Posted by: jleventon | October 30, 2012

On Language, Research and Error…

A couple of weeks ago, I went to a guest lecture given by Professor Robert Chambers here at the University of Leeds.  I suspect that most students of international development will have engaged with Chambers’ work on participation, and I am no exception.  He was visiting Leeds to talk about Professional Error and Ignorance in International Development.  The lecture was every bit as interesting as the title suggests; Chambers asked us all to consider the varying ways in which we introduce error in understanding through our positionality, understandings and interpretation of data.  Its certainly got me thinking about my own data collection, analysis and use; the sources of error therein; and how I deal with it.  Being fairly nomadic, and working in a range of countries, one of the things I have focussed on is language.

I cannot escape the fact that language is a potential source of error in my research.  Amongst others, I’ve conducted research with participants who speak Hungarian, Greek and Bemba; all languages from different roots.  For my own research, identifying policy documentation and stakeholders (and participants) is hard going when everything is written in a different language, or even a different alphabet.  Therefore introducing the potential to miss participants, documents and contexts.  Its not such a problem for executing methods if I am interviewing policy elites at the national level as English is their working language.  Finding the right contact details, getting past the receptionist and making an appointment can all still be challenges though!  Below the national level and with community participants, conducting a meaningful data collection method in a language that is not my own is truly difficult and potentially leads to mis-understandings or missed data.

Whether or not we should learn the language we are researching in seems to be a bit of a divided topic.  I’ve witnessed a candidate, in his thesis defence, be interrogated on why he chose to stick to case studies where he already spoke the language.  The questioner felt that the candidate should have based their selection entirely on academic justification and learned the necessary language. Like the candidate, I think I disagree with this stance.  I speak other languages to a good standard.  But it takes a huge amount of time, more than is usually covered by PhD stipends or research grants.  Furthermore, I wonder whether I will ever understand the deeper meanings and cultural foundations that underpin another language.  Intonation, humour and meanings behind words vary between languages; more so where the language has a different root (e.g. Slavic vs Latin).  I think it takes years of immersion in a language to understand all these things.  Without such immersion, I would argue that misunderstanding such meanings, or not noticing them, would be a source of error in research.

Perhaps the most obvious and common solution is to use native language research assistants.  During recent work in Zambia the rapport that my research assistants had with community participants was incredible and gave me confidence that the data collected was rich and meaningful.  But this was better achieved when I gave them more autonomy and allowed them to be assistants rather than just translators.  Thus diminishing my control over the direction of data collection and my deeper understanding of the data collection process and in turn reducing my ability to accurately incorporate that data once I got it back to the UK, even when the research assistants had been carefully managed.  Such separation may also be beneficial; in my PhD research, I felt like having a translator between me and the participant allowed me to remain impartial and prevented me from becoming emotionally engaged with participants.  I believe this allowed a dispassionate and therefore less erroneous analysis of the data.  However there is obviously a balance to be struck in maintaining both separation and sufficient involvement to understand the data and its context.

My approaches for dealing with the language issue target both research design and its practical implementation.  On a practical level,  I like learning languages so I try to learn what I can.  Depending on how much I am able to learn in the time available, I can use this to start looking for documents and stakeholders.  It also allows me to get some rapport with participants, even if I can’t conduct the entire interview in their language.  In my PhD I was fortunate to have sufficient time to learn enough to be able to identify themes in responses before my translator spoke, allowing me to remain responsive.  However, learning to this degree is less possible in my current post-doc research where I am not spending such long amounts of time in the field, and where formal language training is not accessible.  Instead, I am trying to design-out potential issues, making sure that the detail of analysis required by my methods/research questions does not exceed the detail that I am actually able to gain in light of my language abilities.  For example, at the community level participatory mapping and creating picture cards are decreasing my reliance on verbal communication and are appropriate for the data that I need to collect. Where I have verbal data collected in a different language, I will be using it for coarse themes rather than attempting to perform a detailed discourse analysis.

Hopefully, through these measures and by remaining reflective on these topics I will be able to avoid some of the errors that were so articulately discussed by Prof. Chambers.  I am open to other advice, tips and thoughts though!



  1. Hello!

    Very interesting Julia, I had not thought of these things relating to language. I guess it depends on the area in which you conduct research. If you are British and you want to study ancient Greek archeology and history then I guess learning the language is vital, but also expected. In your case, you work in several countries in Africa, with people speaking all kinds of languages. Unless you said you would focus on one country or tribe, then it is impossible to master all these languages.

    Thanks for the “food for thought”….

    P.S. if you want to learn Greek let me know… 🙂


  2. I like this and recognise it as a real practical problem. I was always glad to have a minimum grasp of the local language so I could follow th egest of the conversation betwee nassistant and interviewee. It also raises a conceptual puzzle for me: I think it is useful to clearly distinguish between facts/data and meaning/interpretation. If you are looking for meaning, as you do, then I wonder whether ‘error’ is an appropriate term any more? It all becomes interpretation, and that is not a clear cut issue even when you work in yor own culture – but obviously a higher risk when in another culture – so maybe it’s culture not language that we need to address?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s


%d bloggers like this: