Mental Models

Mental Models

On several occasions, I’ve mentioned “effortful retrieval” as key to longterm learning.  I think that if I were to criticize myself  as a classroom teacher, I’d have to say that failure to review was a primary weakness in my teaching repertoire.  The key is too allow enough time to pass between study sessions that some forgetting has taken place, but not so much that competing memories have completely obliterated the retrieval routes to a learning.  Take a math fact.  Taught today, it should be reviewed tomorrow or the next day, then after about 5 days, then 10, then about 20 days, and after that, at random.  For the “average” person, this would be adequate for making the memory of a math fact or vocabulary word, reasonably permanent.

Repeated retrieval causes the brain to strengthen the routes of recovery, making them more resilient.  As we retrieve learning, we also improve it by reconsolidating it.  That is, we combine that knowledge or skill with newer information and skills, literally remaking the memory into a richer and stronger understanding.  Bit by bit, as we revisit memories we not only reconsolidate that specific memory, but we develop mental models.

Mental models, as distinct from a reconsolidated memory, might be thought of as clusters of memories.  A simple example might be related to driving a car.  When we learn to drive, individual skills like breaking or shifting become reconsolidated as we do them repeatedly and gain sophistication with their use, applying the skills to new and slightly different situations.  At some point, we move from “breaking” or “steering” or “shifting” to “driving a car.”  At that point, we’ve developed a mental model of driving a car and it contains all the separate skills necessary for driving.  With that model, we can move from vehicle to vehicle, applying our model with great skill.

So it is with learning.  It’s why we try to get students to recall things “like” what we’re introducing in class.  The attempt is to see the new information within the model students have already developed.  Research says that if we probe students for “What is it like?” rather than telling them what it’s like, the effort that they expend identifying the model will strengthen the learning and make it more resilient.