Thinking like a linguist
Our tasks in this chapter will be to explain what it means to ‘think like a linguist’ and to show how linguists’ assumptions about language often differ from those of the layperson. One might assume, for example, that a linguist would be the first person to turn to when seeking advice on good speech or writing. In fact, few linguists would see it as part of their role to prescribe how language should be used, preferring instead to describe the facts of language as it is used. As we will see later in the chapter, linguists are quick to point out that the bases for our linguistic value judgements generally turn out to be arbitrary, spurious and inconsistent.
In literate societies, we are also used to equating language with its written form, and treating speech as somehow deviant. Linguists make precisely the opposite assumption, reminding us that we all learn our mother tongue at a very young age without the aid of books, and if we learn to read and write in that language at all, we do so only after we have mastered speech. As we will see in this chapter, language looks radically different when we start from a spoken language perspective. It will also become clear that some everyday assumptions we take for granted – for example, the difference between a language and a dialect, or the notion of ‘beautiful’ or ‘primitive’ languages – become highly problematical once our linguistic prejudices are stripped away.
The science of language
It makes sense to start by asking what the term linguistics actually means. The following definition is taken from Collins English Dictionary:
‘Linguistics, n. (functioning as sing.) The scientific study of language’
As a working definition, ‘scientific study of language’ will probably do, but the word ‘scientific’ might appear problematic in this context, because language doesn’t seem to belong to the realm of science in its conventional sense. One certainly doesn’t imagine linguists in laboratories wearing white coats, and it isn’t immediately obvious how one could undertake experiments on language, something that resides ultimately in the head of a native speaker.
It might help if we construe ‘scientific’ here to mean something like ‘objective’, but achieving ‘objectivity’ in linguistics is far from a straightforward task, not least because speakers’ judgements about the same data can differ hugely, making reliable conclusions difficult to draw. For example, while most British English speakers would probably reject the sentence ‘I didn’t do it though but’, it’s perfectly acceptable in some British dialects. Likewise, many English speakers accept ‘innit?’ as a contraction of ‘isn't it?’ but reject it (often vehemently) as a tag question in sentences like: ‘We’re seeing him on Saturday, innit?’ – now commonly used in some varieties of British English. Even for a question as apparently innocuous as ‘Do you speak language X?’, native speaker intuitions may be contradictory or difficult to interpret: responses may be influenced by informants’ attitudes to the language in question (‘Do I approve of X, or even think of it as a proper language? Would I want people to think I use it?’) or to their understanding of the question, which might range from: ‘Do I speak this language every day?’ to ‘Can I understand it, even if I don’t speak it?’, or even ‘Can I manage a few words if the need arises?’ So linguists need to be especially careful when claiming ‘scientific’ objectivity for their findings.
To approach their subject matter objectively, linguists need first to shed a number of everyday assumptions, or ‘language myths’: we’ll be looking at some of these below. The good news is that learning to think like a linguist isn’t difficult: in a real sense, it’s a bit like releasing your inner child, as we’ll see in the next section. A further piece of good news is that, as a native speaker of any language, you’re already in possession of some ‘expert knowledge’! But before you start, you need to grasp two fundamental principles that underpin everything linguists do and that go some way to explaining what ‘scientific’ means for the study of language:
Principle 1: ‘The spoken language comes first.’
Principle 2: ‘Linguistics is descriptive, not prescriptive.’
Principle 1: The spoken language comes first
As we saw, thinking as a linguist does is like ‘releasing your inner child’. The following thought experiment will help get you started.
Spotlight: Try to forget you can read
Imagine what your world would be like if the written word were completely alien to you, and letters on the page no more than meaningless squiggles. Since you’re already reading this book, you’re probably finding that quite difficult, but this is of course a world you once knew, albeit when you were rather younger, probably before you started school.
For most adults, the written word takes up a significant proportion of our lives, whether we be reading a novel or daily newspaper, consulting an instruction manual, updating our Facebook status, catching up with the latest Twitter feed or texting a friend. If you’re at university or college, the written word soon becomes a prime focus: you read for a degree, which may well involve writing notes at lectures, where you may be given handouts, and you’ll be asked periodically to commit your thoughts to paper in the form of written essays. Writing is all around us, and modern life and the technological advances we take for granted would be impossible without it.
For linguists, however, writing takes second place to speech. Linguists are not uninterested in the written word: indeed, written material, particularly from earlier, pre-mass media eras, can offer important clues to language structure and linguistic change. Linguists working in the field of literary stylistics devote much of their time to the analysis of written texts. But generally linguists follow the principle of according primacy to speech, for a number of very good reasons:
All the world’s existing and extinct natural languages have had native speakers, but only a minority of them have ever had a written form.
While languages such as English, Mandarin, Hindi or Russian all have a long written tradition, many others, particularly those with small numbers of speakers, do not. Many African languages (e.g. Ewe, Wawa, Lugbara), Australian aboriginal languages (e.g. Dyirbal, Warlbiri, Guugu Yimidhirr) and native American languages (e.g. Arawakan, Hopi, Miskito) are not generally used for writing. We know little of the Gaulish language, which was spoken in what is now France before Roman occupation, because Gauls had no written system, and much of what we do know about the language comes from attempts to transcribe it using Latin characters, which were not designed for Gaulish.
Speakers of minority languages in unsympathetic nation states have often been taught that writing is acceptable only in the dominant or ‘official’ language, making it harder for their supporters to develop an accepted written standard if and when those same states later adopt more tolerant attitudes.
Cockney, Brummie, Geordie and Glaswegian (see Case study on next page) have no written form and their speakers are dependent on the conventions of standard English for writing. Estimates put at around 6,000 the number of different languages spoken throughout the world, of which only a fraction have a written form: it would seem perverse – not to say ‘unscientific’ – for linguists to limit their inquiry to this group.
<Image src="OPS/images/common5.jpg" alt="image" />
Case study: Dialect or language?
One might object here that Geordie, Cockney, Glaswegian and Brummie are dialects rather than languages. But this argument is a difficult one to sustain, as linguists are unable to find a watertight distinction between the two. One criterion might be mutual intelligibility: while we wouldn't expect to understand another language, we might well understand a different dialect of a language we do speak. But this criterion soon poses problems. The ‘dialects’ of Chinese (e.g. Mandarin, Hokkien, Cantonese) share a writing system but are mutually unintelligible, whereas the Scandinavian ‘languages’ Swedish, Danish and Norwegian are similar enough to be mutually comprehensible (sometimes with a little effort). The difference in practice is generally determined on socio-political rather than linguistic grounds: we tend to associate languages with nation states where they are spoken. Or, as cynics would have it: ‘a language is a dialect with an army and a navy’. To avoid problems of this kind, linguists talk of language varieties.
Even where a writing system exists, not all adults acquire it.
Few advanced societies come close to Finland’s near 100 per cent literacy rate. But almost everyone learns to speak at least one language from a very early age, and children’s remarkable ability to make sense of oral language data is a puzzle which has long fascinated linguists, particularly those working within the generative paradigm (see Chapter 8).
Writing derives from speech (not the other way round), but is rarely a faithful or consistent representation of it.
In ideographic writing systems, for example Egyptian hieroglyphics or modern Chinese characters, the symbols used offer no clue to pronunciation, but even where alphabetic systems are employed, in which letters or graphemes purport to correspond to speech sounds, the relationship between writing and speech is a complex one.
Writing is so ubiquitous and familiar that we rarely even notice its conventions and oddities. If you learned to write in English, for example, you’ll expect a capital letter at the start of every sentence, but only occasionally elsewhere. so this Sentence Looks a bit odd. If your mother tongue is German, you’ll expect nouns to have initial capitals as well, e.g. das Tier (the animal). More significantly, there is often a mismatch between the way we write and the way we speak. Why, for example, is the h of hope pronounced, but not that of hour or honest? Why is night spelt with a gh sequence which isn’t pronounced? It is precisely these anomalies that are most obvious to us as children learning to write, and discovering that it’s far from a simple matter of converting speech sounds to letters.
<Image src="OPS/images/common12.jpg" alt="image" />
Spotlight: Spelling and speech
The relationship between spelling and speech can be ridiculously idiosyncratic, as seen in this example from English:
ought \[ɔ:]
through \[ʉ:]
cough \[ɔf]
thorough \[ə]
Lough \[ɔx]
hiccough \[ʌp]
though \[əɷ]
drought \[aɷ]
rough \[ʌf]
All nine words contain the same orthographical sequence ough: in every case its pronunciation (indicated in International Phonetic Alphabet, or IPA, symbols, which will be explained in Chapter 3) differs, and in no case is the g or h ever pronounced, at least not in standard English. Worse, the variation seems largely arbitrary, so if you’re a non-native speaker attempting to learn English from a book, you'll have little to go on when a new word, say, trough, bough or chough (pronunciations 3, 8 and 9 respectively), comes along. A visiting Martian, informed that the woefully inconsistent sound-symbol relationship demonstrated above forms part of the accepted written convention for the modern world’s most powerful and prestigious language, might reasonably conclude that humans had taken leave of their senses.
French spelling has few of the illogicalities of English (though one might mention in passing the case of aulx – ‘heads of garlic’ – which has four letters, but only one sound ‘o’ \[o], which doesn’t appear in the word!). It does, however, have a number of arcane grammatical spelling conventions, which few French citizens ever completely master. A case in point is the preceding direct object agreement rule, which requires past participles to agree in number and gender with a direct object (not an indirect one), but only if it precedes, so J’ai **vu** la montagne (‘I saw/have seen the mountain’) but Je l’ai **vue** (‘I saw/have seen it’), with a final e to indicate that the pronoun l’ (elided form of la) is feminine, because it refers to la montagne. The complexity of this rule, which takes up four full pages of the French grammarians’ bible Le Bon Usage (plus a further 12 pages on special cases), is compounded by the fact that in most cases it has no effect on pronunciation. In his book Talk to the Snail: Ten Commandments for Understanding the French, Steven Clarke bravely attempts to explain to a layperson why J’adore les chaussures que tu m’as offertes (I love the shoes you gave me) requires an agreement in es, and warns his readers:
In France, you can’t ever throw away your school grammar book. It would be like taking the airbag out of your steering wheel. You never know when it might save your life.(Clarke 2006: 103)
This observation is true enough, no doubt, for the prescriptive written language, but French people have no more trouble talking to each other than any other nationality does, as anyone who has witnessed heated intellectual debate in a French café can testify.
The above are, admittedly, extreme examples, but everyday inconsistencies in the relationship between speech and writing are not hard to find. The same letter (or grapheme) will often have more than one sound value (think about the pronunciation of c in code and ice) or, conversely, the same vowel or consonant may be represented by different letters or letter combinations (take for example the ‘k’ sounds in cabbage, back, charisma, Iraq, flak, accord, bacchanalia, extent, mosquito, Khmer, biscuit). Little wonder that the sequence ghoti has been facetiously proposed as an alternative spelling of ‘fish’: gh as in rough, o as in women and ti as in nation.
English is far from alone in its poor fit between speech and writing: all languages with alphabetic writing systems present inconsistencies of this kind to a greater or lesser degree. The reason, in a nutshell, is that pronunciation changes too rapidly for spelling to keep up, with the result that writing systems are often a better guide to the way languages used to sound than to the way they are spoken now. The initial k of knave, for example, reflects an earlier state of English in which it was actually pronounced (it still is in its German cognate Knabe, ‘lad’). Other oddities, too, give clues to previous states of the language. The first vowel of mete sounds more like the vowel in ski than that of led because spelling hasn’t yet caught up with changes that occurred during the Great Vowel Shift of the fourteenth and fifteenth centuries, which we’ll consider in Chapter 13 on language change.
The seductively linear nature of writing can also engender some false assumptions about speech. We tend, for example, to think of ‘words’ as things separated by convenient orthographical gaps, as they are on a page. The reality, of course, is that in speech, whatwesayisrolledtogetherinsequenceslikethis (only robots in low-budget science fiction movies actually mark a pause between words when they speak). As we’ll see in Chapter 6 on morphology, from the spoken language perspective, watertight definitions of ‘words’ prove elusive. Is blackberry, for example, one word or two? What about Jack-in-the-box: one word, or four? Do short, unstressed items like a or the qualify as words at all? When we consider such questions, as we usually do, from the perspective of the written word, they seem quite trivial, but they are important for our understanding of how children break down and make sense of the language data they hear when learning their mother tongue. It’s easy to forget, as adults, that we were at our most successful as language learners when we were infants, and there wasn’t a grammar book, verb conjugation table or dictionary in sight.
For a variety of reasons, then, linguists accord primacy to speech, and work primarily with spoken language data. This will often require speech sounds (rather than letters or graphemes) to be noted down, a task for which, as we’ve seen, conventional orthographies are clearly ill-suited. To address this problem, the IPA, first published in 1888 and regularly updated since, provides a common set of symbols which enables linguists to transcribe the sounds of all languages precisely and consistently (conventionally between square brackets, as for the ough examples above). As we’ll see in Chapter 4, this demands a strict principle of one-to-one correspondence between sound and symbol: what you see is always exactly what you get and, unlike with conventional spelling, a change in pronunciation necessarily entails a change in transcription. Fortunately, IPA symbols are mostly familiar and easily learned, because they have been largely taken from the Western alphabets with which its founders were most familiar.
When the focus of enquiry is shifted from writing to speech, as linguists argue it must be, many of our common-sense assumptions about language are called into question. For example, most English speakers, if asked the question ‘How many vowels are there in English?’ will probably answer ‘Five: a, e, i, o, u’ (some might add a sixth: y). But this is a statement about the number of vowel letters in the English alphabet, not the number of vowel sounds. In fact, the number of vowel contrasts used by English speakers to distinguish words is considerably higher. Consider, for example, the different pronunciations represented by a alone in cart, cat and Kate, or the eight different vowel sounds rendered by the sequence ough above. In total, there are 21 vowel phonemes, i.e. sounds which are used to contrast words, in Received Pronunciation (RP), the standard British English accent favoured by BBC newsreaders, though not all English speakers use all of them: Northern English speakers of English do not contrast put and putt, for example, while Southern speakers do; many British English speakers no longer contrast paw and pore.
Key idea: Spoken language first
Without the principle of primacy of the spoken language, modern linguistics could never have developed as a serious academic subject.
Principle 2: Linguistics is descriptive, not prescriptive
The above statement can be found in many an introductory linguistics textbook, and enjoys a status akin to Article One of the Faith among linguists. But why is it so important? Once again, a thought experiment may help here.
Spotlight: The problem of value judgements
Picture yourself for a moment back in the bookshop where you (demonstrating great wisdom and intelligence) decided to purchase this volume. But imagine that the book you thumbed through on the shelf had been called Astronomy: A Complete Introduction, and that your eye had fallen on the following paragraph:
Some people behave as if it is perfectly acceptable for the moon to orbit the earth every 24 hours, but any sensible judge will tell you they are wrong. Nor should the earth’s orbit of the sun take a slovenly 365 days: a 300-day orbit would be neater and more efficient. In fact, it’s purely through idleness that the earth orbits the sun at all: early astronomers who saw the earth at the centre of the universe in fact had a very good idea of how things should be. The stars in the night sky are scattered disagreeably, and the less said about Jupiter’s ugly moons, the better.
Clearly, no series editor would ever publish such drivel, but had one done so, there’s little doubt that the book would have stayed on the shelf. You would quite reasonably have objected that, instead of describing the universe as it is, the author has chosen to tediously rehearse his personal prejudices about how it should be. Arbitrary aesthetic judgements are peddled (‘Jupiter’s ugly moons’) and the universe is ascribed negative moral traits, like laziness or slovenliness, for which there can be no possible justification. This work is grossly unscientific, you would surely have concluded, and cannot be taken seriously.
And yet, surprisingly, when it comes to language, we readily accept thinking of this kind. Prescriptive judgements are so common, in fact, that they often pass unnoticed. When we hear, for example, that ‘standards of English are declining’, or that a speaker has ‘slovenly speech’, we rarely to stop to question the basis on which such judgements are made.
In Britain in particular, linguistic value judgements find expression in the entrenched view that some accents are ‘better’ than others. During the Second World War, the broadcaster Wilfred Pickles was asked, apparently in an attempt to confuse the Nazis, to read the news in his native Yorkshire accent rather than in RP. The experiment was soon ended when it became clear that listeners were objecting, and in some cases no longer trusting the information they were being given. As recently as 2006, Olympic gold medallist turned broadcaster Sally Gunnell left the BBC following criticism of her ‘awful estuary English’.
In cases like these, the yardstick for acceptable speech is a social rather than linguistic one: speakers are condemned for using what are perceived to be low-status accents rather than the prestige standard pronunciation. But in linguistic terms, there is nothing inherently superior about RP, nor any reason to favour any one accent, or language variety, over another. Associated primarily with educated, middle-class speakers based in the Home Counties around London, the prestige of RP merely reflects the social advantages its speakers tend to possess. That wealth and power in Britain are largely centred around London is a matter of historical accident: had the UK capital been Gateshead, Dundee or Bristol, then British conceptions of ‘correct’ pronunciation would be very different, and what we now call ‘Received Pronunciation’ – if it existed at all – would be just another low-status accent which purists would enjoin us to avoid (or, for a small fee, offer to ‘cure’ us of).
That linguistic value judgements have a social rather than linguistic basis is quite simple to demonstrate. In what are termed ‘matched guise’ experiments, participants are asked to listen to recordings of speakers saying the same thing, but in different regional accents, i.e. as far as possible all factors except the speaker’s pronunciation are held constant. When native speakers of British English are asked to evaluate the accents of other Britons, whom they cannot see, in terms of intelligence, friendliness, trustworthiness, etc., there is a remarkable consistency in their responses. City accents, particularly those of London, Birmingham and Liverpool, are negatively evaluated, whereas those associated with less densely populated areas, notably the West of England or South Wales, are viewed more positively. Speakers of RP are generally seen to be the most intelligent, though not always as friendly as speakers of some regional accents.
When the same recordings are played to non-native speakers of English, however, this remarkable consensus evaporates, and there’s no agreement at all about which accents are ‘beautiful’, or connote friendliness, honesty or intelligence. Similar findings have been obtained elsewhere, notably in North America, and it’s hard not to conclude that informants are responding not to any linguistic qualities but to social and regional stereotypes associated with the accents in question.
‘Advanced’, ‘beautiful’ and ‘primitive’ languages
Just as some accents are evaluated more highly than others, many people believe that some languages are ‘better’ or ‘more beautiful’ than others. Many a French president has commented on the supposed ‘clarity’ and ‘precision’ of French, as if clear thinking could not be expressed equally well in another language. Others cite, for example, the richness of Shakespeare’s poetry as evidence for the supposed superiority or inherent beauty of English. Matched guise tests again refute claims that any one language is more beautiful than any other: when played to hearers unfamiliar with European languages, no consensus emerges regarding the aesthetic superiority of any one language as opposed to another. More generally, arguments for the superiority of a given language tend to confuse the rhetorical or linguistic dexterity of some individuals with the qualities of the language itself. The obvious problem here is that speakers do not possess these skills in equal measure, as can be seen in the following examples:
Rarely is the question asked: Is our children learning?George W. Bush (US President 2000–8) 11 January 2000
When you take the UNCF model that, what a waste it is to lose one's mind, or not to have a mind is being very wasteful, how true that is.J. Danforth (‘Dan’) Quayle (US Vice-President 1988–92) 9 May 1989
So I think the basic point that it is necessary in order to have private capital in our industries to get the extra resources that we do want that you have to be privatized is not borne out by the facts, in other countries, and neither should we have it here also and if he’s any doubts about that go and have a look at the reports that talk it.John Prescott MP (UK Deputy Prime Minister 1997–2007) 18 May 1992
A language – any language – is as precise an instrument as its native speakers need it to be for the expression of complex ideas or feelings, and will be used more effectively by some speakers than by others. Don’t blame the language if the thinking it expresses is muddled.
Another common belief is that there are ‘primitive’ languages, just as there are ‘primitive’ societies. Here again, a widely held perception has no basis in linguistic fact. Indeed, if our criterion were grammatical complexity, it might be easier to make the opposite case, namely that languages spoken in isolated, ‘primitive’ societies are often more complex than those used in technologically advanced societies which have been subject to high levels of contact (we’ll look more closely at the effects of contact and isolation on linguistic change in Chapter 13).
This is not, of course, to say that languages will be equally rich in all areas of their lexicon, or vocabulary. We would not expect a language such as Pirahã, spoken by a remote Amazonian tribe of about 250 people, to be as rich in information technology vocabulary as, for example, English: for now at least, Pirahã speakers have little need for such terms and consequently have not developed them to a high degree. But this does not mean that so-called ‘primitive’ languages spoken in less developed societies are unable to acquire new resources when they do need them: in fact they do so with remarkable ease, often by borrowing from other languages. A case in point is English following the Norman Conquest, which borrowed heavily from Norman French: estimates have suggested that around 30–40 per cent of modern English vocabulary is ultimately of French origin. It’s certainly true that English, French, Russian and Spanish are more widely spoken, and more prestigious, than Pirahã, Inuit or Guugu Yimidhirr, but again this reflects socio-political realities rather than any superiority in linguistic terms. To a linguist, all languages (and dialects) are equal.
Linguistic purism
So ingrained is the habit of making linguistic value judgements that it can be difficult to distinguish descriptive statements from prescriptive ones. To a linguist, a grammatical sentence is one that a native speaker either produces or accepts as possible in his/her language. Purists, on the other hand, see only a prestige or standard variety as acceptable, and condemn transgressions against its norms. Because purists often present prescriptive rules as if they were descriptive ones, statements like ‘X is not English’ can be ambiguous: they appear to mean ‘No English speaker would ever say X’, but often in practice mean ‘Some English speakers do say X, but I don’t think they should do’, which is a different claim entirely. Purists, moreover, often justify their strictures in terms that have little to do with language, as the following examples of prescriptive English rules will demonstrate.
• ‘“He hasn’t got none” is ungrammatical.’
The word ‘ungrammatical’ here is immediately problematic, as there are clearly many English speakers whose non-standard grammars allow such constructions, which is why we regularly hear examples of constructions like this. The justification for this stricture in standard English, however, is that it employs a double negative, which amounts to a positive, so this really means ‘he does have some’.
Purists’ reasoning here is appealing: in mathematics, taking away a negative is the same as adding a positive, so the two negatives might be seen to cancel each other out in this sentence. And it’s certainly true that this sentence could be construed in that way, though this would normally require a marked stress pattern (He hasn’t got none). But this superficial logic in fact begs a number of questions. Are languages generally like mathematics? Should they be?
Unlike mathematics, grammar is riddled with idiosyncrasies that defy ‘logic’ in any conventional sense; grammatical rules, moreover, are subject to variation and change, whereas mathematical formulae express (or purport to express) logical truths that are universal and timeless (e.g. A = πr2, where A = area of a circle, r = radius and π = approximately 3.14). So it’s odd, to say the least, that language should be evaluated according to mathematical criteria, and intuitively unlikely that a principle like ‘two negatives make a positive’ will have much importance cross-linguistically. In fact, the double negative construction turns out to be very common not just in non-standard English dialects, but also in the standard varieties of many major languages. No one complains about their use in French (je **ne** sais **pas**) or Spanish (yo **no** tengo **nada**), and double negatives occur regularly in Russian, Italian, Hungarian, Arabic, Breton and Portuguese, to give but a few examples. The double negative stricture is a good example of an arbitrary prescriptive rule being dressed up in logical clothes.
• ‘You can’t split an infinitive in English.’
The objection here is to placing an adverb between the two parts of an English infinitive, e.g. in to generally agree. Again, such structures are commonly used, as Star Trek fans of a certain age will recall:
These are the voyages of the starship Enterprise. Its five-year mission: to explore strange new worlds. To seek out new life, and new civilizations. To boldly go where no man has gone before! (Emphasis William Shatner’s.)
As Captain James T. Kirk opened each new episode with those immortal words, legions of prescriptive grammarians would rage in their armchairs at the incalculable damage inflicted on the nation’s youth by such grammatical barbarism: ‘boldly to go’ or ‘to go boldly’, fine, but ‘to boldly go’? Never.
The reason for their ire goes back to the eighteenth century, and an analogy drawn by grammarians between Latin and English. English infinitives supported by the preposition to were required to remain fused as single unit, like their one-word Latin counterparts, e.g. amare (to love). The grammar of English was for many years described using the same categories as those applied to Latin, and many of our prescriptive rules (e.g. that one should not end a sentence with a preposition, or that one should say ‘It is I’ rather than ‘It is me’) derive ultimately from Latin. But it’s patently nonsensical to require one language to follow the rules of another, and English is very different from Latin in almost every respect. Unlike Latin, English doesn’t have (among other things) noun gender, case marking of nouns (apart, arguably, from genitive ’s), adjective agreement or a fully marked verb paradigm, so it seems perverse in any case to focus on this particular construction.
Spotlight: Changing norms
Did you notice the split infinitive (‘to tediously rehearse’) earlier in this chapter (p. 10) and did you mentally ‘correct’ it? If your answer to both questions is ‘no’, that’s perhaps an indication of how linguistic norms change over time. Purists often identify and resist developments they dislike, but find themselves as powerless to stop them as Canute was to hold back the tide. Yesterday’s ‘slovenly speech’ is often today’s standard, and the split infinitive nowadays attracts far less condemnation than it used to.
Once again, the superficial logic behind a prescriptive rule is based ultimately on a failure to compare like with like: double negatives are condemned on the basis that ‘language should be like mathematics’, while split infinitives are ruled out on the grounds that ‘English should be like Latin’. Linguists prefer to focus on what native speakers of English or other languages actually do say, and to leave prescriptive judgements – and their often spurious justifications – to others.
Key idea: Descriptive or prescriptive?
You have learned in this chapter that linguists see their task as describing language and attempting to find explanations for real language data, rather than telling people how they should speak: linguistics is descriptive, not prescriptive. In this connection we saw that there are two kinds of rule. It’s a descriptive syntactic rule of English that the house is grammatical, but not \_house the_ (a structure which is fine in Swedish and Bulgarian), because no English speaker would naturally say this. Similarly it’s a descriptive rule of English phonology that no word can begin with the sequence \vdr-, which is a perfectly acceptable word-initial sequence in Russian.
Rules of the ‘don’t split infinitives’ kind, on the other hand, are prescriptive, in that they set out what purists think speakers ought to do. Prescriptive rules are generally associated with the usage of a dominant or prestige group, and are generally reinforced by the formal education system: that they need to be, of course, is a sure indication that they are often transgressed. A linguist’s first task is to describe the rules that a native speaker unconsciously obeys, whether or not these correspond to those of standard usage. In doing so, linguists accord primacy to speech because children acquire language through hearing and speaking before learning to read and write, and because conventional writing systems are at best an inconsistent and poor reflection of that speech. These two principles are fundamental to linguistics, and if you’ve grasped them, congratulations! You’re starting to think like a linguist.
1By convention, forms which do not occur are marked with an asterisk.
Fact-check
[1](answers.mdx#rfn1-1) Linguistics is …
about improving languages
about teaching people how to speak their mother tongue properly
about improving people’s writing skills
none of the above
[2](answers.mdx#rfn1-2) Which of the following are linguists unlikely to believe?
the only languages that don’t change are dead languages
Latin stands out above all others as a perfect language
linguistics is a science
spelling reform in some languages is desirable
[3](answers.mdx#rfn1-3) Linguists often talk of ‘language varieties’ because:
diversity is a good thing
everyone speaks differently
what constitutes a ‘language’ or ‘dialect’ tends to be decided on political rather than linguistic grounds
some languages are more difficult to learn than others
[4](answers.mdx#rfn1-4) Spelling and pronunciation are often out of step because:
the written language rarely keeps pace with changes in speech
spelling often reflects a standard usage used only by a minority
spelling sometimes incorporates grammatical information not realized in speech
all of the above
[5](answers.mdx#rfn1-5) The International Phonetic Alphabet (IPA) is important because:
it provides linguists with a means of accurately transcribing the sounds of all languages
it underpins all modern spelling systems
it enables speakers of different dialects to communicate with each other
it’s useful when spelling out words over the telephone
[6](answers.mdx#rfn1-6) Prescriptive rules of English often derive from Latin because:
Latin is an excellent template for all modern languages
English is closely related to Latin, and its grammar follows the same pattern
purists often had a Classical training and applied Latin strictures uncritically to English
English naturally acquired Latin rules during the Roman occupation, and has never lost them
[7](answers.mdx#rfn1-7) How many vowels does English have?
5
6
about 21, but not all speakers use the same set
26
[8](answers.mdx#rfn1-8) Which of the following is grammatically ill-formed in English?
he ain’t got none
the house beautiful
the chases dog a cat
all the above
[9](answers.mdx#rfn1-9) Received Pronunication (RP) is:
a prestigious accent
grammatically correct English
an English dialect
accepted by most linguists as objectively the most beautiful English accent
[10](answers.mdx#rfn1-10) Standards of spoken English are declining because:
journalists and TV presenters think sloppy speech is trendy
children don’t learn Latin in school any more
basic grammar is no longer valued
who says standards are declining? Languages change – live with it!
Dig deeper
The language myths discussed above, and many others, are entertainingly dissected in Bauer & Trudgill (eds) Language Myths (Penguin, 1998): see in particular Cheshire on double negatives, Giles and Niedzielski on ‘beautiful’ languages and Evans on ‘primitive’ ones. Milroy’s chapter probes the often-implicit appeal to Latin in prescriptive grammarians’ rules of English, as does the first chapter of Palmer’s excellent Grammar (Pelican, 1971) and Chapter 9 of Trask’s Language: the Basics (Routledge, 1995).
For a good recent introduction to the principles of the linguistic science, see Blake’s All About Language (Oxford University Press, 2008), Chapter 1, or Chapter 1 of Fromkin, Rodman & Hyams, An Introduction to Language (10th edition, Wadsworth, 2013). The first two chapters of Lyons’ Language and Linguistics (Cambridge University Press, 1981) also address many of the issues raised here in greater detail.
The ghoti for fish proposal is often attributed to the playwright George Bernard Shaw, a strong advocate of spelling reform, but in fact has been traced back to 1855, a year before his birth: for details, see http://languagelog.ldc.upenn.edu/nll/?p=81.
Wilfred Pickles’ use of a Yorkshire accent in broadcasting is reported by L. Mugglestone, ‘Talking Proper’: The Rise of Accent as Social Symbol (2nd edition, Oxford University Press, 2003), pp. 270–2.
For more on matched guise experiments, see Wardhaugh’s Introduction to Sociolinguistics (6th edition, Blackwell, 2010), pp. 110–12, or Fasold’s The Sociolinguistics of Society (Blackwell, 1984), pp. 152–8.