Norm, standard, deviation
Richard Hudson
Abstract
A language is a collection of norms shared by the members of a community, so different communities have different norms. These differences matter because shared language norms don’t only allow efficient communication but also the signalling of shared social identities. As communities get larger, new norms evolve including standard languages, which may develop either by “natural” processes or by deliberate official intervention. I question the anglophone preference for “natural” market forces rather than planned intervention. I also explain why it’s important to distinguish different kinds of deviation from any linguistic norm according to whether they’re caused by ignorance or carelessness. I then turn to specifically educational issues. Language is central to education, so education needs a policy on language education. I outline two recent changes in this area of the UK’s education: towards greater tolerance of diversity, and towards more explicit teaching about language, including grammar. But these very healthy trends in official policy are seriously undermined by weaknesses in teachers’ knowledge about language. Finally, I turn to educational linguistics. I explain why linguists need to know more about how education affects language, then I pose two research questions for each of the three subject-areas covered by the conference. For linguistics: what is the standard language (in contrast with non-standard alternatives) and how does language mature during the school years? For didactics: how should teachers teach languages, and how should they themselves learn more about educational linguistics? And for cultural studies: how does the target language relate, in teaching, to the target culture, and how are language learners influenced by the culture they live in?
1Norm: why language is normative
I have been asked to consider the topic: „Norm, standard, deviation“ – an interesting challenge which takes me well outside my comfort zone. I can’t say I have ever thought hard about this particular triad of ideas, so I hope you will accept some fairly unpolished thoughts. Having discussed each of the three ideas on its own, I shall try to bring them together in some thoughts about a more familiar topic: education, and in particular, educational linguistics.
The word norm isn’t one that I use a great deal and even when I do use it I’m not completely sure what it means, so I’d better say what I think it means – that is, what I want it to mean in this chapter. I assume that a norm is a cultural rule which defines „normal“ behaviour in some community – behaviour which is accepted by that community. Norms define the clothes we wear on different occasions, the things we talk about, how we treat other people and so on (and on). A norm can be expressed either as an „if-then“ procedure(if you’re in public, wear a shirt) or as a static declaration (being in public is linked to wearing a shirt); in the following I shall assume a „declarative“ approach.
Norms are different from ideals; so even though we admire people who make great sacrifices for other people, we wouldn’t say that altruistic behaviour is required by a norm. But at the other extreme, they’re also different from mere descriptions of human behaviour. We all know that half of marriages end in divorce; but that doesn’t mean that divorce is required by a norm. A norm, therefore, is what we expect of a typical member of the community; but we also know that, for all sorts of reasons, actual behaviour may not conform.
Why do we have norms, and why do we trouble to follow them? Because they make our behaviour predictable to other people. With norms, I can predict your behaviour and you can predict mine, and we understand each other; without norms, we’re unpredictable to each other, and life is a constant battle between isolated individuals. We all know how hard it is to negotiate our way through life when we move into a community where a few norms are different from those we’re used to; but those problems are trivial compared with the problems we would face without any norms at all.
Turning now to language, this is a very large collection of norms in exactly this sense – what we might call „normal norms“. Take a very simple example, the link between the meaning „cat“ and the various forms that have this meaning in different languages: {cat} in English, {chat} in French, {katze} in German, and so on. Each of these links is a norm in the sense defined above. Meaning „cat“ is linked to saying {cat}, so if you say {cat} I know you’re meaning (i.e. thinking about) „cat“, and I know that by saying {cat} I can not only tell you that I’m thinking „cat“ but also get you to think „cat“ as well. As Stephen Pinker says, thanks to language „we can shape events in each other’s brains with exquisite precision“ (Pinker 1994: 15). If you know English, I can literally control your mind (albeit in a very small way) by saying {cat}; in that situation, it’s very hard for you not to think „cat“.
Norms, then, are a tool for coordinating social behaviour by ensuring that we all follow the same rules; and in language, they’re a tool for coordinating mental activity – in short, for communicating. We learn the norms of our community by a combination ofobservation and instruction. A child observes people saying {cat} while looking at a cat, and hears instructions from adults such as „Look, a cat“ while we’re looking at a cat; and we draw our own conclusions. By the age of about two, we draw a more general conclusion: that whenever adults say words, they have some specific idea in mind, a meaning. The discovery that words are systematically linked to meanings opens the way to the famous „vocabulary spurt“ when children actively seek meanings instead of merely noticing correlations in others’ behaviour (Hudson 2007c: 220-2).
Learning in this way allows us to coordinate our norms to a remarkable degree, so that all the members of a community agree not only on the meanings of tens of thousands of words but also on fine details of pronunciation and grammar. It’s true that coordination is never perfect. For one thing, norms are constantly shifting – a word here, a vowel there, so young people always sound different from their grandparents. And for another, communities have their own complexity so that most of us belong to a range of different overlapping communities, none of which have precise boundaries. Nevertheless, years of experience teach us how linguistic details correlate with social details. Consequently, the linguistic norms we build in our minds match in complexity the models we build for society, and the two models are tightly interlinked. In short our „I-language“ is integrated with our „I-society“, where in both cases „I“ stands for „internal“ and „individual“ (Hudson 2007a).
Language is undoubtedly the area of social life where norms matter most. This claim is hard to prove, but where else do we find tens of thousands of tightly interconnected items of behaviour, where details are so vital? A single phoneme can turn one word into its opposite, and subtle changes in word order can change meaning radically. As long as the speaker and hearer share the same norms, information flows both smoothly and astonishingly fast. We take it for granted that listeners can listen at the same speed that speakers can speak – about two words per second – but we shouldn’t, because both feats are extraordinarily complex. For instance, if you have an average vocabulary of about 30,000 words, it’s a miracle that you can even link a particular meaning to a particular sound-pattern (or the other way round) in half a second, not to mention the extra decisions relating to word order, pronunciation and so on. Whatever mental apparatus this miracle may require, it surely requires extremely close coordination of the speaker’s and hearer’s norms.
Norms blur the distinction between what is and what ought to be – between science and ethics. On the one hand we learn our community’s norms partly by observing other people’s behaviour, i.e. by observing what „is“. On the other hand, we’re looking for „normal“ behaviour, so we ignore exceptional examples such as speech errors – occasions when people say {cat} when they really mean „dog“ or „hat“. That’s already a step towards what „ought to“ be, but another step is our own behaviour. We know that the link between „cat“ and {cat} isn’t simply something that other people recognise, but something we ourselves ought to recognise; so I accept that when I say {cat} I ought to mean „cat“. Why? Because I accept Grice’s cooperative principle that I shouldn’t mislead my hearer, which is a matter of morality. Moreover, I know that if I flout the norms that other people follow, they won’t like me; and I want to be liked. Another moral issue.
In short, norms are „normative“ – descriptions of how things ought to be.
These questions about the status of norms are important for linguists because we’re all anxious to be „descriptive“ rather than „prescriptive“ (saying what should happen) or „proscriptive“ (saying what shouldn’t). How can we take a descriptive approach to norms which are basically prescriptions or proscriptions? The answer is obvious: norms are socio-psychological data which can be described, just like any other part of our cognitive world. What makes a linguist’s work descriptive is its impartiality, in contrast with prescriptive linguistics which says that some norms are better than others. But however obvious this principle may be, it’s easy to forget that the language we’re describing impartially is itself fundamentally normative.
2Standard: why nations adopt standard languages
The price we pay for the almost unlimited expressive power of language is the need to learn many thousands of norms; and since we learn these norms from whatever community we happen to belong to, they inevitably differ from community to community. The biblical story of the Tower of Babel explains the diversity of language as a punishment by God, but it’s actually the inevitable consequence of our social life in communities. The very communities that allow communication among their own members also prevent it between members of different communities.
There are two reasons why this is inevitable, one negative and the other positive. The negative reason is simply that contacts between people in different communities are, by definition, more limited than those between members of the same community, so opportunities are also more limited for learning norms. The positive reason is more interesting. We each have a personal identity which defines the kind of person we are, and an important component of this identity is a list of the communities we belong to. In simple societies this list may be short: a family, a village or tribe, a sex and some kind of age-defined group. In our more complex societies the list is much longer, but regardless of complexity, the same principle applies: language is an excellent badge of social identity.
Every word we use offers the possibility of signalling allegiance to some group or other – a slang word here, an educated word there, a local vowel here, a regional consonant there. Modern sociolinguistics has revealed astonishingly subtle and complex links between very specific bits of language and social categories (Hudson 1996, Hudson 2007a), but of course the same principles also explain gross differences between entire languages which act, in the same way, as social distinguishers. As long as there are communities that need to be distinguished, language will remain an important distinguisher.
This principle raises obvious problems. One problem is in communication across community boundaries, which is harder in proportion to the differences between the communities’ norms. But another problem is in the building of communities. How can larger communities bind together members of smaller communities whose members pride themselves on being different?
Both these problems face large communities such as nation states, and it has always been obvious to nation-building politicians that language is crucial. Linguistic diversity can divide one region from another, but a single language can also unify them into a single super-community – which is why nations typically recognise a national language, in a standardized variety which reduces regional diversity. In other words, a standard language.
Standard languages solve both kinds of problem mentioned above. They allow easy and efficient communication between people from different regions, and they provide a form of language which distinguishes that nation from other nations, but without distinguishing one region from another – just the kind of emotionally charged symbol of citizenship that any nation-builder would wish for.
But what about the local dialects and non-national languages? If governments listened to linguists, they would take a liberal approach which celebrated diversity and recognised the possibility, and even possible benefits, of multilingualism and multidialectism. Some governments do take such an approach, but there are notorious examples of blind governments which create unnecessary problems both for themselves and for their citizens by proscribing all alternatives to the national standard (Spolsky and Lambert 2006).
For a descriptive linguist, a standard variety is just like a non-standard one, neither better nor worse. It’s tempting to think that because standard languages are promoted in schools they must be in some sense less „natural“ than non-standard varieties; and some linguists have indeed given in to this temptation, as shown by this quotation from Noam Chomsky:
I would certainly think that students ought to know the standard literary language with all its conventions, its absurdities, its artificial conventions, and so onbecause that's a real cultural system, and an important cultural system. ... It's notbetter, or more sensible. Much of it is a violation of natural law. In fact, a good deal of what's taught is taughtbecause it's wrong. You don't have to teach people their native language because it grows in their minds, but ifyou want people to say, "He and I were here" and not "Him and me were here," then you have to teach thembecause it's probably wrong. The nature of English probably is the other way, "Him and me were here," ... So a good deal of what's taught in the standard language is just a history of artificialities, and they have to be taught because they're artificial. (Olson and others 1991: 30)
This must surely be wrong because some people speak standard English as their native language, learned in the normal way from parents and family. Such speakers naturally use „he and I were here“, as would anyone who speaks a language that has inflected cases such as German or Old English. And if the spoken standard changes (as it has done in this case), this is a natural process of change, rather than a switch from unnatural to natural.
The advantages of having a recognised standard language are beyond dispute. On an emotional level, it allows us to identify with the national community; and in practical terms, it improves communication. For instance, just think how slowly we would read if spelling and grammar reflected the writer’s local norms. However, there’s a great deal that I, for one, don’t know about standard languages, and I believe we need serious research on questions such as the following.
How should standardization affect pronunciation? In the English-speaking world, linguists insist that pronunciation is not part of standard English: standard English is a dialect, defined only by its vocabulary and grammar, in contrast with Received Pronunciation (RP), which is an accent – a way of pronouncing the standard dialect. We teach our students not to confuse the two, by pointing out that standard English can be spoken with a wide range of regional accents including RP (which is regional if we consider England as a region within the English-speaking world). Given the close link in England between RP and social class I believe this is an important distinction, but it leads to a serious paradox because pronunciation is a much more serious barrier to spoken communication than either vocabulary or grammar. Is this impression correct, and if so, what implications does it have? And what general principles can we draw for standardization in other countries?
Another research question involves the two main models for standardization: the standard-dialect arrangement found in countries such as the UK, and diglossia, whose most famous European exemplar is German-speaking Switzerland. In Zurich, for example, everyone speaks the local Zurich non-standard variety at home, and the standard variety – High German – is used only on more formal, public occasions. At least to a UK linguist, diglossia looks much fairer because nobody has the advantage of learning the standard at home. But is it really so much better? We don’t know, but it would surely be possible to imagine relevant research projects.
And finally, what’s the best way to define a standard language? Is it best to have an official language academy or to leave everything to market forces, as in the English-speaking world? It’s easy to see the political and ideological issues lurking behind this question, but it’s also a research question. We in the anglophone world tend to rejoice at our independence and to make fun of the Académie Française (Mugglestone 2006); we like to think it’s better to let „natural“ processes of change run their course rather than to try (in vain) to channel them. But are academy’s really as ineffective as we like to think? And, for that matter, how „natural“ are the processes that influence English, ranging from publishers’ style-sheets to advertisers’ balance-sheets? Once again, there is a great deal of important research waiting to be done.