People-focused knowledge management

Working Memory
Working memory, or short-term memory, is considered to be the
center of consciousness. Conscious and explicit reasoning takes place
in working memory at the top level of attention. Hence, when I think
about how I will drive across town to a particular address, I reason
about it in my working memory. To help me, I retrieve from long-
term memory images knowledge of the road conditions, different
routes that I am familiar with, and similar material and use that to
“think” consciously about what to do.
A noteworthy aspect of the present model is that people have the
equivalent of many working memories which can operate on differ-
ent levels of consciousness or attention. Of major interest is the
working memory, which mostly operates on the conscious level.
However, people have the ability to perform reasoning and other
mental functions nonconsciously as when “suddenly getting a good
idea” without having thought explicitly about, it or when they sud-
denly remember the name of a person that they met five minutes
earlier but could not name at the time. Similarly, people perform a
large number of activities in automatic or semi-automatic ways. For
example, we can carry on a deep and highly conscious conversation
with someone while driving a car, we can discuss difficult topics with
someone while walking, or we can open and close doors automati-
cally and analyze obstacles and make semiautomatic decisions on
where to step.
In reality, the brain does not have a physically separate set of
neurons that constitutes working or short-term memory. Instead,
AP.qxd 5/3/04 2:29 PM Page 314

---

selective parts of the brain are activated when focusing on topics 
or functions handled by that area. When the brain is activated in 
this way, the target area becomes the focus of consciousness and will
remain active for about five seconds after the conscious mind has
shifted its focus elsewhere. That is conscious short-term memory.
To illustrate the way conscious or working memory functions,
think of the following model: You are in a dark room with a flash-
light. Wherever you shine the flashlight on the wall is where activity
takes place, where consciousness is focused, and where you can
perform short-term memory tasks. After you move the light beam
elsewhere, the wall continues to glow with an afterglow that lasts
some five seconds.
Another model — the “pinball machine model” — has been pro-
posed to illustrate how people shift attention from one subject to the
next. The pins are different subjects or concepts. The ball is the con-
scious focus. The focus bounces from one concept to another driven
by associations, and reasoning results from the mental activities at
the last stop.
We do not know what changes occur in the brain to transfer
mental objects from working memory to long-term memory. It
appears that the retention of mental objects in working memory is
caused by temporary chemical changes in neurotransmitters and
receptors. The initial chemical changes are later replaced by new,
semipermanent, or permanent neural connections, which then result
in long-term memory. What happens in between — that is, when the
medium-term buffer memory is in effect — is not known. However,
what is known is that much processing and integration of new knowl-
edge takes place during sleep.
Working memory has been demonstrated to be a “serial pro-
cessor,” in general capable of handling only one issue or stream of
conscience at the time. However, people often perform simultaneous
multitasking by pursuing several lines of thought nonconsciously at
any one time (as when we activate several of our nonconscious
working memories to perform different tasks). Working memory is
quite rapid, with access time and “object manipulation cycle” in 
the hundred millisecond range. We now consider this to be slow 
compared to modern computers; it allows us to make only a small
handful of reasoning steps every second. However, although the
capacity of working memory may only be of the order of five to nine
“chunks”
2
or mental objects at a time, a chunk may be quite
complex, ranging from a symbol, two to four digits in a group, an
Memory and Knowledge Categorizations
315
AP.qxd 5/3/04 2:29 PM Page 315

---

abstract concept, an image, or a phrase. Hence, since a vast amount
of understanding and meaning may be encapsulated in a chunk,
people can process tremendously complex reasonings in the blink of
an eye. In situations where the mental objects that are reasoned with
are complex abstract concepts, people adopt reasoning that is qual-
itative (“fuzzy,” “approximate,” or “inexact”) rather than using crisp
and precise logic and arithmetic, as when we add 2 
+
2 and find that
the result is 4.
When we speak deliberately about what we think or remember,
we use working memory to select what we want to say. The mater-
ial that we wish to communicate — facts, concepts, relations between
them, etc. — is recalled from long-term memories into working
memory, often facilitated by priming memory and primed by the con-
cepts present in working memory. We consciously weigh and select
what we want to say; we may even “think in words” as most, but
far from all, people do. Next, we encode what we have selected to
say into words and sentences and utter those as speech. Since working
memory is a serial processor, and since speaking is a “linear process”
that only allows presentation of one word and line of thought at a
time, this is a relatively slow process.
When we consciously pursue one line of thought and recall related
memory objects for processing the “next items,” we may recall many
objects, some of which have direct, while others have indirect, 
associations with the “thought” we want. As a result, our working
memory may be presented with many simultaneous memory objects
and become burdened by the need to process this wealth of facts,
perspectives, and concepts that often are at different levels of 
abstractions and, therefore, require considerable processing to be
compatible. The working memory, in effect, becomes a significant
bottleneck in our attempts to communicate what we know when we
are fortunate enough to recall the relevant memory objects. The first
time we have to communicate something that we know well, we may
find it difficult to select which concepts to present first — what we
want to say. If we talk repeatedly about the same issue, it becomes
easier for us to explain. We may even have remembered a sequence
of statements (a “party line”) that expresses what we wish to com-
municate. In the end, that tendency may lead to inflexible, rigid
explanations and positions that can be a liability when the world
around us changes.
When we know a subject extremely well, we normally have auto-
mated or compiled our knowledge. Therefore, it is no longer neces-
sary to access what we know in detail. In fact, the details may no
316
People-Focused Knowledge Management
AP.qxd 5/3/04 2:29 PM Page 316

---

Memory and Knowledge Categorizations
317
longer be available to working memory, making it impossible for us
to explain what we know. How this happens has not been 
established.
Conscious working memory is considered to be divided into three
specialized functions: (1) The central-executive function controls
what we think about. (2) The visuo-spatial scratch pad holds the
memory units (chunks and associations) while we work with them.
And (3) the articulatory loop (or “phonological” loop) is the mech-
anism we use to speak, write, and express other physical behavior,
including nonverbal gestures and facial expressions, to the external
world (Baddeley 1992a). The articulatory loop is coupled very 
closely to the motor system. We may have the capacity to operate
what appear to be several parallel working memories, which we 
use at different levels of consciousness to perform more or less 
automatic mental functions and tasks, some of which can be very
complex.