality to a following bilabial nasal /mi, as in you cam have mine [ju ko hhasm mat in}. In this sequence also the vowel of ean probably disappears (see further below on reduction) and the final alveolar nasal /n/ assimilates towards the glottal articulation offv and becomes velar. Finally we turn to assimilations involving voice. Word-final lenis (voiced) fricatives ‘may become fortis (voiceless) when they are followed by an initial forts consonant, especially ifthe two words are part ofa closely knit group. Examples: / becomes 10) in with thanks [0 Oaey}s), breathe slowly [bliB stv sr; /2! becomes is in these socks [is sok*s], he was sent {In was sent]; 1 becomes /t/ in of course [ot k*9s], ‘we've Jound it (wil fav nd 11}; fg! becomes (tf] in bridge score [bUrtf ska ‘Assimilation isa widespread phenomenon in speech, The instances that we have discussed represent the main ones in English. Quite obviowsly they do not necessarily occur all the time; but statement about assimilation implios that when people speak, and the more rapidly they do so, then these kinds of alteration may {ake place. Since this is so, these variations will need to be taken account of in a harrow transcription of speech, Indeed, it may be argued that the phonemic38 Analyzing English variation resulting from ass should even be accounted for in a broad transcription. But for practical purposes we shall assume that a broad transcription represents words as if spoken in isolation. By way of further illustration, here is an imagined short conversation between a policeman and a motorist in narrow transcription including assimilations (under- lined). . Good morning, sit {om mo:nig 9 Good morning, officer. m ma :niy ofis Your stop lights not working jp stop [ais nop wat teas when left he house. tp woz wen a1: ke f.bo havs —That’s not quite now. Oss nok Kwart nao Which one might tbe? with vam malt 1p bi —The left light doen fate LU have to get the thing fixed. aH hak to get 60 Oxy fist} Buercse 6 Make marow transition of the following sequences of wort, ineuding any poste assimilation, 8 Shut the yard gate behind you [fa D2 jag getp tuhat:ndsus) A. They train cows to jump 2. A badd picture of rose tree 3. Who do you want? 4 He was shipwrecked 5. He's found seven fake pitures Elision Besides assimilation, there are two further processes that cause variation in the pronunciation of sounds at word boundaries: elison and liaison. The process of élision involves the complete disappearance of a sound and takes place only in fairly rapid speech. In such speech it is very common for word-initial fh’ to disappear—and this isa characteristic by no means restricted to cockney speakers! Example: up the hill (4p 01 1], In word-fnal postion it isthe alveolar plosives that are most likely to undergo the process of elision, particularly if they are preceded by another consonant and the following word has a consonant in initial position. The pairs of word-final consonants for which this isthe case can be divided into two groups. The frst of ‘these comprises continuant consonants followed by /t or a: st, ft, nd, 1d, 2d, 8, val, Examples: nest day [nek's det:), lef turn [lef than}, mashed potatoes [maf Potertov:z], bend back [bem bak), hold tight (hav t tart], refused both {ufju 12 bov0], loathed beer [lov :6 bx9 +], moved back {mu : v bk], The second group comprises plosives and affricates followed by /V oF /d: ip, kt, tft, bd, gd, 5d. Examples: kept quiet [k'p! K>wat :a1) liked jam [laik* dseem), reached Paris [ph red bor Lob boo) lagged behind (eg brat: changed colour ter sods kb) Sounds in connected speach 99, \Word-final/U may undergo elision in two further circumstances. I he initial sound of the following word is iV or /d), then the resulting cluster of /t/ of /t! is often simplified to or (respectively. Examples: 've got 1 go (at:v gp to gov: |, what do you want [wo «gu : wont]. The second case involves the negator not when itis reduced to nt, and the /U/ undergoes elision before another consonant. Examples: ‘you mustn't Tose itv masn lu:z.1t, doesn’t she know [dazn Ji:nav:)- This similar to the continuant +/t, d/ cases discussed above. Liaison Liaison, the other process occurring at word boundaries, involves the insertion of a sound that is by and large unmotivated from the phonetic context. The sound is. ‘And it is by and large unmotivated, because, although an /r/ occurs in the written form of some of the words involved, reflecting an earlier pronunciation, the use of sking /rf has been extended to contexts where there is no r in the orthographic form. In the sequences here and there (ial an Sea), father and son {fa :69.1 20 san), fand the far east [Ba fa: 1 ist] an r occurs in the written form, although not if the words concerned are pronounced in isolation: [ha], (fa:60), [fa:}. But in the sequences law and order [lo: 1 an ada} andche idea of i [1 a1:419: J av it) there is no in the orthography and yet liaison takes place Reduction ‘There is one further process that occurs as the result of rapidly articulated speech, but itis not one that occurs at word boundaries. It is called reduction and involves the substitution ofa ‘weak’ central vowel, especially // ori, for a peripheral vowel fe front or back. Reduction occurs particularly in unstressed syllables of words with ‘mote than one syllable, and in single syllable words that have a grammatical (rather than a lexical) function, eg words like is, are, have, of, his, the, does, and, a, an (Syllables are discussed in Chapter 6 and the distinction between grammatical and lexical words in Chapter 9.) Ic is impossible to give a detailed list of all the possible reductions, but the following examples will provide a fair idea of the kinds of cases that may occur. Take the ‘word of, pronounced in isolation as /ov. In connected speech this would commonly be pronounced [av], and may even be reduced further to [v] or [3] eg in wo of mine {tu: o(v) main}. In isolation have would be pronounced /hiev/, but in connected speech the /hy will undergo elision and the front vowel /a/ will be reduced, giving [ov], eg in shey have eaten [Set : ov iton]. The word the, pronounced in isolation as ‘ix/, may in connected speech be either [O1], if it occurs before a word beginning with'a vowel, or [6a] before an initial consonant. In isolation and is pronounced ‘endl, but in connected speech as [ond], or more frequently as [an] or even [n]. In unstressed syllables that end in lor the vowel is normally the unstressed mid ‘central spread vowel /9/ ep in boule fbotot or buuon Vaan’. But in connected speech a further reduction often occurs which causes the vowel to be dropped completely, giving (bot*4] and [btn]. Normally a syllable has to contain a vowel (Gee Chapter 6), but in this case the /V and /n/ are doing service for the vowel and40 Analyzing English ‘are called syllabic A and syllabic ini respectively. The symbol for a syllabic consonant is the small stroke bencath the consonant symbol: fj) [2] As in the case of assimilation, a narrow transcription of a connected sequence of speech will have to take account of the processes of elision, liaison and reduction By way of summary, hete is list of the processes described in this chapter that ‘occur when words are put together in connected specch. Assimilation Word-final alveolars become dental before dental fricatives, Bilabial and alveolar nasals become labio-dental before labio-dentalfricatives. Word-initial /W and /r! becomes voiceless after fortis consonants. nal labio-dental fricatives may become bilabial before bilabial plosives. inal /V is non-velarised if followed by an initial vowel inal lenis plosives and Iricatives are not devoiced if followed by a vowel or voiced consonant. inal /, d/ become bilabial before bilabial consonants. inal /t, d/ become velar before velar plosives inal /n/ becomes bilabial before bilabial consonants. inal /n/ becomes velar before velar plosives. inal /nt, nd/ both become bilabial before bilabials and velar before velars. Word-final /s, 2/ become palato-alveolar before palato-alveolar fricatives and the palatal frictionless continuant. Word-final /t,d, s, / become palato-alveolar affricates (It, dl) or fricatives (Js, z/) before / and /j/ disappears Word-final /d) becomes a nasal before a nasal, at the place of articulation of the nasal. Word-final /v/ becomes a nasal before a nasal Word-final lenis fricatives become fortis before an initial fortis consonant Elision Word-inital Jh frequently undergoes elsion, Word-final /t, d’ preceded by another consonant and followed by an initial consonant undergo elision. Word-final t’ followed by /t, d/ undergoes elision. Liaison Word-final vowel followed by an initial vowel are often linked by insertion of /!. Reduction Vowels in unstressed syllables, or in monosyllabic grammatical words are often reduced to // of i ‘Complete reduction, ie elision, may oceur in unstressed syllables ending in or /n/, giving a syllabic consonant, ‘Sounds in connected speech 41 Exercise 7 Make s narrow transcription ofthe following uiterences, taking account of any asain, stision ltion and reduction, * an fg He'oe the of fof ta: greengrocer i019 oe tra of Sk pteloes) 1. What cid you sy his name 2. We discusted the idea of Twas ast Tuesday 5 He dnt come tote lecture or he would have known a 5 ‘They camped by the river on the other side ‘The masked gunman held the hostage at gunpoint6. Combining sounds Phonemes themselves have no meaning: their only function is to combine together to form higher-level mesningful units; that is, words. But between phonemes and words we have to recognize an intermediate level of phonological organization: the syllable. Many words in English have only one syllable; they are ‘monosyllabic’. But ‘many have more than one syllable; they are ‘polysyllabic’. And yet it is clear that the same principles operate in the combining of phonemes into syllables, whether the syllable itself constitutes a word, or whether the syllable in turn combines with another or several other syllables to form a word. Like phonemes, syllables are in principle meaningless eg sim /sum/ and ple /pal/ of simple, even though in some cases they do eaincide with a meaningful unit, eg in hard—ship, un—do or, of course, in monosyllabic words, Syllabie structure A syllable has a basic three-part structure, composed of a central part and a peripheral part before and after the central one. The central part is almost always a vowel sound and the peripheral parts are consonants. Each syllable has only one vowel, either a pure vowel or a dipthong, at its centre; but the number of consonants in the peripheries may vary. It is possible that one or other or both of, the peripheral parts will not be realised ie there may be no consonants either before or after the vowel, as for example in out favt, foe /tav/ /at!. In English there may, in fact, be up to three consonants before the Vowel and up (0 four consonants after it eg in sirengihs /strepk0s!. Phonemes, then, combine to form syllables, and syllables combine to form words. The maximum number of syllables possible in an English word seems to be about seven or eight eg in in—ter—natio-—na-lizar fon. We have said that a vowel is almost always at the central part of a syllable. The exceptions to this statement are the so-called syllabic consonants, discussed in Chapter 5, which arise from the reduction of a weak vowel to extinction, The consonants which most often become central io the syllable for this reason are il/ ‘nd /n, asin the examples given in Chapter 5: bottle bo, button ib ty In some other languages syllabic consonants are normal central elements, eg a form of rin Czech, Vowels in English never function at the peripheries of syllables, and this provides the reason for treating the ‘semi-vowels’/w/ and /j as consonants: inspite of their vowel-like quality they pattern in syllable structure like consonants, ie they ‘occur in the peripheries of syllables and not in the central part. One practical reason for recognizing syllables as intermediate units between phonemes and words is that they represent the most convenient level at which to describe the way in which phonemes combine together and the restrictions that Combining sounds 43, ‘ccur on combinations of phonemes. The possibilities of phoneme combination are not unlimited: any phoneme may not combine with just any other. There isthe obvious restriction imposed by the nature of syllable structure Consonant-Vowel- Consonant (CVC); a syllable could not be composed of just consonants, nor of more than one vowel or diphthong. But even taking ths restriction into account, some phoneme combinations would be impossible to pronounce eg /pgnust, except perhaps with extreme difficulty. Other combinations, although in principle quite pronounceable, just do not occur in English, eg nufR/. English does not use all the possible pronounceable combinations of phonemes to make syllables and words ‘That isto say, there are certain rules of phoneme combination for English which can be deduced from the syllables and words actually occurring Initial periphery Since the nature of the syllable itself specifies that only one vowel phoneme may ‘occur in each syllable, we can describe the rules of phoneme combination in terms (of the consonants and their combinations that may occur in peripheral positions in syllables. We shall look first at the consonants in initial position in the syllable. Tt is possible that no consonant at all will occur before the vowel (as in oud), and that is the case in English before all the vowel phonemes except two: / and /oa/ never ‘occur without a consonant preceding. All the consonant phonemes of English may coccur singly in syllable initial position except two: / and iq/ never occur by themselves, nor for that matter in combination wit another consonant, in the initial periphery of an English syllable. The permissible combinations of two consonants ia syllable initial position may be expressed by the following diagram: Figure 28: Sllable-inital ewo-consonant combinations als eens f According © this diagram the following syllable initial two consonant combinations are possible: /smv, sn, /st, ski, sp sf, fw, Ow/, lt, fr, er, hw, k,l, pr, Hl, fot, axl, ipl, MU, POV, IgV, Ife, ISw/, VS, making a total of 26. This obviously represents only a very small proportion of the arithmetically possible combinations of the 24 consonants in English, which would amount to 576, assuming that each consonant would be allowed to occur with itself. The restric- tions are even more remarkable when we turn to combinations of three consonants in syllable initial position. Here the arithmetical possibilities would be 13,824, but the actually occurring combinatiqns number six, of which one is found in only one44 Analyzing English ‘word. The following diagram represents the permissible three consonant combina tions: Figure 26: Syllable-initiatthree-consonant combinations The six three-consonant combinations are, then: /spV,/spr,/st, st! /kr/,/skw ‘The combination /sfr/ occurs only in the word sphragistics, the study of engraved seals xercie 8 Find English words to illustrate exch of the permissible two-and three-consomant combi tions in syllable initial position in English Final periphery ‘We turn now to the rules of phoneme combination that apply to the final periphery of the syllable, Here also itis possible for no consonant to occur: this is the case after any vowel phoneme in English, except the following five, which are always followed by a consonant: fe; fa, /o/, ial. All consonant phonemes may occur singly in syllable final position in English, except the following four: i, wl 11 is more difficult to find precise formulas to express the permissible clusters in syllable final consonant combinations than in syllable initial combinations, As stated earlier, there may be up to four consonants in the final periphery; and as in the case of syllable initial clusters, the fower are the permissible combinations, the larger the clusters become. Although not expressing the restrictions in the greatest possible detail or with complete precision, the following chart provides a neat representation of permissible combinations in syllable final position: Figure 27: Sllable-inal consonant combinations Position A_ Position B _ Position € Position D__| T ‘Any consonant @ t mao except ta 5 3 hewj se eeesae sees ‘Two consonant combinations may comprise either a consonant from Position A combined with a consonant from Position B, or from Position B combined with Position C. The 48 permissible three-consonant syllable final combinations are either Position A + Position B+ Position C, or Position B + Position C + Posi tion D. The seven permissible four consonant combinations are as follows: /Ikts, Impstl, Impts/, Aksts/, MDs), Os, /ntOs!, as in mudets, glimpsed, tempts, texts, ‘welfihs, sicths, thousandihs. Obviously, these sound clusters are likely to be Combining sounds 45, reduced under the pressures of connected speech. The reader is let withthe task of discovering examples for the two-and three-consonant combinations, and indeed of discovering precisely which combinations are found in contemporary English. "The phoneme combination rules that we have been discussing are established on the basis of monosyllabic words. In polysyllabie words itis not always clear, when a number of consonants ovcurs at syllable boundaries, to which syllable they should be assigned. For example, in the word exira fekstro/ there are obviously (wo syllables, as two vowel phonemes occur, but itis not clear how the syllables should be divided. Three possible divisions may be made, all of which would give permissible consonant combinations: k-str, ks-tr and kst-r.The method proposed to solve this kind of problem is a statistical one, A statistical survey of syllable initial and syllable final positions has produced an order of “favoured” syllable beginnings and endings, and the break between syllables that conforms to this most favoured principle isthe preferred one. At the beginning of a syllable the order of favour is: CV, VC, CC (C= consonant, V = vowel); at the end of a syllable the order is: VC, CV, CC. So in a word like aster /esta/ (VCCV) the preferred break would be as-7. For extra one would presumably have to choose between VC-CCCV and VCCC-CY, ie ek-stra and ekstr. Stress Another good reason for recognizing syllables is that there isa further phenomenon that scems to be associated with them, that of stress placement. When we hear polysyllabie words spoken we are often aware that some part of the word sounds louder or seems to be uttered with more force than another part. Or, to put it another way, one syllable is more prominent than another. This is achieved by the fact that syllables have different degrees of stress, and this is one of the phonetic factors that ensures that speech does not sound completely monotonous and level ‘Stress is perceived as difference in loudness, but from the articulatory point of view stress variation arses from force of articulation and more especially from difference in pte In English words we usually recognize three degrees of stress: primary stress, indicated by a stroke above the line at the beginning of the syllable so stressed ('); secondary stress, indicated by a stroke on or below the line atthe beginning of the syllable $0 stressed; and unstressed syllables, which are not marked. Each English word, spoken in isolation, has one and only one primary stress, So monosyllables have just a primary stress. In connected speech some monosyllabic ‘words may become unstressed; this applies particularly to ‘grammatical’ words, those that are also subject to vowel reduction (see Chapter 5). Polysyllabie words may have, though not necessarily, one secondary stress, and in some fairly rare ‘cases Iwo secondary stresses. All remaining syllables will be unstressed, Each English word, then, has @ characteristic stress pattern which normally remains unchanged in connected speech. Unlike in some languages, primary stress in English does not fall on the same syllable in every word (eg last sylable or penultimate); it may fall on just about any48 Analyzing English syllable. However, for any particular word primary stress usually fall on the same syllable whenever the word occurs. So in every occurrence of reserve, primary stress always falls on the second syllable. In English, then, stress placement is free in the sense that there is no uniform pattern for every word, and itis fixed in that the stress patterns of individual words do not change according to context, Here now are some examples of stress patterns i English polysyllabic words: secondary + primary jun known unstressed + primary llone primary + secondary ‘profile primary + unstressed 'ta ble Secondary + unstressed + primary un derstand primary + unstressed + unstressed "ba che for Primary + unstressed + secondary ‘pho tograph unstressed + primary + unstressed im'por tant secondary + primary + unstressed jun'cer tain unstressed + primary + two unstressed n'po ssi ble secondary + primary + two unstressed ,un'for tu nate unstressed + primary +-unstressed + secondary eli maize primary + three unstressed ‘me fan cho ly Primary + unstressed + secondary + unstressed ‘educa ted unstressed + secondary + unstressed + primary misjun der'stand unstressed + secondary + unstressed + primary + unstressed comsi de'ra tion secondary + unstressed + primary + unstressed + secondary in ca'pa ciate unstressed + primary + three unstressed ad’mi nis tra tive unstressed + secondary + unstressed + primary + two unstressed con,ver sbi lity secondary + three unstressed + primary + unstressed yaa tio na lea tion secondary + unstressed + primary + theee unstressed yn dis'tn gui sha ble secondary + unstressed + Secondary + unstressed + primary + two unstressed vu rei abi lity Uunstressod + secondary + two unstressed & nrimary + two unstressed impe ne tra bi lity ‘A mere difference in stress pattern does not usually serve to distinguish one word from another. Stress is not contrastive in the sense that phonemes are: the substitution of one phoneme for another results in the creation of a different word eg ban'panitanican etc. But for stress this is not the case: stress is contrastive in the sense that it contrasts one syllable with another in the linear sequence of a word, However, there is small number of words in English where the substitution of one stress pattern for another does result in the creation of a different word. Very often, though, the difference in stress pattern is uecompanied by a difference in vowel quality, since unstressed syllables tend to be associated with vowel reduction and centralized vowels. Compare the following pairs of words: ‘trans fer trans fer ‘digest di'gest ‘con duet con'duct ‘protest pro'test Combining sounds 47 ‘The words in the first column, with primary stress on the first syllable, are nouns; while those in the second column, with primary stress on the second syllable, are verbs. We have said that each word in English has a characteristic stress pattern. These stress patterns may, however, change in the course of time. For example, the word ‘momentary has the stress pattern in modern English of a primary stress on the first syllable, with all the others unstressed, In the seventeenth century the primary stress was on the penultimate syllable, with a secondary stress on the first syllable—mo men'ta ry. This older stress pattern is retained in present-day American-English pronunciation. In modern English there are a few words whose stress pattern is variable, even within the speech of an individual, and it could be that these are words whose stress pattern is currently in the course of changing, eg “im te graliin'te gral; ‘con tro ver sy,con'tro ver sy Rxercse 9 Work out the stress patterns for the following English words: eg control, fant a s},,sup po 's ton. 1, machine 2. female 3, magazine 4. quantity 5 telephone 3. thinoceros 7. Gowardliness 8. satisfactory 9, inferiority 10, personification 11, unilateralism a7. Intonation When we speak, we do not do s0 on a monotone, Indeed, we should all sound very boring if we did. Speech does not just consist of a sequence of phonemes combined into syllables and words. Accompanying these combinations of phonemes, occur- ring simultaneously with their articulation, are two features of speech that have linguistic importance. One of these is stress; the importance of which we sav in the previous chapter in contrasting one syllable with another in the phonological structure of words. Stress has a similar {unction in longer stretches of utterance, in according prominence to one syllable (and so to the word of which itis a part) as ‘against another. For example, the utterance The cat chased the mouse could be stressed in one of three ways: The CAT chased the mouse The cat CIASed the mouse The cat chased the MOUSE. As in the case of words, stress is produced in utterances not especially by greater force of articulation on one syllable rather than another, but more particularly by a change in pitch on a syllable, ‘Tone groups ‘The other linguistically important feature accompanying the phoneme combina tions is intonation, This isthe musical or melodie accompaniment of speech. On the level of intonation each utterance is composed of a number of tone-groups, Each tone-group represents a unit of information, which isto say that intonation has the function of dividing an utterance up into information units. Viewed phonetically, tone-groups consist ofa series of rises and falls in the pitch of the voice. Since, as we have said, pitch movement is probably the most important contributing factor to stress, intonation and stress can be seen to be closely interconnected in this area of speech production. Besides structuring the utterance in terms of information units, intonation is often important grammatically in distinguishing one type of sentence from another; and it is also important in signalling the attitude of the speaker in what he is saying. Nucleus Every tone-group has a nucleus, which i its most prominent part. The nucleus is, in other words, the most heavily stressed syllable in a tone-group, and the one which has associated with ita pitch movement. In a tone-group without contrastive or ‘emphatic stress the nucleus falls on the primary stressed syllable of the last lexical Word (noun, verb, adjective or adverb, as against pronoun, preposition, ete which 48 Intonation 49 are ‘grammatical’ words—see Chapter 9) eg The cat chased the MOUSE. Where there is contrastive or emphatic stress, the primary stressed syllable with such stress is the locus of the nucleus, eg The cat CHASed the mouse, The CAT chased the Intonation is a notoriously difficult area of phonetic investigation, and phoneticians vary considerably inthe way that they analyze it. Among other things, they disagree ‘on the number and types of nucleuses that should be recognized for English. We will restrict ourselves to four types of nucleus, which seems adequate, They are: falling, ising, flling-rising ond rising falling. We can illustrate these, together with the notation that is frequently used, on the uterance No: Falling ‘No (a matter-of-fact statement) Rising ,No (questioning) Fallngzrising ‘No (expressing incredulity) Rising-falling "No (emphatic probbition, solding). “The notes in brackets are an attempt to provide a meaning for each of the nucleus tones, and to help the reader identify and produce them for himsel. Tail [Any lables in a tone-group coming after the nucleus are called the tall The pitch ‘on these syllables continues the direction of the nucleus. So if the nucieus has a rising or @ fallng-rising pte, that on the tail will be ising; and it will be falling i {he nucleus is falling oF rising-falling. Compare the following utterances: He wants some “ICE-cream on top off (statement) He wants some ,ICE-cream on top of i (qiestioning). Head “The part of the tone-group coming directly before the nucleus is called the head. This extends from the primary stressed syllable of the first lexical word in the tone-group up to the nucleus. The tones of the head are level, rather than having. pitch movement, 50 that two kinds of effect may be found. The head may all be ‘tered on the sume pitch, usually low, or a stepping effect may be produced, either from relatively high to relatively low pitch, or from relatively low to relatively high pitch, Compare the following utterances: Pat it on the “TABle (nonchalant instruction) Put it on the “TABle (emphatic, impatient command). Pre-head ‘Any syllables coming before the head are called the preshead. They willbe part of “lexical” words of “grammatical” words, and will be relatively unstressed. The pre-head may normally consist of high level tones or of low level tones. Compare the following, utterances: He dest kinow Row io 'SAY 3 (matter-of-fact statement) Fe doesn't kilaw how fo “SAY it (mocking tone).50 Analyzing English ‘The foregoing isa sketch of one possible way to analyze and describe the intonation of English. When one tries to observe intonation, it is often difficult to recognize ‘what isa falling tone and what isa rising one, but it soems clear that in conversation ‘we do perceive and react to these differences, and that they play an important part in signalling meaning. 1t is partly from the intonation of an utterance that we Understand what a speaker intends by what he is gaying. That means his intention both in respect of linguistic meaning, and in respect of his attitude to what he is saying, We will now consider these two aspects of the function of intonation in English. Intonation and meaning The contrast between a falling intonation pattern and a rising intonation pattern often correlates with a distinction between a statement and a question. Compare, {for example, the intonation patterns that might normally accompany the following: Cats chase mice.—statement; Do cats chase mice?—question expecting the answer yes of no; Where did the ball go?—question seeking information; itis possible for this kind of, question to be spoken with either rising or falling intonation, Indeed, with information seeking questions falling intonation is probably the norm, In any case, for either kind of question it is not just the intonation that indicates whether itis & {question or a statement: the order of words (do before eats, id before the ball) and the presence of a question word (where) in information seeking questions are ‘enough (o make the distinction clear. Where the intonation invariably carries the contrast between statement and question is if the utterance has a statement form but a rising intonation pattern eg Cats chase mice? Another meaning associated with a rising intonation pattern is that of incomplete~ ness. A rising intonation at the end of a tone-group may indicate that what is being said is not finished, that there is more to come eg First she went to the Cupboard, then she tried the Sideboard, and finally she found itin the table draver Or, in offering a list of items—Would you lke tea, o coffee or milk?—a falling intonation on the last item implies that the choice is limited to those items mentioned, whereas a rising intonation on the las item implies that the fist is not closed and that the speaker is open to further suggestions ‘The meaning of an utterance may also be affected by where the nucleus of a tone-group is placed: this may be erueial in deciding which itoms in an utterance bbelong together grammatically. The by now famous example to illustrate this point is the utterance He fed her dog biscuits. Ifthe nucleus falls on BiScuifs, then itis the ddog that he is feeding, If, on the other hand, the nucleus falls on DOG, then the lady participant is being fed with dog biscuit. In a similar way, how the utterance is divided into tone-groups, ie where the ‘one-group boundaries are placed, may also affect the sense of what is being said. A. traditional example is the utterance He washed and brushed his hair, If this ltterance is spoken as one tone-group, ie with only one nucleus, then it means that Intonation 51 the actions referred to applied only to the hair. If, on the other hand, itis spoken as two tone-groups, with a tone-group boundary between washed and and, then the washing may be taken to refer to other parts of his body and the brushing alone applied to the hair. Intonation may, then, be & contributory factor in determining which words belong together grammatically in an utterance, by tone-group boundary placement, by the placement of the nucleus within a tone-group, or by the kind of nucleus and general intonation pattern that is used, Intonation and attitude Perhaps more important than the use of intonation to make distinctions in grammar is the function of intonation to signal how a speaker intends his utterance to be taken or what his attitude is as he speaks the utterance. But this aspect of intonation is probably the most difficult of all to characterize, not least because we scem not to have a handy or agreed list of labels to apply to different attitudes. For example, how could we characterize the differences between the following, utterances: ‘That's fine (falling)—? matter of fact That's sine (rising) —2 encouraging That's “fine (falling-rising)—? what have you been worrying about. Part of the difficulty is that the same intonation pattern may have different ‘meanings or signal different attitudes in different contexts. It seems to depend on the meaning of the lexical items as much as on the meaning of the intonation. Compare the following utterances: No (rising-faling)—indignant ‘Really (rsing-falling)—ineredulous, sarcastic He "always is (cising-talling)—impatient and the following: No (falling-rising)—doubtful but encouraging If you want to (falling-rising)—grudging acceptance. It is not just the type of nucleus that may be important in conveying the attitude of the speaker, butalsothe kind ofhead, For example, Good Morning isnormally spoken with a falling nucleus on MORNing. But the head good may be characterized by cither a high level pitch or a low level pitch: the high pitch signals the attitude ‘cheerful’, while the low pitch signals the attitude ‘routine’ Bxercise 10 For normal rendering of the following utterances, consider where the nucleus would fall, land what kind of nucleus it might be: eg thar's a STUpid way to do i Hell telephone 19’ MORrow. 1, Don't do it lke that. 2. He came home yesterday 3. What did hd say after that. 4. His suggestion ix simply ridiculous, 5, You ean” go out,82 Analyzing English (6. He seems alright to me. 7. Is that al he talked about? 8, He won't find me there, 9. Lan understand it 10. You're never here on time. Notes ” "The etttude labels are taken from A C Gimson’s An introduction to the pronunciation of English Edward Amold, 1970 # Labels trom Gimson, op cit For a dettiled investigation ofthe attitudes conveyed by different kinds of tone-group and intonation patetn, the reader is refered to O'Connor, J D, and GF Arnold Intonation of colloquial English Longmen, 1961. | Part Two: Structures8. Grammar and sentences ‘A linguistic description, and consequently a language, is often regarded as being ‘composed of three parts: phonetics/phonology, grammar and semantics. And there is a sense in which grammar links phonology and semantics: phonemes combine into words (phonology), words combine into sentences (grammar), and sentences refer to events, actions and states in the world (semantics). However, phonology ‘and grammar are different from semantics in that they deal exclusively with the internal structure of language (form), while semantics is concerned largely with the external relationships of language (meaning). Phonology and grammar The relationship between phonology and grammar is not as simple as has been implied. There is no automatic progression from phoneme to word to sentence There are, in fat, two parallel sets of hierarchical relationships: phoneme—syll able—word (—tone-sroup); morpheme—word—. sentence. It so happens that the unit ‘word occurs in both, and not without reason. “Word” i the level at which the greatest amount of congruence is found between units in phonology and units in grammar. For the majority of phonological words there are equivalent grammatical words in # one-to-one relationship. But there are several interesting cases that demonstrate the separateness of the two sets of hierarchical relationships. These are cases where there is no one-to-one match between phonological and grammati- cal word. The uneven match between phonology and grammar is welt illustrated by the so-called homophones, where different grammatical words have the same phonemic structure or pronunciation, Take the phonological word /bav!. It represents a ‘number of lexical items (grammatical words): bough, bow (of a boat), bow (verb, “to incline the head or body"), bow (noun, ‘inclination of the head or body’). Besides being homophones, the last three items mentioned are also homographs, ie they are written the same. Or take the phonological words /stend! and /marnd/ teach of them represents two grammatical words, one @ noun and the other a verb Uneven matching between phonology and grammar is also found the other way round, though more rarely. The past tense of the verb dream may be represented phonologically as either Jdremv or /drimd/. And whether integral has the phonolog. ical structure /ntagraV or /in'tegral, itis till the same item as far as the grammar is concerned. ‘The separateness of the phonological and grammatical sets of relationships is sometimes referred to as the ‘double articulation’ of language. This implies that language is patterned simultaneously on two levels: sound and syntax. Phonemes pattern into words, words pattern into sentences, and the link between the two