Most of the books I’ve been reading are assisting my understanding of the topics that are central to my study. Others are in areas which I hadn’t initially considered, but might inform my research. ‘Discourse of Twitter and Social Media’ (Zappavigna, 2012) perhaps surprisingly falls into the latter group. Yes it may be on Twitter, but I hadn’t really thought about discourse analysis as a possible method, mainly because although I was aware of it, I didn’t really appreciate what it entailed … which is why I chose the book.
Dealing with the issues of corpora for social media and then outlining the specific features of microblogging, Zapavigna then uses successive chapters to explore in detail how people share their experiences and enact relationships by adopting a social semiotic perspective. Covering memes, slang, humour and political discourse, she provides illustrative examples under each category, together with the analytical linguistic techniques used. A particularly resonant chapter for me was that on ambient affiliation, seen as an alternative mechanism through which people with a shared set of values and interests become associated, other than as a community. This bonding is achieved through common behaviours and simple yet powerful semiotic markers like the hashtag. Affiliations may be more transient than communities, virtual or otherwise, and be time-bound to specific incidents or issues. Whilst others form, stick and show durability over time. We begin to see here the importance of what Zappavigna calls ‘interpersonal search’ where people employ the affordances of the medium to actively (and passively?) seek out other people with whom to bond around (or clash over!) certain issues.
Another aspect of the book I found valuable was the way in which many of the technical elements of microblogging were scrutinized. In addition to the aforementioned hashtag, the functions served by the @ and RT were also discussed. This naturally prompts me to consider here what ANT role these components play; actors, mediators, intermediaries? Although the part played by emoticons was also discussed, I wondered whether these markers worked at a different level than #, @ and RT, instead serving more as semiotic replacements for the visual cues lacking in online settings. Do emoticons perform a significant functional role in professional learning exchanges? I suspect not, but it might be interesting to find out.
One small topic which caught my attention was ways in which we can classify Twitter users as information sources, friends or information seekers (Armentano et al, 2011), or indeed as informers or meformers (Naaman et al, 2010). This seems particularly relevant to my study where people often claim that one element of the professional learning they derive from Twitter is that it is a wonderful source of resources. Are they the information seekers who rely on the information sources? Are some of those information sources also meformers? Should we extend or amend the categories to resource sources and resource seekers?
An area which has been flitting in and out of my mind and one which I will need to discuss at much greater length is my approach to sampling. Zappavinga highlighted how episodic the nature of the Twitterstream is and how that influences the choices you need to make when devising a sampling strategy. The 100 million word HERMES corpus used throughout most of the book was collected over a relatively short period from the whole Twitter population and therefore is to some extent skewed by what was topical at the time (among other things, Valentine’s Day apparently!). Since I’m focusing on a particular issue, the whole population is of less interest to me than those who are teachers or involved in education (need to think more about what my target group will be). How I establish that target, devise a meaningful time-period and thereby compose my Twitter snapshot will require careful and informed consideration, as indeed of course will the subsequent analysis.
One finding in the book which I found reassuring was that the most common 3-gram (3-word triplet) in the HERMES corpus was “Thanks for the”, which suggests that people are indeed getting something from their fellow Twitter users. Perhaps that 3-gram might be worth investigating within the samples I choose to use, as an indication of resource and information sharing? It might be interesting therefore to follow that up with what appears in the R1 position i.e. is the next word in the sequence.
What I learned
- Some of the techniques and processes involved in discourse analysis.
- A better understanding and appreciation of the importance of some of the specialised lexis which Twitter users employ.
- How the medium was set up mainly to allow broadcasting and that engaging in dialogue or co-ordinating action may be consequently be more difficult to achieve.
What I need to do next
- To give further consideration to distinctions, similarities and crossovers between affiliations, communities and actor-networks.
- To explore further texts which will help me think about sampling strategies.
- To learn more about discourse analysis and whether there are elements (like n-grams) I might be able to fruitfully use.
Armentano, M.G., Godoy, D.L., Amandi, A.A., 2011. A topology-based approach for followees recommendation in Twitter, in: Workshop Chairs. p. 22. (PDF)
Naaman, M., Boase, J., Lai, C.-H., 2010. Is it really about me?: message content in social awareness streams, in: Proceedings of the 2010 ACM Conference on Computer Supported Cooperative Work. ACM, pp. 189–192. (PDF)