Audiences in discourse

[Christopher Brooks]

Hi,

I wanted to dig in a bit to the idea of an audience in discoursedb.  In an asynchronous form, is the audience the list of people who could have opened the message but didn't (e.g. all of the students enrolled in the course)?  A list of people who viewed an index where the message was listed but didn't open it (e.g. in an online forum where I'm a member, instead of "the whole internet", or all of the followers of a given person on twitter)?  Or just those who actually opened the message and read it?

Is it important to distinguish between these?

Regards,

Chris

[Oliver Ferschke]

Hi Chris, 

this is a very good question. I think audiences can be understood at two levels - what was intended at what actually happened ( a little bit like illocutionary and perlocutionary acts - but of course that's comparing apples with oranges).

On one level, we need to represent the intention of the author and take into account the affordances of the source system. For example, in Twitter, you can send a reply to a tweet (explicit recipient), which will automatically appear in the streams of your followers and the followers of the recipients (implicit audience) and at the same time it can potentially be read by anybody who happened to find the link in a search. These audiences could all be considered differently in an analysis. The types of audiences that exist largely depend on the options the source platform provides.

On another level, we need to be able to take into account what actually happened in the discourse - i.e. who engaged in a discussion either actively or passively. This includes information about who actually opened a message or visited a thread, who actively engaged in a thread or replied to a message and who of the potential recipients did not open a message. Within a closed system that we fully control, it will be possible to gather all this information while we will have gaps if we want to represent data for which this information is not available. However, in this context, we are getting very close to the clickstream level. I don't think we want to attempt to represent actual clickstream data in DiscourseDB. In the MOOC context, for instance, this is handled by MOOCdb. We collaborate with the developers of MOOCdb, so we are eager to find connections between the two projects. And this would be one - getting the information about what actually happened. In contexts outside of MOOCs where clickstream data might also be available, we have to discuss how this data could be incorporated.

-Oliver