This month’s issue of The Economist included an article entitled ‘Why central and eastern European children lag behind in British schools’. Here, Professor Yaron Matras responds to the article and discusses the difficulties of using official statistic to record languages.

• The article looks mainly to languages as being a factor in differential educational achievement, but age of arrival and the level of English are also important factors
• Interpretations of ‘main’, ‘home’ and ‘first’ languages vary
• This is due to several factors including demographics and languages’ status within families and communities
• Recording a pupil’s non-English ‘first language’ is not just a token of heritage recognition but a mark that alerts staff to anticipated or presumed problems
• Ideally, we should implement a comprehensive School Language Survey

The Economist newspaper prides itself for basing its reports and analyses on scientific surveys wherever possible. But when it comes to assessing language diversity, survey tools can be misleading: They may be designed to capture data on language use, but they are often ineffective because they remain framed within a monolingual mind-set.

In its article from 15 July 2017, the Economist relies on Lambeth Council figures to cross-reference GCSE achievements with pupils’ first language. Speakers of Asian languages like Hindi, Cantonese Gujarati and Urdu seem to perform well: they tend to be near and often above the national average of 57% who achieve five or more GSCEs at grades A* to C. By contrast, speakers of Eastern European languages, including Polish, Latvian, Slovak and Czech, do poorly, with fewer than 45% showing a similar level of achievement. Speakers of Middle Eastern and Western European languages, including Arabic, Kurdish, Dutch, and Italian, are somewhere in between, closely orbiting the national average.

Misguided analysis

The report is cautious not to attribute this to language itself but instead to its social correlates, pointing out that speakers of Lithuanian and Polish tend to live in poor areas. However, it then strays into a rather essentialist cultural narrative, attributing the lower end of the scale (Czech and Slovak) to the performance of Eastern European Roma and repeating the misguided, seemingly progressive mantra that due to their past discrimination, Roma are reluctant to engage with public services, and that Roma parents are reluctant to allow their girls to mix with non-Roma in schools. Only in passing does the article mention the age of arrival, and the level of English, with which it invariably correlates, as possible factors.

‘First’ or ‘main’ – language matters

Left out of the equation is the sociolinguistic dimension and the meaning of asking about pupils’ ‘first language’. The 2011 population census was the first to ask about language (other than Welsh and Gaelic), but the question ‘What is your main language?’ resonated with respondents in very different ways. Many chose their language of education and literacy, or their language of work, over the language in which they had heard or sung their first lullabies, while in Manchester one each claimed ‘Manx’ and ‘Cornish’ to be their main language, both considered extinct (except in language revivalist circles) and clearly not a primary vehicle of everyday chats if there is only a single speaker in the whole city.

‘First language’ is usually perceived as heritage language, with school officers recording either what they think they know about a pupil’s family background or else what parents are willing or eager to disclose. It is telling that the Economist links Czech and Slovak to Roma, whose first or home language is in fact usually Romani (of which officers are often unaware, and parents not keen to disclose); many Roma children who were raised in Britain are not even proficient in the national language of their country of origin.

Measuring the interpretation gap

The volatility of existing measuring tools for languages is best revealed by triangulating datasets, a method pioneered a few years ago by the Multilingual Manchester research unit and subsequently outlined in an article on the city’s language policies and practice.

Schoolchildren make up on average between 15-17% of Manchester’s population. Comparing the numbers of pupils in Manchester schools whose ‘first language’ was recorded in 2010-2011 as German, Czech, or Polish, to name a few, with the numbers of individuals (including all household members) who provided the same languages as their ‘main language’ on the 2011 census, we find that the proportion of schoolchildren is generally close to the average. For Chinese, however, the number of schoolchildren drops to around 4%, while for South Asian languages such as Bengali, Urdu and Punjabi the proportion is 45%, and for Yoruba a staggering 75%.

So what’s going on?

The answer is that several factors play a role, one of them being demographics, another the language’s status within the family and community. Central European immigrants tend to have similarly sized families to average Britons, are literate in their home languages (which are also national languages) and, being fairly recent arrivals, they tend to maintain those languages in the home. Many of Manchester’s Chinese speakers are elderly, or students; neither group has many school-aged children.

South Asians tend to have larger families, but they are also more likely to under-report their home language in the population census; this can be an acknowledgement of the colonial legacy of English as a second national language in the origin countries, or simply a factor of the length of residency in Britain.

Yoruba speaking Nigerians tend to use English as their only medium of literacy and will often regard their home language as a ‘dialect’ that is not worthy of mentioning in formal settings. Schoolteachers, however, often overrate the likelihood that South Asian and African children who have a home language other than English will face learning difficulties; recording a pupil’s non-English ‘first language’ is thus not just a token of heritage recognition but a mark that alerts staff to anticipated or presumed problems.

We must therefore be careful with numbers and what they represent when talking about ‘main’, ‘home’, or ‘first’ languages. These notions are entangled in both internal and external narratives; not only can they mean different things to different people, but they are also permeable, taking on different contextual meanings.

In an ideal world we would implement a comprehensive School Language Survey where pupils and their parents are asked to describe their full repertoire of language practices in detail. That would require time and resources, but also more openness toward our multilingual reality.