O n f e r e n c e

Methodological Discussion 
Colpaert (2013) discussed four methodological challenges to CALL research. They were: 
research design, replication, slow research and transdisciplinarity. The biggest design 
problem, naturally, in this kind of study is how to map processes of norm setting. The 
process is very much a slow one, and needs the long-term study of the same subjects in 
the same setting under the same conditions 
–
a real challenge for researchers. It is 
demonstrated here that such long-term study is not always needed, and norm-setting 
can be seen in shorter-term, smaller sets of data. Variations in the data are a challenge 
for researchers in the recognition of norms. However, it has been argued in Complexity 
Theory (Larsen-Freeman and Cameron 2008, Larsen-Freeman 2012) that variation is a 
natural part of complex systems like language, and can be a sign that change, like the 
setting of a norm, may be about to take place. 
Transdisciplinarity depends on the nature of the research. Traditions of analysis can be 
brought in from different areas of linguistics, both qualitative and quantitative, as well as 
from other research areas like anthropology or ethnography. As a discipline, CALL 
develops by researchers bringing their different research traditions, and we gain different 
perspectives on the same object of study. 
Finally, we see replication as really the main methodological problem, as data is always 
contingent on its conditions. Thus, it is extremely rare to have two sets of data that can 
be considered truly comparable. This is both positive and negative for the discipline, but 
we see it as a sign of the richness of our data, language in use. We now turn to the data 
we will analyse in the remainder of the article. 
Our data 
The data consists of a series of chatlogs from an MA course in English linguistics run by a 
Swedish university. The students were non-native speakers of English, and needed an 
average IELTS score of 7.0 with no lower than 6.5 in each component to be admitted 
onto the programme. The course chosen for analysis was the first on the programme, an 
introduction to English linguistics. There were eight sessions in total, including core 
linguistic topics like phonetics and also sociolinguistic discussions about the link between 
language and politics, for example. For the three sociolinguistic sessions plus the one on 
morphology, students arranged their own pre-seminars in four groups without the 
teachers being present. The resulting chatlogs were sent to the teachers, and helped 
guide the seminar discussions. The other sessions had only seminars. All discussions took 
place through the Skype textchat function. The chatlogs came to a total of roughly 
160,000 words. The students had limited Internet experience generally even in their 
native languages, Bangla and Vietnamese, and none had taken a distance course before 
this one. Thus, they would have had very little access to English native speaker norms in 
computer-mediated communication. As a result, the processes of them setting norms for 
the discourse community are of interest. We turn finally to the analysis of the data, and 
hope to demonstrate that norms can be identified even in small sets of data like ours.