HookFifty years on, the accent league table has barely moved
In 1970 Howard Giles played listeners a range of British accents and asked them to rate the speakers. Received Pronunciation came out on top for intelligence and competence; urban vernaculars like Birmingham sat at the bottom. In 2019–20 the Accent Bias in Britain project re-ran the exercise at scale — over 4,000 respondents rating 38 accent labels — and found the league table almost unchanged: 'Queen's English' style speech still at the top, Afro-Caribbean, Multicultural London English and Birmingham still at the bottom. Half a century of social change; the same hierarchy.
Now flip it round: in those same fifty years the language itself transformed. A brand-new London variety was born, Estuary English spread along the Thames corridor, wicked and sick flipped polarity, and English acquired hundreds of millions of new speakers who never set foot in Britain. That pairing — a language in constant motion, judged by attitudes that barely move — is the whole of this topic. English varies across space (diversity) and across time (change), and the Section A essay asks you to evaluate ideas about both: with named research, real numbers and each side weighed.
DataAccents, dialects and the map of the British Isles
Get the terms straight first, because AO1 does: an accent is pronunciation only; a dialect includes lexis and grammar too. Received Pronunciation is the odd one out — a prestige accent tied to class and education rather than to any region, and spoken in its classic form by only a small minority (commonly estimated at around 3% of the population). Everyone else is somewhere on the map: Tyneside's gan doon for 'go down' and negative divn't; Yorkshire's definite article reduction ('down t'road'); West Country rhoticity, sounding the /r/ in farm; Scots lexis like outwith and ken. The Survey of English Dialects (1950s, led by Harold Orton) sent fieldworkers into more than 300 mostly rural localities to record exactly this variation before it eroded.
Erosion is the modern story — but a complicated one. Dialect levelling is the smoothing away of locally distinctive forms as mobility mixes speakers: Paul Kerswill's Milton Keynes research found children of in-migrating families building a levelled koiné that matched none of their parents' varieties. David Rosewarne coined Estuary English (1984) for the middle ground spreading out from London — glottal stops and l-vocalisation without RP's social baggage. Yet levelling coexists with brand-new diversity (MLE), which is your ready-made evaluation point: British English is becoming less locally distinct and more socially and ethnically varied at the same time. Attitudes are measured with Giles's matched-guise technique — same speaker, different accents, listeners rate the 'different people' they think they hear — and explained by his accommodation theory: speakers converge towards audiences they want approval from, and diverge to mark distance.
ModelGlobal Englishes — whose language is it now?
Braj Kachru mapped English as three circles: an inner circle of native-speaker countries (Britain, the US, Australia), an outer circle where colonial history left English as an institutional second language (India, Nigeria, Singapore — the English of courts, schools and parliaments), and an expanding circle where it is learned as a foreign language. On current estimates around two billion people use English, and second- and foreign-language users outnumber native speakers several times over — which quietly demolishes the idea that Britain owns the standard.
Outer-circle varieties are not failed British English; they are systems. Indian English prepone (the opposite of postpone — filling a genuine lexical gap, now in the OED); Singapore English's pragmatic particle lah; Nigerian English go-slow for a traffic jam. Edgar Schneider's dynamic model tracks how such varieties mature in five phases, from foundation through nativisation to independent norms of their own. Jennifer Jenkins pushes further: most English conversations today happen between non-native speakers, so she studied English as a lingua franca and proposed a Lingua Franca Core — the pronunciation features that actually block intelligibility — noting that famous shibboleths like the 'th' sounds mostly don't. The essay-ready debate sits on top: a pluralist World Englishes reading (varieties as legitimate local assets) against Robert Phillipson's linguistic imperialism (English's spread as a bulldozer flattening smaller languages — David Crystal estimates a language dies roughly every two weeks). Strong answers can argue both with the same data.
Model1600 to now — what actually changed
Put a text from the 1600s next to one from this morning and every language level has shifted. Grammar: the intimate second person thou/thee collapsed into universal you by around 1700; the -eth inflection (hath, goeth) gave way to -s; do-support regularised ('I know not' became 'I don't know'); and multiple negation, unremarkable for centuries, was outlawed by 18th-century grammarians. Orthography: the long s (ſ) survived in print until about 1800, u and v were positional variants, spelling was unfixed, and 17th- and 18th-century printers capitalised nouns as a matter of style.
Then came codification. Samuel Johnson's Dictionary (1755) — 42,773 entries, nine years' work — set out to fix the language and famously conceded in its own preface that no lexicographer can hold a living tongue still. Robert Lowth's grammar (1762) supplied the school rules, double negatives banned. The OED (from 1884) flipped the philosophy: record on historical principles rather than legislate. The 20th and 21st centuries brought informalisation — public writing drifting towards speech — plus new electronic modes, waves of American influence, and the mainstream revival of singular they.
Decode this invented shop notice, written in authentic 1711 style: "To be SOLD at the Sign of the Golden-Lyon, a Quantity of fine Bohea Tea, lately arriv'd; whoſoever hath a Mind to purchaſe the ſame, let them repair thither with all convenient Speed." Feature by feature: ſ is the long s, a positional print variant dead by about 1800 (graphology/orthography). Capitalised common nouns — Quantity, Mind, Speed — are 18th-century typographic fashion, not grammar. arriv'd marks an elided vowel once pronounced (phonological change fossilised in spelling). hath shows -th surviving longest on high-frequency auxiliaries (morphology). whosoever… let them is singular they in 1711 — three centuries before style guides re-admitted it, a lethal fact against 'sloppy modern usage' arguments. repair meaning 'go' has since narrowed to mending things (semantic change), and Bohea is trade-route borrowing from Chinese for black tea (lexical change via contact). The exam move each time: name the feature with the right framework label, date it roughly, and explain it contextually — incomplete standardisation, printers' conventions, empire and trade.
MechanismWhere new words and meanings come from
Word-formation processes, with one clean example each: borrowing (pyjamas from Urdu and Persian, bungalow from Hindi and Bengali, tsunami from Japanese); compounding (doomscrolling); blending (brunch, smog, chugger); clipping (flu, app); initialism versus acronym (FBI spelled out, laser pronounced as a word); affixation (unfriend, the endlessly productive -gate); conversion (to google); eponymy (hoover, boycott).
Meanings drift too, in four canonical directions. Broadening: holiday once meant only a holy day; bird once meant only a young bird. Narrowing: meat was food in general, deer any animal. Amelioration: nice climbed from Latin nescius, 'ignorant', to today's all-purpose approval; pretty began as 'crafty'. Pejoration: silly fell from 'blessed', hussy from 'housewife' — and the fact that pejoration so often strikes words for women (hussy, mistress, spinster) is a ready-made bridge to the sociolect essay. Metaphorical extension powers tech vocabulary: mouse, virus, catfish.
Causes and models are where AO2 evaluation lives. Causes: contact and trade, technology minting lexis with every wave, social change demanding new terms (Ms, singular they), prestige and fashion, economy of effort. Models: the S-curve (change starts slowly among innovators, accelerates through the community, then tails off), Bailey's wave model (change radiates outwards from an epicentre, weakening with social and geographical distance), Halliday's functional view (language changes because users' needs change) against random fluctuation. The strongest paragraphs pair a process with a cause and a model: what changed, why, and how it spread.
ModelPrescriptivism, descriptivism and the war over 'correct'
The prescriptivist tradition has a long pedigree: Jonathan Swift's 1712 Proposal wanted an academy to fix English permanently; Lowth's rules hardened into two centuries of schoolroom law; and the trade still thrives — Lynne Truss's Eats, Shoots & Leaves (2003), a zero-tolerance punctuation manifesto, sold in the millions, and broadcaster John Humphrys branded texters vandals doing to the language what Genghis Khan did to his neighbours. Descriptivism — the working stance of linguistics — describes what speakers systematically do instead of ruling on what they should.
Jean Aitchison's 1996 Reith Lectures named the three images underneath decline-talk: the damp spoon (change as laziness, provoking the same distaste as someone else's damp spoon in the sugar); the crumbling castle (English as a once-perfect building falling into ruin — though no era of perfection ever existed); and the infectious disease (bad usage as contagion — though speakers adopt forms, they don't catch them). James and Lesley Milroy add the complaint tradition — recognisably the same complaints recurring for centuries — and standard language ideology, the belief that one variety simply is the language and the rest are corruptions.
Data usually embarrasses the panic. David Crystal's Txtng: The Gr8 Db8 (2008) found only around 10% of words in typical text messages were abbreviated, and Plester and Wood found children's textism use correlated positively with reading scores. But the sophisticated evaluation move is Deborah Cameron's Verbal Hygiene: everyone regulates language — including linguists, including you when you proofread — so the binary is too neat; the real question is whose norms get enforced, on whom, and with what power behind them.
MechanismThe Section A essay — evaluating an idea, not reciting a history
Section A offers a choice of two evaluative essays — in practice one leaning towards diversity, one towards change — and you answer one in about 45 minutes. The stem hands you an idea: practise against original stems like 'Evaluate the idea that there is no such thing as bad English' or 'Evaluate the idea that English no longer belongs to Britain'. Method: define the loaded term fast (what would 'bad' or 'belongs' even mean, linguistically?), stake a position, then build paragraphs that each test the idea against named evidence — a study or dataset, its limitation, a rival reading.
Move between micro and macro. Micro: a single labelled feature (th-stopping, prepone, t'reduction) used as evidence. Macro: a model or ideology (the S-curve, standard language ideology) used to interpret it. And cross-pollinate deliberately: bring attitudes into diversity essays and diversity into change essays, because levelling is change in progress and accent bias is an attitude to diversity. A model sentence to imitate: 'Kerswill's Milton Keynes children did not inherit their parents' dialects but built a levelled koiné — evidence that mobility, not laziness, drives change; the crumbling-castle reading of the same data mistakes motion for decay.'
VocabularyKey terms the mark scheme pays for
TrapsMisconceptions that cost marks
ExamWhat examiners want
The Section A essay carries 30 marks: 20 for AO2 and 10 for AO1. The AO2 descriptors climb through 'different views and perspectives' towards genuine evaluation — so every claim needs a counterweight, and a paragraph that presents Phillipson's linguistic imperialism should meet the World Englishes reply before it ends. Answer the idea in the stem in every paragraph; the essays that drift into a chronology of English cap out early because narration is not evaluation.
Evidence beats assertion. Quote figures and dates — 42,773 entries, around 3% RP speakers, roughly 10% of texted words abbreviated, an accent hierarchy static across fifty years — and label micro-examples precisely for AO1: koiné, nativisation, pejoration, glottal stop. Register discipline matters too: write 'non-standard', not 'incorrect'; 'variety', not 'slang', unless slang is literally the topic.
Two traps recur in examiners' commentary. First, moralising — arguing that prescriptivists are fools — instead of analysing prescriptivism as a position with social causes; Cameron's verbal-hygiene point is the elegant escape, since it shows both camps regulating language. Second, ignoring half the paper: diversity questions reward change evidence and vice versa, because levelling, MLE and global nativisation are all change happening in real time. Plan five testing paragraphs, about 45 minutes, position stated by the end of your first.