Graphical representation of accent online
The study in a sentence
Language variation and change can be seen in differences in phonetic/phonological form (how something is pronounced), in lexical choices (which words are used) and in syntactic form (which grammatical forms are used). Although you might think that social media posts will only show lexical and grammatical variation, it turns out that social media users often represent the phonetic features of a dialect in written form; spelling choices can be used as a way to project linguistic identity online.
This study explores the following questions:
do spelling choices on social media reflect the phonological features of regional dialects?
if so, how are the phonetic dialect forms represented?
how often are these spelling choices used? and where are they used?
does the spread of regional features in social media posts match where the dialect is spoken?
Sociolinguistic salience is the property of a spoken form which causes listeners to respond to the form in such a way as to indicate that it encodes information about the (presumed) social characteristics and/or geographical origins of the speaker. Relative sociolinguistic salience can be estimated using a Social Category Association Test.
From this study, we can infer that the dialect features that social media users choose to represent in written form are those features of the dialect that are highly salient.
"naturally occurring corpus data - which is not affected by the observer paradox - can uncover how variants can either have a stereotyped function within a speech community or be used to portray an identity" (Nini et al 2020:23)
The authors used a corpus of posts ('tweets') from Twitter, which in some cases are 'geo-coded' i.e. accompanied by information about the geographical location of the user when they posted.
This was a large corpus (2 billion words!) so searches for regional spelling variants were carried out automatically. The results were plotted on maps showing the 'spread' of features across the UK.
Spelling variants reflecting dialect features that were found in the corpus include:
TH-stopping thing -> ting; them -> dem
happY laxing happy-> happeh; city -> citteh
"G-dropping" going ->goin'; singing -> singin'
note that there is no 'g' sound at the end of the word in the pronunciation of the 'ing' ending, so you can only call this pattern 'G-dropping' when talking about changes to the spelling!
In more detail
A longer explanation of the research study
Pre-Workshop Talk [10 mins]
Live Webinar Talk [22 mins]
Slides from the Live Webinar talk