Within this discussion I can find references to 12 nodes. I indicate primary
links, but there are others.
Environment-Do > -Be --------------------------------------- Input-Do > Output-Do > Outcome-Do -Be -Be -Be --------------------------------------- register\ indicator\ result\ ^ interpretation\ \---------------- action------------------------/
This model allows anybody to be right who proposes a positive mechanism
for learning. We can argue over which links are better, but affirming
one does not rule out another.
1. I agree with McMaster that a lot of Input-Do fails to produce anything
good, but it might.
2. I don't say that something is unmeasurable. If I can refer to it, I
can find a way to characterize it, hence I have measured it. The measure
may be unreliable, and not satisfy sticklers, but there it is. A lot of
life is run on these measures, and if we say they do not exist, then we
give up on improvement, and I would say learning.
3. If you only had two "learning detectors", where would you place them
in this system? Mulligan places a detector in Outcome-Do. McMaster
discounts Outcome-Do and seems to place his detectors at Input-Be and
Output-Be.
I propose placing one detector at interpretation and another at
Outcome-Do. That might be cheating since, in one sense, interpretation is
'outside the system'. But the point of LO is that we should be
reflective, and be aware of how we are reflecting. A lot of the
improvement in Be has to do with meta-system judgments. I define
learning, in this context, as taking action on the basis of whole
performance, and adjusting action based on changes in whole performance.
(For performance, I can select from any node, not just the Do nodes.)
I put my other detector at Outcome-Do, not because it has any overwhelming
significance, but because it summarizes the most and will tell me if I am
out of business. (I'd want detectors everywhere else. In particular, my
register/indicator/result links are highly vulnerable.)
I just read Thomas Landauer, The Trouble With Computers. He has a very
nice treatment of user-centered design. He divides his detections into
two modes, formative evaluation and summative evaluation. I prefer this
to the input vs. outcome terminology, because it focuses on what I am
doing with the information and less on unmotivated system description.
Kent Myers
Alexandria, VA (I'm not allowed an incoming email address where I am!
So much for professional development.)
--myersk@postoffice.sdf.sbis.com
Learning-org -- An Internet Dialog on Learning Organizations For info: <rkarash@karash.com> -or- <http://world.std.com/~lo/>