From hpstapp@lbl.gov Tue Sep 4 18:59:35 2007 Date: Tue, 4 Sep 2007 18:59:35 -0700 (PDT) From: Henry P. Stapp To: GeorgeWeis@aol.com Subject: Rovelli's Relational Interpretation. Dear George, You wanted to have a talk about Rovelli's "Relational Interpretation". I looked at the two papers you sent and believe that a "talk" about them needs to be preceded by a written exchange: there are just too many points that need to be made about them for a talk to be profitable: A lot of sorting out needs to be done. So let me begin by trying to clarify one basic point: perhaps the basic point! Rovelli's core claim (page 1 of RQM--Relational Quantum Mechanics) is that the fact that "distinct observers give different accounts of the same events" entails "that the notion of observer-independent state of a system is inadequate to describe the world beyond the h/bar --> 0 limit..." The notion of "inadequate"is somewhat obscure. But Rovelli gives later a clearer claim (section C) "Quantum mechanical description...CANNOT BE TAKEN as an "absolute"(observer independent) description of reality. Rovelli refers to the "unease" that quantum mechanics generates as his reason for turning to his "relational" interpretation. But if this "unease" is the motivation, then it might be argued that the best way to eliminate the "unease" is to do precisely the opposite of what Rovelli recommends, namely to explain how all the data, including the different descriptions given to the same events by different observers, fit coherently together into one absolute and objective (observer-independent) global reality. Such a solution would be a far more "adequate" to dispell "unease" than the profoundly obscure claim that there are no ultimate facts. Rovelli certainly does not show that such an absolute/objective solution is impossible. The fact that different observing systems have available to themselves different data, and have different systems of accessing and storing the information that they get, and of expressing/conveying that information to others, is quite sufficient to account for the fact that different observing&describing systems give different descriptions of the same events. The differing descriptions given by different observing systems is a completely insufficient reason for concluding that no underlying true facts exist. Such a conclusion would be far more radical than what is logically required by the fact that neither WE knowing/calculating/theorizing human beings, nor anythings essentially like us, can know everything. On the contrary, it can be argued that the cause of this "unease" is precisely the reluctance of of the afflicted thinkers to go beyond the inherently limited pragmatic/scientific (as contrasted to a frankly ontologically constructive) understandings. For a conception of reality that embraces the idea that there are "true facts that lie beyond the bounds of direct human knowledge" is far better able to quell "unease" than the essentially pragmatic interpretation of the founders of QM, and of Rovelli, which self-consciously restricts itself to the very tiny part of reality that is known or knowable to our biological human selves. Rovelli starts his argument by reminding us (p.3) that "it is important in the following discussion to keep in mind that the observer can be a table lamp." Rovelli then considers a state of a system S that is in a superposition of states |1> And |2>, and a situation in which the observer O (e.g., a table lamp) measures the state of the system S, Rovelli considers a *given specific measurement E by O of S that yields outcome |1>. But another observer P (another table lamp) who observes the combined system S&O only before the measurement by O of S, if it knows the Schroedinger equation, and how to compute its consequences for the evolution of the initial product state of S&O, will compute that this state evolves into a superposition of two quantum states, the first containing |1>, and the second containing |2>. Of course, table lamps do not make measurements or describe outcomes. But in any case there is no difficulty with the idea that if P makes no bservation on S&O after observing the initial superposed state, and has a computational capacity to make a QM theoretical prediction about the state of S&O at the later time, then that computation, based on the info available to P, can, of course, take no account the collapse that an objective account could say actually occurred. The fact that an observer/computer would compute a projection into the future that does not take account of happenings of which it has no knowledge is completely natural, hardly grounds for concluding that nothing happened. We normally believe that all sorts of things are happening of which we are unaware, and which our expectations can therefore take no account of. But the pragmatic Copenhagen interpretation does gives us a good *sufficient condition* for the objectivity of a happening: normally, what a human being describes as sense data describable in "everyday language refined by the concepts of classical physics" can be assumed to be objectively true. This is merely a sufficient practical rule, and We would like---meataphysically---to have a necessary and sufficient general rule. But our lack of knowledge of such a rule certainly does not entail that there are, actually or necessarily, no actual facts of the matter. In any case, Rovelli's "Main observation" [In QM different observers can give different accounts of the same sequence of events.] is certainly neither surprizing nor open to question. But Rovelli concludes from this (Section C) "Thus a QM description of a certain system...CANNOT BE TAKEN as an "absolute" (observer independent) rescription of reality." The truth, rather, is that there are many possible mutually computible QM descriptions, and the "pragmatic" descriptions are always relative to some actual or imagined possible set of circumstances. But the existence of these many different "relative" QM descriptions does not preclude the possibility that there is some absolute QM description that is compatible with all of the valid partial descriptions that represent the partial knowledge available to different "observing systems", and arise from the capacity of the different observers to formulate and express the mathematical consequences of the knowledge that they are able to acquire, and then to develop into future expectations via the knowm-to-them QM rules, by means of their own computational powers. But it is one thing to recognize the limitations in human knowledge, and the variety of representations of the limited knowledge of actual observers, quite another to extrapolate these easily understood limitations in our human-based representations to the universe as a whole, in order to dispell our "unease" with a pragmatic theory that is manifestly metaphysically/ontologically incomplete, though not in any way necessarily uncomplete-able within the general quantum mechanical framework, which explicity accommodates, within the density matrix formalism of von Neumann, a multiplicity of QM descriptions, some encompassing others, but all mutually compatible with the existence of a growing set of absolutely true facts, most of which are, as to be expected, unknown to any one measuring/observing/computing/describing system, or to the collection of all physical systems that have these enormous---yet still minuscule on the cosmic scale---powers. I have a huge number of additional comments scribbled on the margins of my copy of Rovelli's paper. But my present point is that it is useful to distinguish/isolate what Copenhagen quantum mechanics actually does for us, which is to allow each of us to form our own expectations on the basis of our own knowledge, and to allow each of these quantum mechanically expressed descriptions to be partial descriptions a single evolving density matrix of the universe, which specifies a well-defined rationally coherent growing set of absolute facts. This absolute ontology conforms to the conditions that the various personal descriptions of sense data at the level of "everyday language refined by the terminology of classical physics" are mutually compatible. This limited feature of absoluteness "suggests" that there is in fact a deeper level of absolute facts, even though any attempt to ascribe "certainty" to the existence of such an underlying reality is blocked by the problem that the empirical foundations of our represenations of our knowledge of the physical world lie wholly in the pragmatic-level description [which I think encompasses the essence of Rovelli's proposal] which, absolutely understandably, falls far short of being a description of every true fact. But is the demand for "certainty" in such matters reasonable? A description that encompasses only what thinking subsystems of the universe can know and form expectations about is likely to be grossly incomplete. So why not accept as a tentative "best available understanding" what the QM density matrix formalism so generously offers? Why straight-jacket our thinking by an unreasonable demand for certainly? i hope this note will help to keep our further discussions on a productive course, by identifying in advance the central issue; "certainty", restricted a tiny part of reality, versus the simple, reasonable, rationally coherent, and mathematically natural conjecture about the objective structure of the underlying whole that the quantum mathematical formalism automatically offers us.