On Mon, 20 May 1996, Paul Bains wrote:
Would someone have the patience to explain to me the status of the 'observer' in the measurement problem. I am told by a little birdie that it is an error to interpret the 'observer' as a pyschological subject. But if the observer isn't 'a subject' what is it? Is a camera an observer....
I'm writing a paper for JCS which deals with this subject at length (and some related issues and their implications for consciousness, AI, etc.) The question of the observer in quantum mechanics (henceforth QM) is highly controversial, partly because it was treated for a long time as if it weren't. (A notorious example: A student at Princeton tried to give a talk in the 50's on new ideas in the interpretation of QM, but the topic was immediately squelched by Oppenheimer, who said that the problem had been solved and could not profit by further discussion -- I read this in one of Jeremy Bernstein's books, I think.)
First Bohr's approach: (His writings are generally considered opaque, so trying to interpret "what Bohr really meant" is also highly contro- versial.) Bohr certainly believed that the conceptual difficulties of QM were related to the problem of subjectivity. A sample quote: "A close connection exists between the failure of our forms of perception, which is founded on the impossibility of a strict separation of phenomena and means of observation, and the general limits of man's capacity to create concepts, which have their roots in our differentiation between subject and object." ["The Quantum of Action and the Description of Nature" (1929), reprinted in Atomic Theory and the Description of Nature (Cambridge U.P., 1934, reprinted 1961) p. 96]
I will now go out on a limb and try to "clarify" Bohr. Physical analysis makes use of two kinds of descriptive concepts or "modes of description" -- on the one hand the historical, empirical relating of facts in our common experience of the physical world (things defined by reference to experience) and on the other abstract dynamical descriptions of causal processes and relations (things defined in terms of abstract conceptual systems, as in geometry). These two are clearly closely related and interdependent. Newtonian physics has conditioned us to expect the two kinds of description to converge into a mechanical picture of the cosmos in which the abstract, mathematically defined concepts of dynamics can be combined to create a geometrical "map" of the physical universe in which there are structural features which clearly correspond to every observable physical fact in our experience. In atomic physics, however, this expectation is frustrated. The dynamical models have a non-historical quality -- they cannot be interpreted as events in physical (empirical) space and time, but have only a probablistic relation to the outcome of experiments. For this reason, we cannot use quantum dynamical models as the basis for an ontological picture of the world. We will always require in addition a "classical" description of the apparatus in order to give "an unambiguous description" of the results of an experiment -- i.e. we must use the empirical, historical terms we have developed to describe the facts of the physical world as it features in our common experience. Curiously, only by making this concession to the "subjective" (or at least empirical) can physics retain its objective (factual) character. A purely abstract (axiomatic, geometrical) account of the physical world (such as Descartes was striving for) is not possible.
I would not agree with Bohr on all points (he took the need for "classical" concepts a little too literally and was not open to Feynman's useful conceptual innovations) but I believe his insights were quite profound. Unfortunately, physicists preferred to enshrine a few of his ideas as dogma rather than encourage continued discussion of them in depth, so that the best of his thought became little known and less understood. I think most of those who later rejected Bohr were really rejecting a textbook caricature.
The interpretive problem was enormously confused by the (otherwise laudable) contributions of Dirac and von Neumann. Dirac's formulation of QM is powerful and mathematically beautiful. Von Neumann's Mathematical Foundations put Dirac's theory on solid logical ground. The disaster occured when von Neumann tried to show the generality and conceptual consistency of QM by applying the dynamical formalism to the apparatus itself. This _seemed_ like a good idea. The apparatus is a physical system, too -- shouldn't it also obey quantum laws? Wrong! This was the opening of Pandora's Box, the fall from the Garden of Eden. Von Neumann had missed Bohr's point entirely! [In fairness, remember that von Neumann's purpose and the bulk of his book concerned mathematical issues, not philosophy.] The immediate result was infinite regression, facile talk about "psycho-physical parallelism", the "collapse of the wave function", and the Measurement Problem, which has haunted quantum physics for over 60 years.
The apparatus is qualitatively different from the quantum microsystem not because it is larger or obeys a different set of physical laws, but because it plays a different logical role in the physical analysis of the experimental situation. The "apparatus" is simply the experiment approached from the historical empirical point of view. The microsystem is the experiment approached from the abstract dynamical point of view. These are two aspects of the same thing, a duality as inescapable as left and right, and -- as Bohr said -- related to the Subject-Object duality.
Now, it is sometimes said by physicists that a human observer or conscious mind is not necessary to constitute a completed experiment, but only a macroscopic recording device to register the result. This is quite true. However, such a device is necessarily an entity defined primarily in empirical-historical terms. It cannot exist as an abstraction in the world of operators and state vectors. So the physical world is objective, but not in the sense of Descartes' realm of pure "extension", "geometry made real" (or a more general modern equivalent). Physics presumes the subject as well as the object.
There are other approaches (Many Worlds, Bohmian mechanics, etc.) but none widely accepted among physicists or (IMO) at all appealling. They all seem to me to derive from an inability or unwillingness to accept that it is not logically necessary for the two modes of description to converge to a Newtonesque mechanical model of the physical world. One can have an intelligible non-mechanical theory and in fact we have one. But this simple and straightforward idea is so uncongenial to many that it has been villified as Bohr's "mysticism". Furthermore, viewed in the wider context of epistemology and the problems of consciousness, the non-mechanical quality of QM is not only not mysterious, but actually natural, convenient, beautiful, economical, and satisfying.
jcs-online
menu