Communicating cognitive schemata to novices
There isn't much of a question that, in comparison with novices, experts have richer cognitive maps. In other words, not only they have a larger number of interconnected conceptual nodes, but the richness of the communication among their conceptual nodes is greater. There are some excellent resources about this field out there, the book by Stephen Downes being one of my most recent discoveries.
What follows is probably described and even demonstrated by others, but here are some thoughts that occurred to me while going over this material:
Experts are likely distinguished from novices not only by the number of conceptual nodes and their connections, but also about how they display the hierarchy among these nodes. So, for example, a certain node clusters are centered around a concept that is considered more important by the expert, and it's this central concept that then connects to other regions of the network.
This central node could well be considered a latent node, since it connects a number of items that are directly measurable
A big bunch of the literature doesn't make use of the extensive visualization resources available through social network analysis software such as R and its amazing task view on graphical models
Although classical elicitation methods such as think aloud and problem handicapping are used in the cognitive task analysis literature - see for example Schraagen's book, it is interesting that these evaluations are not placed within specific situations or contexts. In other words, would be interesting to see the situated cognition literature join the wagon.
Fuzzy cognitive maps should definitely be part of this discussion.
But still, the most important and unanswered question is how, once these cognitive maps are appropriately prepared and represented, how do we use them to train novices?