"Can Organizations Learn?" Depends on how we define learn. My short
definition of learning may have generated more questions than answers. I
suggested "learning as changing future behavior based on the response from
previous behavior".
Fred Nickols' comments on my definition suggest how easy it is to be
pulled into the trap of anthropomorphism (note Fred's use of pronouns):
"...It seems to me you're saying that learning is reflected in changed
behavior and that these changes in behavior occur as a result of our
experience. Thus, we act differently in the future because we learned
something in the past.
"It seems to me that we also learn in ways that confirm existing behavior
patterns. Thus, we act roughly the same way in similar situations in the
future because we have learned that doing so is effective, that is, it
produces the outcomes we desire.
"As you can no doubt deduce, I'm using "learn" in the sense of reflecting
on, analyzing, and becoming aware of relationships between ourselves and
our world, including other people and, especially, the relationships
between our actions and our goals. What I learn might or might not be
reflected in changed behavior. "
In addition to anthropomophism, by the third paragraph Fred provides a
different definition of learning than where I started. My definition
requires no reflecting, analyzing or awareness. My definition requires
changed behavior while Fred can learn without changing behavior. No
wonder we have different answers for the question "Can Organizations
Learn?"
If I was smarter I would have defined learning to include a feedback
process that enables systems to "modify their patterns of behavior on the
basis of past experience so as to achieve specific anti-entropic ends."
(see Norbert Wiener, "The Human Use of Human Beings: Cybernetics and
Society"). But I haven't thought through the problem of organizations that
learn harmful behaviors. That is, learnings which generate increasingly
entropic ends in the short run. (For example, organizations that become
addicted to downsizing.)
I. Market Mechanisms (Population and Selection Mechanisms, "Darwinian
Learning")
Wiener doesn't limit learning to only individuals. "Biological
reproduction is a phenomenon that occurs in the life of the race, in
phylogeny, but the race learns even as the individual does. Darwinian
natural selection is a kind of racial learning, which operates within the
conditions imposed by the reproduction of the individual." (From "God &
Golem, Inc.: A Comment on Certain Points where Cybernetics Impinges on
Religion" by Norbert Wiener)
To avoid risking more confusing anthropomophism, individual finches don't
"act differently in the future because" they "learned something in the
past". Finches don't "reflect" on their sexual prowness, they are not
aware of "goals", they are not too concerned about "anti-entropic ends".
Over time a "race of finches" learns by changing their behavior as they
adapt to environmental changes. What the incumbent individual finches
"know" may change over time (their specific behavior to specific
stimulus). But specific finches have not learned anything.
I proposed "Population and Selection" as one form of organizational
learning mechanism. Fred wrote: "I'll also acknowledge that intake,
selection and retention mechanisms are consciously, deliberately used to
maintain and/or change those populations."
Fred's answer suggests there has to be a purposeful "Watchmaker" to manage
the population and selection processes. I think these mechanisms can be
engaged without anyone "watching". And even when watched, the
organization's intake, selection and retention mechanisms can generate
both unintended and unknown learning consequenses.
I previously wrote: ".... In some firms... Successful business units grow
... while less successful business units may shrink,.... What the
organization has learned about business success is stored in the business
unit boundaries."
Fred writes: "Your last sentence immediately above is troubling. ... I
certainly see how some people, maybe even many people, might learn
something through their experiences of the kinds of events to which you
refer but I fail to see how the organization has learned anything."
Again, I define "learning as changing future behavior based on the
response from previous behavior". As a result of growth and success of
their mainframe business the IBM organization "learned" to focus on the
mainframe market. I assume the mainframe folks got the first pick of
talent, the most resources, and provided the largest talent pool to
executive offices. Business units with long, successful track records are
more likely to shape corporate decisions than untested startups.
Fred also writes: "...That the people who people the organization are new
and different people with new and different skill sets is pretty obvious
but it sounds a lot like you're saying that, at this level, organizational
learning and individual learning are indistinguishable. "
No. I'm just saying changing populations (finches or employees) can
produce organizational level learning, it does not mean any specific
individual member of either population (old or new) had to learn anything.
Fred also writes: "... are you saying that the "parent organization" is
some kind of collective intelligence, composed of the unit populations,
and sentient on its own? I suspect you are using "perceive" in a
figurative way owing to your use of quotation marks but I can't be
certain. "
Yes, I'm using "parent organization" as composed of all the unit
populations, including corporate staff organizations. Yes, the parent
organization is a "collective intelligence" in the sense it has learned
(used the results of past behavior to change future behavior). With 20-20
hindsight, I don't think "parent" was a good choice of words to represent
the collective organization. Yes, I'm using "perceive" in the sense of
the parent organization's reaction to various facets of the environment.
No, if "sentient" means conscious or self-aware, I am not talking about a
"sentient" organization. (I am reminded of the Danny Hillis quote,
something like "Someday I would like to build a computer that could be
proud of me" :-)
II. Clan Mechanisms (Neural Network Mechanisms and Skinnerian Learning)
Fred wrote: "... I read and re-read the neural network example and I
concluded that it doesn't give me enough information to understand what
you're talking about..."
I can't remember where I read about the student neural network. It may
have been in "Apprentices of Wonder: Inside the neural network revolution"
by William F. Allman. The book is 10 years old. I assume more current
material is available.
Hopefully we can agree neural networks have the capacity to learn (i.e.,
we are not distracted by vitalism). The example didn't provide details so
consider it a thought experiment. Imagine a simple neural network could
be built using some wires, batteries, meters and a gym full of students to
serve as synapses, with special banks of students to serve as input and
output receptors.
In the extreme example of the population and selection mechanism, the race
of finches learned without any specific finch learning anything. In this
extreme neural network example: 1) the organization could learn without
the students learning anything (their instructions and conditional
behavior didn't change), 2) no student knew what the organization knew, in
fact, 3) no one could know what the organization knew without observing
the organization's response to stimulus and 4) the process of testing
what the organization knew would most likely change the organization's
behavior.
Fred wrote: "...The example below seems to me to encompass some well-known
groupings such as "cliques," "old boys network," "the in crowd," and so
on. Clearly these kinds of subgroupings exist, and many more besides, but
I don't see their connection to organizational learning...."
At a minimum these informal networks shape information flows through the
organization. I suspect the denser the fabric of informal relationships
the greater the organization's capacity to adapt to environmental change.
III. Hierarchy Mechanisms (Symbols and Rules, "the Organization as a
Stored Program Machine")
Fred wrote:
"I don't think organizational analogs are difficult to find. How
about policies that require officers to step down at age 65? ...There
are all manner of policies, rules and regulations that shape an
organization's population and thus the capabilities it can be said to
posses. How about the requirement for an MBA for certain positions?"
Yes, there are a lot of hierarchy, symbols and rules in organizations.
But which ones are the key levers that shape the organization's capacity
to learn at the organizational level.
>From another string, Kim Ridgway wrote:
"I work for a large ($2.5 billion in sales) company that is very old (100
years) that is trying to change how it does business in the future. ...As
a company, we are successful, making money in the traditional lines and I
would suppose the majority of employees are quite content and happy with
the way things are - no urgency to change in their minds.
"Research, experience and my basic instincts (which I try not to overlook
anymore) tell me that a small team of people (maybe ten) working in some
"skunkworks" physically separate from the larger company could
successfully develop new business lines more efficiently and effectively.
.... Eventually the larger organization needs to acquire some of the same
characteristics of the new team. "
What I have been wrestling with is how to shape processes in both the
"skunkworks" and the larger organization so that the "parent" can
successfully learn from the "child."...."
Does anyone think Kim can find the answers she is looking for by focusing
on the learning of the individuals in the organization?
Doug Merchant
What Learning-Organization theory can Kim look to for help with her
question?
Doug Merchant
--"Doug Merchant" <dougm@eclipse.net>
Learning-org -- Hosted by Rick Karash <rkarash@karash.com> Public Dialog on Learning Organizations -- <http://www.learning-org.com>