Language investigation ideas
A good investigation will cover each of the Assessment Objectives.
To get a strong AO1 mark an investigation needs to use demanding terminology accurately and incisively, but the specifics will be different for different investigations - ideally cover a range of features and keep an eye open for surprising or unanticipated features in your data.
Starting from features that researchers have discussed is a sensible approach. Any investigation should be built on sound linguistic concepts and ideas from research to cover AO2 - one way of designing a good investigation is extending or challenging work that others have done (O'Barr and Atkins' work in courtrooms was an excellent example of this, taking Lakoff's idea of 'Women's Language' and testing it in a specific context).
AO3 covers understanding the context the investigation is taking place. For good AO3 marks, be as open-minded as possible in considering what's making the data come out the way it has. If your investigation is focused on gender, don't ignore other factors like age, occupation, personality, professional role and so on. A level investigations are by their nature short investigations and cannot come to definitive conclusions about grand topics, so always be tentative in stating your conculsions and consider other factors that may be at play.
Welcome!
Englicious is totally free for everyone to use!
But in exchange, we ask that you register for an account on our site.
If you’ve already registered, you can log in straight away.
Since this is your first visit today, you can see this page by clicking the button below.
Language investigation ideas: Accent and dialect
Does shop assistants’ speech converge with the speech style of their customers?
A version of Labov’s study can easily be done anywhere there are a variety of similar shops (or publicly accessible institutions like sports centres or libraries).
Labov gathered data from three department stores (one expensive, one a discount store and one in the middle – their status was measured by the price of a ladies coat in the store) by inducing assistants to say the phrase ‘fourth floor’. As the rhotic /r/ is a prestigious feature in US accents, the question was whether assistants in the more expensive stores would use the prestige form more often (they did).
This idea has worked well for A level students going into different supermarkets and collecting data on a specific variable (something that can be pronounced in more than one way). In one example, students induced the shop assistants to say the phrase ‘cottage cheese’ (to see whether they glottal-stopped the /t/ sound in the middle of ‘cottage’), but you could design other phrases to test other forms of pronunciation that are socially marked (i.e. seen as high or low status) – th-fronting, for example, could work well.
To collect data on which supermarkets in the investigation might be considered most or least prestigious, choose a specific item that every store sells and compare the prices (some stores have a wide range of e.g. cheese, so it might be worth looking for the absolute cheapest and/or most expensive item of a specific type in each store).
The data collection would need to be consensual, so ask at the store’s customer service desk whether they mind data being collected.
AO1 – this is complicated by how the A level investigations are marked. Labov focused on one variable (rhotic /r/) but that’s too narrow to get good AO1 marks in an A level investigation, so this idea would need much more data than just that. If the interaction involves being taken to the product in question there’s often some chat on the way to the relevant aisle, and that can have interesting elements like facework (e.g. asking/instructing the customer to follow, the assistant representing themself as completely willing to leave whatever they were doing), weak modality to give the customer options (for things like ‘you could try this one…’) and spoken features like phatic content to fill time while walking, and question/answer adjacency pairs for the transactional parts of the conversation.
AO2 - Labov is the obvious and central person for this idea, but discussing facework could involve Brown and Levinson, and both Malcolm Petyt’s work in Bradford and Trudgill’s work looking at non-standard features in Norwich fit well when discussing high and low status features and their association with class. Worth especially looking at the detail of Trudgill’s findings, where people spoke differently (shifted their style) according to whether they were casually conversing with the interviewer or reading a word list aloud. Could arguably get into social network theory research like Jenny Cheshire’s Reading study or the Milroys’ work in Northern Ireland when discussing factors that might be producing whatever results come up (people who have worked in a shop full-time presumably converge to some extent). Ideas from language and occupation research might also be relevant – institutional roles and allowable contributions could probably both be applied to the assistants’ speech.
AO3 – There are lots of possible factors to look at here. The prestige of the store is the obvious one, but the time of day might affect how friendly or chirpy the subjects are (or how close they are to the end of their shifts), the assistants’ experience and familiarity with the locations of items in the store, the age of the assistants (with a possible feeling of more social closeness if they’re similar ages to the investigator), how common the item asked for is, the accent the investigator uses to ask the question, and the social networks of the assistant would all play a role (though some would be hard to get data on without full interviews of the assistants after the experiment).
Language investigation ideas: Ethnicity and social networks
Do second and third generations of immigrant families converge less with local Anglo English dialects?
Research by Sharma and Sankaran measured the use of non-standard British Asian and Cockney features in the speech of two generations of British Asian people. They found that the older generation were converging much more dramatically with people they spoke with. The younger speakers seemed to be confident in their identities, in that they didn’t dramatically shift towards Cockney features at any point in the data, producing Asian-language-derived features more consistently across all the situations in the data. It’s possible that the climate of racism that first generation immigrants arrived to in the 20th century may have affected their language by making the ability to converge essential for them, whereas their children and grandchildren may be more willing to display Asian-British features in all contexts, having grown up more confident in their identity in a more accepting environment. It would be interesting to see how widely these findings apply.
To collect data for this, you would need access to first and second (and third?) generations of a family who moved to the UK in living memory. You would also need to decide on features of pronunciation to focus on – some would need to be associated with Anglo speakers, and some with British-Asian (or British-whatever-else, there's no reason this couldn't be done with speakers of any non-UK heritage) speakers.
AO1 – Sharma and Sankaran identified features of pronunciation to focus on, and besdies these you could look at discourse markers, slang, or any other features that vary between the two groups.
AO2 – Sharma and Sankaran, social network theory (the Milroys, Cheshire in Reading, Cheshire, Fox, Kerswill, Khan, Torgersen...), Trudgill (for convergence, and demonstrating code-switching), Labov on Martha’s Vineyard (demonstrating sounds becoming associated with social groups and certain values).
AO3 – data on things like how the speakers feel about where they’ve grown up and their ethnic heritage could be interesting. Even a simple questionnaire could open interesting questions when compared with the data – e.g. do speakers who feel closer to, say, an Asian family background produce more Asian-British features? Are there differences in how the first and second generations perceive the country they live in, and are there resulting differences in their speech? Also contextual factors like places of work and the social networks of the speakers could be worth considering. The linguistic competence of the speakers would also be worth examining (how many languages does a given speaker speak? Did they grow up learning each of them or were some acquired later?). The amount of time spent in the country of origin could be worth looking at too.
Language investigation ideas: Language and gender
Do female MMA fighters use ‘masculine’ speech features?
Given combat sports are a violent and competitive environment, they might be thought of as a stereotypically masculine environment. An investigation could look at whether female participants use features that researchers have claimed are more commonly used by men.
The obvious core of the data would need to be transcripts of interviews with female MMA fighters, but then data for comparison could include transcripts of interviews with male MMA fighters, or transcripts of interviews with female athletes in other (non-violent or not stereotypically masculine) sports. It might also be interesting to look at differences in how fighters speak after wins or losses, or in interviews from the start of their careers compared to when they’re more established. If they've appeared in more casual media environments like podcast discussions or chat shows, those might offer a contrast to ringside interviews.
AO1 – features that researchers have claimed are masculine or feminine (e.g. all of Lakoff’s ideas for ‘women’s language’, having their topics taken up, face-threatening turns, interrupting and overlapping etc…)
AO2 - work by people who’ve claimed genderlects exist – Lakoff, Tannen, Fishman, Trudgill (for differing male/female use of non-standard forms) etc.., Zimmerman and West re overlapping/interrupting, people who support the Gender Similarities Hypothesis (Cameron and Hyde), Bassiouney for non-standard gendered speech in Arabic, and O’Barr and Atkins’ work that suggests that context and social role affects speech style more than gender.
AO3 – important contextual factors would include the idea that combat sports are violent and competitive, something that would be seen as male, traditionally, but could also consider things like what sort of media interview is it? Is there an obligation to hype an upcoming fight (perhaps by pretending there’s real animosity between the competitors)? How established are the fighters involved? How old are they? Are they winning a lot or losing? How high-status are they within the sport? Do they have a comfortable relationship with whoever’s interviewing them? Have they had media training that recommends a certain style of interview? Do they speak differently in different sorts of interview (e.g. maybe a ringside interview after a match versus how they may speak on a podcast).
Language investigation ideas: Social networks, ethnicity and brokers
Do different social networks adopt new features at different rates? (or: Are specific individuals spreading novel features through their social networks?)
To work well, this idea would need a sharp ear for new features and access to specific sorts of social networks.
Research by Cheshire, Fox, Kerswill, Khan and Torgersen discusses the presence of certain non-standard features (including h-dropping and the pronunciation of the vowel in ‘goose’) in the speech of people in Hackney (inner London) and Havering (outer London). It found that the non-standard features were most common in the speech of non-Anglo speakers, and least common in the speech of elderly speakers. The Anglo speakers in the study were classified as having either an Anglo network (i.e. mostly white friends) or a non-Anglo network (i.e a largely non-white friendship group), and the research found that the speakers with non-Anglo networks were more likely to use the novel features that those with Anglo networks. The researchers recommend that people interested in change look to minority-ethnic groups in cities for the source of new features.
Beyond that, they also discussed the ideas of ‘brokers’ – high status individuals with large social networks who seemed to be responsible for taking features from one area of their network and spreading them to another.
So an interesting investigation is possible if the investigator has access to social networks of minority-ethnic speakers and Anglo speakers. Any element of pronunciation or non-standard grammar could be used as the focus of an investigation – for example, the Cheshire et al. study found that the inner London teenagers were h-dropping (that’s my ouse) much less than the elderly speakers they measured. So if you had access to enough Anglo and non-Anglo speakers in an area you could measure the levels of h-dropping for people from each group. This could also work for things like goose-fronting, ‘dark’ L, the substitution of /d/ for ‘th’ (dem, dere for ‘them’ or ‘there’), more traditionally Cockney features like th-fronting, or any other features currently associated with change (you could do things like slang as well but any specific item of slang will appear much less than a phoneme does so you would probably end up with much less data).
AO1 – this would be based around whichever features you decided to focus on (discussed above). It would be sensible to look at several to give maximum options in analysis (and it might be hard to know which feature will prove most varied or interesting before actually collecting data, so it could be best to include several possibilities).
AO2 – the Cheshire, Fox, Kerswill, Khan, Torgersen research is the obvious core of this, but Cheshire’s work in Reading and the Milroys in Belfast would both also be good for discussing social networks and how they affect language. Trudgill’s work in Norwich demonstrates that people associate certain pronunciations with class, and that people can code-switch. Labov’s work on Martha’s Vineyard shows a specific form of pronunciation becoming associated with the identity of a social group and gaining in prevalence because of that – much as some features may become associated with individuals who seem high-status and therefore worth imitating (however unconsciously).
AO3 – a massive range of factors could be considered here – a map of the social networks involved could be very useful in demonstrating which people know each other and allowing a mapping of the features onto that. And details of speakers’ backgrounds, interests and influences (e.g. someone very into grime might use MLE pronunciations more often than someone who likes baking or lawn green bowls) other languages spoken, and even attitudes to sources of influence like London could be revealing.
Language investigation ideas: World Englishes and gender
If genderlects exist, do people who learn English as a second language display features that have been identified as gendered?
If men and women speak in distinct ways (as has been claimed by Lakoff, Tannen and others), it would be interesting to see whether people who have learned English as a second language display these traits – presumably if the (claimed) difference is primarily due to biological factors (e.g. perhaps men are more aggressive or assertive in their speech as a result of having more testosterone) then non-native speakers will produce these gendered features. If any differences result primarily from socialisation, then perhaps non-native speakers will notproduce those features.
The idea for this study comes from a student who had Asian family members who visited the UK but had never lived here, and who were non-native speakers. When they visited, she had access to their conversations with English speakers who lived in the UK and recorded and transcribed some of those conversations. Partly because of less competence in the language, some of her female relatives seemed quite rude and assertive when speaking to restaurant staff in English.
Gathering detailed information on the language backgrounds of each subject would be important (how long have they been speaking it? Do they have any formal qualifications in English? How often do they speak it? What other languages do they speak? Which do they speak most often?). One approach would be to count up (and analyse/label) errors in their English and use that as a way of measuring their competence in English (be sure to take the amount they speak into account too – average number of mistakes per turn would be better than just a raw total).
With that information gathered, comparing the rate of gendered features to each speaker’s level of competence in English could be fruitful – if female subjects are consistently producing ‘feminine’ features whatever their level of English language ability, that’s interesting. If male speakers are producing lots of ‘feminine’ features then that would be interesting, and if a high level of competence in English is associated with a high level of ‘feminine’ features then perhaps stereotypically feminine speech is something people are acquiring as they learn English.
AO1 – all the suggested genderlect features from research (like Lakoff’s ‘women’s language’ features, topic shifters being taken up, face-threatening turns, interrupting and overlapping, using standard forms etc..). Facework seems worth discussing in light of Tannen’s assertions about how men and women speak (‘orders vs proposals’).
There would be scope for some discussion of transfer errors (mistakes caused by approaching a new language like your native language) that could introduce all sorts of AO1 terms like subject-verb agreement, morphology, pluralisation, auxiliary verbs and so on (depending on the specific mistakes). Examining a few of these could allow an assessment of the level of English-language competence of people in the data.
AO2 – genderlect researchers like Lakoff, Tannen, Fishman, Zimmerman & West etc.. and supporters of the Gender Similarities Hypothesis (Hyde, Cameron). O’Barr and Atkins. Could get into Zimmerman and West’s Doing Gender (the idea that gender is something we perform) to discuss gender roles in different cultures, how changeable supposedly gendered behaviours are and perhaps whether different languages might be priming different behaviours.
AO3 – The lives of the people involved would provide good rich context to discuss. If the non-native speakers do produce gendered features at a similar rate to native speakers, it would be worth considering whether learning a language also involves learning certain attitudes typical of speakers of that language. Everything to do with their backgrounds could be considered as well – depending on the results, all sorts of things could be relevant. For example – if older speakers are ruder and perform less facework, could that be because they’re more established in the world, or higher status in that family and so don’t feel the need for politeness so much? Or are their personalities just like that in both languages/cultures? If the investigator knows about their behaviour in their native country, that could be useful background for their general levels of, say, politeness, and stereotypically feminine or masculine behaviour.