OMG! Can we use social media to investigate language change?

For the last class of Introduction to Minority Languages II the first-year students (of BA M&M) were asked to do a small investigation of language in social media. One of the larger debates concerning language in our everyday lives concerns the effect that texting and social media might have on our language use. Most people will have seen Facebook posts and “news” articles about the derogatory effects of “text speak” on teenagers’ language. I think it’s important for students to view these criticisms in light of what they have learnt about language change: that the natural state for language is to be continually changing – there is no such thing as a static (living) language. Students should also realise that social media language is yet another register, or style, that our amazing human linguistic capacity can deal with alongside all of the other styles we use on a day-to-day basis.

The language we use on social media can be extremely interesting. One might say that it’s the style of writing that lies the closest to speech. For sociolinguists this presents the possibility of using social media to investigate language change in progress. Importantly, social media presents us with a plethora of data – most of us tweet, Facebook and text with one another throughout the day.

For their seminar students considered language variation and change on Twitter, Facebook and on WhatsApp. They were free to pick their own linguistic feature and community to research, but I had given some ideas for topics. One was to see whether women are at the forefront of introducing “OMG” (oh my god) as an expression in European languages. We’ve been talking about gender differences in speech, and the classic sociolinguistic principles when it comes to language change and gender. One of these hold that women are leaders in the use of innovative forms.

The students who picked this topic looked at Dutch and German and all found pretty much the same thing going on. When looking at the last 100 tweets in which OMG occurs on Dutch twitter there was a consistent trend of 70-75% of these being produced by women. German is less straight forward: one study found that 69% were produced by female twitter users, while another found a 52% usage rate. As always, there’s lots of possibilities for future research here: what are the interesting regional or social differences in the use of OMG within German? Would we categorise the use of ‘OMG’ as code-switching or as borrowing – How does the word function linguistically? Can we take a more refined approach to gender and social identity to understand the social meaning of ‘OMG’ in Dutch and German speech (or writing)?

Of course the social media class gave students to look at other linguistic forms: One student looked at the use of ‘slay’, in the sense of being successful or “killing it”. There are also indications that the use of this lexical variant (with this particular semantic meaning) in English is led by women, at least in a sample of the latest 100 tweets from American English. Another student looked more locally and tried to figure out whether the Frisian word ‘fertuten’ (‘use’ or ‘achieving results’) is used predominantly by older or younger twitter users. The largest user group for this word consists of older males. Since Frisian is a lesser used written language, this presents us with interesting follow up questions: is this word in fact used predominantly in this group, or do older men make up the largest proportion of Frisian-speaking twitter users?

There are plenty of interesting questions to explore for the first years in the courses still to come in the BA. More research can be done, also at a BA level, as this year’s cohort so brilliantly show.

For people who are interested in learning more about social media and language use you can check out this great TED talk by John McWhorter about the topic.

Did you know that the Computational Linguistics Dept at the RUG collects twitter data on a daily basis? you can have a look at the corpus of Dutch tweets (billions of them!) here.

If you’d like to see what our students produce on a weekly basis, have a look at the short report written by student Sybrand Grasdijk below. The purpose of doing these assignments and writing these reports is to become familiar with linguistic research questions, and to understand the fundamentals of quantitative sociolinguistic work.


‘OMG’ on Twitter: gender distribution

The purpose of this study was to examine whether male and female Dutch Twitter users, tweeting in Dutch, vary in their use of the expression ‘omg’ (oh my god). Using a list of 100 tweets, I looked at the number of times ‘omg’ was used by males and by females.

In order to obtain data on this subject, I used the advanced search function of Twitter. The search term I used was simply ‘omg’ and to contain the limit the results to only Dutch Twitter users, I set the location as ‘The Netherlands’, which lets Twitter only show tweets that are sent from the Netherlands.

In the list of search results, there were still a number of tweets in English, because some Dutch Twitter users use English as their Twitter language. I did not use those tweets in my final results. I made a list of the 100 most recent Dutch tweets containing ‘omg’.

Making the list, I looked at the Twitter user of each tweet containing ‘omg’ to determine whether the Twitter user was male or female. Sometimes it became clear from their profile picture, other times from their account name, but from time to time I also had to look to their tweeting history to determine whether this user was male or female. The few times the gender of the user was not traceable, I did not use those tweets in my final results. Also, I only used each Twitter user once. Some of them appeared more than once in Twitter’s search results, but I only used one single tweet per user in this study.

Females (71)
71 / 100
Males (29)
29 / 100


This shows that significantly more females use ‘omg’ in their tweets than men. To get a clearer view of the actual use of ‘omg’ among males and females, more research is necessary. This research should focus on the gender distribution among all Dutch twitter users and then look at how many of them use ‘omg’.

There is no data of the gender distribution of tweets sent from the Netherlands, but according to Sloan (2014), 48,8% of all Twitter users is male, 51,2% is female. This is quite even, as opposed to my results of the ‘omg’ use.

So to conclude, it looks like female Twitter users use ‘omg’ much more often than male users, but more and more extensive research should be done to get a clearer view of this gender distribution.


Sloan, L. (2014). Who Tweets? Deriving Demographic Information from Twitter . Accessed January 13th, 2016 via