Variation in Speech
In traditional phonetic description, it has been usual to describe the characteristics of one particular type of speech. Where possible, analysts have looked for a “standard” or “model” accent. In Spain, for example, the accent of the region of Castilla (Castilian Spanish) has for centuries been treated as the “purest” form of Spanish, and the one which foreigners should attempt to copy. In Britain, a similar standard traditionally known as Received Pronunciation (RP) was used for most of the 20th century, though it is becoming more and more difficult to define what this is and to explain why it should be treated as special. As I explained in Chapter 1, an increasing number of writers (myself included) now prefer to refer to the standard English pronunciation as a BBC accent. Perhaps the most common procedure is to adopt as a model the accent of educated people in the capital city of the country where the language is spoken. In some cases, however, descriptions are based simply on the pronunciation of speakers who were available to act as “informants” to the analyst – this is frequently the case in studying languages spoken by only a few people, and which are likely to die out soon. In a few cases, a language has been recorded and analysed using the last surviving speaker of it, and the work done quickly because of the risk of the informant dying before the study was completed.
While this concentration on a single variety of a language is a convenient way of keeping one’s description clear and simple, we should never forget that there is an enormous amount of variation in how a language is pronounced, and in this chapter we will look briefly at some of the types of variation.
The study of regional variation is probably the best-known and longest-established form of the study of variety, and most of us have a stereotyped picture of the earnest dialect researcher roaming through the countryside to seek out ancient rustic characters and elicit information about vocabulary and pronunciation. It is usual to make a distinction between the study of dialect (which looks not only at pronunciation but also vocabulary and grammar) and accent (which is purely a matter of pronunciation). We have already met the word ‘accent’ in a completely different use (to refer to distinctive pitch patterns), and this sometimes gives rise to confusions.
Regional variation can arise from many causes. One cause is invasion or colonisation: parts of Britain, for example, were colonised by Norse and Saxon invaders while other parts remained unconquered, and there are still recognisable differences in English language and pronunciation due to this fact. Historically speaking, we can see that separate varieties were most likely to emerge when there were barriers and frontiers between the areas in which a language was spoken. Countries in which isolated communities have been separated from each other by mountains or sea often have greater differences in pronunciation than countries where there has been free and easy movement among people. We can see something of the sort on a very large scale in the way in which American English moved away from the pronunciation of English in England where it originated, thanks to the barrier of the Atlantic. When the first American “talking” films were shown in Britain in the 1930’s, the distributors had to consider putting subtitles on the films because most members of a British audience had virtually no experience of listening to an American accent. In the present day many people are able to communicate by phone with others who have very different varieties of the language, and can hear many different language accents on radio and TV. Now spoken communication between British and American speakers seems relatively straightforward and misunderstandings are likely to be due to cultural differences rather than linguistic or phonetic ones.
There has been a recent growth of interest in a related area of the study of English pronunciation, sometimes referred to as English as an International Language. Since English is now used by so many people around the world for international communication, it is possible to see pronunciation varieties emerging which are not based primarily on the native-speaker accent of some part of the English-speaking world. Instead, they show English as a global “common property” in which key phonetic and phonological distinctions are retained, but choosing to sound, for example, English or American seems irrelevant.
To consider the complex ways in which social factors affect variation would take us into the domain of sociolinguistics and beyond the scope of the present book. But for our purposes we can broadly distinguish between three different kinds of variation. One is related to social class: in some societies (but certainly not in all) people have a pronunciation which identifies them as a member of some social class, or as being at some point on a scale from low to high social class. A good example is “h-dropping”: a well-known study in Bradford showed that speakers were more likely to pronounce the h sound in words like ‘house’, ‘hat’ etc. if they were of higher social class, and more likely to omit it if they were of lower class.
A second type of social factor is speakers’ tendency to use different pronunciations in different social situations. Not everyone does this, and many people who do it are reluctant to admit that they do. Many people can and do speak something like the standard accent in their professional life, but switch to a different accent (either a local regional variety or a lower-class pronunciation) when they are with family and friends.
Finally, there are social divisions in society other than class ones. Many languages show differences between the speech of men and women; different professional groups (for example, teachers or members of the armed forces) often have some distinctive accent or speaking style, and many societies, though with a common language, have strong differences of religion which are reflected in the variety of the language spoken.
We are all capable of changing the way we speak when this is necessary for successful communication. Everyone can vary between speaking rapidly or slowly, or between quietly and loudly in a way that is appropriate to the communication situation, (though some people make such adjustments more successfully than others). Phonetic descriptions of languages have tended to be made on the basis of a slow, careful speaking style, and this creates major problems when one comes to study more natural speech and discovers that it does not fit many of the “facts” stated in textbook descriptions.
Teachers, priests and politicians are good examples of people who need to be able to speak in a range of styles: public speaking is something which does not come easily to everyone, and in some cases people even take lessons in how to address a large group of people.
Age and variation
Everyone knows that young people speak differently from older people. It is not likely that this is due to physical causes. We do not know how much of age-based variation is due to individuals changing as they grow older and how much is due to the pronunciation changing from year to year. It is likely that a major factor is the wish of young people to speak in a different way from their parents, and in the present day this is strengthened by broadcasting aimed specially at young people. Some changes happen rapidly, while others emerge only over a very long period. Two changes in English have been specially noticeable since I began to work in phonetics in the mid-1960’s. One is the growth in the use of the glottal stop (for which the phonetic symbol is [?] ), either as a replacement for the t phoneme in words like ‘getting’, ‘better’ (so that ‘getting better’ is pronounced [ge?IN be?«]), or in conjunction with /p/, /t/, /k/ or /tS/ where the glottal closure precedes the oral closure in words like ‘captive’ [ka?ptIv] , ‘cats’ [ka?ts], ‘accent’ [a?ks«nt], ‘butcher’ [bU?tS«]. The other change is the fronting of the /uù/ vowel, most noticeably after /j/. In the early twentieth century, the “Received Pronunciation” version of this vowel was back and rounded in all contexts, but in the speech of younger English people from the South-East this vowel in words like ‘union’ /juùni«n/, ‘human’ /hjuùm«n/, ‘usual’ /juùZu«l/ has become more of a front vowel (nearer to the /iù/ vowel), and it now has very little lip-rounding. The word ‘used’ in ‘I used to’ thus sounds almost like the word ‘yeast’.
The pronunciation of a language, then, is liable to constant change, and at any time there are many varieties which are found in different places and situations.
Ideally, we would like to study speech which is as natural and as close to “real life” as possible, but it is often very difficult to collect carefully-controlled material for scientific analysis in an everyday context. In most of our daily life, for example, we are surrounded by a lot of extraneous noise which can make a recording difficult to study accurately, so it is often felt preferable to record speakers in a studio (usually within a speech laboratory). However, one of the common complaints about laboratory studies of speech is that speech recorded in this way does not sound natural. The use of what is often referred to as “lab speech” has a number of disadvantages. Often, speakers have to read what they say (from a written text or from a computer screen), instead of speaking spontaneously. This can have serious effects: for example, a common mistake is to let the speaker see the end of the list of items that they are reading – intonation and speaking rate tend to change as one gets near the end of a list. In a long recording, fatigue is another problem – while some people can happily talk to their friends for hours, recording prepared material in a recording studio becomes very boring (and therefore tiring) after twenty or thirty minutes. Another problem is that speakers tend to be selected from the people who can be found near the laboratory, and are thus usually students or staff of a university; it is astonishing that experiments using such people are almost always described as using “normal” speakers. In relation to the population as a whole, people who teach or study in universities and who volunteer to be recorded are not normal. Finally, everyone who knows that there is a microphone near them that is picking up what they say tends to speak more carefully than they would if they were speaking spontaneously. The well-known “observer’s paradox” is based on the fact that we want to know how people speak when we are not there, but we (or a recorder) have to be there in order to observe what is said. This has led to some researchers recording people secretly when they are talking naturally, thinking that they are speaking in private. This seems to me to be completely unethical, and I would never recommend doing it. How can natural speech then be recorded for us to study? One possibility is to develop interviewing skills, as the sociolinguist William Labov has done, to the point where after some time the speaker becomes relaxed and absorbed in what they are saying and forgets about the presence of the microphone. Another widely-used way of eliciting fairly natural speech in studio conditions is to give two or more speakers a task to complete using only speech. A typical example is the “maps task”, where two people are given maps of the same area, but with information missing. The speakers cannot see each other’s map, so they can only discover how to plan a journey by discovering through verbal interaction the information that is missing. In my own work, I have used the sort of puzzle that you sometimes see in magazines, where two pictures are printed which are identical apart from a number of very small differences. One picture is given to each speaker: they cannot see the other’s picture, so they have to talk to each other to find out where the differences are. Often they become so absorbed in this task that it is difficult to stop them when enough speech has been recorded.
The most important point about this discussion is that we should never ignore the variation that we find in different speaking styles, and should always plan carefully in deciding how to record data that is going to be used for the scientific study of speech.