Who speaks Scots? The 1996 survey revisited

As part of the 2011 Census, for the first time the people in Scotland were asked to say if they could understand, speak, read and / or write Scots (see the Aye Can site). The results will be available some time next year and will no doubt be the subject of much debate as people argue their validity, irrespective of the result. 

In the meantime the only data available is the 1996 Scots Language Survey. Although the survey is 16 years old it is until next year the only data we have avaiable on how much and where Scots is spoken.

In the summer of 1996 the General Register Office for Scotland, GRO(S), carried out a survey on the Scots Language to investigate the feasibility of including in 2001 census a question similar to the question in the 1991 Census asking the respondent’s proficiency in Gaelic. The survey was the result of a long campaign by Scots language proponents.

How was the survey carried out?

The research was done in two phases, firstly the reaction of a focus group of likely Scots speakers to a Census type question was tested (their responses were correlated to an assessment of their language ability) and secondly commercial polling organisations were asked to test three questions variations on three representative surveys of 1000 people. Thus in all some 3111 people were polled.

What did they conclude?

Not surprisingly the survey found “The language used in Scotland today retains a lot of traditional speech forms, though there is a continuum of speech type in the Scottish population ranging from clearly English to clearly Scots.” The main problem the researchers found was that after years of suppression, people were poor at defining their own abilities as Scots speakers. Thus in the survey report, GRO(S) concluded that the responses to a Census question would be “a self-assessment of a concept which is poorly defined by the public and would measure the prevalence of a speech tradition derived from Scots rather than a particular ability in the Scots language. Given this qualification the results would probably be of limited use to those seeking to meet educational needs.” The report recognised that “the inclusion of a census question would undoubtedly raise the profile of Scots“, but suggested that surveys using some (as yet undeveloped) linguistic test/index would be more useful and appropriate.

How many Scots speakers did they find?

Even given the caveats above the survey found about 30% of the population of Scotland will respond “Yes” to a question of the form “Can you speak Scots or a dialect of Scots?“. This would correspond to about 1.5 million affirming they speak Scots, a figure still widely quoted.

What were the regional variations?

Note there is no data from Orkney and Shetland. Given the strong influence of attitudinal factors (see below), the researchers felt that some of the regional differences may be do to local differences in acceptance of Scots. For example Doric now has a reasonably high status in Grampian and people may be more ready to confess to speaking it.

As an aside, the researchers noted that although the idea that Scots dialects are strongly differentiated was very prevalent, older people who had known people from all over Scotland during the War “denied the existence of dialect differences strong enough to hinder communication”. This notion of major dialectical difference in Scots remains a persistent and harmful myth today.

Age and class

The evidence suggests that Scots speaking is related to age and class. Younger age groups are less inclined to assess that they speak Scots and people in lower socio-economic groups were marginally more likely to say they spoke Scots. However the relationships were surprisingly weak related to region.






How did the census define a Scots speaker?

A person classified as speaking with a Scots accent would use the same words as an English-speaker but sound different; a person speaking with a dialect would chose words that are local variants of the ‘mainstream’ language; a person whose speech was classified as being a different language would use constructions of the language as well as vocabulary“.

The researchers “came to the view that ‘Pure Scots’ did exist at one end of a continuum to English and that many people’s speech could clearly be placed as predominantly stemming from one or the other of the two languages, and that there was much maintenance of traditional speech forms and vocabulary.

Attitudes to Scots

Many attitudinal views about Scots were expressed, and the researchers suspected this may interfere with respondents willingness to respond to the questions. “In Britain – where accent, dialect and class effect language – language ability is very difficult to assess, especially since in an assessment situation, language readily changes“. The report quotes J Menzies’ 1991 paper suggetsing “‘Code-switching’ or ‘dialect sliding’ will be exacerbated:

  • If a mode of speech or dialect is considered slang;
    Where the interrogator does not share a speech code (The researchers found that the presence of a Scots expert in the assessment team “drew out the Scots speaking ability”);
  • Where the language is undergoing a process of assimilation or corruption by a neighbouring (powerful) language – since its inception;
  • Where the differentiation between the two languages (in this case from Anglo-Saxon into English and Scots) was never ‘complete’;
  • Where the language is unrecognised by some of its speakers
  • Where, superficially, the language is not of any apparent use in daily transactions to an outsider and finally, and most powerfully;
  • Where English is viewed as a lingua franca for communication outside the local and family community.

While many thought Scots should be encouraged more in schools, English was commonly viewed as the prestige language of communication. “Good spoken English was related to improved employment opportunities”, indicating a continuing perceived or real prejudice against the language. There was much support for learning more about Scots history and literature, some support for Scots medium teaching, but also for teaching Gaelic as the “true and prestigious language of Scotland”.

How important was the survey?

It should be emphasised that the aim of the survey was not directly to identify Scots speakers, but to test potential Census questions. One problem is that each of the three main surveys used a slightly different question wording. This turned out to have a marked influence on the responses. Nevertheless as the largest survey of its type attempted until the 2011 census, the GRO(S) project yielded useful information, and the use of commercial polling organisations to carry out the surveys made it reasonably objective. The figure of 1.5 million Scots speakers has been widely used, even in official documents.

However, while emphatically recognising the existence of the Scots language as one end of a linguistic continuum the report warned that for survey purposes ‘Scots’ could at the time only be a broad concept because of firstly the linguistic diversity within the Scottish population and secondly the lack of general education and information on what ‘the Scots Language’ actually is. It is likely both these themes will raise their heads when the analysis of the 2011 census starts.


Aye Can (2011) Scots in the census site

Máté, I (1996) Scots Language Research Report, The General Register Office for Scotland, Ladywell House, Ladywell Row, Edinburgh EH12 7TF (Tel 0131 334 4295)

Menzies , J (1991) An investigation of attitudes to Scots and Glasgow dialect among secondary school pupils, Scottish Language, 110 (Winter), Association of Scottish Literature Studies (Ed McLure D)

Scots as a ‘social mechanism’

This wee scrieve frae the kenspeckle Marxist lang-heid Eric Hobsbawm (that deed yestreen) in the Guardian the day caucht ma ee.”The destruction of the past, or rather of the social mechanisms that link one’s contemporary experience to that of earlier generations, is one of the most characteristic and eerie phenomena of the late 20th century. Most young men and women at the century’s end grow up in a sort of permanent present lacking any organic relation to the public past of the times they live in.” It made me think aboot Scots, shuirlie ane o Scotland’s maist important ongaun ‘social mechanisms’ linkin oor current sels tae oor fore-fowk.

[Vocabular: scrieve – piece of writing, kenspeckle – famous, lang-heid – thinker, yestreen – yesterday, ee – eye, ongaun – continuing, sels – selves, fore-fowk – ancestors]

“The Broons” – how ‘Scots’ are they? (Part 2)

To start the analysis I collected the data was in two corpus collections;

  • The 1940s collection of 5292 words (15 texts)
  • The 2006/7 collection of 3073 words (13 texts)

The research also had access to the Open University’s version of the British National Corpus (BNC-OU) , a spoken (i.e. conversational) corpus of 1,035,176 words of spoken English for comparison. I used ‘concordance’ linguistic analysis software to look for systematic variation (and similarities) in the two vernacular collections in comparison with each other and the English BNC-OU. Here is what I found.

  1. Overall structure: The comics can be interpreted as ‘pseudo-conversation’ in that they attempt to tell a story. Moreover, they are aimed for the most part at children, so some clarity and brevity in the storytelling might be expected. For example, there is little ellipsis of whole words and phased, though there is much shortening of functional words e.g. “ye’re”, “he’s”. Questions are quite common, in the first 10 comics in V1 they occur in 12% of the speech bubbles. There is almost no use of tags (“is it no?”), presumably for brevity. Disfluency is frequently represented by false-starts, hesitators and pauses (indicated by dashes) e.g. “It’s like this – look!” and “Oh – er – p-pleased tae meet ye.” Inserts such as “Ach!” and “Oh!” are likewise frequent. The use of “an’” (and) as a link between utterances is particularly noticeable, mimicking the real-time production of natural conversation. The main characteristic of the test as representing “speech”, though is the vernacular range, and this will be addressed below.
  2. Orthography: The lowland Scottish phonological system is distinct, complex and challenging to transcribe authentically as fiction (e.g.Corbett 1997, Hagen  2002). Watkins adopts a what would now be considered a non-standard though consistently-applied orthographical convention (although spelling conventions are often hotly debated), for example using Burns-like apostrophies to indicate ‘missing’ letters in Standard English e.g. “o’”, “a’richt” and “Gran’paw”.
  3. Phonology: Some of the Scots vowel system are represented e.g.”hame” (home), “lang” (long),“cauld” (cold), “oot” (out), “tae” (to) and “ony” (any) but common pronunciations are missing such as “ah” for “I” and “oo” more generally in words such as “soond”, “poond”(but “mooth” and “hoose” do appear) and “ai” in “caird” (though “pairt” appears). The tendency for Scots to clip some final consonants is represented in e.g. “bein’” (being – the “ing” is never used in the early comics), “an’” (and), “no’” (not) and “a’” (all), hence “a’richt” and “Gran’paw” above. The tendency for the hard “ch” sound in “nicht”,(night), “brocht” (brought) and “michty” (mighty) is sometimes represented.
  4. Grammar: The most noticeable grammar differences is the replacement with the shortened form “-n’t” in standard conversational form with “-na”, hence “dinna” (don’t), “canna” (can’t). There is a trace of the distinctive Scots use of the definite article in “the nicht” (tonight). The irregular “strong” verb forms are generally not used apart from “kent” (knew) and “telt” (told). Modal verbs, apart from “hae tae” (have to) are generally in the standard form, although there are common vernacular equivalents. Watkins does use interrogative forms, however, such as “wha?” (who?), “whaur” (where?) and “whit?” (what?).
  5. Vocabulary: The Broons uses a distinctive vocabulary but by no means uses all the lexical repertoire available to many of its readers. Typical vocabulary is  “aye” (yes), “but an’ ben” (country cottage), “bonny” (pretty), “bairn” (child), “bide” (stay), “dub” (puddle), “claes” (clothes), “ken” (know), “mind” (remember), “Maw” (Mum), “Paw” (Dad), “ower” (too as well as over), “picters” (cinema) and “lassie” (girl). The exclamations such as “Jings!” and “Crivvens!” have already been mentioned.

Were there differences between 1940s and 2006/7 text? In this study while slight changes are noticeable, they are not simply indicative of a loss of vernacular forms.

  • In the original 40s strips there was considerable variation in vernacular use, some were more ‘anglicised’ than others, for no apparent reason.
  • Much of the core “functional” grammar and lexis seemed unchanged (although the “-na” endings had changed to “-nae”, suggesting an influence of Glasgow dialect).
  • The 2006/7 strips dealt with more modern subjects (television shows, cars, sky-diving, telephones, survivalism etc) so more general lexis is to be expected.
  • There was an influence of modern phraseology e.g. “cheap as chips”.
  • Some “new” Scots phrases appeared “Weel!”, as well as more “Scots” spellings “feenished” (finished), “groond” (ground),”puir” (poor), “eediots” (idiots), but some were tending towards anglicisation e.g. “juist” (just) was spelled as “jist” or often “just”. “Sae” (so) was morphing into the standard form and some of the rarer words in the early strips such as “dub” (puddle) were missing from the later sample.

How different is this from the ‘standard’ English collection?Using the concordance program, the top 50 corpus items were listed from the BNC-OU conversational corpus and compared with the top 50 items in the 1940s collection. Starting with the BNC-OU benchmark, the 1940s list shared 58% of the terms, but 18% were equivalent vernacular forms (e.g. “tae/to”, “ye/you”). In other words there was only a 40% match in the most common words. In the 1940s corpus, 14 items (28%) had a distinctive Scottish orthography

tae, an’, ye, o’, no’, a’, oot, ma, Paw, wi’, juist, Maw, Jings, wis.

Interestingly in the newer corpus, this had climbed to 18 items (i.e just over a third).

tae, ye, an’, o’, whit, a’, ma, oot, aye, no’, hae, ye, jist, dinnae, ken, wis, aff, aboot

In other words, by this criterion The Broons had actually become more vernacular.

I then compared the most frequent 100 words. The results were similar. 33% of the top 100 words were identical between the BNC-OU corpus and the combined 1940s and 2006/7 corpus and 22% were the vernacular form equivalents. Again looking at the Scots forms in the two corpora, in the 1940s collection again 28% had a Scots orthography but in the contemporary collection a striking 37% of items were Scots. Again the trend is apparently quite opposite that identified by Hoyer.

How were the words used? Bronislaw Malinowski (cited in Hewings 2005 p14) argued that choices of words and grammar were dependent on culture and (sociocultural) context. These choices may be subconscious, but are nonetheless motivated. In the case of Scottish vernacular language, diglossia and accommodation theory may be highly influential, and the Broons can be considered a case where lexicogramatical choice (initially by Dudley Watkins) was used to construct an identity for his characters. By using concordance software, the quantitative extent of this becomes clear, but what is the qualitative role of vernacular variants? The distribution may of course be arbitrary or historical, but that would undermine the idea of linguistic choice. Looking at the most frequent vernacular words in the combined Broons corpus, is it possible to determine their function?

tae, ye, an’, o’, whit, a’, ma, oot, aye, no’, hae, yer, jist, dinna(e), ken, wis, aff, aboot, weel, mind, ane, wee, Paw, ye’ve, twa, cannae, fae, Hen, nae, doon, Jings!, oor, mair, efter, ower.

Most are functional words; noticeably the only verbs are mental processes “ken” (know) and “mind” (remember). There is also the affirmative “aye” (yes), the affectionate diminutive “wee” and the exclamation “Jings!”. Personal and possessive pronouns are represented; “ye” and “yer”, “ma” and “oor” as is the question “whit?” (what?). There are three negatives “nae’” (no), “dinna(e)” (don’t) and “canna” (can’t) and a judgmental adverb “ower” (too). Running a concordance reveals that four times out of 10, “dinna” is a negative command. Similarly “hae” (have) often  – 22% of the time according to a concordance search – appears in the modal expression “hae tae” (have to). Two words are personal nouns (“Hen” and “Paw”).

Thus even this small set of words there is a strong sense of personal interaction and judgement. Much of the dialogue is in question form, personal pronouns are common The modal verbs almost always have vernacular elements e.g. “ye’ll need tae”, and “hae tae”. Likewise, attitude is indicated by asserted evaluations such as “This’ll no dae”, “Jings!”, “Ye auld gowk!” and vernacular phrases such as “wee” and the diminutive suffix “-ie” as in “hoosie” are commonplace. Informal constrictions, particularly contractions such as “that’s” (a phenomenon shared with colloquial English) further reduce social distance are use as are intensifiers, particularly the evaluative  “ower”. This linguistic features are used to reduce the social distance between the characters and the characters and the reader.

The comic book mode is quite highly structured. A theme is usually set out in the first few panels, and thematic progression occurs through repeated language and visual clues, leading to the inevitable denouement. Each strip is a mini-morality tale, usually variations of “pride comes before a fall”. The importance of word choice here is that complex situations have to be explained quickly, so using the readers’ linguistic and cultural references can make this process more efficient, aid identification and reduce communicative distance (i.e. seem more immediate).

In my view the debate about Scots language and writing has suffered from the neglect of this rich linguistic field from established linguists. The curious role of The Broons and Oor Wullie deserves more attention and it is encouraging that linguists (from outside the British Isles, it should be emphasised) are now beginning to work with this resource. I believe for example the continuous debates on Scots orthography have (deliberately?) overlooked this enrmous collection of popular texts.

Maybe the argument is that the vernacular of The Broons is simply not ‘Scots’. The use of concordance software allows us to say that The Broons is about 50% Scots, considerably more than many Scots speakers would use in everyday conversation. However the intriguing idea emerges whether the “differences” i.e. vernacular words have some special communicative value. By looking at the most common non-standard terms in I suggest that they cluster around fundamental mental and processes relating to personal interaction. Moreover I found that far from eroding, vernacular forms in The Broons, although they had indeed changed, were actually about 10% more prominent, quite a significant increase.


Anon. (2007a) The Broons, D.C Thomson & Co Ltd, London
Anon. (2007b) The Broons and Oor Wullie 1946-1956 The Golden Years, D.C Thomson & Co Ltd, London
Corbett, J. (1997) Language and Scottish literature Edinburgh: Edinburgh University Press.
Hagen A I (2002) Urban Scots dialect writing, Peter Lang AG, Bern
Hewings A (2005) Corpus and grammatical description in K A O’Halloran and C Coffin, Book 1 E303 English grammar in context, The Open University, Milton Keynes

Scroll to Top
Scroll to Top