Date: Wed, 4 Nov 1998 14:31:20 -0800
Subject: Re: PKF: Where is the discussion?


>     I have been on the list serve a week without  receiving any
>discussion.  Is this usual?  I want to get some  information
>about Feyerabend's idea of "proceeding counterinductively"  (AM) Barbara
>Kellam, RN, MS PhD student, Augusta, GA "...with trailing clouds of glory
>do we  come from God who is our home."" Wordsworth

Barbara -

The notion that science does NOT proceed by cumulative inductions was first
articulated by Popper.  PKF built on that.  (The rest of us spent decades
trying to explain why induction was justified --  assuming all along that
that was how science proceeded.)

Popper argues further that the most productive experiments (and associated
research programs) will be those that are the least likely to succeed as
reasoned from what we know at the moment.  Popper argued that we stand to
learn the most from the success of the least likely to succeed experiment.

What is meant by "productive"? (See below)

Reasoning (so to speak) from a classical inductive framework one would
expect that the best (most productive) experiments would be those most
likely to succeed as reasoned from what we know at the moment.  Herein you
grasp the contrast with Popper.

Counterinduction has a twist, not always appreciated.  The "counter" is
really qualitative and not just a simple negation (viz. a versus not-a).
The "counter" direction is more like a dialectic.  To generate the
dialectic counter research program you look at the deepest assumptions you
can articulate of the current, established and dominat research program and
reverse them (note: how ever it is we do this it is not formally
algorithmic).  Then look for confirmations of your newly defined "counter
research program".  (This is the best way to get a sense of what Popper
meant by the-least-likely-to-succeed-as-reasoned-from-current-understanding
experiment.)

So the tricky part is to understand that when you are reasoning about the
probability of one hypothesis versus another -- in this context -- that the
alternatives are really qualitatively distinct or incommensurable (i.e.
there is no algorithm within the established theory to lead us to the new
hypothesis of the counter theory).  So quantitative comparisons can be
misleading.  Popper was NOT clear about this.  His talk about the least and
most probable experiments was understood by many (most) to entail that
these alternatives existed within one logically homogeneous experimental
framework.  (And unfortunately, in the larger-than-science framework of
inquiry they probably do make sense in one context.  But not in any
formalizable positivistic sense.  Sorry if this is confusing.  Many ships
were lost in these waters.)
________________

What is meant by "productive" here is controversial.  For the positivists
the model of scientific inquiry was one of convergence -- the theories
converge to a correspondence with reality.  And that correspondence is what
they suggested as the meaning of "truth".  However, this model doesn't
correspond (sic; noting the self-reference; science looking at itself) with
the history of science.  Kuhn et al. pointed out that since the overall
Scientific Research Program develops qualitatively (viz. contains
revolutions), it is impossible, or illegitimate, or wrong to try to
represent it as one qualitatively homogeneous convergence.  (And we have no
independent access to "reality" to check our progress (viz. old Kantian
point).

For Popper and Feyerabend progress or a productive step in inquiry meant
that you learned something that you didn't know (or understand) before.
Popper/Feyerabend came to understand that this (rather obvious point
(obvious after you think about it)) meant that the later theory must be at
least partially incommensurable with the earlier; some qualitative
differences.  This means that the later theory can't just be a logical
extension of the earlier.  The "cumulative inductions" model of science
entails a homogeneity of the accumulated inductions that is equivalent to a
conceptual homogeneity (note: otherwise they can't be compared,
mathematically).

Popper and Feyerabend argued that if we really learn, then the conceptual
system must "develop" such that conceptual homogeneity is violated:  to
learn is to make a conceptual advance.  This was frequently argued in terms
of the notion of "meaning variance" (= conceptual inhomogeneity).

John Dewey and C.S. Peirce, it seems to me, made this point a generation
earlier.  Look particularly at Dewey's essays in "Experimental Logic".  I
like their aphorism:  an advance in science is an advance in logic.  (Also
this means that an advance in science is an advance in scientific method -
in the logic of reasoning about the next potential experiment;  so the
perceived problem of science here develops, emerges.  --  But not in the
positivist convergence model.)

This is also quite in tune with the musings about quantum logic where the
logic of the theoretical framework at any moment is linked to the nature of
the elements (viz. objects of the concepts; particles, forces etc.) at that
time in the history of the theory.  So as our conceptual understanding of
the electro-weak force changes so does the logic of the theory of science.

If these last remark are correct then there is no "time-space invariant
logic of science", no universal scientific method.  If the "cumulative
inductions" model of science were correct then there would be a
"conceptually independent, universal, t/s-invariant logic of science."
This is where Feyerabend and Lakatos realized that they were on the same
path with Thomas Kuhn.

ONE of the reasons this line of thinking has had political problems is that
it entails that science is not and never has been an autonomous method of
inquiry.  It is simply part of the larger human enterprise.  --  If you
call yourself a scientist and that is how you get your money in life, then
this is not a warm thought.

WHERE does all this lead????  I think that it is fair to say that Popper
"backed" his way into his eventual position.  He argued in a variety of
ways that the positivist model was wrong.  Popper in this sense is himself
an example of the "counterinductive inquirer".  If you look at the BIG
PICTURE of Popper's intellectual development you see that the positivists
are committed to a "Closed Universe" model.  Popper gradually discovered
that if he were to be systematic about his opposition to the positivist
model that he must, in effect, propose an "Open Universe" model.  Briefly,
a closed universe is qualitatively closed; nothing new really comes into it
(sorry no revolutions here; no incommensurability); just algorithmic
transformations amounting in the final analysis to nothing (absolute
symmetry).  (Note:  John Barrow (U Sussex is currently writing a book about
this aspect, entitled (guess) "Nothing";  cf. his current "Impossibility".
An open universe allows for -- indeed requires -- qualitative change.

MY best guess is that you can't make consistent sense of either of these
models without the other.  They are complementary, suggesting a quantum
metaphysics.

Terry Bristol


**********************************************************************
Contributions: mailto:feyerabend-AT-lists.village.virginia.edu
Commands: mailto:majordomo-AT-lists.village.virginia.edu
Requests: mailto:feyerabend-approval-AT-lists.village.virginia.edu