Trial and Error

CSGnet Archive CSG1998
1

[From Rick Marken (981201.1520)]

Me:

But I think learning can be done more efficiently than trial
and error; I think that's the hope, anyway, that underlies
the fact that we have an educational system.

Bruce Gregory (981201.1700 EDT)

It's not obvious to me that we learn _anything_ except by trial and
error (isn't it obvious that this is the way I'm learning PCT?) Gary
Cziko has argued this point and I happen to believe he is right.

Yes. I agree. All learning is trial and error. What I meant was
that I don't think all learning needs to be _random_ trial and
error.

The "trial" part of learning is selection of a output _function_;
a trial _relationship_ between error and output. The "error"
part of learning is (as you note) the change in system error
(control ability) resulting from this trial output function.

In _random_ trial and error a trial output function is selected
_randomly_. I think education is a _system_ for selecting trial
output functions; it's a way of helping a person select the
output functions that work. Education (I think) should be about
making learning a _non-random_ trial and error process; a more
efficient trial and error process becuase it increases the chances
that the correct output function will be selected by the learner.

Of course, as you note, you can't tell people what outputs
to produce, except as those outputs can be experienced as lower
level perceptions by the learner herself. So when you tell
a learner what _lower level perceptual variables_ to control
you are doing something like telling the person what output
function to use to control the higher level perception.

So there are two different ways to show a person what to perceive;
these correspond to two different levels of perception. Usually,
a person knows what they want to perceive at the higher level
(I knew I wanted to perceive myself riding a bike) they just
don't know what lower level perceptions to control (what output
functions to try) in order to produce those perceptions. The
teacher's job (I think) is to figure out what lower level
perceptions have to be controlled in order to control the higher
level perception; and then the teacher has to figrue out how
to communicate that information to the student.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

2

[From Chris Cherpas (981201.1830 PT)]

Rick Marken (981201.1520) --

Education (I think) should be about making learning a
_non-random_ trial and error process; a more efficient
trial and error process because it increases the chances
that the correct output function will be selected by the
learner.

I generally agree. While there's plenty written about
philosophy of education, I'd like to make a fuzzy distinction
here between education and training. In the training, you
are satisfied with raising the probability that the correct
output function is "selected" by the learner to the degree
that your teaching technology allows. In education, you
are also interested in the multi-year developmental process
whereby what the student learns raises the probability that
future learning is more probable. On the training end of
the continuum, the highest rate of getting from Y to Z is
the goal. At the education end, the process is optimized
by seeking higher accelerations: increasing the rates
of learning J to K, after learn A to B, and so forth.

An analogy to development within the womb is also appropriate,
I think. During that process, structures are built up
and essentially thrown away, as they serve as the
"scaffolding" for developing the complex structures that
we eventually see as an intact, functioning organism.
Likewise, in education, we may build up control systems
which are relatively generative of the systems we eventually
want to see, but, in the meantime, we are trivial, or even
antithetical to a fully operational adult repertoire. It
may be inevitable that along the way, we develop systems
which we know are going to run into conflict, that this
is the only way we can put together the pieces of a
structure that operates with beautiful synchrony ultimately,
but, in a linear developmental process, requires building
up and then tearing down what was often so painfully built.

Rick Marken (981201.1520) --

Usually, a person knows what they want to perceive at the
higher level (I knew I wanted to perceive myself riding a bike)
they just don't know what lower level perceptions to control
(what output functions to try) in order to produce those
perceptions.

In a training context, yes. In education, the higher level
perception is often sampled for the first time (who knew I
wanted to see myself thinking in terms of proportions?).

Rick Marken (981201.1520) --

The teacher's job (I think) is to figure out what lower level
perceptions have to be controlled in order to control the higher
level perception; and then the teacher has to figrue out how
to communicate that information to the student.

The curriculum developer's job is, in addition, to provide
processes that converge to acquiring new, higher-level
perceptions the student had barely a hint of previously.
At graduation, the student should ideally be able to
solve the problem of learning: how do I go about learning
something? Both "problem solving" and "learning to learn"
have been addressed in various ways by the educational
community, but I think these are often attempts to compensate
for poor curricula, which have not set up the processes for
teaching what should not require problem-solving or learning
to learn. The culture has some refined processes that really
do address these issues in the practices of research communities.

For the individual learner, the big breakthrough (if there is
to be such a thing) will come in the form of technology which
allows the learner to model his/her own repertoire, which monitors,
records, and analyzes the learner's history of learning, as
well as providing tools for communicating with other learners
and providing access to teachers and cultural artifacts that are
useful for learning. So far, we've made a lot of progress on the
last item, but not much else. What I have long advocated, but
done little to accomplish thusfar, is the development of what
I have called "user expert" systems -- systems whose job is
to help the user understanding him/herself. We currently
rely on other people for much of this, but I think automation can
be especially useful for disentangling others' self-interests from
what help they actually provide, teachers included. On the other
hand, we might better clarify what role others may have in helping
us learn that which is beyond what internal conflicts would prevent
us from even trying, no matter how much self-understanding we
may accumulate.

Best regards,
cc

3

[From Bruce Gregory 9981201.2053 EDT)]

Rick Marken (981201.1520)

So there are two different ways to show a person what to perceive;
these correspond to two different levels of perception. Usually,
a person knows what they want to perceive at the higher level
(I knew I wanted to perceive myself riding a bike) they just
don't know what lower level perceptions to control (what output
functions to try) in order to produce those perceptions. The
teacher's job (I think) is to figure out what lower level
perceptions have to be controlled in order to control the higher
level perception; and then the teacher has to figure out how
to communicate that information to the student.

Time to get _really_ nervous. I agree with this and everything else in your
post. This post and Bill's last post have been very helpful. I appreciate
your indulgence.

Bruce Gregory

4

[From Rick Marken (981202.0810)]

Chris Cherpas (981201.1830 PT)--

I'd like to make a fuzzy distinction here between education
and training.

I like the distinction and the way you make it. My thoughts about
"trial and error" learning were all in terms of your "training" (or,
equivalently, "coaching"). Maybe the distinction between training
and education can be made in terms of levels of control.

In training the student knows what she wants to do (the higher level
goal); she just doesn't know how (the functions that set the lower
level goals as the means of achieving the higher level goal).

In education, the student doesn't know the higher level goal; the
student can't even see the world in terms of these goals. The
student may know the lower level goals that _could_ be used to
achieve the higher level goal; the student just can't have the
higher level goal yet. So the educator has to teach the student
about the existence of new way to perceive the world. Very tough.

For example, the student may know how to control for rules: thou
shalt not kill, thou shalt not steal, etc. But she doesn't yet know
how to control in terms of principles. Education would be about
teaching the person that there are higher level goals -- in this
case principles -- that can be acheived by _varying_ the references
for lower order perceptions -- the rules. I think Jesus was trying
to teach precisely this: that the rules can be use to control for
higher level percepteions -- principles, like respect for other
humans ("love thy neighbor as thyself"). Obviously, education can
be a thankless task;-)

Best

Rick

···

---
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

5

[From Bruce Gregory (981202.1150 EDT)]

Rick Marken (981202.0810)

In training the student knows what she wants to do (the higher level
goal); she just doesn't know how (the functions that set the lower
level goals as the means of achieving the higher level goal).

In education, the student doesn't know the higher level goal; the
student can't even see the world in terms of these goals. The
student may know the lower level goals that _could_ be used to
achieve the higher level goal; the student just can't have the
higher level goal yet. So the educator has to teach the student
about the existence of new way to perceive the world. Very tough.

Worse than tough, impossible in my view. If students do not know what
the goal is there is absolutely no way that they can tell if what they
are doing is taking them closer to, or further, from the goal. It is
like searching for something in the dark and having no way to tell when
you have found it! If students can't see the immediate goal as a way of
satisfying their own higher level goals, they are just humoring the
teacher.

Bruce Gregory