In my parallel blog I highlighted some [current work in epistemic cognition and games]1 – particularly ones I think could be fruitful for further investigation. This blog expands that, if you’re interested some of these ideas appear in a different form in my CSCW workshop paper on [tracking epistemic beliefs and sensemaking in information retrieval]2. I want to suggest that, if we’re thinking about how people conceptualise ‘knowledge’ and the knowledge they need, we need to think about this in terms of their actions, and their use of language ‘to do’ (as opposed to ‘to represent’). While Osterholm [zotpressInText item=”{DBFA2ZA2}”] has explored this to some extent, that work did not focus on the collaborative elements of dialogue as a tool for joint action and co-construction. Where Shaffer’s work on epistemic games has looked at interactions between agents, I think this is generally more focussed on the roles of the agents than on their use of language to think together. I suggest we should focus on action, and language-in-action, moving from a desire to discover underlying “beliefs” and cognitive constructs, to a more pragmatically oriented post-cognitive perspective on epistemic action – those actions which are oriented around epistemic goals, tools, and desiderata. # Conceptualising a model Given the characterisation of epistemic beliefs as “a lens for a learner’s views on what it is to be learnt” ([zotpressInText item=”{DBFA2ZA2}”]), I propose that one highly fruitful way to explore epistemic behaviour is to explore the ways in which users negotiate, and enact, the information search retrieval and processing activity. That is, the behaviours they engage in to: work out what they need to know; seek to find this information; evaluate the information; construct some perspective, product, or use from the information. One such activity is information retrieval using a search engine; this is a great activity to explore because: 1. It is often collaborative or cooperative (I won’t put refs in here, but there is research to back this up – although it is obviously not always collaborative) 2. We have access to a searchers search terms (their query) 3. We know what pages they open, and if we can see their final products (or track dialogue on the way) we have at least some access to information gleaned 4. People engage in various other broadly epistemic behaviours in such contexts including sharing links, asking people for advice, copy-paste, structuring results (see e.g. my paper on [tracking epistemic beliefs and sensemaking in information retrieval]2). Classic models of information retrieval, highlight the shift through stages of identifying an information need, seeking information, and evaluating that information.  These bear some commonalities with at least the model of epistemic beliefs and self-regulated learning presented by Muis, who proposed “four phases of self-regulated learning…(1) task definition, (2) planning and goal setting, (3) enactment, and, (4) evaluation” [zotpressInText item=”{5VUIDIHN,307}”]. Even at this descriptive level the relationship between this and models of search, as presented below, are striking, indicating that we can conceptualise the first phase as that in which the ‘need’ is defined, the second as that in which those needs are translated into search queries, and enacted (with more or less use of the sophisticated search engine tools), and the third as the ways these results are treated.  The fourth could denote the iterative nature of search, both in terms of need redefinition, and refinement.  The nature of these similarities may further indicate that IR is a good activity through which to explore epistemic beliefs. However, such a model is not to imply that the process is linear, or uni-directional.  Users will move back and forth between stages, and iterate on stages in loops which encompass a larger or smaller number of the stages.  Throughout each, the ‘information need’ is mediated by the tools, collaboration, and the information found. The exploration of this is exciting for two reasons. 1) to explore how use of tools allows us to extend beyond a simple “did you get the factual recall?” assessment, to use of search as an epistemological tool [zotpressInText item=”{P8385ZVT}”] 2) as a means to explore epistemic behaviour, as a new post-cognitive method.

 Some Initial Mapping – Search Actions as Epistemic Indicators

(Epistemic move first [no.], then the indicators underneath [letters] – sorry, seems not to like inserting tables into this) 1. Assess source credibility (justification?) (certainty?) a. Open results? b. Open results based on those results (through ‘similar to this’ or references) 1. Talk re: credibility 2. Seek information, define information need (simplicity?) a. Ask other people b. Enter queries c. Use reading lists as a source d. Talk to people about the IR process (‘search together’) (exploratory (interthinking) v. other sorts of talk?) 1. Make suggestions to other people regarding articles, authors, and keywords 2. Build, connect, filter ideas? (simplicity?) – a. Modify queries b. Make connections between ideas (use concepts from one paper/search repeatedly) – cohesive tie, and use of central nodes in Cohere c. Look up keywords (engage in the language of the community) 3. Assert (source?) – a. Cite sources b. Ignore suggestions (from recommendation systems, peers, and seniors [reading lists or explicit suggestions]) 4. Simplicity/certainty? – a. Go past the first page of results b. Spend a long v. short period searching c. Search for opposing perspectives and stability (geographic, time) (proxy for which is, structure results to express semantic relationships; support, critique, etc.). 5. Multiple – a. Summarise perspectives (create nodes) Bibliography [zotpressInTextBib style=”apa” sortby=”default” sort=”ASC”]


  1. http://

  2. 2