The Census, the Coronavirus and the ESOL Strategy

By James Simpson, University of Leeds

This post is about the census which came around recently on its 10-yearly journey, and how it interlinks with monolingualism, the Coronavirus pandemic, and the need for a national strategy for ESOL. I was inspired to write when I read this interesting piece by the historian David Olusoga on the role and purpose of the census, and why we need to engage with it. Olusoga regards completing the census as a civic duty, saying, ‘We need to understand society so we can better plan. How do we build infrastructure, how do we plan for hospitals, the provision of education, without knowing who’s in the country?’ And he’s right – the information from the census is hugely important for planning of all kinds. I’m interested though in the data about language that’s generated from the census, and what’s done with it. 

First, there’s the question of the question. A question about languages other than indigenous Celtic languages appeared in a census for the first time in 2011. It is reproduced in the same form in the 2021 census, at least on the England and Wales version. The census in Scotland, to be distributed in 2022, has more questions about language, in slightly more detail, but is not dissimilar. The question is 'What is your main language?' The guidance reads: 

Only give one answer.

If you're not sure what your main language is, think about the language you use most naturally. It could be the language you use at home.

If you select "Other", you will be asked how well you speak English in the next question.

For ‘speakers of multiple languages’, the further guidance is: 

If you have two or more main languages, not including English, enter the name of the language that you prefer to use.

The census is designed and administered by the Office for National Statistics (ONS). You can read its account of how the 2021 census questions on national identity, ethnicity, language and religion were developed here.

There was a good amount of debate about the wording of the question about language in the years running up to the 2011 census, with organisations such as the British Association for Applied Linguistics contributing to a long consultation. The finalised question was widely criticised at the time by academics and others as not being able to capture Britain's linguistic diversity. The criticism clearly did not hit the mark, given that precisely the same question is being asked this time around. The debate continued, though, in the run-up to the current census. Here is a sample contribution, from academics associated with the Multilingual Manchester project. They say: 

the way the question was asked was based on the assumption that one person can only speak one ‘main’ language (or, in the Scottish version, just one language other than English) at home. Accordingly, it did not allow participants to name different languages. The division into ‘main language’ versus English does not reflect the diversity and complexity of the current linguistic landscape in the UK.

I completely agree. At the very least there should be the opportunity to name more than one main language other than English. That might go some way to showing that the government recognises not only linguistic diversity but the nature of multilingualism on the ground. As I and many other researchers have maintained over several years (e.g. in the TLANG project) in their everyday practice people draw upon multilingual repertoires in flexible and fluid ways that not only involve the use of more than one ‘named’ language but do not adhere to the boundaries between languages. Capturing some of this reality would be difficult for the ONS, but surely not impossible. As it is, the difficulty is handed to multilingual respondents, who must decide which ‘main language’ they ‘prefer to use’. Moreover, the census will generate a very partial picture of Britain’s multilingualism. To use Makoni and Pennycook’s (2005) terms, the orientation of the census question is segregationist: languages are understood as bounded separate and individual entities. Actual language use on the ground (apart from – notably – in much institutional and educational discourse) is integrationist. 

Multilingual and translanguaging approaches in pedagogy recognise this. In bilingual education programmes translanguaging is beginning to gain a foothold. (This article describes a good example from New York.) In ESOL, this is as yet less the case perhaps, though projects such as Our Languages (Cooke, Bryers and Winstanley) contribute to lodging multilingualism into the ESOL consciousness. 

This brings me to how the findings from the language question on the census are used. Even with the partial and imprecise picture of language use that the census paints, where only one ‘main language other than English’ is recorded, it becomes clear that we live in a country where – in some places at least – multilingualism is the norm. This is very useful to know. It seems to me though that this data has not been employed for productive language planning in the years since 2011 at national scale – certainly when it comes to provision of opportunities for non-expert adult language learners to actually learn English. Rather, in the intervening decade the data has been abused or ignored by politicians and in media and public discourse.

Here are the figures for the 2011 census question about language, as reported by the ONS: ‘While 7.7 per cent of the population (4.2 million) of England and Wales had a main language other than English, only 1.3 per cent (726,000) of the population reported that they could not speak English well and 0.3 per cent (138,000) reported that they could not speak English at all.’ In terms of percentage of the population, the proportion who reported that they cannot speak English well or at all is pretty low, at 1.6%. In raw figures this is of course a large number: over 800,000 is a lot of people. And that figure, or something like it, has been wheeled out time and again by politicians, especially when they are exhorting people to learn English for social integration. Here the Guardian reports Sajid Javid, quoting census figures for England, associating a lack of English with the ‘problem’ of Pakistani and Bangladeshi women: 

We estimate that there [are] 770,000 people that live in [England] that speak hardly any or no English. That figure will be in the report. And most of those people – we estimate 60 to 70% – are women. And most of those women, in turn, are of Pakistani or Bangladeshi origin.

These comments gloss over the difference between someone self-reporting that they ‘do not speak English well’ on the census, and a claim that a person speaks ‘hardly any English’. They reproduce and reinforce the link made consistently in political discourse over the last two decades, between a lack of competence in English and an inability to ‘integrate’. Javid, incidentally, was promoting the MHCLG Integrated Communities Green Paper (2018), which drew on the Casey Review into Opportunity and Integration (2016) for its position on social integration. You might recall this review, commissioned by the Conservative Government and written by Louise Casey, which linked social and economic exclusion to lack of access to the English language, termed by its author as the ‘common denominator’ (2016: 94). Insufficient competence in English was positioned firmly as a social problem connected discursively to crime: ‘Central and local government should develop a list of indicators of a potential breakdown in integration. These might include incidences of hate crime or deficiencies in English language’ (2016: 167). Prominent in the Casey review were Pakistani and/or Bangladeshi Muslim women, seen as being uniquely challenged and problematic because of cultural, religious and social barriers to integration, including their failure to learn English. They in particular are deemed to need ‘targeted’ ESOL provision with funding for community-based English language classes. The report was criticised (see e.g. Bassel 2016 for an excellent critique) not least for its adoption of an othering stance. Othering is the creation in discourse of in-groups and out-groups (we and they). It is a negative stereotyping often associated with race, ethnicity, foreigners or minority groups. Othering is strongly evident in much current policy discourse, where a language-based deficiency is identified in the Other.  

The Casey review, and some of its problematic content, is extensively cited in subsequent government and policy review documents, most recently in the report of the Commission on Race and Ethnic Disparities released on 31 March 2021. This document recycles the familiar and by now rather lazy-sounding tropes about ‘adults from a Bangladeshi and Pakistani background’ being ‘the most likely not to speak English well or at all.’ One reason for this being most pronounced among people from Pakistani and Bangladeshi backgrounds, according to the CRED report (p. 43), is: 

that they tend to live somewhat more separately from the mainstream, both physically and in terms of social norms, and are two of the groups most likely to bring in spouses from their ancestral homes, especially the Pakistani group. This produces the so-called ‘first generation in every generation’ issue, with full integration constantly being restrained by one parent with a foot in another country.

It is shocking to read this in 2021, and suggests that the political mainstream hasn’t managed to move beyond the talk of migrants leading ‘separate lives’, from around 15-20 years ago. For example Tony Blair spoke in 2005 about ‘a separateness that may be unhealthy’, associating a lack of English with terrorism. 

So how might the census data support planning for ESOL? The number of people who do not speak English well or at all should clearly be an issue for the Government, given the importance it places on people being able to use English. And it is an issue for us too: no-one denies that access to the dominant language of the country matters, least of all new arrivals who do not speak English, or their teachers, or indeed ESOL researchers. This is stressed in report after report, most recently in the ETF report on ESOL learners and functional skills English, written by NATECLA (see p. 24). The difficulty lies in the English language being constantly and consistently used an emblem of loyalty and national identity by senior politicians, and its non-use as being a problem and an indicator of an inability to ‘integrate’. Here is the Prime Minister speaking in 2019 just before he attained his current high office: 

I want everybody who comes here and makes their lives here to be and to feel British, that's the most important thing, and to learn English. And too often there are parts of our country, and parts of London still and other cities as well, where English is not spoken by some people as their first language, and that needs to be changed and people need to be allowed to take part in the economy and in society in the way that that shared experience would allow. 

There are so many things wrong with this statement – not least that it supposes that people can change their language history – that it’s difficult to know where to begin. The response from SNP MP Angus McNiel, that ‘Boris is moronic and clueless,’ probably sums it up for many.

Beyond the rhetorical passion for English and Britishness though, a national language policy for adult migrant learners of English – that is to say, ESOL policy – is difficult to find these days. I’d argue that national ESOL policy around access and funding is somewhere between incoherent and non-existent. There is neither a UK-wide strategy to support provision of ESOL, nor an England-specific one. In fact central Government appears keen to bleach out linguistic diversity from policy altogether. NALDIC reported at the end of March that the post of Ofsted lead for EAL and ESOL is being cancelled. Similarly, any efforts in Government to design and develop a coherent ESOL policy – as promised in the Integrated Communities Green Paper– appear to have fallen by the wayside. 

So why not use the census data as the basis of a sensible strategy to ensure that people have opportunities to learn English, rather than wield it to berate people for not conforming to a particular notion of Britishness? But the statistics are not currently being used well to enhance language policy at national scale, to address the language needs of adult migrants, for the benefit of society as a whole. Since 2010, Government support for ESOL has fallen through the floor. For example, central Government funding for ESOL provided by Further Education colleges, which comes via the Education and Skills Funding Agency's adult skills budget, fell from £203 million (2009-10) to £90 million in 2015-16 (Martin 2017). Much ESOL provision therefore sits outside central Government education policy, in the hands of non-governmental and non-profit-making organizations, including charities, voluntary and community groups. Any Government funding that does reach this third-sector ESOL is typically channelled through the Ministry of Housing, Communities and Local Government and the Home Office, rather than the Department for Education, in the form of piecemeal small-scale local project funding. And as we all know, the demand for accessible, freely available classes consistently far outstrips supply.

At local scale there are signs that the census data are used more productively. Ideally, local governments will look closely at the responses to the language question on the census, and support ESOL provision in areas where there are large numbers of people who report not being able to speak English well or at all. Something like this is done in Leeds, where the data collated by the Leeds Observatory helps the Local Authority target translation and interpretation services and material in languages other than English. This is of course crucial, now more so than ever, living through a pandemic as we are. With the development of COVID-19 vaccines, there is light at the end of the tunnel. A major effort is needed, though, to get reliable information about the safety and efficacy of these vaccines to everyone in society. Local authorities and the NHS report the importance of this for people whose main language is not English, and especially those who – in census terms – do not speak English well or at all. A group of NHS doctors write in the BMJ about ‘Cultural competence in covid-19 vaccine rollout’:

Information about the vaccines should be widely distributed in multiple languages in both written and graphic formats to enable people from all ethnic backgrounds to make fully informed choices. No community or individual should be stigmatised for initial low uptake, as occurred earlier in the pandemic in Leicester and other areas with large ethnic minority populations.

ESOL is crucial in this regard too. I would never suggest that ESOL classes can replace translated information. One is not the same as the other, and both are needed, as Mike Baynham said in The Conversation a few years back. Yet for the adult population to have access to the dominant language in the country, the one that’s normally used for institutional and medical encounters, is surely a good thing. Senior politicians recognise this, apparently, if their rhetoric is to be believed. Actions speak louder though. The 2021 census data will show that many adults do not have access to that dominant language. A reasonable response would be to develop a coherent national strategy to ensure that the people who need one have the opportunity to attend a freely available, high quality and appropriate ESOL class, at a place and at a time that suits them.