skip to content

Cambridge Language Sciences

Interdisciplinary Research Centre
Milton Keynes to Cardiff (via New Zealand)

Research by David Willis in the Department of Theoretical and Applied Linguistics into syntactic variation in present-day Welsh has recently been in the spotlight in Wales. As part of a long-term collaboration with researchers at Newcastle University and the University of Essex, Dr. Willis has been seeking Welsh speakers in Cardiff to take part in a project to produce a Syntactic Atlas of Welsh Dialects.

Why are you focusing on Cardiff in particular?

We’re looking at the whole country, but it’s a lot easier to find informants in places where everyone speaks Welsh than in a large city where they are a small minority. I’m focusing publicity on Cardiff because I haven’t got enough participants from there. My main research interest is in mechanisms of syntactic change, and this is really what the project is about. Syntactic change has many manifestations and the project deals with a number of them. The fieldwork has been covering most of Wales (northwest, southwest and southeast) because different regions exemplify different scenarios: in the northwest, where the language is changing continually under its own steam, I’m looking at how and why internally motivated innovations emerge and spread, while in the south the interest is more in contact with English. In the southwest, the question is how well traditional dialect features are being maintained in the face of the general weakness of the language, while in Cardiff and the rest of the southeast, an area where there’s no longer any continuity with the original Welsh spoken in the region, the interest is in how a new variety emerges when people come together from different areas, with the added complication of language revitalization.

Cardiff is interesting because speakers from all over Wales come together. Many migrate there for work – increasingly so in the last 30 years. The result is that, for the first time in generations, a significant number of people have started growing up speaking Welsh as a native language in Cardiff. Some are the children of migrants from Welsh-speaking areas, while others are the offspring of non-Welsh-speaking parents who have chosen to send them to schools where Welsh is the medium of instruction. Working in Cardiff has allowed me to work with speakers from all over Wales and also have access to speakers of the new emerging Cardiff variety.

How unusual is it for a new variety of a language to emerge like this?

New varieties of language emerge all the time: you only have to think of relatively newly-populated cities such as Milton Keynes, or areas where a new variety of an immigrant language has emerged, as with English in New Zealand or Ireland. One of the most interesting things about Welsh, though, is that the syntax of different dialects varies significantly, much more than in English, so that in Cardiff today we can observe the creation of a new syntactic system in progress.

Who are you interviewing?

Traditionally, dialect atlases focus only on older people who’ve grown up and stayed in the same area all their lives. People like that are becoming increasingly rare, and the impact of dialect mixing on language change is also becoming more and more evident. The current project is a radical departure from traditional methods because we allow any Welsh speakers who began to acquire Welsh in a naturalistic environment before the age of 5 to take part. The ability to acquire a language natively declines with age, and by adulthood people have lost the ability to acquire a language just by being exposed to it and have to be taught. Even within this group, though, there are differences between people who were exposed to Welsh in the home and those who spoke only English at home and acquired Welsh at school.

What about the family set-up – surely that makes a difference?

One of the aims of the project is to investigate the effect that the way in which people acquired Welsh has on the actual structure of the language. Specifically, are sequential bilinguals (people who acquired English first, then Welsh) responsible for introducing aspects of the syntax of the language they learned first (English) into the language that they learned second (Welsh)? And does increasing re-integration of Welsh into the education system lead to the spread of standard or school Welsh? Although the Welsh-medium education system was dismantled in the mid 19th century, today, as the result of language activism from the late 1960s onwards, there’s a large network of nurseries and schools where children are taught in Welsh: currently around 20% of Welsh children are in Welsh-medium education, although only 8% speak Welsh at home. These statistics don’t tell the whole story, though, and the reality is much more complex: for example, one informant who grew up in Cardiff has a mother from the southwest who he spoke Welsh with, and a father from the southeast who he spoke English with; another older informant spoke English with her parents at home and at school, but acquired Welsh from her wider family and community.

You mentioned New Zealand earlier, and in your piece for S4C.  Could you explain a bit more about that?

Research on New Zealand English gives us the main theoretical model for how a new dialect emerges. Recordings were made in the 1940s of the first generation of English-speaking New Zealanders (who were by then in old age). We can hear from these recordings that their language involves idiosyncratic and non-systematic combinations of features from different dialects of the British Isles. For instance, the same speaker might drop the 'h' in harm (a London/southern feature) but also produce a 'rolled' 'r' in the same word (a Scottish feature). Only after another generation has grown up hearing this mixed language as their input does the variability subside, and a variety that sounds recognizably like the New Zealand English we know today emerge. Although this model was developed for the emergence of new sound systems, we can test whether it applies also to syntax by looking at southeast Wales today. Already in the fieldwork interviews, I've found speakers who mix aspects of the syntax of northern and southern Welsh dialects within the same sentence, for instance, one speaker who deletes auxiliaries (as in the English You coming? for Are you coming?) according to the rules found in southern Welsh dialects but forms tag questions (as in the English You're coming, aren't you?) according to the rules for northern Welsh dialects.

So you’re interested in how language in general is evolving, not just Welsh?

Yes, I’m interested in the mechanisms of language change. We think that language acquisition plays a pivotal role in introducing innovative patterns into a language, some of which go on to spread and change the language. Children unconsciously work out the often quite abstract structure of their language by hearing lots of examples of it around them. While this normally leads to children successfully replicating the adult system in their own language, sometimes children create new rules or over-apply existing ones. Contact with other languages seems to accelerate this process. Welsh is an ideal environment to investigate how this works because the syntax of the spoken language is much more varied and there’s less pressure to conform to a spoken standard language.

Any surprises in the findings so far?

The project has revealed that there’s a lot of internally-motivated change underway in the north today – more than we anticipated. We knew that people in the north use a word that originally meant refuse (cau) to mean ‘won’t’, saying the equivalent of ‘The door refuses to open’ for ‘The door won’t open’. But what’s interesting is that existing features of Welsh grammar are being extended onto this. The most important is negative concord, a general feature of Welsh syntax, where you express a negative multiple times, saying ‘I haven’t not seen no one’ for ‘I’ve seen no one’. In the north, speakers split into two groups, some extending negative concord to ‘refuse’, saying (literally) ‘The door doesn’t refuse to open’ for ‘The door won’t open’, while the other group stick to the traditional pattern. The next stage is to work out if this innovation occurred once and then spread or whether it has occurred repeatedly (and therefore in some sense was ‘bound’ to happen).

What comes next?

The full atlas is a long-term undertaking, but some of the early results for specific areas of grammar (pronouns, negation) will be published as articles over the next year, and the data will be made available via the project website as maps and examples sound files at the same time. I’ve used a British Academy mid-career fellowship to work on the project this year, and we're aiming to transform it into a larger research project over the next five years.

Where can people find out more?

What we do

Cambridge Language Sciences is an Interdisciplinary Research Centre at the University of Cambridge. Our virtual network connects researchers from five schools across the university as well as other world-leading research institutions. Our aim is to strengthen research collaborations and knowledge transfer across disciplines in order to address large-scale multi-disciplinary research challenges relating to language research.