Date: Tue, 15 Sep 1998 22:10:57 -0700 From: Stuart Hameroff Subject: [q-mind] Classical vs quantum - Henry Stapp MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" >From Henry Stapp Subject: Can Classical Physical Theory Provide An Adequate Foundation For The Scientific Study Of Consciousness? I shall argue here that the answer to this question is "no". The premise of the argument is this: For a physical theory to be adequate for the scientific study of a phenomena, every aspect of that phenomena must lie in the domain of phenomena that the theory can in principle cover. I shall argue that classical physical theory can in principle cover only a certain limited domain of phenomena, and that some aspects of conscious experience lie outside this domain. Our conscious experiences are highly unified, and not easily separated into different parts. It took perhaps the genius of Descartes to carve out a part that could serve as the foundation of a rationally coherent physical theory. Descartes was the founder of analytic geometry, and the name "Cartesian coordinates" reflects that lineage. Reflecting perhaps on Kepler's ellipses and Galileo's cylinders rolling down inclined planes in accordance with mathematical rules, he singled out the geometric aspects of our experiences as something that an appropriately conceived theory, based on conceptions of mathematical structures existing in a spacetime framework, might rationally account for. Newton, Maxwell, and Einstein completed the job. Newton provided a way for accounting for locations and shapes (and changing locations and shapes) by assuming matter to consist of tiny objects moving on trajectories in conformity to mathematical rules, and Maxwell and Einstein filled in the details of these rules. The physical objects around us, and in the heavens, were, according to this theoretical proposal, to be considered to consist of huge numbers of these tiny objects. Hence all sorts changing shapes and trajectories of objects appearing in our phenomenal worlds could in principle be accounted for---within the theory---by having the tiny objects follow appropriate trajectories. So, broadly speaking, all *geometical* aspects of phenomena could in principle be covered by classical physical theory. If the phenomena of "a pain in your left big toe" occurs then that phenomena certainly has a geometric or spatial aspect, namely its "in-the-left-big-toe-ness" . But the pain aspect is neither a triangle-experience, nor a tetrahedron-experience, nor any shape-experience. The tiny objects in classical physical theory have no properties beyond those expressed by the trajectories they follow: masses and charges etc. become superfluous if the trajectories are specified. The only aspects of phenomena that CAN be accounted for by classical physical theory are locations and shapes (and their changes). Because the pain aspect of a pain-in-the-big-toe experience is not experienced as a location or shape the pain aspect lies outside the domain of phenomenal aspects covered by classical physical theory. I stress at this point that the subject under discussion is the adequacy of a THEORY to cover PHENOMENA. This is important, because a theory is something that is under theoretical control: it is exactly what its principles say it is. Classical physical theory is that theory that is taught in university courses in classical physical theory. I am not talking about some presumed real world that may somehow exist outside of all human experience, and that may have all sorts of imaginable and unimaginable properties. I am speaking of a mathematical model that is defined by stated principles, and is neither more or less that what those principle assert it to be. A physical theory must have some rules that define how it is related to human experience. This rule in classical physical theory is that to first order our experiences of shapes and locations are closely related to the shape and locations specified by the trajectoriies: if under reasonably ideal circumstances an experience occurs that someone describes to himself and his colleagues by saying a triangle-experience has just occurred, then, according to the postulated connection between the classical model and human experience, some distinguishable set of trajectorties is forming a figure similar to what would be defined in analytic geometry as a triangle. Under ideal circumstances the shape you experience, and can describe to yourself and your colleagues, is asserted to be close to the shape formed by the trajectories that constitute the objects you are examining. That's the assumption that is used to tie classical physical theory to phenomena. The fact that the picture of the world provided by classical physical theory must be regarded as a theoretical model, not a picture of the real world out there, was not as well understood by physicists at the end of the ninteenth century as it is understood by physicists at the end of the second millenium. We now know that things just aren't what Newton imagined them to be! So let's not mix up ontology with classical physical theory. My assertion that (re)started this exchange with Pat&Aaron was: Claim X: "Conscious experience is not logically within the classical conception of the universe" Now, I have just mentioned the link that ties classical physical theory to human experience. It was of the form "If E then C(E)" , where E is an experience and C(E) is its image in the classical model. Claim X is about whether connections in the other direction---from the classical model to human experience---are entailed or necessitated by the principles of classical physical theory. Claim X asserts no such connections are entailed or necessitated by the principles of classical physical theory: the principles of classical physics are completely compatible with the possibility that every possible physical world is completely devoid of conscious experience. So if we imagine some classically conceived world as being the classical model of the world we live in, then that very world, with no trajectory altered, could be devoid of consciousness without disturbing principle of classical physics. In his posting of Sept 13 Aaron says: Henry denies this. He says that in a classical world it is *impossible* for any physical combination of conditions to suffice for (i.e. ensure) the existence of consciousness. I.e. as Pat says he is making a strong impossibility claim. This claim of Henry's implies that EITHER (a) the patterns of biological evolution which produced human beings, chimps and other things we agree are conscious could not occur in a purely classical universe OR (b) if such forms of evolution were to occur in a classical universe they would not guarantee the existence of conscious organisms, for the human-like creatures which evolved in such a world could be so-called ``zombies''. I have no idea whether Henry believes (a) or (b) or has even noticed that he has to believe one or the other, for consistency. I certainly have noticed: (b) is exactly what I have been asserting repeatedly, as clearly as I could, over these dozens of exchanges with Pat&Aaron, provided one clarifies (b) by specifying that the "they would not guarantee..." in "they would not quarantee the existence of conscious organisms" are " the forms of all the systems formed from all the trajectories would not quarantee, by virtue of the principles of classical physical theory, ..." [It should not be necessary to add this clarification since these are the only principles available here to guarantee anything.} I am startled to learn that Aaron could imagine that I could not know what I meant my repeatedly asserted and defended words to be claiming. Since the only things that the principles of classical physical theory do is to fix trajectories, the only way in which these principles could guarantee the existence of consciousness in these organisms is for these principles to entail in some way that if certain patterns of trajectories exist then conscious experiences exist. But I have just noted that the only things the principles of classical physical theories entail are conditions on the forms of the trajectories. [I am taking the action-at-a-distance form of classical mechanics: the fields can be brought back in by extending the meaning of trajectories.] In the way I am using "entail" any consequence of the trajectories that follows from the principles of classical physical theories are "entailed", and this includes what follows from the "linkage principle" that links the model to human experiences. But, as noted above, this linkage principle pertains only to the content of existing experiences: it does not assert the existence of any conscious experience. Let me reply to Pat's arguments. His main argument is his identity-thesis argument: he simply asserts that conscious experience IS a brain activity, and demands that I disprove it. But my claim is about what can be denied within the framework of classical physical theory: the existence of conscious experience in every possible classical model can be denied. I am not denying that one can consistently add onto the classical theory some postulate that adds consciousness onto the classical model. If we stay in the framework of theories and (experiential) phonomena, where things are under control, and do not descend into murky ontological realms, then it seems clear that we must first define what the principles of classical physical theory are, and see what can be denied without violating these principles. Within this context Pat's challenge can be stated as follows: "I claim property C is entailed by the principles of classical physical theory: prove me wrong." The ace up his sleeve is that he does not define property C, and rejects any definition I propose. But he says it is specified by patterns of trajectories. He says C will by defined in a thousand years, but wants me to prove him wrong today. If I say C is never mentioned, and is not defined in classical physical theory, and hence cannot be entailed by that theory, he replies that the physical phenomena does exist, even though it has not yet been identified. To cut through this line of argument I say: If C is some "conscious experience" then the claim that C=M(C), where M(C) is a pattern of trajectories, means that M(C) must have every property that C has. But M(C) has only geometric properties, whereas C can have non-geometric properties, such as "pain". Aaron, citing the weakness in Pat's identity-theory approach that I have just exploited, takes a different tack: implementation theory. An analogy between "temperature" and "consciousness" is often cited by proponents of the idea that conscious is simply a property of a physical system that can be adequately represented in terms of the concepts of classical physical theory. But the phenomena covered by thermodynamics, such as readings on thermometers and barometers, and shapes of containers, are geometric qualities, and hence in the realm of things covered in principle by classical physical theory, whereas conscious experiences have other dimensions. A similar consideration holds for "implementation theory". I first quote Aaron's recent remark on this issue: [Aaron] >I have so far not been able to make any sense of Henry's notion of >"geometric objects". When I create a virtual machine in a computer >(which might, for all I know be a classical machine, though it probably >isn't nowadays...) then I am going way beyond geometry. > >I am harnessing all sorts of causal relations in the virtual machine >into an intricate web which constitutes a chess playing machine, or a >word processor, or a spreadsheet, or a model of a tornado, ... and those >(non-physical) causal relations are implemented in a very complex set of >physical causal relations and structures. > >Anyone who thinks that is just a matter of geometry and particles moving >in accordance with rules has a view of computers and physics which I >guess I have no hope of understanding. > >Causation, not rule-following, is central to my notion of physics, >whether classical or quantum mechanical. > >Henry seems to think classical physics is just a matter of geometrical >patterns of motion of particles (maybe he is a Humean about causation?). > >I see classical physics as being concerned with causal powers and causal >interactions of physical objects, states, events and processes (not >necessarily only particles e.g. why exclude things like electromagnetic >waves?) Within such systems configurations with appropriate boundary >conditions can produce new structures and processes with their own >causal powers not logically derivable from the physical specification. > >Unfortunately it is notoriously difficult to analyse the concept of >cause. So unpacking all this requires further work. >[Fin Aaron] Computer scientists do a huge amount of conceptual work. How that is achieved is part of what is being studied here. But it certainly does occur. Then they "implement" some conceptual structure in a physical process running on some physical system, which for present purposes, is assumed to be adequately described by classical physical theory. So this classically conceivable physical system S is "implementing" some complex idea in the mind M of some human being. That fact surely does not make S conscious. So what is the relevance of all that to the issue of whether some similar but perhaps even more complex physical system S' that is adequately described by the principles of classical physical theory would, just by virtue of these principles, be conscious. If no matter how complex this implementing system S' is it could, without conflicting in any way with the principles of classical physical theory, be devoid of consciousness, then consciousness would seem to be, within this classical conceptualization, a superfluous appendage. So how does one pass from the fact that a system that conforms to the principles of classical physical theory can implement the complex concepts of its builders to the conclusion that some such machines must by virtue of the principles of classical physical theory be conscious? One difficulty is that all of the implemented properties must be expressed in terms of the causal properties of the classically conceived system S' and these are geometrically expressible. So even if the pattern of trajectories in S' are implementing the programmer/builder's conception of "pain" one is faced with the question of how one is to pass from the geometric facts about these trajectories to the conclusion that these geometric facts necessitate, by virtue of the principles of classical physical theory, the existence an experience of pain in S'. Note that in pragmatic quantum theory no such problem arises: the primary elements are experiences of any kind that can be described to one's self and one's colleagues, and the brain correlates of such experiences.