You are on page 1of 12
24 Are all phonetic differences equally important? Do all sounds play the same role in communication? The function of sounds in the language system is to identify words, that is, to distinguish one word (or morpheme) from another. Some sounds in English do have this function: for example, tie is not the same word as die, send is not the same as sand, and wash is not the same as watelt, These pairs show that in English the sounds /V and /dl, /e/ and fee, Mf and Hf are capable of distinguishing words. Their differences are not purely phonetic, but have a distinctive function in the language. Putting it differently, we may say that the difference between // and /d/, etc. is not just a surface difference but an underlying one. Such distinctive (ie. contrastable) sound units, which represent an underlying difference, are called phonemes. The difference between Wand /d/ is a phonemic or distinctive difference, so /¥ and /d/ are different phonemes. (Pronounce forni-m/, /forsni:rmk/.) ‘A phoneme is an underlying sound unit capable of distinguishing one word from another in a given language. It is called underlying (H. “mégittes”) because one phoneme may have different realizations, that is, the same underlying unit may appear as different sounds on the surface in actual promunciation. The number of phonemes in a language is normally around 40, much smaller than the number of sounds. ©) Homophones. The opposite may also happen: two words are spelt differently but pro- nounced exactly the same (just like H hdncs = hdnts, fojt = folyt). These look like minimal pairs in writing, but in reality they are homophones and not minimal pairs at all. For example: meat (mi: ~ meet way Ave = weigh war Iwai! = wore there [eal = their son isan! = sun two Itu:! = to0 berry beri’ = bury practice 'ptekus! = practise red ited) = (he) read nose !nowzi = knows guest /gest! = guessed plane plein! = plain Minimal pairs 25 To find out which sound-differences are distinctive, we must take two words that differ in one paxticular sound only, e.g, send and sand. If we find that the two forms are judged by native speakers to be two different words (and not just the same word pronounced differently), then the two sounds in question represent two different phonemes, and the word pair is a minimally differing pair. (Such a pair is usually called “minimal pair” for short, but of course it is not the pair that is minimal, but the difference!) To put it more technically: we take an environment, say /s_nd/, and try both /e/ and Fo into it to check if the meaning changes, We find that it does: /send/ is a different word from /seend/. This proves that in English the /e <> e/ difference is distinetive: we conclude that /e/ and /e/ are not only two different sounds but two different phonemes as well On the other hand, if we replace the English-type /r/ in room with a Hungarian (or Scottish) type trilled r, the result will not be a different word, just a curious or dialectal rendering of the same word. There is no minimal pair in English whose difference would hinge on the two kinds of r, 2.7 Not all sound-differences have such phonemic function. If we replace the unaspirated {p] sound of sleepy with the aspirated [p*] sound of pig, the meaning will not change. The result is just a strange, slightly un-English rendering of the same word sleepy. This shows that though phonetically [p] and [p"] are different sounds, their difference is not distinctive in English. It is not a phonemic difference but only a subphonemic one, The aspirated! unaspirated difference has no function in distinguishing one English word from another. No ‘minimal pair can be found where the meaning would depend on one word having an aspirated consonant, the other having an unaspirated one, Such sounds, which are different phonetically, yet do not produce minimal pairs (ice. they have no distinctive function), are regarded as variants (or altemants) of the same phoneme, and are usually not distinguished in transcription. In technical writings — including the present hook — it may still be necessary to use notation which is capable of showing cither sounds or phonemes, as the case may be. The convention is to use different types of brackets for the two purposes, namely [square brackets] are used to symbolise sounds, and ‘slants/ are used to symbolise phonemes. For example, we found that the sounds {e] and [2] belong to two different phonemes in English, so we can write them as /e/, /z/; but the sounds [p] and (p"] are just variants of the same phoneme /p/, so we do not put [p"] between slants. When a phoneme has several variant realizations, the simplest or most frequent of these is chosen as the “chief” member to stand for all the others. Thus the phoneme symbol (p/ includes both [p] and fp) 2.11 It should be clear from the above that the sounds of a language can be approached and described in two different ways, leading to quite different results. One approach is that of phoneties: it examines the physical properties of sounds (cf. 1.8). In a phonetic investigation we ask how the different sounds are produced, transmitted through the air, and heard (perceived) by the hearer; what are the articulatory differences between the sounds of a language, or between the sounds of two different languages Phonetics deals with the “material” aspect of sounds. It describes them very much like the natural sciences describe the phenomena of nature. The other approach is that of phonology: it examines the function of sounds within the system of one particular language (cf. 2.4). In phonological analysis we ask wether certain sounds (or rather the differences between them) are capable of distinguishing words; whether their occurrence in certain environments is predictable; what restrictions exist on their distribution; which sounds are the variants of the same phoneme; and so on. Phonology deals with the “organizational” aspect (a5 we often say, the “behaviour”) of sounds. It tries to discover their values in the system. Phonological analysis is, fundamentally, phonemic analysis, 2.12 Phonology is more abstract than phonetics, because phonology can ignore those features of a sound that are non-distinctive, and concentrate on the distinctive features only. For example, Hungarian phonology can claim that d is “the long pair” of a, even though phonetically they are quite different vowels: d is [a:], but a is [p]. English phonology can claim that /i is always long, as in mean, even though it loses its length in eeriain positions, e.g. heat is phonemically /hi:t/ but phonetically [hit], with a shortened vowel. But since this length difference is non-distinctive (it is predictable, caused by the environment of a following /U)), phonology can claim that underlyingly the vowel of heat is as long as that of mean, and neither of these has length as a distinctive feature. Tt is often quite controversial what the real distinctive features are. For example, what distinguishes heat /hi:t/ from Ait ‘hv’? Is it the length feature, the vowel quality feature, or both (as our transcription suggests)? Or what distinguishes // from /d/? Is the voice- less/voiced difference distinctive here, or is it something else like aspiration? It falls outside the scope of this book to discuss theoretical details of phonemic analysis, but we shall con- sider some of these problems as we go along, Allophones: complementary distribution 28 Very often it can be predicted which of the variants will appear in a given environment, that is, their distribution can be specified. For example, at the very beginning of a word or of a stressed syllable, the phoneme /p/ is articulated as an aspirated [p"], as in pig compare, occupation, permission. In all other positions /p/ is articulated as an unaspirated Ip], as in sléepy, speak, company, sleeps. Such alternants of a phoneme, whose oceurrenc regularly predictable from the eavironment, are called the allophones of the phoneme. ‘The fact that the choice between the allophones of a given phoneme is automatically prescribed by the environment means that in certain positions you have to use allophone A, in others allophone B. Allophones are in complementary distribution: where one appears, the other does not, and vice versa. To return to our example, in one group of environments (at the beginning of a word or of a stressed syllable) the speaker must choose (p"], elsewhere he must choose [p]. The two allophones thus “complement” each other, and together they make up all the occurrences of the phoneme /py. It follows from the above, of course, that the allophones of a phoneme can never appear in cach other's position, since they mutually exclude each other in any given environment. Therefore they can never be in contrast with each other — consequently there are no minimal pairs with them. This is why they are not phonemes, just allophones. Free variation 2.9 Less often we find cases where the speaker can freely choose between variants, that is, where there is no distributional rule prescribing which one has to be used. For example, the vowel in go (and home, road, blow, etc.) is usually pronounced [os], a diphthong sound whose first element is [0]; but it can also be pronounced [an], with a first element [a] similar to H 6. Both are correct: [go] or [go], though the second is more current in BrB, the first is normal in AME. The sounds [ou] and [a0] are free variants of the phoneme /owy. Aspiration 4.11 The voiceless stops (/p t k/) are aspirated whenever they start a stressed syllable, e.g.: tin, compare, vacation have their underlined sounds aspirated. There is, however, no aspiration after /s/, ¢.g.: stay, despite, skin have their underlined sounds unaspirated, Aspiration is realized in one of two ways: a) when the following sound is a vowel, aspiration is realized as a short [hl] segment: fin is pronounced as (t'in}. —— b) when the following sound is a sonorant (/1 rj w/), aspiration is realized as the de- voicing of the sonorant, e.g. play is [plei] and not [p*lei] (see 5.2). Aspiration is a potential of /p t k/ only. Al other English sounds are unaspirated. 4.12 The following table shows the distribution of aspirated and unaspirated allophones of the voiceless stops. (Aspirated stops are shown in boldface, unaspirated stops in italics.) Aspirated Unaspirated Followed by Followed ty || Preceded by Elsewhere siressed vowel | — sonorant plus if stressed vowel 7) | pie play speak ‘police parody supply spray company “occupation impression explain simple compare computer response map 7 | sime twist stand tomorow entire between mistake eter return tune exrend sit activity attract strike activity vm | Kill ‘lerk shy Took chemist request school orchestra according gross discreet banguet engourage sube _[excuse combine Note 1. Some authors say that there is aspiration (though of a smaller degree) in word initial stops even if the first syllable is not stressed, e.g, police, tomorrow, combine Note 2, In BrE the expression af ail is pronounced as one word /s'to:V, and thus the it is aspirated — exactly as in a tall man. And yet, it is widely held that English pronunciation is very difficult: more difficult than Hungarian or German or Spanish pronunciation. This is true, but the unusually difficult nature of English pronunciation is not due to its phonetics (which is like that of any language) of the spelling. Written English often confuses the learner and suggests the wrong pronunciation. This is far less frequent in other languages, where you can more safely rely on the spelling when you want to pronounce something, When a Hungarian learner pronounces debt (correctly /det/) as */debU, this has nothing to da with phonetics: he is simply deceived by the spelling (just like with gone or wilderness). It is often said that words like debt have an “irregular pronunciation”. This statement is true only if by “pronunciation” we mean the relationship of spelling to sounds. Actually, it would be more precise to say the opposite: the word /det’ — which is, in itself, a perfectly easy and regular phonetic form — has an irregular spelling, with a superfluous silent letter & init Phonetic spelling 3.6 If the pronunciation of a word is unpredictable (- ambiguous), we need some method to disambiguate it. The best way to do this, of course, is to transcribe it. However, for readers who do not know transcription (most native speakers don’t!) the IPA means abso- lutely nothing. For them there is another, more traditional method: phonetic spelling. This, means writing down the word according to the usual spelling conventions of the language. For example, the word move is pronounced unpredictably /mu:v/, since it might equally be */moov/ (ef. stove) or perhaps */mav/ (ef. Jove). We can disambiguate its pronunciation by giving a phonetic spelling: “moov”. Similarly, we can show that orchestra is pronounced “orkestra”, debris “det”, gone is “gon” (and not “gun”), cello is “chetlo" forms show what spelling the word ought to have if it were spelt phonetically Phonetic spelling is often used among native speakers to disambiguate unphonetic spellings. You will read things like “Salisbury is pronounced Sauisbury”, “Dylan is pro- nounced dillan, not die-lan’’; “Nike shoes are called niky shoes”, ete, (In transcription these words would be /'so:Izbari/ Pdilan/ /naiki/.) The same can be done in Hungarian too, e.g egvestilt = “eggyesult”, Attila = “atilla”, dllj félre = “dj fore”, scherzo ~ “szkercé". ‘The phonetic spelling of a word may coincide with a word already existing in the language, e.g. ovo = 00, colonel = kernel, suite ~ sweet, Korea ~ career, bury = berry. Such pairs are of course homophones whose second member is pronounced predictably but the first is not. ete. The re-spelt ‘THE FUNCTIONS OF TRANSCRIPTION 3.10 In Europe, textbooks and dictionaries of English as 4 foreign language use a system of transcription called the International Phonetic Alphabet (IPA). As this system is almost exclusively used in Hungary today, we shall limit our discussion to it, The IPA has several versions, which differ slightly in the choice of vowel symbols and the use of the length mark 11. The consonant symbols arc practically the same in all versions ‘The most widely used version of the [PA for English was introduced by A. C. Gimson in 1977, it is therefore referred to professionally as “Gimsonian” wanscription. Hungarian textbooks and dictionaries usually use this (or an older, slightly different version of it, instroduced by Daniel Jones). Ia the present book I use the Gimson version, for it is now the most widely used one in textbooks and dictionaries deseribing BrE for foreigners. It must be noted, though, that some books (chiefly those that are not produced for foreigners) use —_button-ing /bA.tnan/ listen Pus’ > listener Phsa.of > final "fainl/ final-ly /fainla/ civil Pv > civileize sivlauzs (in finally, as in all words in -a!1Zy, the /V of -y is dropped, so only one /I/ is pronounced.) Of course, in all such words the syllabic /t/ or /V has a free variant pronunciation with /an, all: /oatanny/, Misana/, /famaly, /‘sivalarz/. Word paits like finally (=final-ly) and finely (=fine-+ly) are different in pronunciation: the first is 3 syllables, with a syllabic Dark-L; the second is only 2 syllables, with a non- syllabic Clear-L: finally Maz.l./ # finely Pfain.tul, When a syllabic consonant stands before a vyowel, as in finally, it must always be indicated in transcription, either with a line under Hal, or with dot or hyphen after it *fanlu, Manl-v; otherwise it could not be distinguished from finely *famlv. $20 Syllabic consonants, like all types of weak syllable, are unknown in Hungarian. They are foirly difficult for Hungarian leamers, who tead to insert a vowel before or after them: isn't /znt/ tends to be rendered as “izdni, izint, izndt, iznit”; faculty /fwexlu/ as “fekiilti, fekeilti, feklout, jektiti”. OF these, the insertion of 6 before n, 1 (“izént, fekoiti") is the most acceptable, since it is similar to the E free variant with /a/. However, natural contemporary E speech (especially in BrE) has a strong tendency to use syllabic /n, l/ wherever possible. The left-hand column in 5.18 is now old-fashioned or pedantic. Therefore Hungerian learners ought to master the syllabic consonants actively. Since these may only occur after a consonant, they can be practised in sequences like [tnna, snnn, [nn] or (blll, fl, kill]. Insist on the “darkness” of the // here: channel should sound more like H “esend or “cseno?” than like “csenal”, “esene?” or “csen!”. ‘When using a dictionary, remember that in transeription any /n/ or // that docs not siand next to a vowel must be syllabic. In some dictionaries an italicized /a/ before /n/ or /U/ shows thet the sonorant may be syllabic: /te1bal/.

You might also like