This is a blog post, I have written about two years ago, but now seeing the relation between the Long-Term Memory and Knowledge Managenent, it is well worth publishing it (as intermediate step towards Storytelling as well as Knowledge Mechanism - thanks for your coment, Ewen).
I came across a very interesting article (actually a whole field of research: Cognitive load theory): J. Sweller "Visualisation and Instructional Design". You can read the article from the presentaton skill perspective and it is rewarding, however (you may call it a professional disfunction) I discovered there a lot of insights for KM also: 2 key words are here refinement and reward for contribution.
I recommend to read the article, but as an "imagination" I try to bring it across in a nutshell:
Our brain, the cognitive architecture has two parts: working memory and long-term memory. The working memory, the processing engine, is highly limited in capacity and duration. Due to the limitations the working memory has problems with complex problems (high element interactivity material) and needs long-term and learning mechanisms. Knowledge is stored in the long-term memory in schematic form (semantic memory) and schema theory describes a major learning mechanism. Schemas allow elements of information to be categorized according to the manner in which they will be used. Schema theory assumes that skill in any area is dependent on the acquisition of specific schemas stored in the long-term memory. (Just think of KM databases as long-term memory and schemas as refined knowledge to get the translation into KM). Automation: High element interactivity material that has been incorporated into an automated schema after extensive learning episodes can be easily manipulated in working memoy to solve problems and engage in other complex activities. In order to support automation you can attempt to imagine the procedures that have been learned (the contribution and formulation of a knowledge asset). Imagining requires the learner (that becomes by expressedly in the KM database imagining a teacher) to mentally "run trough" or visualize the procedure in the working memory. For complex problems that will only work when schemas have been acquired, thus supports automation.