I’ve had this blog sitting in draft for so long that I’ve written a paper about it in the meantime! The working copy of that paper can be found on the KMi tech report site and we welcome any comments or critiques of it.
To give a brief motivation of the paper, we note that both the theory of epistemic frames, and the epistemic commitments theory I propose (motivated by earlier epistemic cognition research) are activity oriented and situated. Indeed, both also involve consideration of epistemic facets of activity – what is to to engage with knowledge? An analytic technique for epistemic frames – Epistemic Network Analysis – provides a technique for exploring the connections between particular modes of working with information (epistemic commitments), and it is in these connections we suggest the key insight lies.
One of the interesting things about ENA is the scope to look not only at which types of epistemic commitment are being made, but how they are paired. So for example, we might expect people to “match” answers to questions more when they’re also engaging with a “simple” perspective on the connectedness of knowledge. More interestingly though (and more nuanced) it might be that engaging in “understanding” (i.e., exploratory dialogue) is informative regarding the utility of the sourcing strategies in any particular context – that is, whether students seek ‘authority’ or ‘corroboration’ isn’t interesting by itself (unless they do one or other very often), but knowing they’re engaging in critical thinking around those sourcing decisions is interesting.
In the paper we describe epistemic commitments as involving:
- Which sources re used – comprised of credibility decisions (from corroboration of sources, to trust in authoritative sources)
- How they are used (in action – to justify claims, to make decisions) – comprised of justifications and source use (from dialogic approaches using talk of an exploratory nature, to attempts to match information to answers and use basic factual information)
- How links between them are created (or not) – comprised of claims, (explicitly in language and through structured environments, as well as implicitly through search patterns) made around connectedness of concepts (from a holistic perspective of knowledge to a piecemeal)
Indeed, taking some data from a particularly nice ‘epistemic’ task – the use of search engines to seek information – and comparing prior analysis (publication forthcoming) and ENA analysis, it turns out that the ENA provides a pretty good insight into the epistemic commitments of that pre-analysed data. Of course, this needs validating across a wider set (this is a sort of proof of concept that such a model can be applied…we discuss the concern around circularity/shoe horning in the paper).
Anyway, thoughts welcome on the paper, I have some other ideas around the need to conceptualise the ‘Stanza’ in these tasks, such that the analysis properly associates those things which should be grouped under that title. We can imagine various ways of doing that:
- All talk around a particular document/query/question (as in this pilot) is a stanza.
- A version of (1), all talk around a cluster of associated documents/queries/questions is a stanza
- All trace and dialogue associated with a particular claim made in an output product is a stanz
Using the notion of ‘uptake’ and contingencies I know Dan Suthers talks about might also be interesting here (which moves are contingent on which prior moves) – this goes beyond transactivity (the uptake of explicitly linguistic ideas) and is very in line with the ‘commitments’ – working on someone else’s nodes is an implicit commitment to the contents of those nodes (and is of course contingent on their creation).