amazon.co.ukLanguage and Gender
A* revision for AQA 7702 Paper 2 and OCR H470/02: deficit, dominance, difference and diversity models from Lakoff to Cameron, gender representation in media texts, and how to deploy theorists for top-band AO2 marks.
🎯 Learning Objectives
- Outline and evaluate the deficit model (Jespersen; Lakoff 1975) with its criticisms of evidence and scope (AO2).
- Evaluate dominance research - Zimmerman & West (1975), Fishman (1983) - including the interruption/overlap distinction and O'Barr & Atkins' (1980) 'powerless language' critique (AO2).
- Assess the difference model (Tannen 1990) against diversity and dynamic approaches - Cameron (2007), Butler's (1990) performativity, Eckert & McConnell-Ginet's communities of practice (AO2).
- Analyse how unseen media texts represent women and men through marked lexis, semantic derogation, agency and multimodal choices (AO1/AO3).
- Deploy two or three theorists woven into an argument rather than listed, matching the AQA AO2 'evaluate the idea' essay demands (AO1/AO2).
- Explain how contextual factors - era, domain (public/private), publication - shape gendered language use (AO3).
📋 Key Definitions and Core Concepts
Deficit model: Early framework holding that women's language is weaker or less effective than men's - Jespersen (1922); formalised as 'women's language' by Lakoff (1975).
Dominance model: Approach treating language as a site where men control conversation and women are silenced - Zimmerman & West (1975); Fishman (1983).
Difference model: Tannen's (1990) cross-cultural account: men and women speak different but equal subcultures (report vs rapport talk), so miscommunication is cultural mismatch, not domination.
Diversity (dynamic) model: Later approach treating gender as one variable among many, with language use varying by context and community of practice - Cameron (2007); Eckert & McConnell-Ginet.
Gender performativity: Butler's (1990) claim that gender identity is constituted through repeated stylised acts - including speech - rather than expressed from a fixed inner essence.
Interruption vs overlap: Interruption intrudes before a current speaker's turn is complete (potentially dominance); overlap is simultaneous talk that supports or completes the other's turn (Tannen: often affiliative).
Report talk vs rapport talk: Tannen's (1990) pair: men's public, information- and status-focused talk vs women's private, connection-focused talk.
Semantic derogation: Schulz's (1975) finding that lexical pairs pejorate for the female term over time (master/mistress, bachelor/spinster).
Marked term: The form that carries extra semantic weight against an unmarked norm - 'female surgeon' marks sex where 'surgeon' defaults to male (Spender's 1980 male-as-norm).
Community of practice: Eckert & McConnell-Ginet's (1992) unit: a group with shared practices and goals, within which gendered styles emerge - not innate, but built jointly.
📌 Key Points
Lakoff (1975) 'women's language': hedges, tag questions, empty adjectives, intensifiers, superpolite forms, hypercorrect grammar, rising intonation on declaratives.
O'Barr & Atkins (1980): courtroom 'powerless language' correlates with social status, not sex - the standard AO2 critique of the deficit model.
Zimmerman & West (1975): men made almost all recorded interruptions in cross-sex conversations; Fishman (1983): women do the 'housework' of talk - topic-raising, questions, minimal responses.
Tannen (1990): report talk vs rapport talk; difference is cultural mismatch, not domination.
Cameron (2007): differences are small, within-gender variation exceeds between-gender; the difference model is a 'zombie' statistic recycled by publishers.
Butler (1990): gender is performed through repeated stylistic acts; Eckert & McConnell-Ginet: styles emerge in communities of practice.
AQA 7702/2 Section A = one 'Evaluate the idea that...' essay, 30 marks; OCR H470/02 Section B (Language in the media) = investigate how language features and contextual factors construct meanings, 24 marks.
Deficit and dominance: the founding frameworks
Gender sits within Language diversity (AQA 7702 Paper 2, Section A) and surfaces in unseen media texts on OCR H470/02 Section B (Language in the media, 24 marks). The models are an evaluation toolkit - but examiners reward critique, not citation-listing.
- Deficit - Jespersen (1922): his chapter 'The Woman' claimed women's vocabulary was smaller and more conservative while men innovated; it set the deficit frame that later research inherited, with no systematic data behind it.
- Deficit - Lakoff (1975), Language and Woman's Place: 'women's language' comprises hedges ('sort of'), tag questions ('isn't it?'), empty adjectives ('divine'), intensifiers ('so'), superpolite forms, hypercorrect grammar and rising intonation on declaratives. Her claim: women are socialised into tentative speech, and the language reflects - and reproduces - social powerlessness.
- Lakoff critiqued: anecdotal evidence with no recorded corpus; a trivial feature list that invites caricature; women treated as a single homogeneous group. The killer AO2 move is O'Barr & Atkins (1980): courtroom language showed 'powerless language' (including Lakoff's own features) correlates with social status, not sex - high-status women spoke like high-status men. Lakoff's markers are markers of powerlessness, not of gender.
- Dominance - Zimmerman & West (1975): in recorded informal cross-sex conversations, men made almost all the interruptions, while same-sex conversations (male and female) showed few. Their claim: conversation is a resource men control, silencing women's contributions.
- Dominance - Fishman (1983): analysing taperecorded couple conversations, women did the 'housework' of talk - raising topics (often dropped by male partners), asking more questions, producing minimal responses ('mm', 'yeah') to keep talk going - while men set the topics that developed. Women's linguistic labour produced men's conversational power.
- Dominance critiqued: small, self-selected samples; and interruption is not automatically dominance - Tannen distinguished intrusive interruption from supportive overlap, so the same phenomenon can be read two ways. West & Zimmerman's later 'Doing Gender' (1987) reframed the data: participants perform gender in interaction, pointing towards the diversity model.
- Spender (1980): 'male as norm' - language encodes men as the standard and women as the marked deviation, a bridge from dominance research into representation analysis (below).
Why the early frameworks still earn marks: deficit and dominance are not museum pieces - they are the foil every later model is measured against, and AQA AO2 rewards exactly that measurement. A paragraph that names Lakoff's features, cites O'Barr & Atkins' courtroom evidence and states the consequence ('the features mark powerlessness, not sex') is a complete AO2 unit: claim, evidence, evaluation. The same paragraph structure serves dominance: Zimmerman & West's finding, Fishman's extension, then the interruption/overlap distinction as the critique. Examiners' reports repeatedly flag candidates who list six models without evaluating any - three well-worn models, each collapsed into claim-evidence-critique, outscore them every time.
Difference and diversity: Tannen, Cameron and performativity
- Difference - Tannen (1990), You Just Don't Understand: the cross-cultural communication model. Men favour report talk - public, informational, status-building; women favour rapport talk - private, connection-building, supportive. Miscommunication ('Different Words, Different Worlds') is cultural mismatch, not domination; overlap, for instance, is reframed as affiliative rather than intrusive. 'Different but equal' - deliberately softer than the dominance model.
- Tannen critiqued: Hyde's (2005) gender similarities hypothesis - a meta-analysis finding males and females similar on most communication variables; and Cameron (2007) argued the difference model is a 'zombie' statistic that keeps being recycled because it sells books: it naturalises stereotypes as science and excuses inequality ('that's just how men/women are').
- Diversity/dynamic - Cameron (2007), The Myth of Mars and Venus: measured gender differences are small; within-gender variation exceeds between-gender variation; style shifts with context, role and audience - the same woman in a boardroom, on the phone to a friend and online produces three distinct registers.
- Performativity - Butler (1990), Gender Trouble: gender is not expressed but constituted through repeated stylised acts, including speech - there is no pre-linguistic 'feminine essence' that language merely reports. This reframes the whole debate: gendered language is something speakers do, repeatedly, in context.
- Communities of practice - Eckert & McConnell-Ginet (1992): gendered styles emerge within groups with shared practices and goals - e.g. a sports team, a workplace, a friendship group - so 'masculine' directness may be a team norm, not a male trait. This unit of analysis allows intersectional work (gender + class + region + age) that deficit and difference models cannot handle.
Exam synthesis: the strongest AQA essays treat the models as a historical arc rather than rival truths - deficit and dominance as early lenses with fragile evidence, difference as a popularisation that dodges power, diversity and performativity as the current lens that handles context and within-group variation. Weigh each model against the data and against its social consequences (AO2), and anchor claims in the contexts where gendered language is produced: public vs private domains, media, workplace, online (AO3). Two or three theorists woven into an argument beat six listed in a paragraph.
Gender and media representation
OCR's Language in the media question (H470/02 Section B, 24 marks) asks you to 'investigate how language features and contextual factors construct meanings in this text' - frequently a media text representing women, men or gender roles. AQA Paper 2 texts about gender appear as discourses to analyse and evaluate. The method is the representation method (see Textual Variations topic) pointed at gender.
- Marked lexis: 'female surgeon', 'male nurse', 'career woman' - the modifier marks sex against an unmarked male default (Spender's male-as-norm). Ask in every text: which terms are marked, which unmarked, and what that asymmetry assumes?
- Semantic derogation: Schulz (1975) showed paired terms pejorate for the female member over time - master/mistress, bachelor/spinster. Where a text uses one side of such a pair, the history of the word rides with it.
- Generic and gendered forms: 'mankind', 'chairman', generic 'he' - and reform attempts ('chair', 'firefighter', singular 'they'), which link gender representation to language change and prescriptivism: a productive AQA link between the diversity and change strands of Paper 2.
- Goffman (1979), Gender Advertisements: the ritualisation of subordination in visual representation - women depicted lying down, smaller, 'licensed withdrawal', averted gaze. Use it as the multimodal parallel: who does the grammar of the image empower?
- Asymmetry of evaluation: media texts typically judge women by appearance and men by action - evaluative adjectives ('glamorous', 'family-first') applied asymmetrically to otherwise parallel figures.
Method for the unseen media text: (1) state what gender group is represented and how ('Text B represents women professionals as exceptional rather than ordinary'); (2) work the levels - lexis (marked terms, evaluative and appearance-focused adjectives, semantic fields), grammar (agency: who acts, who is acted upon; passives concealing perpetrators or crediting subjects), pragmatics (presuppositions about roles - 'working mothers' presumes motherhood is the default that needs qualifying), discourse (who is quoted, foregrounded, given the anecdote), graphology (image roles, captioning, layout); (3) tie each finding to contextual factors - publication era, readership, genre - because the OCR mark scheme awards context for explaining the construction, not for decorating it. End by evaluating the representation: perceptive answers say what version of reality the text is selling and who benefits from it.
Media discourses and attitudes: beyond representing gender groups, media texts also carry discourses about language itself - anxiety about singular 'they', mockery of vocal fry or uptalk, nostalgia for 'proper' English. These are evaluative texts to analyse (how the attitudes are linguistically constructed: evaluative lexis, modality, presupposition) and evaluate (whose interests the attitude serves - AQA Paper 2's discourses questions do exactly this). The link is productive in both directions: gender representation draws on language change (pejoration, reform), and language attitudes draw on gender (which speakers get policed). Weaving the two strands is what makes a top-band answer look synoptic rather than topic-bound.
Exam technique: deploying theory in essays and unseen analysis
AQA 7702/2 Section A offers a choice of two 'Evaluate the idea that...' statements, 30 marks each - e.g. June 2017: 'Evaluate the idea that a person's language use is completely determined by the social groups they belong to.' Gender is one such social group; occupation and region are others, so the same essay mechanics serve this topic and the Language and Power topic.
- Command word: evaluate - a verdict is compulsory. Structure as thesis, counter-evidence, refined verdict: e.g. determinism fails because Cameron's within-gender variation and Butler's performativity show speakers exceeding their categories, but social groups still constrain access to registers and to power (O'Barr & Atkins' status effect).
- Theorist deployment: two or three, woven: Lakoff's features to open the deficit frame, O'Barr & Atkins to collapse it, Cameron or Eckert & McConnell-Ginet to rebuild. Name the study, its finding and its limitation in one sentence each - a paragraph per theorist listing biographies scores AO2 at the bottom of the scale.
- AO3: contexts decide the verdict - public vs private domains (Fishman's couple data is domestic; phone-ins are broadcast), era (1970s samples predate the online audience), and medium (interruption behaves differently on air and in chat). Use context to explain why the research found what it found.
- OCR H470/02 Section B (24 marks): investigate language features and contextual factors - deploy gender concepts (marked lexis, derogation, agency) plus the representation method, and credit the text's production context and audience. About 45 minutes: budget three or four developed analytical strands rather than exhaustive coverage.
- Directed writing (OCR H470/01 Section B, 24 marks; AQA 7702/2 Section B): when writing about a gender issue, avoid recycling the stereotypes the theories critique - the examiner notices when a script about gendered language performs the deficit model it claims to evaluate.
A worked verdict chain for the essay: thesis ('social groups constrain but do not determine language use') - so for ('Lakoff's features show socialisation into tentative styles') - so against ('but O'Barr & Atkins show the same features track status, not sex') - so refined ('groups condition access; speakers exceed the conditions') - so context ('and the 1970s samples that founded the field predate the online contexts where style-shifting is most visible'). Each link is one sentence with a theorist attached; the whole chain is the essay's spine, and the remaining paragraphs flesh out each link with data. That is the difference between an essay that has theorists and an essay that uses them.
💡 Worked Example
Exam-style: Evaluate the idea that a person's language use is completely determined by the social groups they belong to. [30 marks] (AQA 7702/2, June 2017, Question 01)
Mark-scheme logic: Mark-scheme logic (AO1 + AO2 + AO3, 30 marks):<br><br><strong>AO1:</strong> fluent, coherent essay with accurate terminology (hedges, tag questions, interruptions, report/rapport talk, communities of practice) and a clear argumentative line sustained from thesis to verdict.<br><br><strong>AO2 - evaluating the idea (the core of the essay):</strong> the idea must be tested, not merely illustrated.<br>(1) <strong>Evidence for group influence:</strong> Lakoff's (1975) 'women's language' features as socialisation into tentativeness; Zimmerman & West's (1975) cross-sex interruption findings and Fishman's (1983) conversational 'housework' showing gendered patterns of talk; occupational and regional varieties conditioning registers.<br>(2) <strong>Evidence against complete determination:</strong> O'Barr & Atkins (1980) - 'powerless language' tracks status, not sex, so groups do not fix style; Cameron (2007) - within-gender variation exceeds between-gender; Hyde (2005) - gender similarities on most variables; Butler (1990) and Eckert & McConnell-Ginet - gender is performed and built in communities of practice, so speakers exceed their categories.<br>(3) <strong>Refined verdict:</strong> social groups shape access to registers and constrain style (deficit/dominance data are real), but 'completely determined' fails against within-group variation and contextual style-shifting - language use is conditioned, not determined.<br><br><strong>AO3:</strong> contexts decide the case - public vs private domains (broadcast phone-ins vs Fishman's domestic data), era of research (1970s samples predate contemporary audiences), medium (interruption in face-to-face talk vs online chat), and the ideological context in which the deficit model was produced. Context woven through the evaluation, not appended.<br><br><strong>Top-band marker:</strong> an explicit, defended verdict plus theorists used as evidence and critique (two or three, integrated), with awareness that the models themselves carry ideological baggage.
❓ Practice Questions & Mark Schemes
Evaluate the idea that women's language is deficient. [30 marks] (AQA 7702/2 Section A style)
Answer: <strong>Thesis:</strong> the deficit idea is historically important but empirically collapsed - verdict delivered early and refined at the end.<br><br><strong>For the idea:</strong> Lakoff (1975) - hedges, tag questions, empty adjectives, intensifiers, superpolite forms, rising intonation on declaratives; her argument that such features are trained into women and reflect powerlessness. Jespersen (1922) as the deficit frame's origin.<br><br><strong>Against:</strong> (1) O'Barr & Atkins (1980): courtroom study - 'powerless language' correlates with social <em>status</em>, not sex; high-status women spoke like high-status men, so the features mark powerlessness, not gender. (2) Hyde (2005): gender similarities hypothesis - males and females similar on most communication variables. (3) Cameron (2007): within-gender variation exceeds between-gender; deficit claims naturalise stereotypes. (4) Butler (1990)/Eckert & McConnell-Ginet: gendered styles are performed and community-built, not deficient versions of a male norm.<br><br><strong>AO3:</strong> the 1970s ideological context producing Lakoff's framework; the domestic vs public domain split (Fishman's couple data vs courtroom and broadcast data); O'Barr & Atkins' institutional setting.<br><br><strong>Verdict:</strong> deficit describes a position some speakers occupy - of any sex - and Lakoff's lasting contribution is the observation that language encodes power, not the claim that women's language is deficient. 'Completely deficient' fails; 'powerless under some conditions' survives.
Using your understanding of relevant ideas and concepts, investigate how language features and contextual factors construct meanings in this magazine article about women in the workplace. [24 marks] (OCR H470/02, Section B - Language in the media, June 2023 wording)
Answer: <strong>Method:</strong> representation of a gender group + contextual construction of meaning, in 3-4 developed strands (about 45 minutes' budget).<br><br><strong>Model hits:</strong> (1) <strong>Marked lexis:</strong> quoted examples - 'female executive', 'career woman' - with the male-as-norm asymmetry (Spender 1980) they assume; evaluative adjectives applied to appearance vs action. (2) <strong>Agency:</strong> who acts and who is acted upon - women professionals framed as overcoming (active) or as balancing (passive/affected); passives concealing who imposes the burden ('mothers are expected to...'). (3) <strong>Pragmatics:</strong> presuppositions - 'working mothers' presumes motherhood as the default needing qualification; 'has it all' presupposes a fixed hierarchy of roles. (4) <strong>Discourse:</strong> who is quoted and foregrounded; anecdote selection; whether women's achievement or their domesticity is the news value. (5) <strong>Context:</strong> publication era (which decade's assumptions), readership (middle-class professionals?), genre conventions (magazine feature: personalisation, lifestyle lexis) - each explaining a linguistic choice. (6) <strong>Evaluation:</strong> the version of reality constructed (women as exceptional/struggling vs ordinary/successful) and its ideological function; optional link to Goffman's (1979) ritualisation of subordination if images accompany the text.<br><br><strong>Top band:</strong> concepts deployed as an explanatory framework with perceptive evaluation, not name-dropped - 'Cameron (2007) would read this article as recycling the Mars-Venus stereotype' earns more than 'Cameron says gender differences are small'.
Evaluate Tannen's difference model of language and gender. [30 marks] (AQA 7702/2 Section A style)
Answer: <strong>Thesis:</strong> the difference model is explanatory for miscommunication but evasive on power - evaluated on both counts.<br><br><strong>For:</strong> Tannen (1990) - report talk vs rapport talk explains cross-sex misunderstanding as cultural mismatch; reinterprets overlap as affiliative rather than intrusive, correcting a weakness in interruption-based dominance research; 'different but equal' framing avoids deficit language.<br><br><strong>Against:</strong> (1) It sidesteps power: if miscommunication is mutual cultural mismatch, no party is accountable - Cameron (2007) argues this excuse keeps inequality intact and that the model is a 'zombie' statistic recycled by publishers. (2) Empirical weakness: Hyde (2005) - gender similarities on most variables; within-gender variation exceeds between-gender. (3) It treats 'men' and 'women' as homogeneous cultures, ignoring class, region, age and community of practice (Eckert & McConnell-Ginet 1992). (4) Performativity (Butler 1990): 'different worlds' assumes pre-existing gendered essences rather than styles built through interaction.<br><br><strong>AO3:</strong> the popularisation context - Tannen's findings from academic research stretched in the trade books; public vs private domains where report/rapport patterns hold or collapse.<br><br><strong>Verdict:</strong> difference is descriptively useful, analytically limited: strongest where contexts genuinely differ (public vs private talk), weakest as an account of institutional power - the diversity model explains what difference cannot.
Evaluate the claim that interruptions signal male dominance in conversation. [25 marks] (AQA-style AO2/AO3 essay)
Answer: <strong>Thesis:</strong> the claim holds only if interruption is defined as intrusive - and that definition is itself the debate.<br><br><strong>For:</strong> Zimmerman & West (1975) - in cross-sex conversations men made almost all recorded interruptions, while same-sex talk showed few; conversation as a resource men control. Fishman (1983) - the wider pattern of women's topic-raising being dropped supports a dominance reading.<br><br><strong>Against:</strong> (1) Tannen (1990): overlap is frequently affiliative - supportive, completing, collaborative - so the same behaviour can signal engagement rather than domination; intrusive vs supportive overlap must be distinguished before any claim. (2) Sample limitations: small, informal, self-selected dyads - not evidence for all contexts (courtrooms, meetings, broadcast differ; O'Barr & Atkins show status interacts with any such effect). (3) Cameron (2007): aggregate differences are small; within-gender variation exceeds between-gender. (4) West & Zimmerman's own 'Doing Gender' (1987) reframes the data - participants perform gender, so interruption may be one resource among many for doing (rather than expressing) dominance.<br><br><strong>AO3:</strong> context decides - informal dyadic talk vs institutional and broadcast settings; the era of the data; the medium (face-to-face vs online chat, where timing cues disappear and 'interruption' is message overlap).<br><br><strong>Verdict:</strong> interruption is evidence of dominance only after interactional function is established - a claim about one conversational resource, not a law of gender.
Explain how semantic derogation supports the male-as-norm view of language. Illustrate with examples. [10 marks] (OCR ExamBuilder-style AO2/AO1 question)
Answer: <strong>Model hits:</strong> (1) <strong>Define male-as-norm:</strong> Spender (1980) - language treats the male term as the unmarked default, the female term as a marked deviation. (2) <strong>Define semantic derogation:</strong> Schulz (1975) - in paired terms, the female member pejorates over time while the male member stays neutral or positive. (3) <strong>Paired examples:</strong> master (skilled controller) vs mistress (now sexualised); bachelor (unmarried by choice) vs spinster (pitied); governor vs governess; king vs queen in slang usage. (4) <strong>The link:</strong> because the female term drifts negative and the male does not, the language itself encodes women as the marked, devalued category - the pattern is systemic, not incidental. (5) <strong>Application:</strong> marked terms in media texts ('female surgeon') ride the same asymmetry; reform attempts ('chair', singular 'they') are attempts to reset the norm - linking to language change and prescriptivism. A 10-mark answer needs the two definitions, three or four paired examples and the explanatory link - not a survey of every gender theorist.