Nonlocality versus nonreality, by H. Dieter Zeh

One of our new bloggers, Vlatko Vedral, has been posting a lot on the subject of "reality"--raising the hackles of FQXi's Dieter Zeh in the process. In a guest post, Zeh responds to Vlatko and defends the maligned concept of reality.

From Dieter Zeh:

Vlatko Vedral has recently published three blog posts that are concerned with the central interpretational dilemma that has emerged from quantum theory and its crucial experiments: Is Nature nonlocal or can a concept of "reality" not be applied to the microscopic world? Let me therefore reply in the order these blogs were presented--culminating in the questioning of reality.

1. The Parallel Universe Experiment

I agree with Vlatko Vedral when he objects to a proposal to test the hypothesis that parallel universes exist made by Frank Tipler in that no feasible experiment can distinguish between the Copenhagen interpretation of quantum mechanicsand the Everett interpretation. This is so, since the latter is a formal consequence of the Schrödinger equation when the equation is assumed to be universally valid, while the former also uses this equation, although it switches to classical concepts whenever it appears appropriate.

However, we seem to be favoring different conclusions that may be drawn from this experimental indistinguishability. While Vlatko seems to be happy that no multiple worlds are required, I prefer to draw the consequence that we do not have to apply any pragmatic change of concepts (quantum to classical). It may indeed be a matter of taste whether we apply Occam's razor to the facts (at the cost of complicating the dynamical laws or switching between different concepts--as for example in the Copenhagen interpretation) or to the laws (at the cost of complicating the facts--such as by concluding the existence of "many worlds"). In similar situations, one has usually chosen the second possibility--in particular in cosmology: the empirically confirmed laws (or perhaps hypothetically formulated new laws) are extrapolated as far as possible, even when they lead to unexpected or unobservable consequences. For example, in general relativity there may be event horizons that represent limits of observability, but they are usually not regarded as boundaries to a conceivable universe.

So it is mere polemic (though perhaps quite popular among physicists) when Vlatko Vedral starts his blog calling the many worlds interpretation "science fiction and fantasy." Science fiction either neglects the laws of nature or, at least, our technological limitations, while Everett's proposal was precisely derived from the assumption that the Schrödinger equation is exact and universal. While it would certainly be meaningless to argue about elements of reality that are dynamically entirely decoupled from the observable, Everett's parallel worlds are defined as the dynamical successors of wave function components (but not any classical "paths" or "histories") that were part of local (that is, observable) superpositions at some time in the past. Whether or not McCain may have won the election in some parallel world does not just require the conceivability of such a possibility, but that an appropriate component of the wave function has evolved out of some "measurement-like" process that happened at some time before the election. Quantum events of this kind are abundant, but may in this case have led to branches with extremely small norm.

2. The Speed of Spookiness

Here I again agree with the Vlatko's opinion that the recent experiment performed by Gisin et al.,which claims to have shown that information travels faster than the speed of light, does not prove the existence of drastically superluminal "spooky" action. This is so, since the observed correlations can be correctly described using a relativistic Schrödinger equation (such as assumed to apply in quantum field theory). An inconsistency with relativity only arises if you presume a collapse of the wave function as a physical process, but such a violation of the Schrödinger equation is neither necessary nor empirically indicated. I have explained this in detail for physicists here. (See, in particular, Sect. 4. The physical process of decoherence, which mimics a collapse, propagates like a three-dimensional relativistic zipper that is being opened.)

John Bell had shown that such correlations between measurement results at distant places, predicted by quantum theory, are incompatible with the assumption of unknown local variables describing reality. However, the confirmation of their existence does not only leave the mentioned fundamental dilemma undecided, it also allows for two different possible roots of nonlocality. Namely, if entanglement is a property of real states, nonlocality is already part of the kinematics, and these "real" nonlocal states would contain the correlations. No spooky superluminal action is then required. So there is no reason to stop using "archaic language" (as Vedral calls it), such as "reality" and "causality"!

3. Is Reality Really Real?

The answer to this question must evidently depend on what you mean by these words, which are laden with prejudice. It would be quite unusual, though, to identify reality with the world of ghosts and witches (another of Vedral's suggestions)!

The most important prejudice about reality is that it must be defined in space and time. Even Einstein and Schrödinger shared it because of our classical experience. Although Schrödinger was directly led to a wave function in configuration space when he invented it, he first specialized successfully on single particle wave functions, which are defined on space and might thus be interpreted similarly to classical fields. When he later pointed out that entanglement is the strangest property of quantum theory, he spoke nonetheless of statistical correlations, as he firmly believed in some local reality hidden behind the formalism. Similarly, Einstein could only think of spooky action at a distance when he discussed certain consequences of entanglement. So he concluded that quantum theory provides an incomplete description of nature - but he did not yet foresee the full consequences later discovered by Bell.

Nonlocal reality may appear hard to visualize, but it can easily be formally defined. In fact, we have always used such formalism whenever we applied quantum theory: the wave function. The problem is not only that its nonlocality, but also its well established probabilistic interpretation, applicable in certain situations, seem to be in conflict with its representing reality (individual states). However, this dynamical indeterminism can be shown to be compatible with a real wave function in several, more or less satisfactory ways.

As there are many examples which demonstrate that a superposition (such as represented by the wave function) may define individual (that is, observable, hence even "operationally real") physical properties, all serious attempts to save the concept of a microscopic reality have used a universal wave function as a fundamental part of it. Let me here mention three examples:

(i) Collapse theories assume that there are as yet unknown stochastic modifications of the Schrödinger equation which cause all "other worlds" (branches) to disappear. This is a fair suggestion, as, if true, it should someday be experimentally confirmed. (So far, all attempts have failed). Although it may avoid nonlocality on a macroscopic scale, this proposal still assumes microscopic reality to be nonlocal (represented by a wave function).

(ii) In Bohm's quantum theory, the wave function is also assumed to be part of reality, but it merely guides some "more real reality": the trajectories of conventional particles and fields in their configuration space. Statistical properties then arise from the assumption that the initial conditions for these trajectories are not known, and distributed in accordance with the Born rule. The disadvantage of this model is that its trajectories can never be observed, and must in fact behave in an extremely strange way--including superluminal motion--in order to reproduce the results of quantum theory. (See this piece that I wrote on the subject, for details.)

(iii) In contrast to these two proposals, Everett's interpretation adds no speculative elements to the wave function or its dynamics. Its central idea is that any observer must become part of an initial superposition whenever he observes the outcome of a quantum measurement. So he has a definite state by his own in each corresponding component (or "branch", or "Everett world"). All one has to assume then is that these components define separate observers (separate subjects), while the initial superposition has now entirely become dislocalized by means of what is called "decoherence". The observed quantum indeterminism is thus a consequence of the indeterministic history of the observers (their correlated branching). Since decoherence depends on the specific environment, it also explains the emergence and "textuality" of quasi-classical properties in a consistent way within a quantum reality described by the universal wave function. (You can read more about decoherence theory and the emergence of an "objective reality" in Anil Ananthaswamy's article on "Quantum Darwinism", although this picture may be a bit misleading. According to Everett's proposal, _all_ branches are real--but separated from one another by decoherence.)

Clearly, Everett's conclusions appear "weird" from a traditional point of view--but they form a dynamically and conceptually consistent picture that does not contain any speculative input. Why should reality be restricted to what is more or less directly observable to us? Nonetheless, in order to avoid this "extravagant" multiplicity of quasi-classical worlds within one quantum world, many physicists are ready to pay the very high price of abandoning reality altogether. But what do they really mean? Einstein insisted that "these people do not know what they are doing." According to my attempts to understand them, reality is systematically denied in the Copenhagen interpretation in order to circumvent consistency problems (such as "Is the electron really a wave or a particle?"). If there is no reality, one does not need a consistent description! Therefore, I am convinced that "complementarity" will someday be recognized as the greatest sophism in the history of science. In particular, no "wave-particle duality" is required any more in such a consistent description in terms of a universal wave function.

Undoubtedly, Niels Bohr has given us ingenious pragmatic rules of interpretation in terms of presumed "complementary" classical concepts. It is understandable that most experimentalists prefer just to use these rules, which, however, remain unsatisfactory from a fundamental point of view. All those famous "quantum paradoxes" occur only in terms of presumed classical or other local concepts, while they have all been predicted by using the wave function (see Sect. 5 of my explanation for more details). Denying the reality of the wave function, but nonetheless assuming it to carry some mysterious "quantum information" (as done in some modern variants of the Copenhagen interpretation) comes close to arguments used in esoteric circles.

--

H. Dieter Zeh is a Professor Emeritus at the University of Heidelberg, Germany.