The death of dialect: The quirky regional terms dying out
By Dr Iain Brown, Head of Data Science, SAS Northern Europe

As we’re more connected through technology than ever before, our dialects are becoming increasingly merged, and our accents intertwined. However, some differences remain and have even widened between different generations.

The evolution of social media has played a major role in the way we have adapted our language. A study of 2,000 parents found that 86% of its participants felt that young people and teens spoke an entirely different language on social media, with a ‘seismic generational gap’ influencing how modern informal language is shared online

This has only become more apparent with terms such as ‘lol’ (laugh out loud), lmk (let me know) and many more acronyms becoming part of everyday conversation - even the Oxford Dictionary has evolved with technology. In 2015, the word of the year was simply the laughing emoji (😂), while in 2013 it was ‘selfie’. Word of the year in 2023 was ‘Rizz’, which has exploded on TikTok after people used it to describe having charisma.

In the UK, there are several regional dialects - some we’re more familiar with than others, ranging from Birmingham’s well-known Brummie accent to London’s cockney twang. Then there’s the West Yorkshire accent from Leeds and the South Yorkshire version in Sheffield, with their known distinctive features such as ‘chuffed’ and ‘ey up’.

An evolving digital dialect

While our dialect is changing and becoming more intertwined, texting, social media and AI have only accelerated the way we communicate. In fact, new words are being created at a much faster rate, with one emerging approximately every 98 minutes, totaling up to 5,400 new words annually. By contrast, in 2022, only 700 new words were added to the Oxford English Dictionary.

The everyday use of Siri, Alexa and Google Assistant can also affect the way we speak, creating a digital dialect. AI-related tools used for natural language processing (NLP) help to understand and produce human language and if needed can quickly adapt to other language-related tasks.

For example, the word ‘selfie’ originated in 2002, but it wasn’t until the rise of social media and AI-powered image recognition that the word became widely used. Similarly, services such as ChatGPT and company chatbots can engage in dynamic conversations, creating new possibilities for how we communicate with customer services and potentially making it easier to attain new information.

This isn’t just affecting how we communicate online but also changing regional dialects across the UK. Today, we adopt more of a “standardised” language rather than native dialects, leading to less regional variation in words used and accents.

A study by Amazon found that 30% of younger people in the UK are confused by the use of regional phrases or words and sometimes don’t understand what their grandparents' mean. A previous SAS hackathon entry revealed that Africa has more than 10,000 languages, but at least 200 of these languages have fewer than 500 speakers. And yet, for a language to stay alive at least 100,000 speakers are generally thought to be needed according to the study.

With the rapid acceleration in changing dialects everywhere, how will regional terms change over the years and how can they be preserved?

What words have seen the biggest decline over the years?

SAS looked at 100 popular regional terms from different local authorities across the UK. It then analysed each word using the Google Books search tool to establish when certain regional terms may become archaic, based on the usage of each word from 1919 to 2019.

The word ‘Ansum’ meaning ‘nice’ or ‘top notch’, derives from Cornwall and is the likeliest word to die out from the English Language according to our study. This term has seen a 97% decline in usage, despite being a popular word in the South around the early 1900s.

‘Scran’ is next on the list. Originating from the North West of England, but a well-known term across the UK, scran is usually used as shorthand for food, but was originally part of the 18th-century slang for a bar tab. The word has declined in use by more than 96% in the last 100 years.

Our research suggests that regions such as the North East, South East and South West are most in danger of losing their regional terms. Each regional dialect provides insight into the historical background of the region and for the residents, individual localisms can offer a sense of unique identity and community.

The research found that there’s been a 55% decrease in people using the word ‘cob’. This term has caused a great debate for years with the Midlands identifying a bread roll as a ‘cob’ whereas in West Yorkshire it’s often referring to a teacake, in the North West a barncake, and in other places a bap.

But with more Brits speaking in digital dialect, terms such as ‘cob’ and ‘barncake’ may not be understood by AI services like Alexa. Similarly, some words have double meanings like ‘parky’ - some use it to mean fussy but in London the term is used for cold weather.

Regional terms and differences in meaning may not  be “fed into” services like Alexa, Siri and Google which might have a consequence on these localisms long term as more people resort to speaking in a widespread digital language.

A recent report revealed that more than 75% of residents with strong regional accents across the UK have regularly adapted the way they speak to be understood by digital assistants such as Alexa or Siri.

The death of dialect: Regional terms most at risk of dying out

Region	Regional term	Definition	% difference
South West	Ansum	Top-notch	-98%
North East	Scran	Food	-96%
Yorkshire	Thoile	Unwilling to pay	-95%
North East	Parky	Fussy	-84%
South West	Belve	Sing loudly	-82%
South East	Liggle	The act of carrying something too big	-75%
South East	Sing small	Put up with less than was expected	-68%
North East	Grand	Great	-68%
East Midlands	Bostin	Very good	-60%
East Midlands	Cob	Bread roll	-55%

Regional terms that have seen biggest increase

SAS also looked at the regional terms that have seen the biggest increase in use over the last 100 years. The term ‘gammy’, often used in the East Midlands to describe being injured, has steadily increased since 1919, rising in popularity by more than 15,900%.

In fact, the Midlands featured the most for dialect terms increasing in popularity, with recognised terms like ‘bonkers’ and ‘titchy’ also steadily rising – illustrating that dialects and accents will continue to evolve. The Midlands has always been influential and is one of the regional accents that led to how English is spoken today.

The Welsh regional term ‘cwtch’ - meaning cuddle - has also seen a big increase of over 8,000%. The term signifies an emblem of being Welsh, underlining the continued importance of community in use of certain words.

Some local terms have seen a rise in popularity following their use in pop culture. Rapper Dizzee Rascal’s song ‘Bonkers’ being number one in the charts, contributed to a huge 6,000% increase in usage in the last 100 years. Similarly, ‘mardy’, often used in Yorkshire and the East Midlands to describe someone being grumpy, was used by Indie band Arctic Monkeys for their hit single ‘Mardy Bum’ and has seen a more modest 30% rise in the last 100 years.

Culture also has an impact on words in a way that might be unrelated to regional dialect. The term ‘fortnight’ has been used 800% more in the last 10 years alone, which could be related to the rise in the popular online game ‘Fortnite’ which ended in a viral TikTok dance craze. 

Region	Regional term	Definition	% difference
East Midlands	Gammy	Injured	15,968%
Wales	Cwtch	Cuddle	8,088%
Midlands	Bonkers	Crazy	6,169%
Yorkshire	Ey up	Hello	4,948%
South West	Dreckly	Soon	3,608%
North East	Blimey	Surprised	3,026%
East Midlands	Titchy	Small	2,660%
Scotland	Drookit	Soaking wet	2,122%
Yorkshire	Lowance	Allowance	2,044%
North West	Noggin	Head	1,372%

Our analysis also revealed the Scottish word ‘drookit’, the Yorkshire term ‘lowance’ and North-West’s ‘noggin’ are proving popular, with all words seeing a rise in usage of over 1,000%.

As dialect is continuously changing the diversity and richness of our language, the quirks, idioms and phrases often associated with certain communities will continue to change over the next few years and may even accelerate as technology continues to exert more influence.

Why understanding changes in language is important?

With certain terms gaining prominence or falling out of favour over time, especially in a technology-focused world, it’s a priority for businesses to understand language usage patterns.

Natural Language Processing (NLP) is a branch of AI and comprises many different techniques for interpreting human language, helping to break down language into shorter elements to understand relationships between them and how they work together to create meaning.

In particular, NLP can be used to document languages and dialects that are specific to a region or at risk of ‘dying out’. Through speech recognition and transcription tools, our regional nuances can be preserved in digital formats. And NLP-powered language learning can also help in teaching ‘endangered’ languages and dialects to new generations.

Some of the processes can help with:

• Content categorisation. A linguistic-based document summary, including search and indexing, content alerts and duplication detection.
• Topic discovery and modelling. Accurately capture the meaning and themes in text collections, and apply advanced analytics to text, like optimisation and forecasting.
• Corpus Analysis. Understand document structure through output statistics for tasks such as sampling effectively, preparing data as input for further models and strategising modelling approaches.

The evolution of dialects across regions can impact businesses, so using up-to-date data can be important for understanding this constantly changing human language which can be used in many contexts.

Documenting and archiving dialects can help to maintain understanding of words and language habits, even where these are very localised or used sparingly. With NLP we can maintain unique identities for regional dialects, ensuring that different meanings are stored and easy to access. Ultimately this can ensure smoother and better understood exchanges between businesses and consumers.

Methodology

SAS looked at 100 regional terms from across the UK’s local authorities and used Google Books NGram Viewer to find out:

• How these terms have changed in usage from 1919 to 2019.
• The average percentage difference in usage over the last 100 years.

Data correct as of February 2024.

Dr Iain Brown, Head of Data Science, SAS Northern Europe