Date: Fri, 11 Sep 1998 19:46:32 -0700 From: Stuart Hameroff Subject: [q-mind] When does it happen, and why? - Henry Stapp MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" >From Henry Stapp Subject: When does it happen? And why us? Consider the case of the double slit expt where a single photon lands on the film and twenty years later the film gets developed and a human observer looks at it. Question: What is the state of the universe during those twenty years ? (The Schrodinger cat business). Let's review the quantum description. As in my dialog with Stan Klein, suppose the film has a 20X20 grid of 400 locations that can be distinguished by the human observer: suppose after he looks at the developed film an "outcome" appears to him, which will cause him to report to himself, and to his colleagues, and write into his log book, for some number N between 1 and 400, "#N appeared". As I said in my dialog with Stan Klein: The state S_b of the brain of the observer O evolves independently of the progress of the experiment until a causal link to the experiment causes S_b to react: S_b then evolve into a sum, S_b = P_e#1 S_b P_e#1 + ... + P_e#400 S_b P_e#400. Question: What has happened to the rest of the universe during those 20 years? The answer to this question depends on the unresolved issue of what the conditions are for a question to be posed (put to nature). Let me describe three possibilities, which are believed to be empirically indistinguishable to human beings: (1) There are no reduction events except one's occurring in human body/brains, and representing increments in "our knowledge". (2) Reduction occurs when a film is exposed. (3) A relevant question is posed in the interim. The first option is in accord with the Copenhagen stance. It is also in line with Wheeler's dictum; "No phenomena is a phenomena unless it is an observed phenomena!" In this case the film could exist for the twenty years in the superposed state: no (human) experience has ruled out any of the superposed possibilities. The founders of quantum theory were obviously uncomfortable having to say such an outlandish thing to their colleagues. So they developed some linguistic devises. Heisenberg had a dual language. One wing pertains to the quantum formalism, and how it is to be used in practice; the other wing describes "What is really happening." The first wing speaks about the quantum state, which represents "our knowledge", and gets reduced when "we" human beings find out something, just as claimed by Dirac at the 1927 Solvay conference. The second wing speaks of "the transition from possible to actual [which] occurs when the interaction between the (atomic) system and the measuring device comes into play". It is important to understand that the first description is about the PRAGMATIC COPENHAGEN interpretation, which was created for the practical application of the quantum theoretical formulas. It is about the quantum state, which IS a direct representation of "our (human) knowledge", described in the language that we human beings use to describe to ourselves and our colleagues, "what we have done and what we have learned" (read: "How we have set up the experiment and which outcome has appeared to us). It is this pragmatic conception that makes quantum theory so perfectly suited to the scientific study to the mind/brain problem. This pragmatic conception is basically the one I use. The second description is ontological: it is about "What is really happening". Then Heisengerg speaks of "the actual" and transition from "the possible" to "the actual". But we have no empirical evidence about "the actual", if it is divorced from the pragmatic theory. This is because the pragmatic theory is what is used in practice, and it appears to give (and is claimed by Bohr to give) everything that atomic theory can say that can be confronted with empirical evidence. Thus Heisenberg can be quite loose in his use of this second language, and says that this transition occurs "when the interaction between the object and the measuring device comes into play". So he uses, in parallel, both the pragmatic conception, in which the jump occurs when the state of knowledge changes, and also an ontological conception in which the jump occurs at the device. Bohr pursues a similar strategy what he says that the process is "brought to a close" by an irreversible amplification. What the observer says he is describing is what HE has learned, and that is both post-amplification and, at that point, describable by either himself or others in terms of classical concepts. Wheeler similarly changed his dictum to: No phenomena is a phenomena until it is a recorded (registered) phenomena. I bite the bullet, and tie ontology to practice by asserting that: (1) "the reality" IS "growing knowledge", (2) the growth of knowledge has a causal structure, and (3) the quantum mechanical theory that we have created represents certain aspects of the causal structure of the growth of knowledge. In order to stay close to scientific practice and empirical evidence I stick basically to the pragmatic stance, and hence to human knowledge. But from the ontological side it is clear that the theory must eventually be expanded so as to include a more general conception of knowledge: anthrocentrism is ontologically unacceptable, but pragmatically useful. To stay on track science should be pragmatically useful, but aimed at ontological completeness. In connection with this question of "When does it happen?", and the possibility of relevant empirical evidence, I note that my theory of Helmut Schmidt's retroK experiment does rest on the assumption that the reductions in that particular experiment occur mainly in conjunction with human observers, not at the mechanical measuring devices. The Bierman-Randi test is presumably of this type. If that experiment does get performed, and does give a positive result close to the target p=.000001, that would lend support to the possiblity that reductions are delayed: IF the reductions are indeed delayed---and there is no empirical evidence to the contrary---then my theory would PREDICT retroK effects of this kind completely within orthodox theory, simply by using QZE in place of the violation of basic quantum theory that I used in my paper in The Physical Review. In the other two cases (2) and (3) the reduction would occur earlier in time, and at some intermediate place along the causal route from film exposure to its effect on the brain. Then only one term would be present in the sum exhibited above. But even if there is an earlier reduction that fixes the location of the physical spot, the observer could still have an experience of seeing the spot at that location, for there is plenty of remaining uncertainty in his brain, even if that uncertainty is not so great as to allow the spot to appear to be at a different location. ******************************************************************* Question: Why not use the familiar Dirac bra-ket VECTOR |PSI> to represent the state, rather than the operator S= |PSI>"operator", where numbers commute and operators don't, and (2) "Integration over phase space (statistical average)"--> Trace". One can immediately write down, at least formally, within the theory, the connection between experiences and their representations in brains, and the rule for computing relationships between experiences. Thus in the operator formulation one can "cut directly to the chase" so to speak. One does not need to instruct neurophysiologists and philosophers about these abstractions called "vectors": matrices are relatively simple to understand. Moreover, I NEED to use the operator notation. This is because in the actual universe "the observer" regarded as a vector is not well defined, due to the entanglement of observer degrees of freedom with those of the rest of the universe. My whole procedure of defining the state of (the body/brain of) the observer, and of specifying both the connection between this physical state and the possible experiences that it represents, and the computational procedures that connect our experiences to one another, are compactly and accurately represented in the operator notation. *************************************************************** Question: What happens when there are two human observers? Question: How is the operator formalism connected to the "familiar" vector notation. If entanglements are ignored then the vector notation is OK. So, for those familiar with the bra-ket vector notation of Dirac, let me show how the two-observer situation works, with the understanding that the results obtained in this way generalize to the real universe, and our entangled bodies within it, if one uses the operator language. Let 1, 2, and R designate observer 1, observer 2, and the rest of the universe, respectively. Let the initial state vector be |PSI, #0> = |psi1#0>|psi2#0>|psiR#0>, with |psiR#0> = |psiR#1> + ... + |psiR#400>. The essence of the process in which the two observers "look" physically at the film---i.e., direct their attentions to the developed film with the intention of learning the location of the spot, but without any outcome yet appearing to them---is this. First suppose that the spot on the film were in a definite location N, and hence |psiR>= |psiR#N>, and that the state of the brains of the observers 1 and 2 before any signal from the film has reached their brains are |psi1#0> and |psi2#0>, respectively. Then the interaction of observers with the film produces (in accordance with von Neumann's analysis) |psi1#0>|psi2#0>|psiR#N>--> |psi1#N>|psi2#N>|psiR#N>, where |psi1#N> is the state of the brain of observer 1 that would represent---if it were to be actualized---his seeing the spot in location N, and similarly for observer 2. Linearity then entails |psi1#0>|psi2#0> [|psiR#1> + |psiR#2> + ... |psiR#400> ] --> [ |psi1#1>|psi2#1>|psiR#1> + ... + |psi1#400>|psi2#400>|psiR#400>] = |PSI> Now form S=|PSI>|2_i>, the sum being over the indices i and j that label a complete set of basis states for observer 2 and for system R respectively. One finds (if all of our fully labelled states are orthonormal) S_1 = S_1#1 + S_1#2 + .... + S_1#400 where S_1#N = |psi1#N>|psi2#208>|psiR#208>P_e is the correspondence called the "neural correlate of consciousness", which is the focus of a huge amount of current empirical work. Thus the theory ties directly into the empirical work. It is directly about our increments in knowledge, as we describe them to ourselves and to our colleagues. The operator P_e projects onto an entire state of a human body/brain, although only certain features may actually be relevant. I hope this spelling out of the essential philosophical and mathematical features of this pragmatic approach will make it more accessible. *************************************************************** Date: Mon, 14 Sep 1998 21:28:34 -0700 From: Stuart Hameroff Subject: [q-mind] Stapp to Klein on when and why MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" >From Henry Stapp Stan Klein, responding to Henry Stapp's "When does it happen? And why us?" [Stan] Henry raises excellent questions that he proceeds to answer: Question 1) Consider a photon going through a double slit, hitting film and then the film sits for 20 years before being looked at by a human. When does the reduction happen? Henry's answer is that reductions take place when knowledge is incremented. Since he is mainly interested in human knowledge his pragmatic approach is to have the reduction take place when the human inspects the film. Like the famous cat, the film was in a superposition of states for the 20 years until finally someone looked. Then Henry says: [Henry-previous] In order to stay close to scientific practice and empirical evidence I stick basically to the pragmatic stance, and hence to human knowledge. But from the ontological side it is clear that the theory must eventually be expanded so as to include a more general conception of knowledge: anthrocentrism is ontologically unacceptable, but pragmatically useful. To stay on track science should be pragmatically useful, but aimed at ontological completeness. [Stan] 'Pragmatic' is a flexible word. I find it pragmatically useful to say that in the above example the reduction takes place at the film exposure. For me it is pragamatic to avoid carrying along the superposition for 20 years. I have no problem in saying that the universe's knowledge is enhanced when the *film* records the landing spot of the light. The formalism of QM is nicely flexible in that it allows the reduction to take place at either the film or at the human 20 years later. I'd be interested in the opinion of others on this matter of when the reduction takes place. [Henry-reply] I was using "pragmatic" in the way William James [The Meaning of Truth] defines/uses the word, and in the way I used the word in my 1972 Amer. J. Phys article, "The Copenhagen Interpretation", which is reprinted in my 1993 book Mind, Matter, and Quantum Mechanics, section 3.4, "The Pragmatic Conception of Truth". The idea is hinted at in five sentences taken from that section. "James accepts, as a matter of course, that the truth of an idea means its agreement with reality. The questions are: What is the `reality' with which a true idea agrees? And what is the relationship `agreement with reality' by virtue of which that idea becomes true? ... What is the structural form of that relationship between an idea and a transexperiential reality that goes by the name of `agreement'? ... So if we want to know what it means for an idea to agree with a reality we must first accept that this reality lies in the realm of experience." This leads recognition of the fact that science is basically about theoretical models that we human beings create to allow us to grasp some of the relationships that exist among our experiences. This leads me (following Bohr et. al) to identify increments in knowledge as the basic realities of science, and to identify our physical theories as our representations of some of the structural relationships that exist among these increments in knowledge. James held that alternative possible theories should be judged on the basis of their utility: on how well they work for us. It is not clear that any such theory could exactly capture all of these connections in all their fullness. Pursuing this idea I formulate quantum theory as a theory of relationships among our increments in knowledge---described as we describe these increments to ourselves and to others---formulated in terms of connections between such experiences and our brains, as these brains are represented in a quantum mechanical model of the whole universe. von Neumann showed that the state reductions that, according to this theory, occur in conjunction with those particular increments in knowledge that are associated with "good measurements" can be shifted from the brain, which presumably is the physical system directly connected to someone's experience, to the degrees of freedom associated with the measuring device. Indeed, a principal purpose of the posting that Stan cites, was to show how this works. This close connection between the event in the brain and the event in the measuring device, in "good measurement" situations, is exactly what makes the theory work in practice. So the most practical point of view is to hold that, since the two are essentially in ono-to-one correspondence, one may place the jump event wherever convenient. Stan has been advocating this point of view for many years. Most physicists want to think that a real event occurs at the device, and that the experiential event is a consequence of it. That is certainly the natural way to think when one is doing ordinary physics experiments, where the nature of the connection between brain and mind is not the issue. This rough practical equivalence between the two placements is the root of all the disputes about the nature of the reality lying "behind" the phenomena. But until some possible empirical ramifications can be discovered the dispute remains outside science. For the scientific study of the mind-brain connection there are two very good to use this pragmatic theory. 1. The rough equivalence that I have described is merely that: it is not exact. In particular, it is based on the von Neumann idea of a "good measurement". Once that constraint is relaxed the whole situation becomes fluid, and there IS the possibility of specifying significant empirical differences. But first, some order must be created: what is quantum theory without good measurements? Things were surely happening in people's brains and minds before the idea of measurements arose. So what do we take as our foundation? The original basic premise of quantum theory is that reductions of the quantum mechanical state represent increment in human knowledge. Certainly if any experiment should discover that there was a reduction at a device---some absence of an intereference effect that the Copenhagen interpretation predicted---that would be a huge discovery: the basic Copenhagen premise would be overturned. So it is fair to claim the orthodox standard theory is one in which reductions represent increments in knowledge. With this basic theory in place one is in a position to study deviations from the equivalence that holds in the "good measurement" cases. Stan is focussing on the cases where there IS equivalence between different placements. But the basic theory should cover also the general case. 2. More importantly, one should not prejudice our scientific study of the mind/brain problem with notions carried over from Descartes and Newton, about the relationship between mind and matter. By saying that the location of the cut can be pushed out to the devices, or even to "lower" levels of the brain one is prejudicing the issue by making mind hostage to brain. That may indeed be true. But it ought not be assumed at the outset of an unbiased study. Quantum theory, pursued in a not unreasonable way, leads to a completely practical theory of the empirical facts---our describable increments of knowledge---that is parsimonious and combatible with all the empirically confirmed results derived from either classical physical theory or quantum theory. Yet it appears to allow, in a highly limited way, a certain inversion of the direction of causal flow that is not conceivable within classical physics. Question 2: Why not use the familiar Dirac VECTOR |PSI> to represent the state, rather than the operator S= |PSI>"operator", where numbers commute and operators don't, and (2) "Integration over phase space (statistical average)"--> Trace". One can immediately write down, at least formally, within the theory, the connection between experiences and their representations in brains, and the rule for computing relationships between experiences. Thus in the operator formulation one can "cut directly to the chase" so to speak. One does not need to instruct neurophysiologists and philosophers about these abstractions called "vectors": matrices are relatively simple to understand. Moreover, I NEED to use the operator notation. This is because in the actual universe "the observer" regarded as a vector is not well defined, due to the entanglement of observer degrees of freedom with those of the rest of the universe. My whole procedure of defining the state of (the body/brain of) the observer, and of specifying both the connection between this physical state and the possible experiences that it represents, and the computational procedures that connect our experiences to one another, are compactly and accurately represented in the operator notation. [Stan] I suspect that Henry is saying somthing wise and important and I'm just not getting it yet. It seems to me that the vector formalism, |psi>, works just fine in representing brains, observers, and entanglements. To me the operator formalism |psi> = |obs(t)> sum(|film_i(t)>) where film_i means the state of the rest of the universe with the photon landing at position i of the film. If one wants to see more details of what other things are going on in the universe then one expands |film_i(t)>. Period 2. Twenty years later the observer looks at film and becomes entangled. |Psi> = sum(|obs_i(t)> |film_i(t)>) Period 3. The wave function is rapidly reduced to the photon at say i=208. |Psi> = |obs_208(t)> |film_208(t)> Is this a legitimate way to describe what is going on? Henry says that the vector notation can't handle entanglements of the observer, but I don't see the problem. (he said 'in general the vector representing the body/brain of the human observer is not well defined'). There is just a summation over the entangled states. Doesn't something similar happen in the operator formalism? Does the operator formalism handle Period 2 (the brief time of entanglements) in a nicer (basis free) way than the vector formalism? But doesn't one want to exhibit the multiple potential film locations in that entangled period? If so then one must have the film locations (the choice of basis) visible in the operator formalism as well as the vector formalism. What am I missing? In both formalisms we can focus on HUMAN knowledge as being central so that it is only human knowledge that does the collapsings. So the advantage of the operators eludes me. As I recall, most of the beginning QM books (Albert, Feynman) use the vector rather than the operator formalism. So I'm surprised that you think the non-commuting operator formalism is more accessible to novices. [Henry-reply] Let x represent all the variables that are needed to describe your bain, and let y let represent all the variables that are needed to describe the entire rest of the universe. Let PSI(x,y) be the wave function of the universe. The wave function of your brain is a function of x. But PSI(x,y) is in principle a different function of x for each value y: you have a different "brain" for each possible location, individually, of each particle in the universe, and, more generally, for each possible combination of the locations of all the other particles in the universe. How do your propose you make sense of the idea of "your brain" if it is really this smeared out set of possibilities? No hand waving please: I demand an exact definition of your brain in the general case. Then I ask: Does it have, in general, the nice properties that my simple operator formula gives? I am sure that you will end up with my definition, because it is the only correct one. Then given that the central object of the theory must be represented in this way, and is neatly represented in this way, and that every other basic formula is also neatly represented in this way, I think the formalism ought to be formulated in this way, in spite of the fact that we physicists were taught to think in more awkward ways. Your simple examples are about the case in which the brain lacks entanglement with every other particle in the universe. But I stressed that vectors work fine in that simple case. [Stan] Let me ask a different (but similar) question unrelated to operators vs vectors. Consider period 1 when the film is unobserved by any human. Does the unexperienced film have a representation in the world if the world is only made of experiences. [Henry-reply] The rule is the same for every object O; the state of O is: S_O = Tr(O~) S, where S is the state of the universe, and Tr(O~) means partial trace over all degrees of freedon except those if O. The world is indeed made of experience in my model. But that experience does have structure, and this structure is real. Hence a representation of this structure is a representation of reality. Tables and chairs are such realities. Recall from your S-matrix days that a stable particle (such as an electron) is merely the effect of a pole singularity in the S matrix. There is no tiny object, but there is a causal link between the source prepared by an experimenter and the detector observer by one. That causal link is real. So in that important sense the electron is real, even though it is not a tiny object. Henry Date: Wed, 16 Sep 1998 22:27:31 -0700 From: Stuart Hameroff Subject: [q-mind] Reply to Awret - Henry Stapp MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" >From Henry Stapp Subject:Re: Comment on Chalmers objection to Stapp’s theory. From: Uziel Awret, [Uzi] The brain and attention The important point to realize here is that one does not need to go beyond standard QM to explain attention. As a matter of fact one does not need QM to explain attention. [Stapp-Reply] I was using attention, not trying to explain it. Attention is an aspect of the causal mind/brain process. It is caused by aspects of that process, and is part of the cause of other aspects: it is in the causal loop. My point was that quantum theory can naturally give to attention an "extra" causal efficacy, by allowing it to govern the otherwise undetermined Heisenberg choice. [Uzi] Attention and evolution In all fairness to Stapp he introduces "attention" here in a very specific context, yet he does not clarify the distinction between consciousness and attention. In this restricted context attention seems to be associated with the same type of "directed collapse" that Stapp advocates. If I understand correctly the way in which attention influences the collapse in Stapp's model is similar to the way in which an experimenter chooses to prepare the experiment in QM. While the way in which the question (which will elicit a yes or no answer) is chosen is not part of the standard quantum mechanical formulation, it is not clear at all how Stapp's attention operates. [Stapp-reply] Remember that the basic realities are supposed to be the conscious experiences, and each attention is an aspect of an experience. Suppose an experience e occurs that has a certain attentional aspect: my focus is on lifting this heavy rock! This experience generates S--> (P_e S P_e). Now the Schroedinger equation takes over: there is an evolution of the state of the brain [S_b = Tr(b~) S] that begins to create the possinilities for the next (experiential) event e'. Of course, there is a lot of thermal and other noise that will tend to push the next experience off the mark (destroy the focus of attention). But there is also the fuzziness created by the quantum uncertainties, which creates a fuzzy cloud around the slightly off-the-target brain state. Now a question must be posed. I am suggesting that the intention in e to attend to the target, which means essentially to make e'=e, at least as regards the essential focus on lifting the rock, will control the the Heisenberg choice, so that the question posed will be e' = e. not any other one of the possibilities associated with the fuzzy cloud of possibilities. Thus the focus of attention in the earlier e has in effect wiped out the effects of the thermal noise that would be expected to occur in a classical calculation, and also in a quantum calculation in which the attention in one experience does not have this coersive effect on the Heisenberg choice associated with its successor. [Uzi] If evolution shapes some physical substrate so that it can perform preparations or ask questions, [Stapp-reply] The physical subtrate does not pose the question: That "substrate" is a *representation* of certain aspects of the experiential reality that is posing the questions. .... [Stapp- previous] This development gives my answer to Chalmers' objection to my theory, which was that consciousness, per se, was not in the causal loop: this attentional effect does put consciousness per se into the causal mind-body loop, as explained in the target article (Whiteheadian Process and Quantum Theory of Mind)” [Uzi]] Attention and consciousness for Stapp are closely associated with freedom. A freedom to choose and influence the physical state of the brain. Also the collapse of a superposition of brain states which leads to an event is not completely random but governed by a principle which is outside of the standard Copenhagen interpretation. One position here is to say that this "freedom to attend to things" can not be formalized, and not only is it outside the known laws of Physics but can not be formalized as such.( Jean Paul Sartre equates consciousness with this kind of freedom.) [Stapp-Reply] I have stressed that everything is causal [assuming Nature's random choice to be basically causal, though beyond our power to predict], and that the Heisenberg choice is causally determined by attention, which we can describe to ourselves and to our colleagues ["I just focused all my attention on lifting that rock."], and which is therefore an element of a possible scientific description. [Uzi] Stapp believes that although the experiential is not an effect it is nevertheless a cause. [Stapp-Reply] I hold that experience is part of a causal process, and that one aspect of the causal structure is represented by the quantum-mechanically described physical world. That "physical" part has a causal gap, which I suggest can be filled by a causal process describable in psychological terms. From stapp@thsrv.lbl.gov Thu Sep 17 12:55:27 1998 Date: Thu, 17 Sep 1998 12:51:04 -0700 (PDT) From: Henry Stapp To: Stanley Klein Cc: A.Sloman@cs.bham.ac.uk, bdj10@cam.ac.uk, brings@rpi.edu, brucero@cats.ucsc.edu, chalmers@paradox.ucsc.edu, ghrosenb@ai.uga.edu, jmschwar@ucla.edu, keith@imprint.co.uk, patrickw@monash.edu.au, phayes@coginst.uwf.edu Subject: Re: attention and QM On Wed, 16 Sep 1998, Stanley Klein wrote: > > I strongly recommend that we don't ever use the word 'attention' in > our quantum mechanics discussions. The reason is that there is a huge body > of very interesting research on attention. Lots is known about its temporal > dynamics, its spatial properties, its dependence on stimulus, its capacity. > Not much is yet known about its neural correlates but that is also growing > rapidly. It seems we have one talk on the neural correlates of attention per > month here. So to bring in QM for attention seems totally out of place given > how successful the classical paradigm is. Much better to use QM for > 'consciousness' or 'awareness' or 'will' which are fuzzier concepts. > Stan > I believe that the reasons you mention are exactly why we should speak about attention: (1) there is a huge body of very interesting research on attention. Lots is known about its temporal dynamics, its spatial properties, its dependence on stimulus, its capacity. (2) Not much is yet known about its neural correlates but that is also growing rapidly. It seems we have one talk on the neural correlates of attention per month here. Stan says: So to bring in QM for attention seems totally out of place given how successful the classical paradigm is. But we may need to be looking precisely for effects that can be understood to first order in classical terms in order to be able to see that classical ideas do not work exactly: we must look for phenomena that can be understood in enough detail from a classical standpoint to be able to discern the need for another paradigm. Stan says: Much better to use QM for 'consciousness' or 'awareness' or 'will' which are fuzzier concepts. The other idea is to try to stay with the least fuzzy concepts. Pragmatic QT suggests that the causal process that "has just occurred" when someone reports that "a very strong focussing of attention has just occurred" is qualitatively different from the normal mechanical processing in the brain: statistical features should be different. I believe that this effect will be found. Of course, that does not prove that a classical account of the effect could not be proposed: a lot of detailed analysis will be needed to sort out the merits of classical vs pragmatic QT explanations. But first, of course, this prediction needs to be confirmed. Henry