---------- Forwarded message ----------
Date: Mon, 18 May 1998 18:06:52 -0700 (PDT)
From: Henry Stapp <stapp@thsrv.lbl.gov>
To: William Unruh <unruh@silias.as.huji.ac.il>
Cc: finkel@thsrv.lbl.gov, mermin@msc.cornell.edu,
    peres@photon.technion.ac.il, RGRIF@cmu.edu, shimony@buphy.bu.edu,
    unruh@physics.ubc.ca, vaidman@post.tau.ac.il,
    sofia@techunix.technion.ac.il
Subject: Re: Reply to Unruh's Revised version



On Mon, 18 May 1998, William Unruh wrote:

> a) I got your reply printed out alright by removing the 
> \include math-macro
> line. You apparently don't use them
> 
> b) Regarding the content:
> You say that I do not separate truth from proof. My claim is that in
> quantum mechanics proof cannot be separated from proof. (or as I state is
> evidence and truth are not seperable). In all classical systems it is true
> that once one has proven something, how one has proven that something is
> irrelevant. If it is true according to one line of evidence than it is
> true independent of that evidence. But that I would argue is precisely an
> assumption of realism. In QM, you cannot seperate the truth from the
> evidence used to deduce that truth. To the extent that modal logic
> incorporates the assumption of truth as independent from evidence, then
> yes I dispute the conclusions of modal logic as invalidly incorporating
> assumptions about reality.
> 

Dear Bill,

You are right that quantum thinking is distinguish from classical
thinking by the fact that in quantum thinking the truth of a statement
cannot generally be considered to be independent of the empirical evidence
used to establish the truth of that statement. So one must examine
carefully the *details* of the structure of a proposed argument to see
whether it  is concordant with the strictures imposed by quantum thinking:
one cannot rely uncritically on the the rules of classically based
modal logic. 

That examination is just what I did in my Mat 15 paper, in regard
to the challenges that you raised concerning my modal-logic-based
proof. 

As regards the issue of evidence versus truth, you said:
``This interpretation of LOC1 is, I would argue, a form of realism,
in that it claims that the value to be ascribed to L2 is independent
of the evidence used to determine that value.'' 

But I pointed out that in the application of LOC1 the value ascribed to
the outcome observed by the observers stationed in L was based on solid
evidence, namely the fact that in the *actual situation* the observers
stationed in L do observe that particular value of the outcome. Of course,
one needs to use the fact that in the given actual situation, in which  L2
and R2 are performed and outcome g is observed by the observers in R, a
prediction of quantum theory fixes what the observers in region L will
observe in this actual situation. But this is exactly the sort of
inference that the von Neumann analysis of measurements is based upon, and
you emphasized that that analysis was OK within quantum thinking

Given this secure evidence-based actual situation as the premise, the
assumption LOC1 is the ``no mechanical influence'' idea that if one kept
everything exactly the same as in this actual situation except for a
change in the free choice made by the experimenter in region R, and the
consequences of the change in that free choice, then what was already
observed earlier by the observers in region L would not be changed (as it
could be if the later free choice could effect what was already observed
at an earlier time.)

Now this assumption LOC1 could of course be wrong: the source of the
eventual contradiction could lie right here in LOC1. Indeed, all we are
doing here is to set forth some quidelines as to how far one can, within
quantum theory, retain the  normal idea that the decisions that we freely 
make to perform one experiment or another have no faster-than-light
(backward-in-time) influences.  LOC1 is the weak claim that what is
*actually observed earlier* can be considered to remain undisturbed by a
switch of the later-in-time free choice. Here the *evidence* for the fact
that the observers stationed in L observe outcome c is the fact that *in
the actual world* they do observe c. I do not believe that there is any
stricture in quantum thinking that says that the evidence of our senses in
the actual world cannot be believed to invariant under a change in a later
free choice.

The way that LOC1 is formulated is such that the actual world is not
changed to anything else: it makes no sense to say that the actual world
is other than it is. The condition LOC1 starts from the premise that there
is one actual world. One then considers a hypothetical world that is
specified to be exactly the same as the actual world except for a change
in (1), a specified free choice, and (2), everything that follows
causally from the change in that one free choice. I do not see how one can
make any clear good sense by formulating LOC1 in any other way. In
any case, that is how *I* formulate it. 

This formulation ties into the issue under consideration here, which is to
what extent one can imagine, without conflicting with the predictions of
quantum theory, that all causal effects of a free choices as to which
experiment one performs are confined to the future light cone from the
region in which that choice is made. Within this framework the analysis of
the hypothetical world in which L2 and R1 are performed is set within an
actual world in which L2 and R2 are performed and the observers in R
observe g. The fact the in the hypothetical world R1 is performed, instead
of R2, does not, in my formulation of the problem, influence anything, vis
a vis the actual world, except those things that are causally affected by
this change, and the claim of LOC1 is precisely that these changed things
do not include what was already actually observed earlier by the observers
in L.

You say
> My claim is that in
> quantum mechanics proof cannot be separated from proof. (or as I state is
> evidence and truth are not seperable). 

I guess you meant to say truth cannot be separated from proof.

This part, about the separation of truth and proof, referred to the change
from L2 to L1, whereas the part about separation between evidence and
truth referred to the change from R2 to R1. These two changes are dealt
with separately in my proof: the first is justified by LOC2, the second
by LOC1. 

The argument given above shows, in regard to LOC1, that although, from a
general overview, one might have expected some difficulty to have arisen
within quantum thinking from the fact that when R2 is changed to R1 one no
longer has, in the new situation, the fact that R2 has result g, which was
our basis for knowing that L2 has result c, nevertheless, my detailed way
of formulating LOC1 allows that potential problem to be evaded without
violating normal quantum strictures. 
 
As regards LOC2, and the question of truth versus proof, the conclusion
is similiar: although in general one might not be able to separate the
steps in a logical  argument in quantum mechanics in a way that would
separate truth from proof as cleanly as in classical mechanics, and hence
although one cannot just conclude just from conformity of an argument to
the principles of modal logic conformity to the principles of quantum 
thinking, I have, in my argument, ensured conformity to both sets of
principles. I am not relying on the fact that in classical thinking
one can separate truth from proof. So your citing of the fact that
*in general* the separation between truth and proof my be less clean in
quantum thinking than in classical thinking is not enough to invalidate
my argument. For my definitions and procedures have been sculpted to
avoid this problem.

So although from a high-level general overview one might expect these
possible breakdowns in classical-type connections between  evidence and
truth, and between proof and truth to cause difficulties in pursuing
logical arguments within the quantum context, and hence in proving
anything in this realm, I claim that the specific details of the way I
have set up my argument (based on many years of experience) allows me to
obtain rigorous results within the more exacting quantum context.  

In my reply of May 15 I have spelled out how I do this. So any challenging
of that claim must be based on the details of that solution, not on mere
general pronouncements of the existence of the very problems that my
detailed solution claims to solve. In my proof both evidence and truth,
and proof and truth, are separated in ways that allow the proof to proceed
within a (modal-logic) framework compatible with the quantum strictures.


Best regards, Henry