This weekend I was at the 5th International Language in the Media Conference in London, which was a great opportunity to catch up with friends and meet some new researchers too. I was talking about absence (of course!) and got some really good ideas for my project, which I will talk about more some other time. But, basically, I have been looking at how we can investigate what is absent in discourses, and, what discourses are absent in discussion of migrants and migration in the press. I have been working on four main ways into the data:
- Corpus linguistic starting points
- (Critical) discourse analytic starting points
- Meta-linguistic markers as a starting point
- Use of external data as a starting point
And I thought I would start with the last in the list; the use of external data. In this category, I am thinking of non-linguistic data which does not come from the corpus, so the example I gave on Saturday was population statistics.
My idea was this: If we compare the rankings of ‘foreign-born residents in the UK’ with the rankings of mentions of nationalities in newspaper articles that also mention migration, can we start to identify which groups are under-represented? And can this tell us anything about the way migration is represented in the UK?
It is currently a very rough measure which needs refining, but it looks worth working on as a point of entry into the data. Most nationalities in the list occurred within 10 ranking places on each scale, for instance, people from Poland were ranked in second place in terms of population in the UK and 11th in terms of mentions in the newspapers. I then focussed on those that were ranked more than 10 places lower on the newspaper mentions list, than in the population list, as a way of identifying who perhaps was less visible. This is what came up:
- 11-20 rankings lower: Ghana, Jamaica, Kenya, Lithuania, Malaysia, New Zealand, Nigeria, Portugal, Slovakia, South Africa
- 21-30 rankings lower: Bangladesh, Somalia, Sri Lanka, Zimbabwe
- 31+ ranking lower: Philippines, Hong Kong
But having found that these nationalities appear to be under-represented, as a discourse analyst, the problem is interpreting the data because the full answer won’t be/can’t be in the corpus, of course.
In a previous study of the Italian context, I found it useful to return to the population statistics and look at what employment roles dominated for particular nationalities and saw a strong correlation between domestic labour and low visibility. But this is only partial too, because as newspaper consumers we know that assimilation of different nationalities into one broader category, e.g. east European, is a common feature of the discourse, in part, because the nationality may not be known (by text producers or implied readers), in part, because it is a rhetorical device for creating threat.
And so, to return the heading: how do you analyse the thing that you have identified as being absent?