The census has got it wrong on languages

The widely reported census figures published yesterday by the Office of National Statistics supposedly gave us all a snapshot of what languages we speak and where we speak them. But the data is woefully inaccurate, says Professor Yaron Matras from The University of Manchester's Multilingual Manchester project.

"Though yesterday's census data provides important confirmation of Manchester's enormous linguistic diversity, our research shows it has significantly under-represented both languages spoken and the numbers of people who speak them.

"We identified that our city boasts of at least 153 languages, making it one of the world's most diverse places. It is very likely to be top of the list in Europe, certainly when compared to other cities of its size, perhaps only outflanked by London and Paris.

"So the census is way out: Manchester is home to the fourth largest Yiddish-speaking community in the world, yet the census only mentions five speakers.

"Romani is reported to have only 29 speakers, yet the Romani Project, also based here at The University of Manchester, supports schools with hundreds of Romani-speaking pupils. A recent survey by the project confirms that there are several hundred Romani speaking households in the city.

"Only 1,700 people are reported to speak Cantonese, and only 13 are reported to speak Caribbean Creole. But these figures must surely be wrong: both communities have many thousands of members.

"And bizarrely, two languages that are known to have died out many years ago, Cornish and Manx, are reported to have one speaker each in the city.

"One explanation for these woeful errors could be that respondents to the census appear to be under-reporting their use of home languages other than English. This may be partly through lack of awareness or fear of stigmatisation.

"But it is also down to the lack of clarity in the question posed by the census which was "what is your main language?" For many people, English is 'main' because they use it during most hours of the day at work or place of study, but it is not their home language.

"Myself and other colleagues from The University of Manchester are currently piloting a new method of surveying home languages, working with a number of Manchester schools. The first results are expected to be published later this year

"But we already know that our new method gives us a much more accurate assessment of the use of home languages by pupils. It is likely to give a much better nationwide picture of multilingualism, as well as help identify need for English language classes.

"If the census has got it wrong for Manchester, it has also got it wrong for England and Wales as a whole. Health services, businesses and schools are among the institutions which are increasingly embracing multilingualism to reach their target audiences of clients and customers. Having accurate data on the language needs of their communities is an important way to help them."