Darwin-L Message Log 26: 94–122 — October 1995

Academic Discussion on the History and Theory of the Historical Sciences

Darwin-L was an international discussion group on the history and theory of the historical sciences, active from 1993–1997. Darwin-L was established to promote the reintegration of a range of fields all of which are concerned with reconstructing the past from evidence in the present, and to encourage communication among scholars, scientists, and researchers in these fields. The group had more than 600 members from 35 countries, and produced a consistently high level of discussion over its several years of operation. Darwin-L was not restricted to evolutionary biology nor to the work of Charles Darwin, but instead addressed the entire range of historical sciences from an explicitly comparative perspective, including evolutionary biology, historical linguistics, textual transmission and stemmatics, historical geology, systematics and phylogeny, archeology, paleontology, cosmology, historical geography, historical anthropology, and related “palaetiological” fields.

This log contains public messages posted to the Darwin-L discussion group during October 1995. It has been lightly edited for format: message numbers have been added for ease of reference, message headers have been trimmed, some irregular lines have been reformatted, and error messages and personal messages accidentally posted to the group as a whole have been deleted. No genuine editorial changes have been made to the content of any of the posts. This log is provided for personal reference and research purposes only, and none of the material contained herein should be published or quoted without the permission of the original poster.

The master copy of this log is maintained in the Darwin-L Archives (rjohara.net/darwin) by Dr. Robert J. O’Hara. The Darwin-L Archives also contain additional information about the Darwin-L discussion group, the complete Today in the Historical Sciences calendar for every month of the year, a collection of recommended readings on the historical sciences, and an account of William Whewell’s concept of “palaetiology.”

-----------------------------------------------
DARWIN-L MESSAGE LOG 26: 94-122 -- OCTOBER 1995
-----------------------------------------------

DARWIN-L
A Network Discussion Group on the
History and Theory of the Historical Sciences

Darwin-L@ukanaix.cc.ukans.edu is an international network discussion group on
the history and theory of the historical sciences.  Darwin-L was established
in September 1993 to promote the reintegration of a range of fields all of
which are concerned with reconstructing the past from evidence in the present,
and to encourage communication among academic professionals in these fields.
Darwin-L is not restricted to evolutionary biology nor to the work of Charles
Darwin but instead addresses the entire range of historical sciences from an
interdisciplinary perspective, including evolutionary biology, historical
linguistics, textual transmission and stemmatics, historical geology,
systematics and phylogeny, archeology, paleontology, cosmology, historical
anthropology, historical geography, and related "palaetiological" fields.

This log contains public messages posted to Darwin-L during October 1995.
It has been lightly edited for format: message numbers have been added for ease
of reference, message headers have been trimmed, some irregular lines have been
reformatted, and some administrative messages and personal messages posted to
the group as a whole have been deleted.  No genuine editorial changes have been
made to the content of any of the posts.  This log is provided for personal
reference and research purposes only, and none of the material contained herein
should be published or quoted without the permission of the original poster.
The master copy of this log is maintained in the archives of Darwin-L by
listserv@ukanaix.cc.ukans.edu, and is also available on the Darwin-L Web Server
at http://rjohara.uncg.edu.  For instructions on how to retrieve copies of this
and other log files, and for additional information about Darwin-L, send the
e-mail message INFO DARWIN-L to listserv@ukanaix.cc.ukans.edu, or connect to
the Darwin-L Web Server.

Darwin-L is administered by Robert J. O'Hara (darwin@iris.uncg.edu), Center for
Critical Inquiry in the Liberal Arts and Department of Biology, University of
North Carolina at Greensboro, Greensboro, North Carolina 27412 U.S.A., and it
is supported by the Center for Critical Inquiry, University of North Carolina
at Greensboro, and the Department of History and the Academic Computing Center,
University of Kansas.

_______________________________________________________________________________

<26:94>From wmontgom@nasc.mass.edu Tue Oct 24 13:56:14 1995

Date: Tue, 24 Oct 1995 12:10:21 -0400 (EDT)
From: William Montgomery <wmontgom@nasc.mass.edu>
To: darwin-l@ukanaix.cc.ukans.edu
Subject: Re: Selection, "Induction," Language

Gary Cziko quotes Bartley as saying "Every animal is born with
expectations--that is with something closely parallel to hypotheses,
which, if verbalized, expresses hypotheses or theories."  This sounds
important.  After all, there can be no hypothesis testing without
hypotheses.  What do animal behavior studies show about these
"expectations?"  After all, it is theoretically possible that animals
simply engage in random behavior and learn to repeat activities that
prove inductively rewarding.  As a historian I have no knowledge of these
matters, and I'd be interested in what the specialists have discovered?

_______________________________________________________________________________

<26:95>From ronald@hawaii.edu Tue Oct 24 13:56:22 1995

Date: 	Tue, 24 Oct 1995 08:35:51 -1000
From: Ron Amundson <ronald@hawaii.edu>
To: Darwin-L List <darwin-l@ukanaix.cc.ukans.edu>
Subject: Wallace on norm of reaction

Venerated List:

Discussion of norms of reaction is much more common among developmentally
oriented biologists than among mainstream adaptationists.  But I found a
short discussion of norms of reaction in a Bruce Wallace paper (which was
intended to dismiss the relevance of embryology to evolution).  In it
Wallace says:

"The reaction of a genetic program to each of many environments
constitutes the 'norm of reaction' of that genetic program; as far as
possible, success in the past guarantees that each point on the norm of
reaction is an adaptive reaction."   (p. 160)

The context is not in itself a dismissal of development, but rather an
explanation of how "genes and environment interact".  And the expression
"as far as possible" is exactly the kind of hedge that makes it difficult
to pin down what's being claimed.  But that kind of adaptationist take on
the concept of norm of reaction seems odd (and extreme) to me.

Does anyone recognize that account of norm of reaction?  Is it common?
Taken to an extreme it would seem to suggest that organisms are ("as far
as possible") adapted to _all environments_, not just the environments in
which they have been selectively successful.  I realize that a viable
organism had better be buffered against _some_ environmental variation.
But how can we conceive of _each point_ on the norm being an adaptive
reaction (even "as far as possible" whatever that means).  Norms of
reaction are the results of the contingencies of development, and those
contingencies (even adaptationists admit) cannot be perfectly tuned to
all possible environments.  (Insert standard line on the larynx nerve in
giraffes, etc.)

Is Wallace just choosing to _concentrate_ on the adaptive successes
rather than non-adaptive byproducts, or is there some serious line of
thought I'm unaware of that implies a perfectly-tuned norm of reaction?

Regards,

Ron

==Ahouse-style Bibliography==

Wallace, Bruce (1986) "Can embryologists contribute to an understanding
of evolutionary mechanisms?" in _Integrating Scientific Disciplines_, pp.
149-163, W. Bechtel, ed., Martius Nijhoff Pub., Dordrecht.

__
Ron Amundson
University of Hawaii at Hilo
ronald@Hawaii.Edu

_______________________________________________________________________________

<26:96>From gerson@hooked.net Tue Oct 24 14:21:13 1995

Date: Tue, 24 Oct 1995 08:21:47 -0700
To: darwin-l@ukanaix.cc.ukans.edu
From: gerson@hooked.net (Elihu M. Gerson)
Subject: Re: Learning as Variation and Selection

>[from Gary Cziko]
>
>Elihu M. Gerson writes:
>
>>I think it's past time we had some specifics on claims that cultural
>>(and/or institutional) change can be described as an evolutionary process
>>analogous to Darwinian evolution. So far, we don't even have a plausible
>>sample just-so story. In particular, we need answers to these questions
>>just to get started:
>>
>>1) What varies, and how does it do so?
>>2) What reproduces, and how does it do so?
>>3) What selects, and how does it do so?
>>4) What connections, if any, are there among that-which-varies,
>>that-which-reproduces, and that-which-selects?
>>5) What is adapted to what, and how is this measured?
>
>These are good questions.  But, I would like to see two preliminaries
>settled before we start discussing possible answers:
>
>1. Do you now agree that classical genetic algorithms rely on BLIND
>variation and selective reproduction?  As I mentioned earlier, if we can't
>come to agreement on this rather straightforward (in which the answers to
>your five questions obvious), there is no way that we will ever make
>progress toward consensus on the above issues concerning cultural (and/or
>institutional) learning.

Yes. I now agree that classical genetic algorithms rely on bind
variation and selective reproduction (hereafter: BVSR). But many other
computer-based aproaches make use of explicit goals, an/or stored and
accessible informatin about past performance, which makes their procedures
something other than blind variation. I mentioned the work of Simon and his
colleagues a couple of days ago; the journal _Machine Learning_ and the
distributed aritifical intelligence literature have many other examples. I
think that my original point still stands:  BVSR is not the only possible
way to develop adaptive behavior without postulating miracles, even in the
narrow limited sense that abstract formalisms provide. Nor does it play a
significant role in human learning above the level of simple movements, nor
(my principal concern here) in cultural, institutional, and organizational
change.

>2. Do you believe that some aspects (not necessarily all) aspects of
>culture (and/or institutions) are adapted to their environments, that
>adaptive change has taken place (not that all change is necessarily
>adaptive)?  If you don't there is no reason to invoke selectionist
>explanations since selection theory is intended to explain only adaptive
>change.  If you believe such adaptive change does take place, give us a
>clear example so that we can begin to explore the answers to 1 through 5 in
>a particular context where we agree adaptive change has taken place.

Of course some institutions and organizations are adapted to one another
and their environments-- not perfectly, and not always, but often to a
considerable degree.

But I don't understand the insistence on the exclusive efficacy of BVSR.
Why even make the claim? What's gained by eliminating non-BVSR approaches a
priori, before the serious work of evaluating the possibility has been
done? I understand why evolutionary biologists like BVSR models-- non-blind
variation raises issues of plan and purpose in nature. But why is this a
problem in talking about human affairs?

In any case, there are (still) major difficulties with Cziko's use of BVSR
as an explnation. Consider his example of finding the light switch in a
dark room, which he quoted a few days ago. This involves knowing about
rooms in general and the likely location of switches; continuing monitoring
of one's own progress in finding the switch. The behavior thus generated is
not random; once having attained the wall, one doesn't leave it; once
having gotten to the door edge, one keeps track of it. Behavior is hsaped
in terms of what is likely to work; just as airplane designers put both
wings on one side of the airplane.

Which brings me to a class of fallacies which has come up several times.
The history of inventions typically shows a series of efforts in which
several alternative arrangements were tried before one was chosen as
satisfactory. Indeed, the history of most technologies seems to consist of
cascades of such alternatives. Such histories are not arguments for BVSR;
often, in fact, they provide the strongest evidence *against* BVSR. The
fact that multiple trials take place does not mean that the trials take
place blindly. To the contrary, the results of trials are typically used
(both in invention and in scientific discovery) to refine the concepts and
models governing the trial-generating procedure. In basic research, in
fact, the trials have no other point at all. And invariably, the results of
trials are explicitly considered in the process which generates the next
round of trials. Once again: this is *not* blind variation. To the
contrary: it is controlled, organized, and systematic variation. The fact
that we don't know where the light switch is doesn't mean that we can't
find it in an orderly way.

So now it's time to have a test of the hypothesis, and see some concrete
examples of BVSR in organizational, institutional, or cultural change. I
think scientific discovery was mentioned-- perhaps we might start with a
specific case.

Elihu M. Gerson
Tremont Research Institute
458 29 Street
San Francisco, CA 94131
Phone: 415-285-7837   Fax: 415-648-7660  gerson@hooked.net

_______________________________________________________________________________

<26:97>From CSM@macc.wisc.edu Tue Oct 24 15:48:50 1995

Date: Tue, 24 Oct 95 09:17 CDT
From: Craig McConnell <CSM@macc.wisc.edu>
Subject: Einstein and Induction
To: DARWIN-L@ukanaix.cc.ukans.edu

I'm responding to the following [excerpted heavily from Gary Cziko's
"Induction and Selection Theory"]:

>                . . . theories as the creative invention of scientists.
>Theories are not "caused" by the environment by repeated observations and a
>process of induction.  Theory precedes observation, in the same way that
>variation precedes selection in biological evolution.
>Who ever saw a object in motion, never slowing down or stopping in Newton's
>time?  Who ever saw light bent by gravity before Einstein came up with his
>outlandish theory.  How could induction (repeated observations of a certain
>phenomenon) ever explain the origin of these theories?
>Einstein himself said that theory is not based on observation.  I will try
>to find this quote, unless someone else has on hand.

I certainly agree that theories are the creative inventions of scientists, and
that new theories allow us to see new unexpected phenomena.  But I take issue
with the notion that "theory is not based on observation"--Einstein himself
knew and often discussed the role of experience in theory formation.  Einstein
had no "theory of bending light"--it's a theory of relativity, and it suggests
that light is bent, and subsequent observation confirmed this, but the theory
of relativity was devised in response to a vast body of experience--
observations, no?--regarding light and electromagnetic phenomena and
gravitation.  (In his 1916 _Relativity--The Special and the General Theory_, he
says "every child at school knows . .  that [the propagation of light] takes
place in straight lines with a velocity c=300,000 km/sec".  It is this
observation (and its many counter-intuitive implications) that sparked the
creative effort that became Special Relativity.  His other work of 1905--
the photoelectric effect, Brownian motion, etc.--is even more obviously the
result of contemplating observations of natural phenomena.

To say that "theory is not based on observation" totally confounds me.  Is
there some distinction here between "experience" (observation before theory
formation) and "observation" (observation undertaken to test a theory)?
Einstein did maintain that physics was a creative endeavor undertaken by
people, but he also believed that there was only one reality out there that
we are trying to explain (thus nature doesn't hand us theories, but it does
make it possible for us to decide between theories).  To build a theory
independant of observation would be to write mathematics, not physics.

Craig S. McConnell, (608) 238-1352
Internet:  csm@macc.wisc.edu

_______________________________________________________________________________

<26:98>From staddon@psych.duke.edu Tue Oct 24 18:21:52 1995

Date: Tue, 24 Oct 95 19:21:41 EDT
From: staddon@psych.duke.edu (John Staddon)
To: darwin-l@ukanaix.cc.ukans.edu
Subject: Re: Selection, "Induction," Language

Gary Cziko quotes Bartley as saying "Every animal is born with
expectations--that is with something closely parallel to hypotheses,
which, if verbalized, expresses hypotheses or theories."  This sounds
important.  After all, there can be no hypothesis testing without
hypotheses.  What do animal behavior studies show about these
"expectations?"  After all, it is theoretically possible that animals
simply engage in random behavior and learn to repeat activities that
prove inductively rewarding.  As a historian I have no knowledge of these
matters, and I'd be interested in what the specialists have discovered?
------------------

If by "expectations" you mean (a) more than raw physical movements, and (b)
processes that are not entirely random, then G. Cziko is quite correct.  I
speculated about the parallel between learning and evolution by natural
selection in an old article:

_______________________________________________________________________________

<26:99>From bill@clyde.as.utexas.edu Tue Oct 24 23:17:24 1995

Date: Tue, 24 Oct 1995 23:20:51 -0500
From: bill@clyde.as.utexas.edu (William H. Jefferys)
To: DARWIN-L@ukanaix.cc.ukans.edu
Subject: Re:  Einstein and Induction

#I certainly agree that theories are the creative inventions of scientists, and
#that new theories allow us to see new unexpected phenomena.  But I take issue
#with the notion that "theory is not based on observation"--Einstein himself
#knew and often discussed the role of experience in theory formation.  Einstein
#had no "theory of bending light"--it's a theory of relativity, and it suggests
#that light is bent, and subsequent observation confirmed this, but the theory
#of relativity was devised in response to a vast body of experience--
#observations, no?--regarding light and electromagnetic phenomena and
#gravitation.  (In his 1916 _Relativity-The Special and the General Theory_, he
#says "every child at school knows . .  that [the propagation of light] takes
#place in straight lines with a velocity c=300,000 km/sec".  It is this
#observation (and its many counter-intuitive implications) that sparked the
#creative effort that became Special Relativity.  His other work of 1905--
#the photoelectric effect, Brownian motion, etc.--is even more obviously the
#result of contemplating observations of natural phenomena.

Just to bring some historical perspective into this
discussion, Einstein was not the first to suggest
"bending of light" as a result of physical law, nor
was he even the first to suggest "black holes." Both
are consequences of Newtonian gravitational theory
(although it was Laplace who apparently first recognized
that it implied black holes, from which light would not
escape.) The difference between Newtonian and Einsteinian
gravitaional theory is not that one predicts no effect
and that the other predicts an effect, but rather that
they predict effects of _different magnitudes_ in
these cases.

Bill

_______________________________________________________________________________

<26:100>From jim@ling.ed.ac.uk Wed Oct 25 04:51:13 1995

From: Jim Hurford <jim@ling.ed.ac.uk>
Date: Wed, 25 Oct 95 09:47:18 BST
To: Darwin-L@ukanaix.cc.ukans.edu, SOCIOBIO@sjuvm.stjohns.edu,
        corpora@nora.hd.uib.no, corpora@x400.hd.uib.no, elsnet@ed.ac.uk,
        funknet@oregon.uoregon.edu, info-childes+@andrew.cmu.edu,
        info-psyling+@andrew.cmu.edu, linguist@tamvm1.tamu.edu
Subject: Chair of Applied Linguistics

                  The University of Edinburgh

                  Chair of Applied Linguistics

Applications are invited for the newly created established Chair of
Applied Linguistics, to which it is intended to make an appointment from
1 October 1996.

The University seeks to appoint a scholar with broad-ranging interests,
sympathies and experience across the spectrum of the discipline who can
offer strong leadership in research and teaching well into the next
century.  The successful candidate will have an active and
internationally recognised record of published research as well as
substantial experience of effective teaching and supervision at
postgraduate level.  It will be an additional advantage if candidates
have extensive collaborative contacts within and outside the academic
world.

Salary will be within the professorial range.

Please quote reference 590481

Further particulars including details of the application procedure may
be obtained from

	The Personnel Office
	University of Edinburgh
	1 Roxburgh Street
	Edinburgh  EH8 9TB

	Telephone: 	+44 131 650 2511  (24-hour answering service)
	Fax: 		+44 131 650 6509

Closing date: 1 December 1995.

_______________________________________________________________________________

<26:101>From g-cziko@uiuc.edu Wed Oct 25 10:53:11 1995

Date: Wed, 25 Oct 1995 10:55:10 +0000
To: darwin-l@ukanaix.cc.ukans.edu
From: g-cziko@uiuc.edu (CZIKO Gary)
Subject: Re: Learning as Variation and Selection: Finding the Light Switch

[from Gary Cziko responding to Elihu Gerson]

>Yes. I now agree that classical genetic algorithms rely on bind
>variation and selective reproduction (hereafter: BVSR).

Good, although that should be "blind."  And Donald Campbell uses BVSR to
refer to "blind variation and selective RETENTION," but your phrase may
actually be a bit better.  But we cannot ignore the CUMULATIVE aspect of
this process (one step BVSR doesn't get you very far).  So I usually call
it cumulative BVSR.  Your change of belief with respect to GAs suggests
that there may be other adaptive processes which you do not now believe
involve cumulative BVSR, but for which you may change your mind as a result
of our discussion.  So I am encouraged that we can make some progress.  I
will also try to keep an open mind (and offer below evidence that I changed
my mind about neural networks as a result of a reviewer's comments on
_Without Miracles_)

>But many other
>computer-based aproaches make use of explicit goals, an/or stored and
>accessible informatin about past performance, which makes their procedures
>something other than blind variation.

Yes, I agree.  I was pushed off my radical blind variation perspective in
this repect by a reviewer of _Without Miracles_, with specific reference to
backpropagation neural networks.  I have appended a relevant extract to the
end of this message.

>BVSR is not the only possible
>way to develop adaptive behavior without postulating miracles, even in the
>narrow limited sense that abstract formalisms provide.

I suppose it depends on what you mean by "adaptive behavior."  The cruise
control system in my car shows "adaptive behavior" in that it knows how to
maintain a constant speed in spite of varying and unpredictable
disturbances.  In much the same way, a pidgeon will increase its rate of
pecking to obtain food if the ratio of reward to pecks is decreased by the
experimenter.  So if this is what you mean by adaptive behavior, I agree
with you.  But I am interested in the emergence of truly _new_ adapted
complexity, such as the design of the cruise control system in the first
place (or the design of the pidgeon).  For truly new knowldge (novel
adapted complexity) to arise, I don't yet see how any process other than
cumulative BVSR can work.

>Nor does it play a
>significant role in human learning above the level of simple movements, nor
>(my principal concern here) in cultural, institutional, and organizational
>change.

I disagree--the truly simple stuff may not require cumulative BVSR at all
as it is based on already achieved knowledge.  It is the emergence of new
adapted complexity which requires (it seems to me) cumulative BVSR.

But before tackling these tough issues, let's go back to the light switch
example to see if we can come to some agreement with this simpler instance.

>Consider his example of finding the light switch in a
>dark room, which he quoted a few days ago. This involves knowing about
>rooms in general and the likely location of switches; continuing monitoring
>of one's own progress in finding the switch.

Absolutely right.  But where did these constraints on variations (not
looking on the floor for the switch, for example) come from?  If these
constraints are based on previous knowledge, you have to explain where that
knowledge came from.

>The behavior thus generated is
>not random; once having attained the wall, one doesn't leave it; once
>having gotten to the door edge, one keeps track of it.

Yes, once you have reached the wall, you stay there and continue your
groping on the wall.  But now you still need to find where on the wall the
switch is located.  Any further advance in knowledge must come from taking
a chance, from a blind groping now in two-dimensions rather than in three.
And once you find the molding (along which you believe the switch is
located), you can now restrict your groping to a single dimension.  But in
each case further advances in knowledge come from blind variations.  That's
the price you have to pay for new knowledge.

And of course, your contraints can also _prevent_ you from finding the
switch.  If you are convinced that the switch is on the wall, you will
grope forever if there is a light hanging from the ceiling with a
pull-chain switch.  All constraints themselves are fallible.  And when they
do fail, you have to back up and try a new direction and another set of
blind variations.

>Behavior is [shaped]
>in terms of what is likely to work; just as airplane designers put both
>wings on one side of the airplane.

I hope you meant "put one wing on each side of the airplane."  Of course we
use what we believe is likely to work.  But how do we know what is likely
to work in the first place?  Where did this knowledge come from?  We are
born with eyes, lungs, legs, ears, etc. that are all likely to work.  But
what is the source of this adapted complexity?  Cumulative BVSR.  And for a
system to be truly adaptive, such as a brain or immune system, it continues
to rely on cumulative BVSR.

>Which brings me to a class of fallacies which has come up several times.
>The history of inventions typically shows a series of efforts in which
>several alternative arrangements were tried before one was chosen as
>satisfactory. Indeed, the history of most technologies seems to consist of
>cascades of such alternatives. Such histories are not arguments for BVSR;
>often, in fact, they provide the strongest evidence *against* BVSR. The
>fact that multiple trials take place does not mean that the trials take
>place blindly. To the contrary, the results of trials are typically used
>(both in invention and in scientific discovery) to refine the concepts and
>models governing the trial-generating procedure. In basic research, in
>fact, the trials have no other point at all. And invariably, the results of
>trials are explicitly considered in the process which generates the next
>round of trials. Once again: this is *not* blind variation. To the
>contrary: it is controlled, organized, and systematic variation. The fact
>that we don't know where the light switch is doesn't mean that we can't
>find it in an orderly way.

But isn't what you are saying here again is that previous trials provide
information on how to constrain the next ones?  In that case, my above
arguments about finding the light switch hold here as well.  Further
_advances_ in knowledge come from trying out blind (if very highly
constrained) variations.  If you already know for sure that something is
going to work, why try it at all?  Doing so, you would learn nothing at
all.  Ones obtain new knowledge by going beyond what one already knows.  In
this sense the variations are blind.  And more than a tautology is involved
here.  We can find evidence for such cumulative BVSR processes occurring
within organisms, as has been found for the mammalian immune system and the
brain.

>So now it's time to have a test of the hypothesis, and see some concrete
>examples of BVSR in organizational, institutional, or cultural change. I
>think scientific discovery was mentioned-- perhaps we might start with a
>specific case.

No, I don't think we are ready for this yet until we "find the light
switch."  We must find agreement on the simpler cases (as we did on genetic
algorithms) before we can hope to make any progress on the much more
difficult ones.

-Gary Cziko

====================================================================
[from Chapter 13 of _Without Miracles_]

It is of particular interest to note that no trial-and-error and no blind
variation and selection are involved in the actual training and functioning
of a backpropagation neural network. Instead, during training the
difference between the network's actual output and the desired output is
used to modify synaptic weights in a deterministic manner calculated to
reduce error the next time that input pattern is encountered. This is the
first clear example we have seen in this book of adaptive change resulting
from an instructionist process-the instruction provided by that part of the
computer program that trains the network.

But this instructionist approach to training a neural network has certain
noteworthy limitations. First, the designer must initially decide on the
number and organization of the neurodes in the network. Particularly
important is the number of middle-level neurodes-too few, and the network
will not be able to distinguish between subtle but important differences in
input patterns; too many, and the system may not be able to generalize what
it has learned to new input patterns on which it was not trained. And since
there is no way to know beforehand the optimum number of such units, this
initial design is actually the result of trial and error on the part of the
network developer.

Second, the correct corresponding output, that is, the proper
classification for each encountered input pattern, _must already be known_
to train the network. A sonar-detecting backpropagation network cannot be
used to distinguish rocks from mines if it is not already known which
signals indicate rocks and which indicate mines. In other words, the
instructionist learning procedure used by backpropagation neural networks
can work only if the instructor already knows all the right answers for the
training inputs. So this process cannot be used to discover new knowledge
or develop new skills, but is a way of transferring knowledge from one
source to another.

Third, backpropagation neural networks cannot always be trusted to find
just the right combination of connection weights to make correct
classifications because of the possibility of being trapped in local
minima. The instruction procedure is one in which error is continually
reduced, like walking down a hill, but there is no guarantee that
continuous walking downhill will take you to the bottom of the hill.
Instead, you may find yourself in a valley or small depression part way
down the hill that requires you to climb up again before you can complete
your descent. Similarly, if a backpropagation neural network finds itself
in a local minimum of error, it may be unable to climb out and find an
adequate solution that minimizes the errors of its classifications. In such
cases, one may have to start over again with a new set of initial random
connection weights, making this a selectionist procedure of blind variation
and selection. One might also use other procedures such as adding
"momentum" to the learning procedure, analogous to the way a skier can ski
uphill for a while after having attained enough speed first going downhill,
or adding noise to the procedure to escape such traps, again a form of
variation and selection.

Finally, for those looking to research on neural networks to further our
understanding of how real neurons work in real brains, backpropagation
neural networks are not biologically plausible. Unlike the neurodes and
their interconnections in a backpropagation network that conduct signals in
both directions (one way for responding, the other direction for learning),
biological neural pathways conduct signals in one direction only. Also,
much if not all human learning occurs without the direct instruction and
the patient training of the type backpropagation networks require. So
although these networks can be useful in quite a number of interesting
applications, their instructionist nature does impose limitations on both
their adaptive flexibility and their applicability to biological neural
systems.

_______________________________________________________________________________

<26:102>From gerson@hooked.net Wed Oct 25 17:53:59 1995

Date: Wed, 25 Oct 1995 15:54:07 -0700
To: darwin-l@ukanaix.cc.ukans.edu
From: gerson@hooked.net (Elihu M. Gerson)
Subject: Re: Learning as Variation and Selection: Finding the Light Switch

>[from Gary Cziko responding to Elihu Gerson]
>
>But before tackling these tough issues, let's go back to the light switch
>example to see if we can come to some agreement with this simpler instance.
>
>>Consider his example of finding the light switch in a
>>dark room, which he quoted a few days ago. This involves knowing about
>>rooms in general and the likely location of switches; continuing monitoring
>>of one's own progress in finding the switch.
>
>Absolutely right.  But where did these constraints on variations (not
>looking on the floor for the switch, for example) come from?  If these
>constraints are based on previous knowledge, you have to explain where that
>knowledge came from.

Cziko's argument starts off with proximate causes (variation and
retention), and now becomes transmuted into an argument about remote
causes. This won't do. One learns about walls, room, switches, etc from
previous experience with them, from textbooks, and from teachers. I doubt
if anyone learns about light switches without instruction and/or inductive
observation, or simply reasoning it out. We still don't have BVSR,
cumulative or otherwise. The rest of Cziko's remarks repeat the same error.

>Further _advances_ in knowledge come from trying out blind (if very highly
>constrained) variations.

If the variations are random with respect to their selective outcomes, then
they are "blind", in Darwin's and Campbell's meaning. If they are
constrained with respect to their selective outcomes, they aren't "blind"
in this sense. The fact that I don't know the outcome of my action *for
certain* is irrelevant; I need only have some reason for believing that one
course of action is more likely to lead to the result I want than another
is. A major difficulty in this discussion has been this shuffling back and
forth between two senses of "blind" variation: 1) indifferent to the
adaptive value of the variation, and 2) not knowing the selection regime or
circumstances that will "choose" the variation. Only the first is relevant
to Cziko's argument.

Once again, Gary: let's have some data in support of the hypothesis please.

Elihu M. Gerson
Tremont Research Institute
458 29 Street
San Francisco, CA 94131
Phone: 415-285-7837   Fax: 415-648-7660  gerson@hooked.net

_______________________________________________________________________________

<26:103>From g-cziko@uiuc.edu Wed Oct 25 20:52:38 1995

Date: Wed, 25 Oct 1995 20:54:35 +0000
To: darwin-l@ukanaix.cc.ukans.edu
From: g-cziko@uiuc.edu (CZIKO Gary)
Subject: Learning as Variation and Selection: Blind and Constrained

[from Gary Cziko responding to Elihu Gerson]

>Cziko's argument starts off with proximate causes (variation and
>retention), and now becomes transmuted into an argument about remote
>causes. This won't do. One learns about walls, room, switches, etc from
>previous experience with them, from textbooks, and from teachers. I doubt
>if anyone learns about light switches without instruction and/or inductive
>observation, or simply reasoning it out.

I would argue that any such learning from textbooks and teachers (although
cannot recall "lightswitch finding" ever covered in my classes or
textbooks) is also based on cumulative blind variation and selective
retention.  And I am not alone in this outlandish proposal.  Perhaps the
best presentation of a Darwinian, selectionist view of education can be
found in:

Perkinson, H. J. (1984). _Learning from our mistakes: A reinterpretation of
twentieth-century educational theory_. Westport, CT: Greenwood Press.

as well as chapter 12 of _Without Miracles_.

Indeed, the notion of constructivism, which is now taking the field of
education by storm (see the last issue of _Educational Researcher_),
refutes the notion that learning is the absorption of knowledge provided by
the environment, although I admit that few "constructivist" would embrace
BVSR (indeed, the most famous constructivist of this century, Jean Piaget,
rejected even Darwinian evolution, embracing instead a quite Lamarckian
idea which he called the "phenocopy.")

>If the variations are random with respect to their selective outcomes, then
>they are "blind", in Darwin's and Campbell's meaning. If they are
>constrained with respect to their selective outcomes, they aren't "blind"
>in this sense.

You are stating here that "random" or "blind" on one hand and "constrained"
on the other are mutually exclusive terms.  I maintain that the variations
that lead to increases in knowledge (or adapted complexity) are both
constrained AND blind.

So let me ask you two straightforward question to see where we stand on
these issues:

1. Are the variations produced in the course of biological evolution (due
to genetic mutation and sexual recombination) blind?  Yes or no.

2. Are the variations produced in the course of biological evolution (due
to genetic mutation and sexual recombination) constrained?  Yes or no.

>The fact that I don't know the outcome of my action *for
>certain* is irrelevant; I need only have some reason for believing that one
>course of action is more likely to lead to the result I want than another
>is.

>A major difficulty in this discussion has been this shuffling back and
>forth between two senses of "blind" variation: 1) indifferent to the
>adaptive value of the variation, and 2) not knowing the selection regime or
>circumstances that will "choose" the variation. Only the first is relevant
>to Cziko's argument.

These statements have helped me to better appreciate your reluctance to use
the word blind to describe the variations involved in human knowledge
processes.  I will return to them, but I first want to get your answers to
the above two questions.

--Gary Cziko

_______________________________________________________________________________

<26:104>From rasmus@leland.Stanford.EDU Thu Oct 26 01:10:38 1995

Date: Wed, 25 Oct 1995 23:10:28 -0700 (PDT)
From: Rasmus Winther <rasmus@leland.Stanford.EDU>
To: darwin-l@ukanaix.cc.ukans.edu
Subject: Two articles

Two articles may help clear up some of the terms and ideas that are being
used in different ways in the recent discussions on the list.

Ron Amundson's "The Trials and Tribulations of Selectionist Explanations"
in _Issues in Evolutionary Epistemology_ Hahlweg and Hooker, eds. SUNY
Press 1989.

and

Susan Oyama's "Ontogeny and Phylogeny: a Case of Metarecapitulation?" in
_Trees of Life_ P. Griffiths, ed. Kluwer Academic Publishers, the
Netherlands, 1992.

The first article clears up exactly what a selective explanation is. It
provides necessary and sufficient conditions for an explanation to be of
a selective type.

The second article provides clarification on the difference between
variational and transformational models of change. It also attempts
to break down that very distinction.

For people interested in the type of debate occuring on the list now,
these are very informative articles to turn to.

greetings,
Rasmus Winther

rasmus@leland.stanford.edu

_______________________________________________________________________________

<26:105>From gerson@hooked.net Thu Oct 26 06:39:01 1995

Date: Thu, 26 Oct 1995 04:39:22 -0700
To: darwin-l@ukanaix.cc.ukans.edu
From: gerson@hooked.net (Elihu M. Gerson)
Subject: Re: Learning as Variation and Selection: Blind and Constrained

Gary:

We don't need confessions of faith; we don't need commercials; we don't
need appeals to authority; we don't have to agree. All we need is a
plausible example of cultural or institutional or organizational change
brought about by BVSR. You specifically mentioned scientific discovery at
one point-- please give us an example.

Best
E

Elihu M. Gerson
Tremont Research Institute
458 29 Street
San Francisco, CA 94131
Phone: 415-285-7837   Fax: 415-648-7660  gerson@hooked.net

_______________________________________________________________________________

<26:106>From wmontgom@nasc.mass.edu Thu Oct 26 11:24:17 1995

Date: Thu, 26 Oct 1995 12:30:08 -0400 (EDT)
From: William Montgomery <wmontgom@nasc.mass.edu>
To: darwin-l <darwin-l@ukanaix.cc.ukans.edu>
Subject: Selectionism

Let us assume that some form of pan-selectionist argument is correct and
that all cultural phenomena can thus be fully explained.  Why should
anyone care?  No matter what the outcome of any cultural development,
adaptation will inevitably have occurred.  If the commonwealth flourishes
and most of its citizens grow to be wise, healthy, prosperous, and moral,
the optimal adaptational result under the given environmental conditions
will have taken place.  If however, the commonwealth fails and its citizens
degenerate into ignorance, fanaticism, misery, and pestilance, optimal
adapation under the given environmental conditons will, no doubt, also have
taken place.  The theorist of adaptation, like Dr. Pangloss, never gets
it wrong.  Adaptation is only adventitiously a state of success or
failure from the point of view of the actors.  They have other concerns
entirely.  Indeed, they cannot help but have other concerns.  Adaptation
comes their way like grace to the Christian, strictly as a byproduct of
doing something different.  Try as they might, they can never earn it.

Cheers, Bill (WMontgom.nasc.mass.edu)

_______________________________________________________________________________

<26:107>From BIOL-WAJ@nich-nsunet.nich.edu Thu Oct 26 13:51:57 1995

To: darwin-l@ukanaix.cc.ukans.edu
From: "Bill Johnson"  <BIOL-WAJ@nich-nsunet.nich.edu>
Organization: Nicholls State University
Date: Thu, 26 Oct 1995 14:01:41 CST
Subject: Group Selection-bigotry

I think I sense a mood of "agreeing to disagree" relative the topic of
adaptation and natural selection, and wonder if there is any interest in a new
topic.  A comment by John Giacobbe in a recent posting to Darwin-L made me
recall an idea I have been tossing around for a few years. I don't put a lot
of stock in group selectionist arguments (for a number of reasons), however,
based upon observations of human behavior, as well as behavioral work that has
been carried out on other social animals, one aspect of behavior where the
group selection argument might have some merit, in promoting the genetic
integrity of the group over the genetic integrity of of the individual. That
aspect of human behavior, which I find present among other social animals in
a less complex and subtle form, is bigotry. I realize that addressing a topic
such as this in an open forum, potentially poses some problems with
individuals who might see this is an opportunity to use the forum as a
<soapbox for their social agenda,> as John Giacobbe says, but I think there
is ground here for a scientific discussion of this idea as well. Once again
quoting John G., <I think that it very important to understand such social
phenomena to reach a more complete understanding of the evolutionary forces
that influence humans' development.> I would like to hear what other members
of the group have to say about this idea. What happens when discrete groups of
social animals meet in the environment (consider Goodall's Gombe chimps, wolf
packs, ants, etc.)? Consider the deep divisions of various groups relative to
ethnicity, skin color, religion, etc., etc.. Is it possible that instead of
ethnicity being <the phenomenon of popular consciousness that it is>, John G.,
once again, that it might be group selected behavior designed to promote the
genetic integrity and insularity of the group? Is outbreeding more beneficial
to the group than inbreeding, considering the ancestral environment?

Lessez les bon temps rouler,

Bill Johnson
Department of Biological Sciences
Nicholls State University
Thibodaux, LA 70310
biol-waj@nich-nsunet.nich.edu

_______________________________________________________________________________

<26:108>From abrown@lazy.demon.co.uk Thu Oct 26 14:56:35 1995

Date: Thu, 26 Oct 1995 14:56:29 -0500
To: darwin-l@ukanaix.cc.ukans.edu
From: Andrew Brown <abrown@lazy.demon.co.uk>
Subject: Re: Learning as Variation and Selection: Blind and Constrained

>So let me ask you two straightforward question to see where we stand on
>these issues:
>
>1. Are the variations produced in the course of biological evolution (due
>to genetic mutation and sexual recombination) blind?  Yes or no.
>
>2. Are the variations produced in the course of biological evolution (due
>to genetic mutation and sexual recombination) constrained?  Yes or no.

Obviously, original variation at the gene level is random in Elihu's strong
sense; whereas the variation that we see is constrained by the demands of
the organism as a whole. The difference emerges over time, and is to some
degree dependent on when we make the observation.

For a baby whatever to be born viable, mutation must be held within fairly
tight bounds. The corollary of this is not that mutation is constrained by
the needs of the organism, but that there are lots of miscarriages, sterile
offspring and other failures.

If the mechanism of selection were not blind, but purposeful, as it is in
human thought and activity, then these mistakes would not be repeated. In
nature, of course, they are. That is the difference between blind and
thoughtful selection.

Andrew Brown
Religious Affairs Correspondent
The Independent, London
Tel: +44-171-293-2682

_______________________________________________________________________________

<26:109>From g-cziko@uiuc.edu Thu Oct 26 17:51:55 1995

Date: Thu, 26 Oct 1995 17:53:52 +0000
To: darwin-l@ukanaix.cc.ukans.edu
From: g-cziko@uiuc.edu (CZIKO Gary)
Subject: Re: Selectionism; Why Care

[from Gary Cziko responding to Bill Montgomery]

>Let us assume that some form of pan-selectionist argument is correct and
>that all cultural phenomena can thus be fully explained.

I for one would not make such an assumption.  As I believe I have stated
many times before on this list, universal selection theory is concerned
only with the emergence of adapted complexity.  I would not be surprised if
many aspects of culture are not instances of adapted complexity.  But other
aspects of culture (science and technology perhaps being the best
examples--including the Eskimo's sealskin coat and kayak) clearly are (at
least to me).  If you don't find any aspects of culture that are adpative
or adapted, you don't need selection theory.

>Why should
>anyone care?  No matter what the outcome of any cultural development,
>adaptation will inevitably have occurred.

This broad use of the term "adaptation" is quite meaningless to me.  I
don't know anyone who uses the term in this essentially meaningless way.  I
address the "why care" question below.

>If the commonwealth flourishes
>and most of its citizens grow to be wise, healthy, prosperous, and moral,
>the optimal adaptational result under the given environmental conditions
>will have taken place.

This would indeed indicate that the culture is adaptive in at least some
respects.  But it may not be in others.  And there is never any guarantee
that an "optimal" adaptation will have taken place.  Biological evoltion is
more a tinkerer than optimal designer (good evidence that God is not the
designer, as Gould has pointed out).  I would guess that adaptive cultural
evolution is similar.

>If however, the commonwealth fails and its citizens
>degenerate into ignorance, fanaticism, misery, and pestilance, optimal
>adapation under the given environmental conditons will, no doubt, also have
>taken place.

This certainly sounds MALadaptive to me from the viewpoint of the continued
survival of the commonwealth's citizens, and is probably also maladaptive
from many other viewpoints as well, such as psychological well-being,
health, freedom from pain, etc.

>Adaptation is only adventitiously a state of success or
>failure from the point of view of the actors.  They have other concerns
>entirely.

Are you saying that citizens of the U.S., for example, are not concerned
with reducing poverty, violence, drug use, and malnutrition while improving
health care, education, the environment, science and technology?  I must
have very different sources of news than you do (I was going to say "Not
everyone is like  Newt Gingrich," but I decided that I had better not lest
I get the conservative Republicans riled up against me along with the
anti-selectionists!).

>Indeed, they cannot help but have other concerns.

I'd be interested to learn what you believe these other concerns are.  Are
you saying that people are not primarily concerned with making their lives
(and the lives of their families) better?

>Adaptation
>comes their way like grace to the Christian, strictly as a byproduct of
>doing something different.  Try as they might, they can never earn it.

Try telling this to a scientist who has spent 20 years of hard work
determining the structure of a complex protein.  Or to a teacher who
devotes much of her or his waking hours making her or his classroom
conducive (adpated) to learning.  I would say that they have indeed earned
their adaptations.

Why should anyone care?  If you want to understand how adapted complexity
emerges (as it does in biological evolution), it provides explanation which
uses cranes and not skyhooks.  If also provides specific hypotheses about
what to look for, mechanisms of variation, selection, and reproduction (as
in the clonal-selection theory of antibody production, or in the "blooming
and pruning" of synapses in the brain).

It also provides useful clues as why certain systems are NOT adaptive.  For
example, if no one is allowed to challenge current beliefs (that is, no
variation allowed), no further progress can take place (a good example is
the crippling effect that Stalin-Lysenko team had on Soviet biology).

And it provides clues on how to facilitate the adaptation of a system, how
to make things better (for example, plant breeders use radiation and
chemical mutagens to increase the source of variation so that new useful
varieties can be developed; genetic techniques for reducing the time
between variation and selection also facilitates the process).

These are some reasons why I care.  But I can certainly understand that not
everyone has the same interests.--Gary Cziko

P.S.  I must admit that I am a bit disappointed that there are not more
Darwin-L subscribers helping to defend and explain the selectionist
perspective.  Bill Calvin could certainly help out with the neurological
stuff ("brain as a Darwin machine"), but I know that he is working hard on
getting one or more books finished, so I suppose he needs his time for that
(check out his wonderful WWW site at http://weber.u.washington.edu/wcalvin/).
I would have thought a list with a name like "Darwin-L" would attract more
individuals who appreciate the tremendous importance of Darwin's selectionist
insight, and that it is not limited only to biological evolution.  But where
are they?

_______________________________________________________________________________

<26:110>From g-cziko@uiuc.edu Thu Oct 26 20:55:50 1995

Date: Thu, 26 Oct 1995 20:57:35 -0600
To: darwin-l@ukanaix.cc.ukans.edu
From: g-cziko@uiuc.edu (CZIKO Gary)
Subject: Re: Learning as Variation & Selection: Serendipity & Discovery

[from Gary Cziko to Elihu Gerson]

Elihu:

I'm disappointed that you didn't care to answer the two questions I posed
in my previous post.  I thought it was important that we come to some
agreement concerning the meanings of "blind" and "constrained."  Otherwise,
we will continue to have difficulty in applying these concepts to other
areas.

So I will repeat the questions in case you agree now that answering will
help us understand each other better:

1. Are the variations produced in the course of biological evolution (due
to genetic mutation and sexual recombination) blind?  Yes or no.

2. Are the variations produced in the course of biological evolution (due
to genetic mutation and sexual recombination) constrained?  Yes or no.

>We don't need confessions of faith; we don't need commercials; we don't
>need appeals to authority; we don't have to agree.

I agree that we don't need the first three (although I would like to sell
lots of copies of _Without Miracles_, still only $30--oops!, another
commercial just slipped in--and I do like to see what other scholars and
scientists who have considered these issues have to say).  But I think we
do need to agree on what we mean by certain words in order to know if we
agree or disagree on the issues.  We need to have common meanings to even
know if we agree or disagree.

>All we need is a plausible example of cultural or institutional or
>>organizational change brought about by BVSR.

No, we need a plausible example of ADAPTIVE cultural (broadly defined)
change brought about by BVSR--such as a scientific or technological
advance.

>You specifically mentioned scientific discovery at
>one point-- please give us an example.

I can give lots of examples.  A very rich source of such is:

Roberts, Royston M. (1989). _Serendipity: Accidental discoveries in
science_. New York: Wiley.

Serendipitous and pseudo-serendiptious discoveries provide clear examples
of the role of blind variation.  A serendiptious discovery is one where
some solution to a problem or useful discovery is made when the discoverer
was not even looking for such a solution or discovery.  One well-known
example is Fleming's discovery of penicillin:

"In the summer of 1928, Fleming was engaged in research on influenza.
While carrying out some routine laboratory work that involved microscopic
examination of cultures of bacteria grown in petri dishes . . ., Fleming
noticed in one dish an unusual clear area.  Examination showed that the
clear area surrounded a spot where a bit of mold had fallen into the dish .
. . .  Fleming concluded that the mold was producing something that was
deadly to the Staphylococcus bacteria in the culture dish." (Roberts, 1989,
pp. 160-161)

Since Fleming was not even intending at the time to find an antibiotic, it
is clear that this was an accidental, blind variation (followed by very
smart selection).

Roberts uses _pseudoserendipty_ "to describe accidental discoveries of ways
to achieve an end sought for, in contrast to the meaning of (true)
_serendipty_, which describes accidental discoveries of things not sought
for." (p. x).

An example of a pseudoserendipitous discovery is Daguerre's invention of
photography.

"Daguerre prepared plates of highly polished silver-plated copper and
exposed them to iodine vapor, which produced a thin layer of silver iodide
on the surface.  Using the camera obscura he exposed these plates,
producing a faint image.  He tried many ways to intensify the image, but
with little success.  One day he placed an exposed plate, which had only a
faint image and which he intended to clean and use again, in a cupboard
containing various chemicals.  After several days, Daguerre removed the
plate and found, to his amazement, a stong image on its surface!" (p. 50).

Then through a process of elimination, Daguerre determined that mercury
vapor had intensifed the image.

Robertson provides over 35 other accounts of serendipitous and
pseudoserendipitous scientific discoveries.

So, Elihu, what do you think?  Do you agree at least that THESE examples
show the role of blind variation (coupled with very intelligent selection,
of course) can play a role in scientific discovery?  Of course you can
still say "yes" here and quite logically maintain that these are
exceptional cases and that BVSR does not play a role in ALL scientific
discoveries.  But if we can't come to agreement here, it would appear to
make little sense to continue our discussion.

--Gary Cziko

_______________________________________________________________________________

<26:111>From CSM@macc.wisc.edu Fri Oct 27 08:48:47 1995

Date: Fri, 27 Oct 95 08:48 CDT
From: Craig McConnell <CSM@macc.wisc.edu>
Subject: Cziko and Gerson
To: DARWIN-L@ukanaix.cc.ukans.edu

#Elihu:
#
#I'm disappointed that you didn't care to answer the two questions I posed
#in my previous post.
#
#So I will repeat the questions in case you agree now that answering will
#help us understand each other better:

This conversation looks more private and of less general interest with every
repetition.  Gary, if Elihu isn't engaging you on the list, why not hash it
out in private and get back to us?

#I agree that we don't need the first three (although I would like to sell
#lots of copies of _Without Miracles_, still only $30--oops!, another
#commercial just slipped in--

These haven't slipped, they have been systematically present for several days,
and your coyness is neither genuine nor endearing.  Please stop.

_______________________________________________________________________________

<26:112>From gerson@hooked.net Fri Oct 27 10:33:56 1995

Date: Fri, 27 Oct 1995 08:33:34 -0700
To: darwin-l@ukanaix.cc.ukans.edu
From: gerson@hooked.net (Elihu M. Gerson)
Subject: Re: Learning as Variation & Selection: Serendipity & Discovery

>[from Gary Cziko to Elihu Gerson]

>"In the summer of 1928, Fleming was engaged in research on influenza.
>While carrying out some routine laboratory work that involved microscopic
>examination of cultures of bacteria grown in petri dishes . . ., Fleming
>noticed in one dish an unusual clear area.  Examination showed that the
>clear area surrounded a spot where a bit of mold had fallen into the dish .
>. . .  Fleming concluded that the mold was producing something that was
>deadly to the Staphylococcus bacteria in the culture dish." (Roberts, 1989,
>pp. 160-161)
>
>Since Fleming was not even intending at the time to find an antibiotic, it
>is clear that this was an accidental, blind variation (followed by very
>smart selection).

This case is an interesting one. Fleming didn't go on to work with his
discovery; the development of penicillin was left to others, years later.
Is the serendipitous discovery of the antibiotic properties of the mold
adaptive cultural change? What aspects of culture changed just then, and
how?

>"Daguerre prepared plates of highly polished silver-plated copper and
>exposed them to iodine vapor, which produced a thin layer of silver iodide
>on the surface.  Using the camera obscura he exposed these plates,
>producing a faint image.  He tried many ways to intensify the image, but
>with little success.  One day he placed an exposed plate, which had only a
>faint image and which he intended to clean and use again, in a cupboard
>containing various chemicals.  After several days, Daguerre removed the
>plate and found, to his amazement, a stong image on its surface!" (p. 50).
>
>Then through a process of elimination, Daguerre determined that mercury
>vapor had intensifed the image.

Did the discovery come at the point where Daguerre noticed the strong
image, or did it come at the point where he determined that mercury was the
intensifer? If the former, why is the discovery adaptive change? If the
latter, why is it BVSR?

I don't think an unreplicated phenomenon counts for either "discovery" or
adaptive cultural change. Indeed, I don't think that changes in one
person's behavior counts for cultural change at all-- cultural change is a
change in institutions.

So the cultural change associated with scientific discovery takes place
around the time that people start changing their research programs (and
possibly other activities as well) in response to a *reproducible* effect
or finding. But this is well past the point where blind variation might
play any meaningful positive role. Serendipity or just plain accident can
play a role in getting a round of the process started, but this is a long
way from the claim that BVSR plays a significant causal role in cultural
change, analogous to the role natural selection plays in evolution.

Elihu M. Gerson
Tremont Research Institute
458 29 Street
San Francisco, CA 94131
Phone: 415-285-7837   Fax: 415-648-7660  gerson@hooked.net

_______________________________________________________________________________

<26:113>From rasmus@leland.Stanford.EDU Fri Oct 27 11:42:43 1995

Date: Fri, 27 Oct 1995 09:42:24 -0700 (PDT)
From: Rasmus Winther <rasmus@leland.Stanford.EDU>
To: CZIKO Gary <g-cziko@uiuc.edu>
Subject: Re: Learning as Variation & Selection: Serendipity & Discovery

A reaction to Gary Cziko's recent posting:

Perhaps Flemming's discovery of penicillin does not count as BVSR. It
seems that we have to beware about calling something a selective
explanation when variations and selective mechanisms are decoupled, not in
terms of variations being independent of adaptive needs, but variations
occurring in completely differenct loci from the selective mechanisms.
Mutations within an organism and ideas within a head are examples of
(blind or not) variations which then are selected on. But would something
which occurred in an external petri-dish count as a variation with
respect to scientific theorizing. Maybe, but then you are changing your
definition of the organic system.

Organisms and the development of scientific theories (in a Popperian
fashion) may be examples of BVSR, but how can you claim that something
which occurs in a petri dish, independent of all theorizing, can count as
a blind variation with respect to a conscious selective mechanism? I am
_not_ saying that it may not be a blind variation within some
larger-defined system, but it seems that if we want to call this BVSR,
then BVSR losses much of its appeal because we then confuse categories
and definitions of organisms or organic systems left and right. BVSR
becomes something that (by fudging) can explain _everything_. The rule
would be something like: "show that X is a blind variation (at all costs)
and then show that, in that or some other domain, X is selected for."

In this way, panselectionism would become feasible (although misleading).

This pan-selectionist rule seems to be the same one applied to Eli
Gerson's (clearly "instructional" (as opposed to "variational-selective"))
examples of children "learning" to look both ways before crossing the road
or searching the wall for a light switch, by Gary Cziko. I see these cases
as clear examples of non-BVSR, and as instructional learning. But we can,
if we want, redefine the system or go back in time or make some
psychological story about how _it must be_ BVSR. We can say it is previous
BVSR, which I do not see it as being, etc..

I must say that panselectionism seems very dubious. BVSR is a very useful
mechanism to explain many facets of organism ontogeny and phylogeny. It
may even be useful to explain some psychological facts about us (although
Skinner's operant conditioning and other pan-BVSR theories in psychology
also seem dubious). But, as in most other things in science, the truth
lies either somewhere in the middle or is only seen when the spectrum
line is dismantled (see Oyama on this latter point). Pan-selectionism or
pan-formalism or pan-historicism (see Gould "A developmental constraint
in cerion..." Evolution 43 (3), 1989, pgs. 516-539) will probably all
turn out to be too extreme. BVSR explanations can be fudged about to
explain practically everything, but that does not seem a useful strategy.

For some sources for common usage of terms and ideas, please see the
previous posting I sent. Those articles are useful.

sincerely,
rasmus winther

rasmus@leland.stanford.edu

_______________________________________________________________________________

<26:114>From wmontgom@nasc.mass.edu Fri Oct 27 14:49:10 1995

Date: Fri, 27 Oct 1995 15:54:55 -0400 (EDT)
From: William Montgomery <wmontgom@nasc.mass.edu>
To: darwin-l@ukanaix.cc.ukans.edu
Subject: Re: Selectionism; Why Care

I see I have been careless.  Let me restrict myself to assuming that all
adaptive cultural phenomena can be explained by selectionism.  In my own
defense, the line between the adaptive and everything else can
be tough to draw in practice.  Still, since this is all hypothetical,
let's assume it can be done.  As for the breadth of my adaptation
concept, I have no apologies.  I simply mean that everybody is in some
sense adapted to his environment though not to the same degree.  By
"optimal" adaptation I simply mean that everybody is doing as well as he
can, given his heredity and environment, not that anybody is guaranteed
any particular standard of success.  Indeed, the point of the passage is
that outcomes can differ enormously.
Our real differences begin at the point where you react to my vision of a
commonwealth in misery.  In real life, misery is a guaranteed outcome for
at least some evolutionary competitors.  Adaptation cannot take place
without such outcomes.  To be sure, citizens in the U.S. and elsewhere
are deeply desirous of eliminating crime, drugs, poverty, and other ills,
but whether their efforts lead to adaptation in an evolutionary sense is
quite another matter, and one that in practice would be extremely hard to
predict.  A scientist who has spent 20 years trying to determine the
structure of a protein may have earned something, but he/she has not earned
adaptation either personally or for the society that supported the work.
Blind fortune still has too much to say about the matter.  It is easy to
look back and say that certain actors of history took a maladaptive path,
but hindsight is cheap.  Looking forward is a much tougher proposition.
Even hindsight runs into difficulties when we are talking about cultural
(as opposed to biological) events, whose implications are often murky
even long afterward.  One contributer to the list has just suggested that
bigotry might be adaptive.  To whom?  Certain individuals?  Certain
ethnic groups or races?  The species as a whole?  We don't even have a
way to begin finding an answer.  Nor, I suspect, could we make much use
of it if we did.  In short, who cares?  Let us cultivate our garden.

_______________________________________________________________________________

<26:115>From jsl@rockvax.rockefeller.edu Fri Oct 27 15:57:51 1995

To: gerson@hooked.net (Elihu M. Gerson); darwin-l@ukanaix.cc.ukans.edu
Subject: Re: Learning as Variation & Selection: Serendipity & Discovery
Date: Fri, 27 Oct 1995 17:01:56 EDT
From: Joshua Lederberg <jsl@rockvax.rockefeller.edu>

<<<<<<<<
[from Elihu Gerson]
This case is an interesting one. Fleming didn't go on to work with
his discovery; the development of penicillin was left to others,
years later.  Is the serendipitous discovery of the antibiotic
properties of the mold adaptive cultural change? What aspects of
culture changed just then, and how?
>>>>>>

Yes, more than meets the eye.  The mold colony on the plate was
serendipity; but there had been many previous observations of
microbial antagonisms which were ignored, largely in the face of
dogmatic discouragement that differentially toxic antibiotics were
feasible.  Pyocyanin had been tried years before and was a flop.
But Fleming did take the next step, cultured the mold, and proved
that an antibiotic substance of moderated toxicity was present in
the culture fluid.  He simply did not have the chemical knowledge
and machinery to proceed with the isolation of penicillin in  pure
form.

Fleming had been prepared for this discovery by his previous
discovery of another antibacterial, lysozyme.  So while he wasn't
explicitly looking for (a) penicillin, he was egregiously
sensitized to recognize it when it fell into his lap.  Is that
a change in (his) culture?

There was an enormous "cultural change" after Waksman's discovery
of streptomycin and the initiation of a canon of further discovery
that became very big business, a huge industry.

I am not sure what any of this has to do with the central thesis
of cognitive variation.  Any creative insight might, arguably,
be the product of unconscious ringing the changes and selection
of the tunes that best fit.  What observation could be a falsi-
fication of that position?  This is quite apart from capricious
inputs from observation: but I don't have to elaborate how
perception relies on a prior theory of the world: a culture, if
you like.

Reply-to: (J. Lederberg)lederberg@rockvax.rockefeller.edu

Prof. Joshua Lederberg
The Rockefeller University
1230 York Avenue
New York, NY   10021-6399

_______________________________________________________________________________

<26:116>From Catalinus@aol.com Fri Oct 27 16:24:59 1995

Date: Fri, 27 Oct 1995 17:24:46 -0400
From: Catalinus@aol.com
To: darwin-l@ukanaix.cc.ukans.edu
Subject: Selectionism: We Do Care

       Forgive me jumping in at this very late stage of the discussion, but
Mr Cziko's impassioned plea caused me to arise and offer my opinion on
the larger issue.  The issue, while revolving around the utility of a
selectionist perspective in the explanation of cultural behavior, can be
expanded to include the adoption of a Darwinian evolutionary model of
cultural behavior.

Here is my take on the thing.  (sorry, it is quite long....)

       I consider myself a proponent of the use of Darwinian biological
evolutionary theory applied either directly or as an extended analogy to
cultural evolution.  I believe we should take a functionalist and
adaptionist approach to the elucidation of an evolutionary framework that
is scientific and particular as to what aspects of the cultural record it
might apply.  I believe evolutionists seek to utilize evolutionary
approaches to generate historical explanations, and would consider that
cultural evolution is not directly equatable with Darwinian biological
evolution, but that as culture is adaptive, and functions within
environmental parameters, and hence a suitable analogy can be derived
from it.  I would consider that evolutionary theory is a framework for
explaining cultural change as the differential persistence of variation in
cultural characteristics.  Stochastic components, as well as
environmental constraints, play a part in evolutionary development, and
cultural continuity can be observed through the fact that all cultural
phenomena within a society are historically and empirically interrelated.
     I believe the subject of cultural evolution is change.  Similarity and
differences between adaptive strategies are not all important, other that
as a cultural-specific explanations for adaptive strategies.  With the
application of the concept of equifinality, a wide variety of possible
adaptions to the same environmental constraints must be considered.
While this weakens the predictive ability of cultural evolutionary theory, I
believe it vastly increases its explanatory abilities.
     Evolution is a selective process, with its accomplishments observed
as the alteration of the frequency of discrete variables.  Selection acts on
three components of evolution: empirical variation, the transmission of
this variation, and the resulting differential representation of the
variables in subsequent cultural states.
     I would take exception to the propensity of Spencerian cultural
evolutionists to equate change with progress.  Progress has no place in
cultural evolution (nor in biological evolution), and that it is clearly not
the inevitable result of evolution.  Cultural evolution causes changes in
adaptive strategies.  Often these changes involve an increase in cultural
complexity, but this is not mandated.
     Many cultures have been observed to exhibit decreases in complexity
as an adaptive response.  In Spencerian terminology, this would be
labeled devolution, and no explanatory mechanism for such a cultural
systems persistence would be offered, or the cultural system would be
characterized as an anomalous condition.  The Darwinian perspective
holds that this is no different from any other adaptive strategy developed
to deal with an environmental shift.
     Spencerian theory does not appear to be a suitable model for
postdicting and explaining any change in culture that did not involve an
increase in complexity.  Darwinian theory allows for plenty of room within
its theoretical bounds to deal with any shift in complexity, and has the
means to explain it, such as through adaptive shifts towards
homeostasis.  The primary reason for this is the Darwinian focus on
natural selection as the influencing factor for cultural change.  Naturally
selective pressures function to favor those adaptionary choices which
offer better survival benefits for a culture.
     While I believe increases in complexity do often occur, but this
increase in complexity is a likely response to population size increases.
This may be analogous with Stephen Jay Gould and Niles Eldredge's
theory of punctuated equilibrium.  Perhaps the population explosion
which occurred at the end of the Cambrian era [which Gould attributes to
a log increase causing a phase change in population ecologies ], might
be considered analogous to a perceived "Holocene Explosion" for human
cultural diversity at the Pleistocene/Holocene boundary.  What do my
fellow scholars think about this?
     Carl Renfrew has proposed a similar idea in his analysis of the
collapse of the Maya.  He developed a model based on Catastrophe
Theory to explain too-rapid changes in sociopolitical and environmental
structures.  Renfrew considers that the Maya had reached a homeostatic
equilibrium between their cultural system and their environmental system.
The population reached the environmental carrying capacity and went
beyond, buoyed by highly efficient cultural institutions.
     Eventually, after a period of environmental degradation, the
environmental situation changed, and was no longer as predictably
bountiful.  The cultural institutions did not react (adapt) quickly enough,
resulting in a collapse of the entire cultural system.  In terms of
punctuated equilibrium, the cultural organism underwent a prolonged
period of homeostasis, punctuated by a rapid period of shift in adaptive
strategy.  The Mayan cultural system still exists now, in another period of
homeostasis, having reached an adaptive plateau.
     I do recognize that the processes of biological evolution require some
redefinition for use in cultural definition.  I do consider, however, that
sufficient analogy exists to transpose one theory for another in many
aspects of archaeological investigations. The processes of natural
selection and mutation have clearly analogous processes in both
biological and cultural evolution, the specific origin or invention of new
elements is not that important for theoretical development, but we might
equates mutation with the act of invention and innovation.
    Gene flow and gene drift exhibit a clearly analogous relationship with
diffusionary processes. Gene flow and gene drift do not always have an
influence in the evolutionary process, but do so at random intervals.
Natural selection appears to have the same adaptionary stimulation
within both cultural and biological evolutionary models.
     An important aspect of evolutionary theory is the role of adaptation.
Almost all theories of cultural evolution incorporate adaptive responses
into their explanatory frameworks.  Adaptive responses are those
"..features of organisms that have come about by natural selection
because they serve certain functions and thus increase the reproductive
success of their carriers" (a la Dobzhansky).
     O'Brien and Holland have noted that adaptation has been too often
used as a post-facto explanation of the expression of a cultural trait.
They suggest that simply the occurrence of a new expression of a
cultural trait (such as in innovation, invention or diffusion) is not enough.
It is only after that trait is replicated and subsequently influences and is
accepted by selective processes that an adaptation exists as a cultural
element.  Following this definition, all functional variations are classified
as adaptations, but only those influenced by selection are considered
adaptive components of a cultural system.  Conversely, stylistic traits are
considered those that neither affect adaptedness nor are under selective
control.
     To mention an alternative position, Braun has noted that even those
expressions of a trait that would typically be labelled stylistic, and
therefore non-adaptive, may in reality possess some adaptive
significance.  Braun notes that stylistic variation may perform some
emblematic function, and that in a cultural system, emblematic
identification could certainly have some selective, survival, and
reproductive value.
    Kirch suggests that adaptation is the key to the integration of many
disparate methodological orientations into an analysis of the central
theme of cultural change as it is related to the environment.  Adaptation,
according to Kirch "...lies at an intersection point between evolutionary
and ecological theory.. (and) offers a contextual perspective on change".
Kirch considers adaptation the fitting of an organism or culture to the
environment.
     Kirch considers culture as a special kind of adaption and that cultural
adaptions are transmitted via learned, non-genetic and extrasomatic
behavioral patterns.  He considers that cultural transmission of adaptive
fitness is clearly Lamarckian in nature, with ontogeny rather than
phylogeny the vector of transmission.
Kirch suggests that the emergence of the concept of culture as an
adaptive system linking biological human populations with their
environment has served to place culture in an evolutionary meaningful
context.  Kirch has attempted an integration of adaptation to
archaeologically meaningful evolutionary principles.
     This systemic-adaptive theory has certain major points.  Kirch
concludes that: 1) culture is to be analyzed in terms of the relations
between elements, particularly feedback, and of the function of relations
as channels for information flow; and 2) culture is an open system,
coupled with environment as well as with the physical population and its
somatic-genetic system, with feedback occurring internally between
elements of the system and between the environment, culture and the
somatic systems themselves.  From these assumptions, he derives a
definition for cultural adaptation as "a process of alteration of a cultural
system in response to change in its coupled environmental and/or
somatic systems".  Those types of behavior that have proven selective
advantage will be selected for and retained, while those that cease to be
advantageous will be selected against, and dropped from the `cultural
pool" of behavioral patterns.
     Kirch cites three features of culture as an adaptive system theory,
these include: 1) the importance of a source of variation within a cultural
system in order to have available options to respond to the adaptive
challenges posed by the environment; 2) a set of selective criteria that
evaluate this behavioral variation based upon selective advantage; and
3) a mechanism for the transmission of those behavioral strategies that
gave a selective advantage to their carriers.
     Kirch notes that the main difference between cultural and biological
models lies in the nature of each's mechanism of transmission.
Biological transmission of trait expression occurs in the reproductive act
and resides in genetic material stored and translated through DNA.
Cultural transmission of trait expression occurs through an interaction
between the central nervous system and the environment, with trait
expression information residing in the collective consciousness of the
population, in the form of accumulated experience.
     Kirch recognizes four contributing components involved in the
process of changing adaptive strategies.  These components are the
available variation itself, the processes of selection acting on the
available variation, the environmental conditions which determine
selective advantage, and the demographic situation of the population
acted upon.
     Variation is the basis for all possible change.  A major issue of
archaeological research is the relation of observed variation to the
probable selective pressures of the environment.  The amount of
variation in a population can be used as a measure of the potential
adaptedness of a population.  Although behavioral variants originate with
the individual, variation must be disseminated for it to have an adaptive
effect on the population.
     Selection is primarily manifest as natural selection.  Some sexual
selection, however, clearly has an influence on character transmission.
This is the factor that decides how much influence a particular
expression of a trait will have on the overall adaptive strategy.  Selection
acts on the group rather than the individual, as far as cultural evolution is
concerned.  The criteria for selective value are manifold, including the
efficiency of energy capture, survival and reproductive success, and even
perceived satisfaction of needs and wants.  In the long run, however, the
most important aspect of selection is its benefit to the survivability of a
culture group.
     There are three modes of selection.  Stabilizing selection tends to
promote maintenance of the status quo, and functions as long as the
environment and cultural adaptions remain in equilibrium.  Directional
selection acts to guide adaptionary changes to fit an environmental
constraint.  It is through this type of selection that most evolutionary
events occur .  Diversifying selection acts to favor extremes of trait
expression.  It would tend to favor alternate forms of an expression, and
tend to not favor a median form.
     The environment is the primary source of selective pressure.  It is
also the ultimate test by which adaptions are measured for their selective
value.  Demographic conditions, such as population size, geographical
distribution, and life table dispersions are a source of adaptionary
influence. Population parameters and interactions with the environment
are the principal influencing factors towards adaptive success.
      Darwinian evolutionary theory can be incorporated into an
explanatory framework, possibly even as a paradigm, for cultural
processes, and it is eminently qualified to give empirical significance to
the archaeological record. The development of an evolutionary paradigm
may place archaeology in the position to making genuine contributions to
Western thought that go beyond what happened when.

Wow, that was clearly a bit more than I intended to post right now, but I
guess Mr Cziko's passion for the idea inspired me.  I would invite any
and all to offer comment on any aspect of my theoretical perspective.
This is a theoretical germination in progress for me, and I would
appreciate help from those with an interest in development of similar
ideas, or even those who think I am full of dingos kidneys!

Thank You for your time!

John A. Giacobbe
Western Archaeological Services, Inc.
catalinus@aol.com
P.S.  Have a good weekend all, Peace.

_______________________________________________________________________________

<26:117>From abrown@lazy.demon.co.uk Fri Oct 27 16:29:05 1995

Date: Fri, 27 Oct 1995 09:09:06 +0000
To: darwin-l@ukanaix.cc.ukans.edu
From: Andrew Brown <abrown@lazy.demon.co.uk>
Subject: Re: reflections on the current debate

>I am interested primarily in adapted complexity, and figuring
>out the process by which in happened, but not just in biological evolution
>but in cultural change, science, cognitive development, language
>acquisition and all areas in which we find a remarkable fit of one system
>to another (apparent design).  In this sense I am perhaps more ambitious
>than you.  The only mechanism I know of explaining such fit is a
>selectionist one, and no one on Darwin-L has yet offered any alternatives.

How would you regard the increasing sophistication of the arguments deployed
in this debate? Assuming for the moment that they are becoming more
sophisticated, and better adapted to their tasks, they are surely not doing
so as a result of selection among random kangaroos. I mean words. Indeed, I
cannot by introspection discover any stage at which my selection of
arguments for this sentence is not as purposive and as far from blind as I
am able to make it.

>What I object to in your perspective is the implication that
>some process other than blind variation and selective retention can result
>in adapted complexity.  I would be very interested to find an alternative
>explanation, but so far I have not seen one offered free of miracles or
>skyhooks.

I presume that this directing intelligence is what you mean by a skyhook. I
am not sure why you think it is an illegitimate explanation of at least this
complexity.

Andrew Brown
footling at home
abrown@lazy.demon.co.uk
andrewb@well.com

_______________________________________________________________________________

<26:118>From ronald@hawaii.edu Sat Oct 28 05:13:52 1995

Date: 	Sat, 28 Oct 1995 00:13:11 -1000
From: Ron Amundson <ronald@hawaii.edu>
To: darwin-l@ukanaix.cc.ukans.edu
Subject: Re: Learning as Variation & Selection: Serendipity & Scientific
Discovery

On Fri, 27 Oct 1995, Rasmus Winther wrote:

> This pan-selectionist rule seems to be the same one applied to Eli
> Gerson's (clearly "instructional" (as opposed to "variational-selective"))
> examples of children "learning" to look both ways before crossing the road
> or searching the wall for a light switch, by Gary Cziko. I see these cases
> as clear examples of non-BVSR, and as instructional learning. But we can,
> if we want, redefine the system or go back in time or make some
> psychological story about how _it must be_ BVSR. We can say it is previous
> BVSR, ...

Yes, Rasmus.  It's turtles, all the way down.

(Small reward offered for pre-Locke citation of world being carried on
the back of turtle-or-turtles.)

> which I do not see it as being, etc..

Well, then, give me an example of a world which is _not_ carried on the
back of a turtle!

[Bit of jocularity, that.  No harm intended.  Smiley, smiley :-)]

Ron

__
Ron Amundson
University of Hawaii at Hilo
ronald@Hawaii.Edu

_______________________________________________________________________________

<26:119>From satori@scn.org Sat Oct 28 12:24:59 1995

Date: Sat, 28 Oct 1995 10:26:15 -0700
From: satori@scn.org (Ronald I. Sato)
To: darwin-l@ukanaix.cc.ukans.edu
Subject: Re: Selectionism; Why Care

Not sure that all social responses can be termed "selectional" nor
even "survival favorable." May best be described simply as "electional."
ron sato

_______________________________________________________________________________

<26:120>From ahouse@hydra.rose.brandeis.edu Mon Oct 30 12:10:40 1995

Date: Mon, 30 Oct 1995 13:10:27 -0400
To: Darwin-L@ukanaix.cc.ukans.edu (Darwin List)
From: ahouse@hydra.rose.brandeis.edu (Jeremy C. Ahouse)
Subject: Amundson's approach and the force of argument

Darwin-L,

        I share with Craig McConnell the sense that Gary Cziko's passionate
defense is no longer of general interest to this list.  But one of the
essays  recommended by Rasmus Winther (Ron Amundson 1989) may well be.

        I had a chance to read it at the end of the week and I liked it a
lot.  It is presumptuous and somewhat counterproductive (since I really
want you all to have a look at it - esp. Cziko) to recapitulate his
argument.  But I will outline a part.

        Ron, after providing Central Conditions (detailing rich &
nondirected variation + sorting) for a selectionist explanation he offers
us a series of illuminating examples of how they are met and mentions his
attraction to selecionist explanation as the "ultimate explanation for
adaptation." (p424)  He then goes on to make the following point.

        ... Doesn't the primacy of selection now tell us something
        important about the world?
                Not necessarily.  The problem is this: The fact that
        selection has been involved somewhere in the etiology of a
        character is _massively uniformative_.  There is no explanatory
        purpose served by invoking  selection without being willing or
        able to specify some level and some structure  by which the
        selection is asserted to operate. (p424)

and later

        ...It is surely true that that _in some sense_ social and
        scientific theory changes can be called "selective."  But
        does this sense in which this is true _explain anything_?
        Does it carry any force? (p427)

        For Ron, selective explanations carry force "precisely to the
extent that the Central Conditions are satisfied."  I suspect that it may
even be possible to satisfy the central conditions and still find the
selectionist explanation to carry less force than another.

        I suggested in a previous posting that there was a kind of "less
teleological than thou" component to the Dennett/Cziko position.  Since
then I looked at James Lennox's article on Teleology (1992) and am
beginning to shift my position.  I had been snowed by Mayr's bluster about
Teleology = bad, old, externalist, platonic appeal to a creator.  Mayr (and
others) tried to rescue the internalist version with a new word (teleonomy)
but might have tried to resucitate teleology.  (It is difficult to know
when to cut and run.)

        So the possibility exists that one of these internalist
teleological/teleonomic explanations could have more "force" than even the
central condition satisfying explanation.  But I will leave that as an
exercise to the reader.

        I am interested in having your help in unpacking the notion of an
argument carrying force.  I know we degenerated to a philosphy of science
argument about induction just recently.  So to keep us near palaetiology
(as Bob O'Hara via the OED reminds us: the application of existing
principles of cause and effect to the explanation of past phenomena) let me
reflect on "explanatory force" with a view to our recent discussions.

        "Explanatory force" shares gut-level "rightness" (that was awkward)
with the kind of naieve empiricism deployed by scientists when they (we)
say about penicillin "it works."  We want our arguments to do something for
us.  That  something is usually explicitly or implicitly about prediction
or anticipation.  Not unlike what the Star Trek computers can do as soon as
they know that the planet being circled is an M-class planet - then we
guess what kind of life forms we will find (notably, contra Gould they seem
to always be hominids?!).

        I see (in Cziko, in Dennett,...) a short form trivial argument; the
central claim *no magic* when coupled with *the past is causal* results in
a necessary and all pervasive "selectionism".  This doesn't have to be at
odds with strong contingency (Gould in _Wonderful Life_) but is thought to
be so by Cziko/Dennett/Dawkins.  Now this kind of (pan/ultra/universal)
selectionism is true but (drearily?) banal - even if it is claimed to apply
only to the domain of "increases in adapted complexity" (?!, Cziko to
Gerson on 20 Oct 1995).

        But this leaves me with a negative sense of explanatory force - I
know when I see something that _doesn't_ have it.  What thoughts do you for
making it positive.  What does an explanation have to do (or can we ever
get agreement on what an explanation has to do) for us to claim it as
forceful.

        I have gone on long enough, but can't resist including the
(sometimes awesome) force of thought experiments.  Here we have (in the
famous hanging bucket experiment, or the balls tied together with a thin
string and chucked off a building) arguments with strong ramifications -
without clear reference to the external world (note that some people think
that empirical knowledge is smuggled in to these constructs) (for nice
review see Brown 1991).  I suspect that selectionist arm chair theorizing
offers a similar seduction.

        - Jeremy

p.s. I found a terrific site for software in evolution/alife modeling on
the net.  While I think that this approach has a long way to come, I am
attracted to the methodological questions that it raises as computer
scientists plunder caricature arguments in evolutionary biology.  For a
fascinating (and technical) approach see a project called Tierra.
http://alife.santafe.edu/alife/software/

_____

Amundson, R. (1989) "The Trials and Tribulations of Selectionist
Explanations" in _Issues in Evolutionary Epistemology_ K. Hahlweg & C.A.
Hooker, eds. SUNY Press. BD161 .I86 1989

Brown, James Robert (1991) The laboratory of the mind: thought experiments
in the natural sciences. Routledge. Q175 .B7965 1991

Gould, Stephen Jay. (1989) Wonderful life: the Burgess Shale and the nature
of history. W.W. Norton.  QE770 .G67 1989

Lennox, J. (1992) "Teleology" in Keywords in Evolutionary Biology, E.F.
Keller and E.A. Lloyd eds. Harvard Univ Press. QH360.6 .K49 1992

note: There is a book on thought experiments by Sorenson that I have read a
review of but not read myself.

Sorensen, Roy A. (1992) Thought experiments. Oxford University Press.  B105
.T54 S67 1992

_______________________________________________________________________________

<26:121>From lbeltran@servidor.unam.mx Mon Oct 30 17:21:15 1995

Date: Mon, 30 Oct 1995 17:22:31 -0600 (CST)
From: Lopez Beltran Carlos-IIF <lbeltran@servidor.unam.mx>
To: Darwinlist <darwin-l@ukanaix.cc.ukans.edu>
Subject: Piltdown (fwd)

late but...

Date: Fri, 27 Oct 95 14:48:05 EDT
From: Val Pinsky <MNHAN139@SIVM.SI.EDU>
To: Lopez Beltran Carlos-IIF <lbeltran@servidor.unam.mx>
Subject: Piltdown (fwd)

          Department of Anthropology, MRC112
          Smithsonian Institution
Subject: Piltdown

frank spencer, 1990 Piltdown: A Scientific Forgery. Oxford UP
misia landau, 1991. Narratives of Human Evolution. Yale UP

Bitnet MNHAN139@SIVM
Internet MNHAN139@SIVM.SI.EDU

_______________________________________________________________________________

<26:122>From ahouse@hydra.rose.brandeis.edu Tue Oct 31 13:40:39 1995

Date: Tue, 31 Oct 1995 14:40:24 -0500
To: Darwin-L@ukanaix.cc.ukans.edu (Darwin List)
From: ahouse@hydra.rose.brandeis.edu (Jeremy C. Ahouse)
Subject: Supervenience: new essays

        A couple of weeks ago we had a short discussion concerning Sober's
suggestion that fitness is a supervenient property.  This engendered just a
little discussion of supervenience.

        In browsing the new books here at Brandeis I found:

Supervenience: new essays / edited by Elias E. Savellos, Umit D. Yalcin.
Cambridge [England] ; New York : Cambridge University Press, 1995.  BD111
.S86 1995

        The introduction and first essay "Varieties of Supervenience" by B.
P. McLaughlin should be of general interest.

        An important point discussed in the intro is supervenience and
causation.  One reading of supervenience results in _all_ the causal "work"
being done at the subvenient level.  This leaves the supervenient proprties
as epiphenomena.  Since there is a claim (in a traditional Darwinian
framework) that fitness differences are causal can this be squared.  (This
isn't so different from the question asked by John Wilkins
(wilkins@wehi.edu.au) on 11 Oct 1995.)

        - Jeremy

__________________________________________________________
Jeremy Creighton Ahouse
Biology Dept.
Brandeis University
Waltham, MA 02254-9110

        (617) 736-4954 Lab
              736-2405 FAX
        ahouse@hydra.rose.brandeis.edu

_______________________________________________________________________________
Darwin-L Message Log 26: 94-122 -- October 1995                             End