From stapp@thsrv.lbl.gov Wed Jul 31 08:46:48 2002 Date: Tue, 30 Jul 2002 17:51:30 -0700 (PDT) From: stapp@thsrv.lbl.gov Reply-To: hpstapp@lbl.gov To: Ken.Augustyn@veridian.com Subject: Re: latest Chapter 13 On Mon, 29 Jul 2002 Ken.Augustyn@veridian.com wrote: > Dear Henry, > > Although you call this hybrid many-minds, I'm don't see where you ever say > that "many-minds" is explicitly part of your theory. Is it? > > If you consider HMM to be an improvement of your earlier theory, then it > would be useful to have an explicit, point-by-point comparison with your > earlier theory, using the same terminology (e.g., Heisenberg choice, > Dirac choice, etc.) whenever possible and showing what survived, what was > killed, what was added. > > On the other hand, if HMM is a variation of the multi-verse interpretation > that Deutsch and others advocate, then it would be helpful to know why you > gave up on a "single universe" solution. > > Best wishes, > > Ken > > > Dear Ken Thanks for your proddings! I shall have to figure out the best way to deal in my book with this shift. Earlier, I often referred to the theory with both Heisenberg and Dirac choices as von Neumann/Wigner to emphasize the fact that it was Wigner not vN who included the Dirac collapses: vN did not use them. I now avoid the name Hybrid-Many-Minds in favor of "von Neumann's approach". It could be misleading to use the term "many-worlds", because that normally means just Process II (Scroedinger Evolution), whereas I include Process I (Heisenberg choice, applied, however, at the level of the brain of the observer) So I have chosen to label it by von Neumann's name, rather than by HMM. I have not figured out how to best explain the shift to the readers of this book. But it may help me if I explain to you my reasons for dropping the Dirac choice. In the original Copenhagen approach the consciousness of the observer entered in two different ways. First he selected which measurement to perform: which aspect of nature to probe. Then he experiencesd the outcome delivered by nature: he experienced the outcome of the Dirac choice. The first choice was more active---a choice to ACT in a certain way---the second more passive: just a RECEIVING of nature's deliverances. These two aspects, active and passive, are both features of our thoughts. However, I do not think that there are two SHARPLY DIVIDED kinds of experiences. Every experience probably has active and passive aspects. The later purely passive experience of receiving nature's answer was not playing any useful role in the theory, and seemed more like a hang-over from the original Copenhagen formulation, where FIRST one prepares the system and THEN observers an outcome. But in the vN approach where the measuring device and the observer are made one, one does not have a natural division into two kinds of experience: every experience can probably be viewed as an aspect of a ACT. The idea of POSING a question can be viewed rather as decision about how to act. And even the experience of hearing the sudden unexpected clap of thunder can probably be seen to have some element of deciding just how to hear that noise. Certainly there are lots of claims by psychologists that without an act of attending there of no experience. This notion that first there is the act of choosing a question followed by an experience of the answer is awkward, because the question anticipates the form of the 'Yes' answer, which, however, may not ever happen. So a model would be forced to deal with this complex arrangement involving two related experiential events. This might be reasonable when dealing with a high-level organism like a human being, but this idea of building in "anticipation" would be difficult for very simple systems. It is Process I that is essential to the interpretaion (vN uses only it) and that involves the observer/participant's decision to act, and that allows the agent to influence the course of physical events via the QZE. The later passive Process III contributes nothing positive to the causal efficacy of the agents. Rather it works against the efforts of the agent by randomly destroying whatever the agent creates: a chance "No" can obliterate a long constructive process. So instead of trying to formulate some complex rule for Process I that links it in some anticipatory way to a later Process III event that might never happen (because nature returns a negative answer) it seems more reasonable that Process I should be self- contained, and not dependent on Process III. And it should be equally applicable to simple and complex systems. My proposal is that the Process I event be understood as a "feeling" that is the grasping of a state of harmonious equilibrium, and the lifting of this state out of the totality of potentialities: this effects the needed separation into two branches that is represented in quantum theory as Process I. I have tried to express this idea in the July 29 version of Chap 13. I would say that this approach is very different from the usual many-worlds approaches, which have only Process II, but neither Processes I nor III. That M-W approach cannot, I believe, (for the reasons spelled out in my article in the Cannadian Journal of Physics) lead to a well defined physical theory (i.e., a theory with the well defined predictions of QT). One needs Process I, but does not need Process III. One can add Process III, and people who abhor the idea that both possibilities are actualized in some fuller reality, can include it without damaging my theory. But there is an aesthetic neatness to not introducing an extra element that plays no significant role, and demands a global action at a distance that can never be observed. Of course, the projection operator P requires (or at least seems to require) an agent-sized instantaneous action, but that is less troubling than the cosmic action at a distance that would be required if the Process III (Dirac reductions) are included in the model. To add more, let me point out that my theory could lead to a possibility of determining empirically whether the process III collapse occurs or not. Suppose an experiment has outcome 1 or 0. Suppose 1 produces state A in the mind of the observer and 0 produces state B in the mind of the observer, where A is not orthogonal to B. (The measurement/observation was not a "good" measurement.) Suppose the states A and B are "states" in the sense of density matrices, so that the nonorthogonality means TrAB >0. Suppose the initial state is a 50-50 mixture of 1 and 0, so that the state of the observer is S= (A + B)/2. Write A=|a>, where x is the value that maximizes Tr PS. Then x=1 (with proper choice for the sign of |b>). Thus the expectation value of "Yes" is Tr PS/Tr S = 1/2 + (1/2)(/(1+ + )). But this expectation value reflects the simultaneous presence of the two outcomes of the original experiment, and the nature of the coupling of those outcomes to the observer, in case both outcomes actually occur. von Neumann, of course, examined "good" experiments, and the presumption about his Process I was that it not depend upon the system being examined. But those are not valid conditions if the brain is examining itself. So it does not seem to me completely impossible to obtain empirical data bearing on whether or not the Dirac reduction really occurs, EVEN IF WE ASSUME THAT THE ENVIRONMENTAL DECOHERENCE CONVERTS THE TWO STATES OF THE "OBSERVED SYSTEM" INTO A *MIXTURE* OF TWO ORTHOGONAL STATES, WHICH IS HOW I TREATED THE TWO STATES 1 AND 0 IN THE ABOVE CALCULATION. The issue is how the state of the observer, obtained by taking a partial trace over all "other" variables, enters into his choice of the Process I event. That is, the issue is: What conditions determine the FREE (but of course brain dependent) choice on the part of the experimenter. If the observer's perceptions of the outcome are so vague as to leave the outcome unclear to him, i.e., if is not zero, then the above calculation shows that if the "harmonious state" is defined in a certain natural way then the predicted expectation value for the `Yes' outcome would depend upon whether or not there was a `Dirac reduction' before the physical interaction between the MIXTURE that characterizes the observed system and the agent that is observing it (i.e., is interacting with it). The other key ramification is that if there are systems that are simple enough so that one could calculate the expected probability of finding this system, regarded as a potential agent, in a state regarded as a potential "harmonious state", then a significant empirical elevation of this probability above what quantum theory without Process I and the self-induced QZE effect predicts would be evidence that tends to support this theory. Generally speaking, this theory would tend to make certain stable states more stable than what normal dynamical and statistical considerations would predict. The temporal auto-correlation in these states would be anomalous. Henry