From:	THEORM::STAPP        27-JUN-1997 16:40:39.57
To:	@KLEINLIST.DIS
CC:	STAPP
Subj:	Re: Nothing Buttery

Dear Aaron,                                         June 27, 1997

I thank you for your detailed message "Nothing Buttery". 
It deserves a detailed reply. Indeed, I think it addresses the
core issue of the mind-matter problem: others who argue that 
mind lies outside classically conceived matter may formulate their 
arguments in terms of "supervenience", instead of "nothing but", but
I think the point can be better made in terms of "nothing but".

I have included the phrase "classically conceived" in front 
of "matter" because otherwise the whole issue becomes:  "What is matter?" 
or "What does `The Physical' mean?" It is conceivable that---
contrary to the ideas upon which classical physics is based---matter is 
basically mind-like, or is fundamentally entwined with mind right from 
the start. Thus one must specify what "matter" means if the mind-matter 
problem to be well defined: there would be no problem if matter were mind. 

I think nonphysicists generally fail adequately to appreciate
that modern physics gives us, at the fundamental level, an idea 
of matter that is very different from the classical one, and hence 
that it is absolutely essential to say, in this context, what "matter" is 
supposed to mean: certainly a "matter" whose basic law of evolution 
is formulated in terms of our thoughts is essentially different from 
classically conceived matter, so a construal of "matter" in this `modern' 
way changes completely the philosophical character of the mind-matter 
problem.

Back to "nothing but": Although the argument can be formulated more 
abstract the point is made most clearly, I think, by going directly to the 
root level of ontology versus epistemology: i.e., "the physically real" 
versus "our concepts".  Things can be specified cleanly here because the 
ontology in question is the ontology of classical mechanics. Consequently, 
since we know that the CM ontology is actually completely false, we do 
not have to squabble, in formulating a CM ontology, over such irresolvable 
questions as what "nature herself is really like": answers to such questions 
lie forever outside the realm of human (scientifically acquired) knowledge. 
All our deliberations are within the realm of human concepts, which we are 
free to organize as we see fit to meet our needs, which I assume here to 
be the need to rationally understand ourselves and the world we seem to 
inhabit. In particular, we are *free to define* the classical ontology in any 
way that gives a lean but adequate rational foundation for both classical 
mechanics and epistemology.
 
I want to start with, and justify using, the austere CM ontology according
to which the physical world consists of "nothing but" the positions, 
velocities, and masses of some finite collection of structureless bits of 
matter that influence only their immediate neighbors. The entire classical 
dynamics is specified (over a large but finite spacetime region) by 
considering, separately and independently, each of a very large but finite 
set of very tiny (overlapping) spacetime regions: one can imagine an 
independent deity's being in charge of each of these tiny spacetime regions, 
and knowing nothing about anything outside that region, but following a set 
of rigidly specified laws for constructing the motions of the particles in 
his tiny realm. The totality of what is known to "Nature" is just the 
disjunctive aggregration what is known to these individual deities, each of 
which is conceived of as living "incommunicado". Thus Nature, in this 
conception of it, cannot know that there are three apples in a certain 
basket, because having that knowledge would require some sort of 
communication among the `deities', or some sort of "overseer" that can know 
what's in the minds of several deities. But the principles of classical 
mechanics (in particular the locality principle of classical relativistic
field theory) tell us precisely the there is no need for any such 
communication, or any such overseer: the classical dynamical evolution 
of the universe can be specified without allowing any such communications, 
or having any such overseer.  

I have spelled out in my book, and will not repeat here, a classical picture
of brain functioning in terms of patterns of brain activities (neuronal
activities?) that I called "symbols", because within the context of the brain 
in which they are imbedded each one has dispositions to excite other symbols.
And I suggested how these symbols would be organized into templates for action, 
and how they would connect to sensory inputs and motor outputs so that the 
classically described organism could be expected to behave roughly in the 
way human beings behave. Our conscious experiences are then to be associated 
with the occurrence of these templates for action, and they consist of the 
`feelings of the initiation of the actions' that the template for action 
initiates. (Some of these actions might be to record answers to certain
questions about the character of one's sensings.) These knowings are quite 
different from Nature's knowings because on the small scale the human knowings 
are far less detailed, but, unlike Nature's knowings, they can combine 
information coming from different microscopic regions.

Now where does mathematics and number theory fit into this
ontology-epistemology? 

Evidently each of the deities is required to have enormous mathematical 
capacity: even to write down one component of the position of one particle 
would, in general, be beyond the capacity of all the people who ever lived, 
or ever will live, aided by all the computers that they will ever build. 
[The fantastic computing abilities assigned to nature by classical mechanics 
are so prodigious as to make it seem almost "obvious", on the basis of 
that alone, that the classical picture could not actually be correct, 
but that question is not the issue here.] On the other hand, the 
repertory of manipulations that these deities are required to master 
is very limited: they need never understand "greater than", or "less than"
or "integer", or "subset of a set", or "denumerable set" or "prime 
number": they only need to know how to integrate differential equations, 
which can, with only indetectable error, be replaced by finite difference 
equations, on some sufficiently fine scale. If we also allow some truncation 
of the numbers then we can reduce the whole classical ontology (over the 
finite spacetime region) to just a huge battery of imagined computers each 
of which needs to be able to do nothing but simple arithmetic, and never 
does anything but simple arithmetic calculations involving numbers pertaining 
to some associated tiny spacetime region. There is no communication between 
any two computers beyond the common access of neighboring computers to 
certain common numbers.

I would like to defend the idea that the austere classical-physics ontolgy
that I have just outlined does make sense, as a "nothing but" ontology
that provides  a rational basis for physics and epistemology: 
the physical world is conceived of as consisting of "nothing but" this horde
of simple idealized computers, each one behaving in the way needed to 
determine the sets of numbers that will define the trajectories of 
"structureless bits of matter" in the way prescribed by the laws of 
classical mechanics.

If X_1 is the X coordinate of particle 1, and X_2 is the X coordinate of
a far-away particle 2, say at time t=0, them the number X_1 exists in the
physical world (according to this austere conception of the physical world)
and X_2 exists in the physical world. But (X_1 + X_2) does not: there is 
a register that contains the number X_1, and a register that contains the 
number X_2, but no register, in the set of all registers that are needed to 
define the physical world (in this conception of the physical world), that 
contains the number that is defined as the sum of X_1 and X_2. [If some 
devices were designed to try to measure X_1 and X_2 and to put the sum of 
these numbers, represented by some physical configuration of readings of 
the basic registers, into the physical system, then to the extent that 
such devices were actually built, and actually did perform as they were 
supposed to, one would have an exception to the assertion made above, but 
such exceptions would be exceedingly rare, and, in view of the practical 
difficulties, undoubtedly of zero likelihood to actually occur]

Of course, all we can ever think about is our thoughts, so some care
is needed to distinguish the ontological physical reality---as it is 
conceived of in this austere ontology, which consists only
of a set of numbers each of which is supposed to be stored in some one of 
the registers specified in the model---and a conception that someone living
in that world could have of that world: this latter conception could in 
principle include a conception of the sum (X_1 + X_2). Thus physicists
and mathematicians are able to do all of the mathematical manipulations 
that they have invented or devised for manipulating their conceptions of 
possible classical worlds of the kind specified by the austere 
classical ontology. 

I have discussed at length---in my earlier discussions with Pat Hayes---
how the classical conception, being built on idealizations of (mainly
visual) sense impressions is such that someone can compare his 
conceptualizations of the physical world to his sense impresssions and make 
judgements that would bring his conception of the world into rough 
coordination with the classical world that is supposed to exist, according 
to this ontology, and how our language grows to accommodate refinements 
connected to the various physical sciences (e.g., biology). 

So where, in all of this, is the verity that you speak of, namely that 29 is
prime and 30 is not.

More generally, where is number theory? Is it part of the physical world, 
or merely conceptual: i.e., an aspect of thoughts about numbers?

The verities of number theory certainly do not lie in the austerely
conceived classical world specified by this ontology: there is, in the horde 
of computers that define, in this ontology, what is part of physical reality 
none that uses, or stores in any register, the assertion that 29 is prime 
or that 30 is not. This model ontology shows that the properties of numbers 
NOT NEEDED to perform the simple arithmetic operations required to work out
the classical dynamical evolution can be consistently regarded as not 
being part of the austerely conceived classical physical world: most of number 
theory is conceptual; it resides in realms of thought, and to the extent that
thoughts and concepts lie only in minds associated with physical entities,
as naturalistic philosophy demands, most of number theory exists only
in minds associated with physical entities.

More relevant to the present discussion is the fact that (X_1 + X_2)
is, in the same sense, conceptual, not physical. If one were to accept that 
this sum, which---according to this austere ontology---exists only in minds, 
and not in the austerely conceived world of classically conceived matter, 
were actual part of the physical world one would, by virtue of this single 
simple error, bring chaos into one's thinking, because every other claim 
that some conceptual thing is part of the physical world could be justified 
by citing its similarity to this case: in logic the admission of one error, 
however minute, causes the whole structure to collapse in ruins.

It is useful to introduce, in contrast to the austere classical
physics ontology, another concept: the subtle classical physics ontology.
The point is this: if we accept a naturalistic philosophy then the
minds are to be associated with physical structures, or systems, and we might 
as well, for definiteness, focus on human beings, considered as body/brains,
with minds somehow associated. The conceptual things discussed above (e.g.,
prime numbers, (X_1 + X_2), concepts and words pertaining to biology,
and recognizable and nameable perceptions) are supposed to be associated
in some way with activities in the body/brains of these human beings.
So there is some subtle way in which these concepts "belong" to the material
world: various dispositions exist in these body/brains that make it possible
for learned discourse about these concepts to occur between properly prepared
body/brains.  Taking, for simplicity, an electrical curcuit analogy, the
array of resistances, impedances, inductances, and capacitances is arranged so
that system of curcuits is able to engage in learned discourse about 
properties of prime numbers etc.. Hence this part of number theory, namely 
the ability of physical systems to discourse learnedly about it, is 
"in the classical physical system" in this subtle way.

So we now have in place, in quite concrete terms, the essential elements 
of the mind-matter problem: the austere CM ontology that is basically
a "nothing but" ontology; our concepts, as we human beings experience/feel 
them; and the subtle imbeddings of our concepts in the high-level physical 
properties and activities of, in particular, certain biological systems 
that we call members of the species "homo sapiens".

Within this framework we can consider the points made in your
"Nothing Buttery" message.

The issue is how causal mechanisms in our brain are related to our 
conscious experiences.

You approach this problem by emphasizing that we are dealing
with a complex situation that can probably be profitably analyzed
by thinking in terms of a hierarchy of mechanisms, with 
higher-level mechanisms implemented in lower-level ones. 

I certainly concur that this way of thinking should be useful,
and believe that nature, both biological and physical, seems to be constructed
like this. As regards the physical, let me recall that it was only during 
this century that the atomic hypothesis became generally accepted (as a
consequence of the analysis of brownian motion) and in this century we have
seen evidence for several deeper levels (elementary particles and quarks, 
with more probably to come). 

You argue, and I would agree, that the idea of "nothing but" is a bad 
idea if it is construed as entailing the non-utility of an analysis of brain 
function in terms of a hierarchy of mechanisms. For the brain activities 
associated most closely with mind may be the activities of a brain mechanism 
that rides on other brain mechanisms. 

However, I would stress that this analysis is being done by someone, perhaps
by you. And that analysis is conceptual: you can do a tremendous amount
of analysis without building a single exemplar.

But let me grant that in the case at hand we are contemplating and 
analyzing a really existing physical system that is best conceived of 
as consisting of a hierarchy of mechanisms.

And, to get to the point, let us suppose that that system is a human brain
that is "having" conscious experiences. And let me grant that your 
hierarchical analysis of the causal structure of the brain activity is 
apparently giving us a satisfactory account of what seems to be happening 
in the brain. And suppose this brain activity is "correlated" with the 
conscious experiences that this brain is "having" in just the way 
specified by a codification of empirical results on the NCC 
(neural correlates of conscious) formulated by Crick, Koch, Klein et. al..

Then one can ask: What is this relationship "having" that connects
the physical brain activities---that your heirarchical causal theory 
may seem adequately to explain---to the conscious experiences that are 
associated empirically with those brain activities?

My answer is that there is here, from the theoretical point of view, a 
logical gap between the austere CM ontological concept of the brain activity 
and the associated conscious experience. The physical activities 
of the brain are asserted to be, fundamentally, "nothing but" the motions 
of structureless bits of matter, but the classical conception of these 
sets of motions is too impoverished to allow the existence any such set 
of motions to logically entail, by virtue of the specified concepts, the 
existence of, say, an actual experience of a patch of "green". 

But if consciousness is not logically entailed by an ontology
that entails the dynamics prescribed by the classical laws then
consciousness can be left out without affecting the physical
world: it is therefore epiphenomenal.

You replied:

>If you really think I am talking about the addition of any ad hoc
>epiphenomenal appendage then I have totally failed to communicate in the
>last couple of months, especially in all my claims that I want something
>stronger than mere correlation between ontological layers, though not
>necessarily as strong as mathematical or logical implication.
>
>I keep saying I am talking about an implementation relation between two
>equally real ontological levels, but where one X is implemented in the
>other Y (X supervenes on Y). If such a thing is possible for mechanisms
>then saying the system is "nothing but" Y, simply ignores the existence
>and causal powers of X.
>
>If you have an argument that it is impossible for any new ontology,
>including new causal powers, to be implemented in a configuration of
>dynamically interacting physical components, lets hear it. Until then
>the use of "nothing but" is without any foundation.
>
>To just brush that aside and say that if Y is a physical system then it
>is "nothing but" a physical system is either to mistake a trivial
>tautology for a profound truth, or else to to beg the question of the
>possibility of mechanism supervenenience.
>
>Whether and how such implementations are possible is what we have been
>discussing all along. To simply *assume* that it is impossible is
>just circular in this context.

I appreciate that *you* are not talking about the addition of an 
ad hoc epiphenomenal appendage: that you are talking about a hierarchy
of supervenient mechanical levels with causal powers.

I have no problem anything you say insofar you do not assume CM. 
For my whole point concerns a problem with CM and consciousness.
 
Insofar as you are accepting CM, and are talking about levels of 
structures *in the brain*, there is, in fact, NO CONFLICT between 
"nothing but" and "causal powers" of high-level structures: within the 
austere CM ontology the high-level structures in the brain will have 
causal powers precisely because of the arrangements of the matter in 
the brain, and the causal power of the thus-arranged matter.

So I think we can agree that it is logically quite possible for 
there to be a hierarchical causal structure within the austere 
CM ontology---in which this causal structure is ontologically composed 
of "NOTHING BUT" these moving bits of matter---and that, moreover, 
anything over and above these moving bits of matter (including, of 
course, all the causal structures built out of them) is, as far as
classical physics is concerned, gratuitous, and in fact, insofar as 
it is physical, is contrary to the CM conception of physical reality.

So I am NOT ignoring the causal power of any structure in the brain
by accepting the austere "nothing but" CM ontology: I am only insisting
that whereas this "nothing but" ontology does account for all causal 
properties of the brain compatible with CM, it is too impoverished,
in terms of its explicitly specified properties---which include all of 
the properties needed to account for all of CM---to demand, *under 
any conceivable configurational condition*, the existence of an 
experience of a patch of "green", which can therefore be left out without
affecting or upsetting the causal structure that fixes the dynamical 
evolution of the physical.

You wrote:
>If you have an argument that it is impossible for any new ontology,
>including new causal powers, to be implemented in a configuration of
>dynamically interacting physical components, lets hear it. Until then
>the use of "nothing but" is without any foundation.

Well, it certainly is impossible for any new ontology that is compatible 
with the principles of CM to be added to the austere CM ontology, and to
add new causal powers, except insofar as the additions are epiphenomenal
with respect to all physical properties and activities. This is because the
austere "nothing but" CM ontology that I have described specifies completely, 
in accordance with the dynamically complete principles of CM, the motions
of all the bits of matter that constitute the physical reality of CM. One can
add on all sorts extra ontology, ad infinitum, and add causal powers that will
determine the behavior of these extra pieces. But these additions cannot
alter the behavior of the parts covered by classical mechanics because the
austere CM ontology already does that job completely. This completeness of
classical mechanics is, of course, the core of the mind-matter problem. 

(I do not mean by this "completeness" that we necessarily know today
all of the kinds of particles and fields that exist in nature: I mean by
classical mechanics the general classical relativistic local-field
framework, taken to be complete in the sense that all of the *existing* 
different types of fields and particles are supposed to be included, whether 
we human beings have yet discovered them or not)

You wrote:
>To just brush that aside and say that if Y is a physical system then it
>is "nothing but" a physical system is either to mistake a trivial
>tautology for a profound truth, or else to to beg the question of the
>possibility of mechanism supervenenience.
>
>Whether and how such implementations are possible is what we have been
>discussing all along. To simply *assume* that it is impossible is
>just circular in this context.

Again, the first question is whether CM is assumed to be valid. If it is
then there is no possibility of a non-epiphenomenal supervenient 
mechanism: any added level can be left out without altering any physical
activity or property because the austere ontology already fixes these
activities and properties, yet is too impoverished, logically, to demand
the existence of consciousness, which therefore cannot supervene upon
the physical, since it can be left out without changing the physical.

If classical mechanics is rejected then I think it rational to go to the 
replacement that physicists have already dicovered, and worked out, and
confirmed experimentally in highly nontrivial ways to one part in a 
hundred million in some cases, and have never found to be violated,
and to accept this replacement in the form that seems best able to the 
close the logical gap described above.

You write:

>Even if some of the possibilities consistent with the minimal
>assumptions are "weird" there's nothing wrong with trying to see what
>can be achieved with minimal assumptions.
>
>On the contrary, NOT doing that by starting from restrictive assumptions
>is anti-scientific since it prevents the discovery of important new
>features of the higher level mechanisms.

I think the more reasonable program is try to stay within 
scientifically established tradition until some difficulty
appears: otherwise the possibilities are just too unconstrained.
When traditional ideas fail, their very failure give some clue as to
how to proceed. I see no rational reason to reject the laboriously
achieved findings of 300 years of physical science.

>But we had better have a pretty clear idea what question we are asking,
>and the formulation of the question, i.e. the definition of mentality,
>had better not in itself assume anything about QM, otherwise the process
>is circular, and ignores the fact that there was a serious question
>being asked (e.g. by Kant, and many others) long before anyone knew
>anything about QM. So it could not be a way of answering THAT question.

You have  argued before that good definitions come only when one
has a theory. But, in any case, I characterize mind by feelings
and experiencings, not by QM. QM could be a way of answering questions
posed before QM was discovered because it provides a way of evading the
overly stringent logical demands of CM without losing any of the 
valid predictions of CM, and it provides also an already created (by 
physicist) way of consistently introducing into physical reality
our causally efficacious thoughts.

>This leaves open the entirely different question whether, as you've
>claimed, it is impossible to articulate QM fully without involving
>consciousness, which seems so far to be the only remotely plausible
>argument for linking the two.

I never made such a strong claim. My argument is that CM is logically
unsatisfactory because there is an uncloseable logical gap involving 
a lack of causal connection (epiphenomenalism), and this makes it
reasonable to open one's eyes to the fact that physicists have already, 
on the basis empirical results, found CM to be fundamentally incorrect, 
and hugely so. And the (orthodox) physicists' replacement of CM did 
involve bringing conscious into physics, inspite of their extremely 
reluctance to do so.

I have said that no alternative to the orthodox interpretation
has claimed the allegiance of a majority of interested 
physicists/philosophers, and that the orthodox interpretation remains,
in the absence of any lilly-white challenger, the uncomfortably-held 
standard view, any failure of which would be regarded as Nobel
worthy, I am sure (in the good company of many ambitious 
experimentalists).

>However, my impression is that this argument is being undermined by the
>fact that whatever the leading physicists of the early 20th century may
>have thought, the current physicist community is moving towards a view
>that replaces notions of consciousness with notions of particular
>physical configurations that in effect "make measurements". (E.g. see
>the article on schroedinger's cat in the current Scientific American.)

I do not believe the current physics community is moving toward any fixed
view on this point. Many experiments are being performed and all results
are in exact conformity with the predictions of the orthodox Copenhagen
interpretation, which remains the simplest because it is based pragmatically
upon what we human observers observe and report to each other.

I think you have misinterpreted the Scientific American article.
It touts Zurek's decoherence theory as if it were somehow essentially
different from the orthodox interpretation when all it really is is
a working out of some of the details of what the orthodox theory predicts:
it does not claim to give any predictions that differ from what the orthodox 
theory says. Any empirical departures from the approximations that
Zurek's equation represent would call for refinements that would be defined
by the underlying orthodox interpretation. The parameters of the GRW theory
were defined to give good agreement with the orthodox predictions: GRW would
become significant if the orthodox predictions would fail, while those of 
GRW would work for some other values of the GRW parameters. But no
failure of the orthodox predictions have ever been encountered.
There is certainly still widespread dissatisfaction among physicists with
the idea of bringing mind into physics, and many physicsts might
opine that QM doesn't *really* bring in mind. But it certainly COULD, and 
I think that this possibility should be of great interest to anyone 
who is seriously interested in the mind-matter problem.

Finally, a word about supervenience vs identity.
My suggestion that you probably want to identify consciousness with
certain characteristic (kinds of) brain behavior did not mean a 
one-to-one (identity) correspondence: it seems to me that *many* brain
behaviours that represent similar causal relationships should be experienced
in the same way: surely the failure of one single microchannel to open 
could hardly be expected to be able to be felt by anyone. 

Your emphasis on the fact that conscious experiences can be restricted 
in ways that the possible brain states are not is of prime importance
to my QM theory. The situation is this: the problem with interpreting
quantum theory without introducing mind is that there are no natural
conditions within the mathematical structure itself to specify when and
where collapses occur. Introducing consciousness gives one an external
(to the mathematical formalism that is the QM generalization of CM)
criterion: the collapse must be to a state that corresponds in an 
appropriate way to a possible experiential state: this restriction 
is the core of the Copenhagen interpretation, and it is needed to
avoid what must otherwise be an ad hoc and arbitrary stipulation
on the theory, and physicists abhor such things. And it simultaneously
makes the theory eminently practical because the probabilities of the
collapses are the only things that the theory cleanly predicts, and
these collaspes are exactly what is, according to the theory, directly 
experienced. So theory and experiment are linked in an ideal fashion.
This restriction on the possible collapses, via restrictions on possible
experiences, is what injects experience PER SE into the brain dynamics.

Thanks for giving me your nice letter to respond to. Your letter has 
greatly helped me to understand your position: I hope this letter will 
be similarly helpful to you.

Best regards,   Henry