People-focused knowledge management

Long-Term Memory
Long-term memory stores information and knowledge as mental
objects for long periods. It may be created by permanent links that
somehow represent what we know. Long-term memory is thought to
have “limitless” capacity for things to be remembered. It is located
throughout the brain, presumably in the areas where the various
functions for vision, speech, and so on, are located. The memory
objects that we store in long-term memory are not directly accessi-
ble but must be accessed by activating working memory. As indicated
above, this often creates a significant bottleneck.
Long-term memory is organized functionally into at least four 
separate kinds of memory which serve quite different purposes: 
procedural, episodic, priming, and semantic tasks. Semantic memory
is further divided into lexical and encyclopedic memories. Some
researchers also suggest that the episodic and semantic memories
together form conceptual memory. Other researchers suggest that the
distinction between semantic and episodic memory is wrong, that
there is semantic content in episodic memory, that conceptual
memory is episodic in nature, and that lexical memory contains
knowledge of syntax.
The domain of procedural memory is thought to be behavior,
whereas the domain of episodic, semantic, and priming memories is
cognition. All these memory functions have properties that are both
interesting and important when we work with what our coworkers
know, how they reason, how they differ, and how they gain access
to their knowledge, as well as when we consider how we can manage
knowledge in daily situations.
AP.qxd 5/3/04 2:29 PM Page 318

---

Memory and Knowledge Categorizations
319
Procedural Memory
Procedural memory is where we remember how we do things: the
procedures we follow when we perform tasks we are familiar with
and skilled in. Our procedural memories keep such mental objects 
as schemata, scripts, and routines as well as specific ways of acting.
It is believed that our procedural memory holds knowledge of
methods and scripts in the form of mental models that include
sequential steps or as chains of events where the outcome of one event
feeds into and controls the behavior of the next. The sequential steps
can be distinctly discrete steps, or they can be similar to “waves” of
activities whereby one activity wave gradually changes into the next
with seamless transitions.
Typically, we cannot explain what we know from our procedural
memory. When we recall mental objects from procedural memory,
the recollections are expressed as actions rather than explicit thought.
We can only “act out,” or perform, the tasks we have memorized in
this part of our memory. And we may only be able to perform the
task under special circumstances. For example, we may have to be
primed to perform a certain task as when we are led to perform it
as a natural extension of our present situation. We may also be led
to perform the task by a cue such as seeing a particular situation,
hearing a certain word, or thinking about something very specific. In
some instances, an expert can be led to document, or tell, how a pro-
cedural task is performed by recalling a realistic scenario and then
explaining it step-by-step from the beginning. The expert essentially 
simulates the performance of the task.