Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-l] "Unlearning"




It seems to me that this is not evidence of unlearning, here is an
alternative:

On Monday students have stored the memory that the instructors shirt
was blue. On Tuesday they store the image/memory of the shirt
being red. On Wed both are in storage and what they "recall"
depends on which memory they access.

I think that there is a lot of experimental evidence that students
(and faculty) who have learned Newtonian physics will access their
still existing knowledge of Aristotelian/medieval physics if the
situation is presented in a way to elicit that. (For faculty it
requires some time pressure.)

It seems to me very important whether we model learning a new
alternative as involving erasing the old beliers/concepts or instead
as creating a new one which exists alongside the old one.

If you notice I pointed out that imbedded paradigms are more resistant to
unlearning, which means they may not be totally erased. But individual
components of pre-existing ideas may be unlearned. The published ILDs
actually are designed so that students unlearn existing ideas which are
replaced by more consistent physics concepts.

As to faculty members showing that their original conceptions were still
there, they generally were taught by conventional methods. Conventional
methods layer the new ideas over the old, and didn't use techniques which
promote unlearning. To unlearn you have to bring up the old idea and while
it is in short term memory, modify it so that during the reconsolidation
cycle it is effectively unlearned.

Of course you can not unlearn all facets of a complex subject within a very
limited time frame.

There are other experiments which shown that it is possible to unlearn
responses to stimuli. This can be used in treating phobias.

We do not know how well the PER techniques have promoted unlearning, rather
than just layering. To do this experiment we need to have a sample of very
advanced students, or faculty who have been taught primarily by PER and
another set that has been taught primarily by conventional methods. Then a
stress test might reveal how much previous styles of thinking still exist.
But at present this is impossible because the number of students who
experience PER in most classes is probably close to zero.

So the example of faculty members shows that the current methods of teaching
do not promote unlearning. But this does not disprove the idea that there
can be unlearning. In the case of physics this would be to make the use of
the correct concepts instinctual, but with an awareness of the non utility
of the medieval concepts. In other words the old concepts often exist
compartmentalized in memory. But if they are decompartmentalized and the
person has a deep understanding of how they relate to the physics concepts,
unlearning may have happened. One does not have to forget these ideas, but
rather break the automatic connections to them.

The brain is amazingly plastic, and the storage is actually quite limited so
I suspect that unlearning goes on all the time. The simple experiment of
the red/blue shirt is very interesting because the students who remembered
the shirt was red, were apparently very adamant that they had actually seen
it as being red on Monday.

Unfortunately unlearning usually goes in the opposite direction. Students
shown a demo can often tell you about it shortly thereafter. But if it
conflicts with their paradigms, they will remember it much later as being
the opposite of what it showed. But by having prediction, and then
reasoning about the result, this doesn't happen. The UTube Mazur video
talks about one experiment where Mazur showed this.

John M. Clement
Houston, TX