The Editor

Physics Today

24:2:99

Dear Sir

Are the predictions of quantum theory (QT) perfect? This would appear to be the opinion of the majority of your correspondents in the very interesting discussions (February, 1999, pp 11-15 and 89-92) on the current state of the art. It seems to be taken for granted that future theories must necessarily reproduce all the present predictions, including the infamous counterintuitive ones, as it is held that no experiment has ever contradicted a single one.

Investigation shows one area, at least, where the evidence has been interpreted as favouring QT but in reality it is inconclusive. It is my belief that if the motivation were there it would be quite feasible to prove that the realist explanation for the EPR experiments, for example, is correct, the QT one wrong [1]. I have discussed the experimental weaknesses now with many people, including several of the experimenters concerned and the editors of Physical Review Letters and Physical Review A. They do not dispute my facts - that there are "loopholes" - though they may doubt their importance.

The data from real EPR experiments agrees with the predictions so far as I can tell for two reasons. One is that there are real, local, correlations at work. For example, the two signals in Aspect's experiments probably are emitted with parallel polarisation directions. Classical theory tells us that this correlation causes the coincidence curves to follow sinusoidal patterns (not, you will note, straight lines, the usual realist prediction for spin-1/2 experiments). The other reason is a complicated function of the actual experimental conditions, faith, social pressures, and the ability of the theory to adapt.

The theory as set out in many popular accounts refers to abstract cases in which every single pair of "particles" is detected. This has long been recognised as inapplicable to real conditions, which have almost all relied on "photons", only a small fraction of which are detected [2]. The result of modifying the theory to cover this and other imperfections has been to make it so flexible that it cannot be relied upon to predict more than the general trend.

The only difference between quantum theory and realism for experiments involving polarisation is that QT predicts a lower minimum to the coincidence curve. But to use the minimum as evidence in favour of non-locality is absurd! As soon as allowance is made for the actual, far from perfect, conditions, it becomes clear that the modified realist prediction can explain low minima (and consequent high "visibility") with no difficulty. They can arise through adjustment of the data prior to analysis, or to the polarisation angles not being evenly distributed, or to the detectors not following an exact square law, or indeed to several other features of actual experiments. These are the "loopholes" whose presence - but not importance - is well recognised.

So the current state of play is, officially, that the results are inconclusive [3]. New tests are planned, but there seems to be insufficient attention being paid to investigating the loopholes in the old ones. There is no sense of urgency, of recognition that non-locality cannot happen and so "genuine" explanations must be sought, with no stone left unturned.

One reason for this lack of urgency is the belief that entanglement does not, after all, involve non-locality. It is just a matter (see Anton Zeilinger's letter, p 13) of a "change [in the] representation of our knowledge". Now, if experimenters believe this, they are admitting that the world is obeying realism, not QT, for Bell's reasoning shows that the predictions cannot be explained in this way. The right answer is being obtained by the wrong method, which does not score full marks!

The logical arguments here are, it has to be said, hard to follow. Let us try and state them clearly.

1. QT, assuming entanglement - the "singlet state" - predicts high visibility for the coincidence curve for perfect EPR experiments.

2. In polarisation experiments, realism predicts lower visibility.

3. Recent experiments such as Gregor Weihs [4] have shown high visibility.

4. They do not involve polarisation in the same direct straightforward way as the Aspect experiments, for example. The experimental report does not give quite sufficient information to say just what is happening. There are (to a realist) certain to be local explanations.

5. Zeilinger (in whose laboratory the experiment was performed) states his belief that the results arise from changes in representation of knowledge.

6. As a realist, I can well believe this to be true.

7. But "entanglement" has not come into it!

8. So the experiment does not support entanglement and tells us nothing about QT.

9. All that remains is to look again at the facts and publish a full realist explanation.

10. Further experimentation is needed to check the true causes of other EPR experiments, as they are not all identical. If the "rearrangement" argument were applied to Aspect's experiments, say, it would produce the usual low visibility prediction of the standard realist model - unless full allowance were made for imperfections [1].

It's time this debate was resolved once and for all, as I believe that progress is being prevented by attempts to match theories to phenomena that do not happen.

References:

1. C. H.Thompson, Found. Phys. Lett. 9, 357, (1996), available at http://xxx.lanl.gov/abs/quant-ph/9611037; quant-ph/9711044, and web pages at http://www.aber.ac.uk/~cat.

2. J. F. Clauser and M. A. Horne., Phys. Rev. D 10, 526 (1974).

3. A. Afriat and F. Selleri, "The Einstein, Podolsky and Rosen Paradox" (Plenum Publishing Co. Ltd., London, 1998).

4. G. Weihs et al., http://xxx.lanl.gov/abs/quant-ph/9810080.

Caroline H Thompson